Human leukocyte antigen associations in myalgic ...
Transcript of Human leukocyte antigen associations in myalgic ...
Human leukocyte antigen associations in myalgic encephalomyelitis/chronic fatigue
syndrome (ME/CFS) and immune modulating treatment
Thesis for the degree of Philosophiae Doctor (PhD)
By
Asgeir Lande
2020
Department of Medical Genetics, Oslo University Hospital
Faculty of Medicine, University of Oslo
© Asgeir Lande, 2021 Series of dissertations submitted to the Faculty of Medicine, University of Oslo ISBN 978-82-8377-842-7 All rights reserved. No part of this publication may be reproduced or transmitted, in any form or by any means, without permission. Cover: Hanne Baadsgaard Utigard. Print production: Reprosentralen, University of Oslo.
Table of contents Preface ............................................................................................................................................... 1 Acknowledgments ............................................................................................................................. 2 Abbreviations .................................................................................................................................... 4 List of publications ............................................................................................................................ 5 Summary in Norwegian ..................................................................................................................... 6 1 INTRODUCTION ......................................................................................................................... 8
1.1 MYALGIC ENCEPHALOMYELITIS / CHRONIC FATIGUE SYNDROME ........................ 8 1.1.1 Brief history ..................................................................................................................... 8 1.1.2 Chronic fatigue and post-exertional malaise (PEM) .......................................................... 9 1.1.3 Criteria for Myalgic Encephalomyelitis (ME) and Chronic Fatigue Syndrome (CFS) ...... 10 1.1.4 Epidemiology and burden for patients and society .......................................................... 16 1.1.5 Etiology and different disease models............................................................................. 16 1.1.6 Genetic contribution to etiology ..................................................................................... 19 1.1.7 Treatment....................................................................................................................... 20
1.2 GENETICS AND AUTOIMMUNE DISEASES ................................................................... 23 1.2.1 Brief introduction to genetic variation ............................................................................ 23 1.2.2 Genetics of complex disease ........................................................................................... 24 1.2.3 Studying genetics in complex disease: Case - Control association studies ....................... 24 1.2.4 Immunogenetics and autoimmune disease ...................................................................... 25 1.2.5 Autoimmune diseases are complex ................................................................................. 25 1.2.6 The HLA complex and classical HLA genes ................................................................... 26 1.2.7 HLA associations ........................................................................................................... 28 1.2.8 Immune modulating treatment in AID ............................................................................ 29
2 AIM OF THE PROJECT .............................................................................................................. 31 3 SUMMARY OF RESULTS ......................................................................................................... 32 4 METHODOLOGICAL CONSIDERATIONS .............................................................................. 35
4.1 STUDY POPULATION ........................................................................................................ 35 4.1.1 Choice of diagnostic criteria ........................................................................................... 35 4.1.2 Patient inclusion criteria ................................................................................................. 36 4.1.3 Comorbidities in patients ................................................................................................ 38 4.1.4 Patient representativeness ............................................................................................... 39 4.1.5 Representativeness for healthy controls .......................................................................... 39
4.2 STUDY DESIGN .................................................................................................................. 40 4.2.1 Population Stratification ................................................................................................. 40
4.2.2 Relatedness among subjects ........................................................................................... 42 4.2.3 Case - control matching .................................................................................................. 42 4.2.4 Collecting clinical data and measuring treatment effect................................................... 43 4.2.5 Placebo effect................................................................................................................. 44
4.3 HLA-TYPING ...................................................................................................................... 45 4.3.1 Brief history and our choice of HLA typing .................................................................... 45 4.3.2 Next Generation Sequencing (NGS) procedures in HLA typing ...................................... 46 4.3.3 Ambiguities in allele assignment .................................................................................... 48 4.3.4 Quality control ............................................................................................................... 49
4.4 HAPLOTYPES AND LINKAGE DISEQUILIBRIUM ......................................................... 51 4.4.1 Haplotype frequency estimation ..................................................................................... 51 4.4.2 Linkage disequilibrium (LD) - concept and methods for calculation ............................... 52
4.5 STATISTICAL ISSUES ....................................................................................................... 54 4.5.1 Power calculations and significance thresholds - type I and type II errors ....................... 54 4.5.2 Multiple test correction .................................................................................................. 55 4.5.3 Statistical methods in stratification analyses ................................................................... 56
5 ETHICAL CONSIDERATIONS .................................................................................................. 57 6 DISCUSSION .............................................................................................................................. 59
6.1 ARE HLA ASSOCIATIONS PRESENT IN ME/CFS? .......................................................... 59 6.1.1 Positive associations from Paper II compared to previous literature ................................ 59 6.1.2 Previously reported significant associations compared to our results ............................... 61 6.1.3 Negative HLA associations ............................................................................................ 63
6.2 DETERMINING THE PRIMARY ASSOCIATION SIGNALS IN THE HLA COMPLEX ... 64 6.3 VALIDITY OF THE REPORTED HLA ASSOCIATIONS ................................................... 65 6.4 IMMUNE MODULATION IN ME/CFS ............................................................................... 67
6.4.1 Experiences from rituximab intervention ........................................................................ 67 6.4.2 Effect and safety of cyclophosphamide administration .................................................... 67
6.5 CLINICALLY OR ETIOLOGICALLY DISTINCT SUBGROUPS OF ME/CFS PATIENTS 69 6.6 IS AUTOIMMUNITY PART OF ME/CFS PATHOGENESIS? ............................................ 71
6.6.1 Definition of an autoimmune disease .............................................................................. 71 6.6.2 Evidence for immune dysregulation in ME/CFS ............................................................. 71 6.6.3 The impact of HLA associations ..................................................................................... 73
7 CONCLUSIONS AND FUTURE PERSPECTIVES ..................................................................... 75 References ....................................................................................................................................... 77
1
Preface
During this doctoral degree I have read a lot about ME/CFS, and I would like to emphasize a
few points that I have reflected upon. The first is the striking severity of the disease. Stories
from patients and relatives have made me understand how much ME/CFS really impairs the
quality of life. Especially, meeting housebound and bedridden patients, in some cases unable
to speak or move, has made an unforgettable impression. The second is the tense climate that
has characterized the debate about ME/CFS, in Norway as well as internationally. Everybody
seems to have an opinion about this disease, and despite the lack of knowledge, some voices
have seemingly drawn their conclusions nevertheless. There is no scientific consensus on the
causes of ME/CFS, nor has effective treatment been established. It is therefore very important
that researchers are curious about different hypotheses, and that research investigate this
disease widely, from all angles. This certainly includes a thorough biomedical approach, and
many of the patients and relatives I have met express their gratitude for such research being
conducted. At the same time, it is my experience that patients and their caregivers are fully
aware of the complexity of ME/CFS. A complex puzzle can only be understood bit by bit, and
I am humbly aware that the papers in this thesis provide only a small contribution to the sum
of knowledge that needs to be accumulated. Still, I am proud to participate in research on
ME/CFS, a serious disease that affects a large group of patients and urgently needs to be
better understood.
2
Acknowledgments
My part of the work in this thesis was carried out at the Department of Medical Genetics,
Oslo University Hospital (OUS) from 2016 to 2020. I want to thank The Kavli Trust for
funding my PhD position and much of the research in this project. My deepest gratitude goes
to all the ME/CFS patients participating in our research, and to their caregivers.
I want to express my deep gratitude to my main supervisor, Marte K. Viken. Thank you,
Marte, for helpfully and competently guiding me through so many different parts of our
research. I guess we both have an eye for details, but I needed your expertise to focus on the
right ones! I have really appreciated your efficacy and precision, both in lab work and in
writing. I am deeply grateful to co-supervisor Professor Benedicte A. Lie. You have an
inspiring attitude, combining ambition and scientific abilities with being relaxed and kind
(often with a delightful laughter). You are almost always available and very efficient, which
makes you a great research group leader. I am very grateful to co-supervisor Professor Ola
Didrik Saugstad. Your experience and motivation and your care for patients and relatives
have been truly inspiring. I really enjoyed your companionship during travels to meetings and
conferences.
Next, I want to thank members of our ME/CFS research group at OUS: Co-author and
Professor Torstein Egeland (thank you for important advice and feedback), Anne Rydland
(thanks for your help in lab) and office buddy Riad Hajdarevic (thanks for hours of interesting
discussions about ME/CFS and lots of other subjects). I also want to thank Irene Andersen
and Fridtjof Lund-Johansen at the Department of Immunology and Transfusion Medicine and
all volunteers in the Norwegian Bone Marrow Donor Registry.
I want to express my appreciation to all collaborating researchers and co-authors in the
ME/CFS research group at Haukeland University Hospital: Øystein Fluge, Olav Mella, Ingrid
G. Rekeland, Kari Sørland, Karl Johan Tronstad, Kristin Risa, Kristian Sommerfelt,
Alexander Fosså and many others. Your thorough scientific work and care for patients have
been inspiring. Thank you Katarina Lien for sharing the spot on description of dramatically
fluctuating self-esteem during a lengthy "creative process" such as a PhD. My deep
appreciation goes to employees and co-authors at the CFS/ME biobank at OUS, in particular
Elin B. Strand, Daysi D. Sosa and Wenche Kristiansen.
3
I want to thank all members of Benedicte's immunogenetics research group: Line, Kari,
Fatima, Maria, Ingvild, Riad, Anne, Marte, Hanne and Teodora. It has been great working
with you, and sharing our perspectives in regular meetings and presentations. Especially I
want to thank you Siri, for plentiful of help in the lab and interesting discussions. I also thank
the staff at NSC for your impeccable service with NGS sequencing. To all colleagues at the
Department: Thank you for comforting and inspiring me. And thanks for interesting
discussions, for instance about how to interpret bad dreams into constructive plans and ideas.
Thanks to Otto Sundar for extracting DNA.
Thank you hard-working employees and volunteers in ME-foreningen and Kjersti Krisner for
interesting discussions and help with patient recruitment. Thanks to patient representatives
participating in the planning of our research.
Dear friends and family, thank you for supporting and challenging me, and for everything we
experience together. Dear mom and dad, thanks for all the love and care and your belief in
me. Dad, you will always be my hero. Favorite mom, I really enjoy your company; you’re an
irreplaceable discussion partner, and probably the only adult sharing my childish sense of
humor (maybe it's a monogenic trait!).
And finally, the three persons that I love the most, my wonderful wife Anett and our super-
boys Ulrik and Ferdinand. Your wisdom and sincerity and your support in me, Anett, is
fundamental for my being. I guess I've been a more difficult man to live with as the PhD drew
out (and out), and you have made an enormous effort. Despite our differences, we
continuously discover yet new similarities between us (it’s funny, right?), and our love seems
to grow stronger each year together.
Kjære Ulrik og Ferdinand, dere to er den aller største gleden i mitt liv. Jeg er så uendelig stolt
av dere, og gleder meg til alt vi skal oppleve sammen framover. Nå er pappa endelig snart
ferdig med doktorgraden!
Asgeir Lande, November 2020
4
Abbreviations
AID Autoimmune disease
BMM Bio-medical model
BPSM Biopsychosocial model
CCC Canadian consensus criteria
DNA Deoxyribonucleic acid
DSQ DePaul symptom questionnaire
EBV Epstein-Barr virus
GWAS Genome-wide association study
HLA Human leukocyte antigen
HWE Hardy-Weinberg equilibrium
ICC International consensus criteria
KIR Killer immunoglobulin-like receptor
LD Linkage disequilibrium
Mb Mega bases - million nucleotide bases
ME/CFS Myalgic encephalopathy/encephalomyelitis/Chronic fatigue syndrome
NBMDR Norwegian bone marrow donor registry
NGS Next generation sequencing
NHST Null-hypothesis-based significance testing
OR Odds ratio
PCR Polymerase chain reaction
PEM Post-exertional malaise
POTS Postural orthostatic tachycardia syndrome
p_nc P-value, non-corrected
p_c P-value, corrected
RCT Randomized controlled trial
RR Relative risk
US United States
95%CI 95% confidence interval
5
List of publications
Paper I
Asgeir Lande, Irene Andersen, Torstein Egeland, Benedicte A. Lie, Marte K. Viken:
HLA -A, -C, -B, -DRB1, -DQB1 and -DPB1 allele and haplotype frequencies in 4,514 healthy
Norwegians, Hum Immunol 2018 Jul;79(7):527-529
Paper II
Asgeir Lande, Øystein Fluge, Elin B. Strand, Siri T. Flåm, Daysi D. Sosa, Olav Mella,
Torstein Egeland, Ola D. Saugstad, Benedicte A. Lie & Marte K. Viken:
Human Leukocyte Antigen alleles associated with Myalgic Encephalomyelitis/Chronic
Fatigue Syndrome (ME/CFS), Sci Rep 2020 Mar 24;10(1):5267
Paper III
Ingrid G. Rekeland, Alexander Fosså, Asgeir Lande, Irini Ktoridou-Valen, Kari Sørland, Mari
Holsen, Karl J. Tronstad, Kristin Risa, Kine Alme, Marte K. Viken, Benedicte A. Lie, Olav
Dahl, Olav Mella, Øystein Fluge:
Intravenous cyclophosphamide in Myalgic Encephalomyelitis/ Chronic Fatigue Syndrome. An
open-label phase II study, Front Med 2020 Apr 29;7:162.
6
Summary in Norwegian
ME/CFS er en alvorlig sykdom kjennetegnet av medisinsk uforklarlig, kronisk og betydelig utmattelse, ledsaget av en rekke karakteristiske symptomer som smerter, søvnvansker, kognitive problemer og anstrengelsesutløst symptomforverring. Det finnes ingen biomarkør eller annen objektiv markør for å stille diagnosen, og flere ulike kliniske kriterier og offentlige retningslinjer blir brukt ved diagnostisering. Pasientgruppen er heterogen, med stor variasjon i debuttidspunkt og alvorlighet. Sykdommen medfører betydelig lidelse for pasienter og pårørende, og utgjør også en vesentlig økonomisk kostnad for samfunnet. Prevalensanslagene varierer fra 0,1% og oppover, og kvinner affiseres oftere enn menn. Etiologi og patogenese ved ME/CFS er ikke kjent, og mange ulike teorier eksisterer. Sykdommen anses å oppstå i et komplekst samspill mellom forskjellige bakenforliggende, utløsende og opprettholdende faktorer av både miljømessig og genetisk karakter. En sentral hypotese er at immundysregulering eller autoimmunitet er involvert i patogenesen. Signifikante forandringer i immunologiske parametere er påvist i ulike studier av ME/CFS-pasienter, men de fleste av disse er ikke undersøkt med samme metodikk i uavhengige pasientgrupper, og funnene er i liten grad reprodusert. Blant de rapporterte funnene er endret funksjon og antall av NK- , B- og T-celler, økt forekomst av auto-antistoffer, endrede nivåer av sentrale pro- eller anti-inflammatoriske cytokiner, og endrede epigenetiske markører forenlig med immundysregulering. Mange pasienter rapporterer post-infeksiøs symptomdebut, og enkelte studier har funnet assosiasjoner mellom ME/CFS og infeksiøse agens som Giardia Lamblia eller Epstein-Barr virus. Et sentralt element i evalueringen av hypotesen om autoimmunitet i ME/CFS, er grundig undersøkelse av HLA-genene (human leukocyte antigen). HLA-genene, som er svært polymorfe, og fremviser stor variasjon både innad i og på tvers av befolkningsgrupper, spiller en avgjørende rolle i reguleringen av immunresponsen. HLA-assosiasjoner er et sentralt kjennetegn for autoimmune sykdommer generelt. For enkelte autoimmune sykdommer, som cøliaki og narkolepsi, er det påvist sterke HLA-assosiasjoner, mens for andre, eksempelvis Crohns sykdom, er det etablert statistisk signifikante, men likevel moderate til svake assosiasjoner med relativt sjeldne HLA-alleler. Det er utført en håndfull HLA-assosiasjonsstudier med ME/CFS-pasienter, og tross enkeltstående statistisk signifikante resultater, er det ingen gjennomgående funn. Studiene er små (N≤110 pasienter), og har dermed begrenset statistisk styrke for avdekking av moderate eller svake assosiasjoner, med andre ord er det høy risiko for falsk negative resultater. Ingen av studiene har gjennomført statistisk korreksjon for multippel testing, og dermed er det også høy risiko for falsk positive resultater. Ett av hovedmålene i dette prosjektet var derfor å gjennomføre en stor HLA-assosiasjonsstudie med flere hundre ME/CFS-pasienter og etnisitetsmatchede friske kontroller. Artikkel 1 utgjør det første norske, publiserte HLA-referansepanelet, bestående av 4514 friske norske personer, rekruttert fra det norske beinmargsgiverregisteret. Moderne, høyoppløselig HLA genotyping var utført for alle individene. Dette referansepanelet representerer et nødvendig grunnlag for robuste HLA-assosiasjonsstudier i norsk befolkning.
7
I artikkel 2 gjennomførte vi en stor og høyoppløselig genetisk HLA-assosiasjonsstudie med 426 voksne norske ME/CFS-pasienter og 4511 kontrollpersoner fra artikkel 1. Pasientene var diagnostisert i henhold til Canada-kriteriene, i et forsøk på å begrense heterogeniteten i pasientgruppen. Vi undersøkte allelfrekvenser for de klassiske HLA-genene HLA-A, -B, -C, -DRB1, -DQB1 og -DPB1, og samlet inn klinisk informasjon via spørreskjema. Vi oppdaget to nye og uavhengige HLA-assosiasjoner, representert ved allelene HLA-C*07:04 (OR=2,1, 95%CI [1,4-3,1], pnc= 0.0001, pc= 0.001) og HLA-DQB1*03:03 (OR=1,5, 95%CI [1,1-2,0], pnc= 0.005, pc< 0.05). Disse allelene ble påvist hos henholdsvis 7,7% og 12,7% av ME/CFS-pasientene. Ett eller begge av disse risikoallelene ble påvist hos 19,2% av pasientene og 12,2% av kontrollpersonene (OR=1,7, 95%CI [1,3-2,2]). ME/CFS-pasienter med disse risikoallelene hadde signifikant høyere komorbiditet for autoimmune tilstander, uten at vi fant holdepunkter for at komorbiditet i seg selv representerte en metodisk skjevhet. Det var ingen assosiasjon mellom risikoallelene og andre kliniske parametere, som kjønn, alder, sykdomsalvorlighet eller post-infeksiøs debut. Det finnes ingen etablert, effektiv behandling for ME/CFS. En rekke ulike behandlingsformer prøves ut, både av pasienter og i kliniske studier. Knyttet til hypotesen om at autoimmune mekanismer bidrar i etiologien, finnes hypoteser om at immunmodulerende behandling har klinisk effekt ved ME/CFS. På onkologisk avdeling ved Haukeland Universitetssykehus har flere pasienter med langvarig ME/CFS, som også har utviklet kreft, rapportert om tydelig bedring av sentrale ME/CFS-symptomer etter konvensjonell kreftbehandling. Denne behandlingen har i flere tilfeller inkludert cyclophosphamid, et immunsuppressiva som ofte brukes i kreftbehandling, men som også har etablert effekt ved behandling av flere autoimmune sykdommer. Artikkel 3 er en ikke-blindet, ikke-placebo-kontrollert intervensjonsstudie med cyclophosphamid-infusjoner hos 40 norske ME/CFS-pasienter, utført for å belyse gjennomførbarhet, sikkerhet og potensiell klinisk effekt. 22 av 40 pasienter rapporterte signifikant bedring i sentrale symptomer som fatigue og fysisk fungering, og effekten var for de fleste langvarig og fortsatt tilstede etter 4 års oppfølging. De fleste pasientene hadde bivirkninger, men det var få alvorlige bivirkninger, og vi konkluderte med at cyclophosphamid-behandling av ME/CFS-pasienter synes forsvarlig, gitt den mulige kliniske effekten. Et svært sentralt ankepunkt er imidlertid at placeboeffekten kan ha bidratt til resultatene. HLA risikoallelene fra artikkel 2 var statistisk signifikant knyttet til cyclophosphamid-respons, og denne sammenligningen var dobbeltblindet, ettersom HLA-status ikke var kjent for hverken pasienter eller observatører under gjennomføringen av studien. Samlet sett tyder våre resultater på at immunsystemet spiller en rolle i bakenforliggende, utløsende og/eller opprettholdende faktorer ved ME/CFS. Resultatene må replikeres i uavhengige pasientgrupper før de kan anses som etablerte. Cyclophosphamid-behandling bør ikke tilbys ME/CFS-pasienter før resultatene eventuelt er bekreftet i en randomisert, kontrollert studie. HLA-assosiasjoner i seg selv kan ikke dokumentere at autoimmune sykdomsmekanismer er involvert i etiologien, men våre funn representerer et insentiv til videre utforsking av hypotesen om immunforsvarets rolle i ME/CFS.
8
1 INTRODUCTION
1.1 MYALGIC ENCEPHALOMYELITIS / CHRONIC FATIGUE SYNDROME
Myalgic encephalomyelitis (ME) and Chronic Fatigue Syndrome (CFS) are medical
conditions characterized by reduced ability and intolerance to exercise and activity, along
with a variety of additional symptoms. ME and CFS is defined by diagnostic criteria and
official guidelines. Several diagnostic criteria for ME, CFS or both have been published [1-7].
Throughout most of this thesis, I will use the conjoined term ME/CFS for this disease, in line
with relevant official guidelines [8-10].
1.1.1 Brief history
For centuries, different terms have been used to describe and study medically unexplained
chronic fatigue, with varying arrays of additional features. Generally, each term reflects a
specific belief or constraint regarding symptomatology or etiology, and several
environmental, metabolic, infectious, immunologic, and psychiatric causes have been
suggested [11]. The following are examples from recent decades of diagnoses and terms being
used to describe conditions with partially overlapping symptomatology with ME/CFS:
Neurasthenia, epidemic neuromyasthenia, Akureyri disease, Royal Free disease, chronic
Epstein-Barr virus syndrome, post-viral fatigue syndrome, fibromyalgia and "mass hysteria"
[11, 12].
Efforts to construct CFS case definitions can be traced back to the 1950s [5]. The Holmes
Criteria from 1988 distinguished CFS from the so-called chronic Epstein-Barr virus syndrome
and proposed the first case definition in the United States (US) [1]. The Oxford Criteria from
1991 define CFS mainly based on chronic fatigue, accepting many comorbid entities [2]. The
1994 Fukuda criteria for CFS, presenting more specific mandatory symptoms, as well as
exclusionary diagnoses such as primary psychiatric disease, dominated for many years and
are widely used in research [3]. Simultaneously, the name ME gained widespread attention
from the 1980s through publications describing outbreaks of CFS-like illness, with post-
infectious onset and proposed inflammatory pathogenesis [13, 14]. In 2003, the Canadian
Consensus Criteria (CCC) for ME/CFS were published, providing detailed clinical description
and diagnostic procedures, and making post-exertional malaise a mandatory symptom [4].
CCC from 2003 are among the most commonly applied criteria today, although they were
9
revised in 2010 [5]. International Consensus Criteria (ICC) for ME were published in 2011
[6], and in 2015, the US National Academy of Science published a thorough report on
ME/CFS, proposing new criteria and also a new name (Systemic Exertion Intolerance Disease
- SEID) [7].
1.1.2 Chronic fatigue and post-exertional malaise (PEM)
Chronic fatigue or exhaustion of disabling character and post-exertional malaise (PEM), are
both fundamental symptoms in ME/CFS. Both fatigue and PEM need a broader introduction
before we can move on to the definition of ME/CFS.
Fatigue and chronic fatigue
Fatigue is a very common symptom defined as a subjective and overwhelming feeling of
tiredness or lack of energy, associated with a certain degree of impaired functioning [15].
Because tiredness and lack of energy are normal responses to physiological or psychological
strain, some definitions discern between normal fatigue and pathological fatigue [16]. The
latter is often characterized by chronicity, lack of alleviation by rest or relaxation, and
decreased quality of life. Chronic fatigue can be defined by duration and severity, and its
prevalence will vary accordingly. In any case, chronic fatigue is also a common condition,
reported in 8.5% of British primary care patients [17], 11% of the Norwegian population [18],
and close to 30% in a Dutch survey [19]. Although fatigue is clearly a non-specific symptom,
it is often associated with specific medical conditions, e.g. anemia, cancer, rheumatoid
arthritis or depression [16, 20]. Hence, the crucial point in further categorization of chronic
fatigue is whether such an underlying cause of the impairment can be identified. If an
underlying cause can be identified through medical or diagnostic evaluations, the chronic
fatigue will be secondary, and a diagnosis of ME/CFS will generally be excluded. If an
underlying condition cannot be identified, the chronic fatigue is termed primary, idiopathic
(without a known cause) [3] or medically unexplained [4, 5, 21]. In general, different criteria
for ME/CFS are further specifications of idiopathic chronic fatigue [3, 4, 6]. Figure 1 is a
simplified sketch of the relation between fatigue, chronic fatigue, and ME/CFS.
10
Figure 1 Illustration of a proposed hierarchical subgrouping of fatigue.
Post-exertional malaise (PEM)
PEM is defined as an activity-triggered profound fatigue with a decline in function level or an
aggravation of other central ME/CFS symptoms [4-6, 22]. The trigger is often no more than
social interaction or normal daily activities, depending on disease severity. The deterioration
can be delayed with hours or days, and can often last for days or weeks [10, 22]. Many
criteria include PEM, or an equivalent feature, as a mandatory and cardinal ME/CFS
symptom [4, 6, 7, 23]. PEM has been objectively demonstrated to some extent [22, 24], and is
thought to be central in ME/CFS pathogenesis. However, defining PEM is inherently difficult
as it is mainly described subjectively [22].
1.1.3 Criteria for Myalgic Encephalomyelitis (ME) and Chronic Fatigue
Syndrome (CFS)
Many different criteria and guidelines can be applied when identifying ME/CFS in clinical or
research settings [1-7, 9, 23, 25]. Table 1 gives an overview of four central criteria. Common
for all criteria is that ME/CFS is diagnosed on clinical grounds alone, as no aberrant and
objective test result is neither necessary nor sufficient for the diagnosis. Most criteria are
based on two key aspects:
i) Mandatory symptoms and some additional symptoms must be present, and
ii) The fatigue and other core symptoms must be medically unexplained.
Fatigue
Acute fatigue / fatigue of short duration Chronic fatigue
Secondary fatigue Idiopathic chronic fatigue
ME/CFSNon ME/CFS Idiopathic chronic fatigue
Primary or secondary
11
Chronic fatigue is mandatory in all criteria, and PEM is mandatory in most, including CCC.
The fatigue must be substantial (defined in some criteria to at least 50% reduction in
occupational or recreational activity) and medically unexplained, without alleviation as
normal by rest. The fatigue must be chronic (usually a duration of 6 months is required),
although fluctuations or relapsing-remitting patterns are common. Cognitive dysfunction,
disturbed sleep, chronic pain and autonomic manifestations are other central symptoms, either
mandatory or additional. Common objective observations are thermolability, pallor, sweating,
palpitations/tachycardia and fasciculations.
12
Tab
le 1
Su
mm
ary
of fo
ur o
f the
mos
t wid
ely
appl
ied
crite
ria fo
r ME
and/
or C
FS [2
-4, 7
]. M
odifi
ed fr
om [2
5]. C
ontin
ues o
n th
e ne
xt p
ages
.
Oxf
ord
cons
ensu
s cr
iteri
a (1
991)
(C
FS) [
1]
US
CD
C F
ukud
a C
rite
ria
(199
4) (C
FS) [
2]
Can
adia
n C
onse
nsus
Cri
teri
a (2
003)
(M
E/C
FS) [
4]
Inst
itute
of M
edic
ine
(201
5)
(CFS
/ME
) [7]
Req
u-ir
ed
sym
p-to
ms
Eith
er:
CFS
: Fat
igue
, de
finite
ons
et, s
ever
e an
d di
sabl
ing,
affe
cts
phys
ical
and
men
tal
func
tioni
ng (d
urat
ion
> 6
mon
ths,
sym
ptom
s pre
sent
>
50%
of t
he ti
me)
O
r:
PIFS
(a su
btyp
e of
C
FS):
- F
ulfil
ls C
FS
defin
ition
abo
ve
- Def
inite
evi
denc
e of
in
fect
ion
at o
nset
or
pres
enta
tion
(pat
ient
’s se
lf-re
port
is u
nlik
ely
to b
e su
ffici
ently
relia
ble)
- S
yndr
ome
pres
ent >
6
mon
ths a
fter
infe
ctio
n
- The
infe
ctio
n ha
s be
en c
orro
bora
ted
by
labo
rato
ry e
vide
nce.
Fatig
ue:
- Clin
ical
ly e
valu
ated
, un
expl
aine
d, p
ersi
sten
t, or
re
laps
ing
fatig
ue (l
astin
g >
6 m
onth
s) o
f new
or d
efin
ite o
nset
- N
o re
lief b
y re
st
- Res
ults
in a
subs
tant
ial
redu
ctio
n in
pre
viou
s lev
els o
f oc
cupa
tiona
l, so
cial
, or p
erso
nal
activ
ity.
Plus
4 o
r m
ore
of th
e fo
llow
ing
sym
ptom
s tha
t per
sist
or re
cur
durin
g 6
or m
ore
cons
ecut
ive
mon
ths a
nd th
at d
o no
t pre
date
th
e fa
tigue
: Se
lf-re
porte
d im
pairm
ent o
f sh
ort-t
erm
mem
ory
or
conc
entra
tion,
Sor
e th
roat
, Te
nder
lym
ph n
odes
, Mus
cle
pain
, Mul
ti-jo
int p
ain
with
out
swel
ling
or re
dnes
s, H
eada
ches
of
a n
ew ty
pe/p
atte
rn/s
ever
ity,
Unr
efre
shin
g an
d/or
inte
rrupt
ed
slee
p, P
ost-e
xerti
onal
mal
aise
(a
feel
ing
of g
ener
al d
isco
mfo
rt or
un
easi
ness
) las
ting
mor
e th
an 2
4 ho
urs.
New
ons
et/s
igni
fican
tly a
ltere
d sy
mpt
oms
pers
istin
g ≥
6 m
onth
s and
mee
ting
all o
f the
fo
llow
ing:
- F
atig
ue: A
sign
ifica
nt d
egre
e of
new
ons
et,
unex
plai
ned
fatig
ue th
at su
bsta
ntia
lly re
duce
s ac
tivity
leve
l. - P
EM
: Ina
ppro
pria
te lo
ss o
f phy
sica
l and
m
enta
l sta
min
a, ra
pid
mus
cula
r and
cog
nitiv
e fa
tigab
ility
, and
a te
nden
cy fo
r oth
er
asso
ciat
ed sy
mpt
oms t
o w
orse
n.
Path
olog
ical
ly sl
ow re
cove
ry p
erio
d, u
sual
ly
≥ 24
h.
- Sle
ep D
ysfu
nctio
ns
- Pai
n: S
igni
fican
t deg
ree
of m
yalg
ia, o
ften
wid
espr
ead
and
mig
rato
ry p
ain
in
mus
cles
/join
ts, o
ften
signi
fican
t hea
dach
es o
f ne
w ty
pe, p
atte
rn o
r sev
erity
. - ≥
2 N
euro
logi
cal/C
ogni
tive
Man
ifest
atio
ns:
e.g.
con
fusio
n, im
paire
d m
emor
y, im
paire
d vi
sion
, ata
xia,
mus
cle
wea
knes
s/fa
scic
ulat
ion,
ph
otop
hobi
a, h
yper
sens
itivi
ty
- ≥1
sym
ptom
from
≥2
of th
e ca
tego
ries
of
auto
nom
ic, n
euro
endo
crin
e or
imm
une
man
ifest
atio
ns: e
.g.:
orth
osta
tic in
tole
ranc
e,
POTS
, the
rmol
abili
ty, b
ladd
er d
ysfu
nctio
n,
tend
er ly
mph
nod
es, r
ecur
rent
flu-
like
sym
ptom
s, ne
w se
nsiti
vitie
s.
Sym
ptom
s mus
t hav
e be
en p
rese
nt
at le
ast h
alf t
he ti
me
and
have
m
oder
ate,
subs
tant
ial,
or se
vere
in
tens
ity:
- Sub
stan
tial
redu
ctio
n/im
pair
men
t in
the
abili
ty
to e
ngag
e in
pre
-illn
ess l
evel
s of
occu
patio
nal,
educ
atio
nal,
soci
al, o
r pe
rson
al a
ctiv
ities
that
per
sist
s for
m
ore
than
6 m
onth
s; is
acco
mpa
nied
by
fatig
ue th
at is
ofte
n pr
ofou
nd; i
s of
new
or
defin
ite o
nset
; is n
ot th
e re
sult
of o
ngoi
ng e
xces
sive
exe
rtion
; an
d is
not
subs
tant
ially
alle
viat
ed b
y re
st
- Pos
t-ex
ertio
nal m
alai
se
- Unr
efre
shin
g sl
eep
- A
t lea
st on
e of
the
follo
win
g:
Cog
nitiv
e im
pair
men
t or
Ort
host
atic
into
lera
nce.
Opt
io-
nal
sym
p-to
ms
Oth
er sy
mpt
oms m
ay
be p
rese
nt, e
.g. m
yal-
gia,
alte
red
moo
d,
and
slee
p di
sturb
ance
(n
ot d
eem
ed e
ssen
tial
to th
e C
FS d
iagn
osis
)
May
supp
ort C
FS/M
E-di
agno
sis:
Gas
troin
test
inal
or g
enito
urin
ary
impa
irmen
ts, s
ore
thro
at, t
ende
r ly
mph
nod
es, s
ensi
tiviti
es. P
ain
was
no
t inc
lude
d in
crit
eria
due
to
insu
ffici
ent p
ublis
hed
data
.
13
Tab
le 1
C
ontin
ued
O
xfor
d co
nsen
sus c
rite
ria
(199
1) (C
FS) [
1]
US
CD
C F
ukud
a C
rite
ria
(199
4) (C
FS) [
2]
Can
adia
n C
onse
nsus
Cri
teri
a (2
003)
(M
E/C
FS) [
4]
Inst
itute
of M
edic
ine
(201
5) (C
FS/M
E) [
7]
Exc
lus-
iona
ry
cond
i-tio
ns
Excl
usio
ns a
re:
- Pat
ient
s with
est
ablis
hed
med
ical
con
ditio
ns k
now
n to
pr
oduc
e ch
roni
c fa
tigue
(e.g
. se
vere
ane
mia
) eve
n if
this
is
diag
nose
d fo
llow
ing
the
susp
ecte
d di
agno
sis o
f CFS
. - S
chiz
ophr
enia
- B
ipol
ar d
isor
der
- Sub
stan
ce m
isus
e
- Eat
ing
diso
rder
- P
rove
n or
gani
c br
ain
dise
ase.
- Chr
onic
med
ical
illn
esse
s ca
usin
g fa
tigue
e.g
. ane
mia
, ch
roni
c he
art f
ailu
re,
auto
imm
une/
infla
mm
ator
y di
seas
es, u
ntre
ated
hy
poth
yroi
dism
, sle
ep a
pnea
, na
rcol
epsy
, pre
viou
sly
treat
ed
mal
igna
ncie
s, an
d un
reso
lved
ca
ses o
f hep
atiti
s B a
nd C
- P
ast o
r cur
rent
psy
chia
tric
excl
usio
ns: M
ajor
dep
ress
ive
diso
rder
with
psy
chos
is or
m
elan
chol
ia, b
ipol
ar d
isor
der,
schi
zoph
reni
a, d
elus
iona
l di
sord
ers o
r dem
entia
- A
nore
xia
nerv
osa
or B
ulim
ia
nerv
osa
- Alc
ohol
/oth
er su
bsta
nce
misu
se
- Sev
ere
obes
ity (B
MI ≥
45)
Cla
rific
atio
ns o
f am
bigu
ities
re
gard
ing
incl
usio
n of
wel
l co
ntro
lled
cond
ition
s suc
h as
di
abet
es a
nd th
yroi
d di
seas
e [3
] an
d ex
clus
ion
of d
epre
ssio
n,
subs
tanc
e m
isus
e, a
nd o
ther
ps
ychi
atric
dis
orde
rs in
the
past
5
year
s hav
e be
en p
ublis
hed.
Act
ive
dise
ase
proc
esse
s tha
t exp
lain
mos
t of
the
maj
or sy
mpt
oms o
f fat
igue
, sle
ep
dist
urba
nce,
pai
n, a
nd c
ogni
tive
dysf
unct
ion.
- A
ddis
on’s
dis
ease
, Cus
hing
’s S
yndr
ome,
hy
poth
yroi
dism
, hyp
erth
yroi
dism
, iro
n de
ficie
ncy,
trea
tabl
e fo
rms o
f ane
mia
, iro
n ov
erlo
ad sy
ndro
me,
dia
bete
s mel
litus
, can
cer;
treat
able
slee
p di
sord
ers,
rheu
mat
olog
ical
di
sord
ers,
lupu
s, po
lym
yosit
is, p
olym
yalg
ia
rheu
mat
ica;
imm
une
diso
rder
s suc
h as
AID
S;
neur
olog
ical
dis
orde
rs su
ch a
s mul
tiple
sc
lero
sis,
park
inso
nism
, mya
sthe
nia
grav
is an
d B
12 d
efic
ienc
y; in
fect
ious
dis
ease
s suc
h as
tube
rcul
osis
, chr
onic
hep
atiti
s, Ly
me
dise
ase;
prim
ary
psyc
hiat
ric d
isor
ders
and
su
bsta
nce
abus
e.
- Exc
lusi
on o
f oth
er d
iagn
oses
, whi
ch c
anno
t be
reas
onab
ly e
xclu
ded
by th
e pa
tient
’s
hist
ory
and
phys
ical
exa
min
atio
n, is
ach
ieve
d by
labo
rato
ry te
sting
and
imag
ing.
- I
f a p
oten
tially
con
foun
ding
med
ical
co
nditi
on is
und
er c
ontro
l, th
en th
e di
agno
sis
of M
E/C
FS c
an b
e en
terta
ined
if p
atie
nts
mee
t the
crit
eria
oth
erw
ise.
Not
spec
ified
; the
com
mitt
ee
note
that
a th
orou
gh m
edic
al
hist
ory,
phy
sica
l ex
amin
atio
n, a
nd ta
rget
ed
wor
k-up
is re
quire
d to
rule
ou
t oth
er d
isor
ders
that
co
uld
caus
e th
e pa
tient
’s
sym
ptom
s.
14
Tab
le 1
C
ontin
ued
O
xfor
d co
nsen
sus c
rite
ria
(199
1) (C
FS) [
1]
US
CD
C F
ukud
a C
rite
ria
(199
4) (C
FS) [
2]
Can
adia
n C
onse
nsus
Cri
teri
a (2
003)
(M
E/C
FS) [
4]
Inst
itute
of M
edic
ine
(201
5) (C
FS/M
E) [
7]
Com
or-
bidi
ties
(not
ne
cess
arily
ex
clu-
sion
ary)
The
follo
win
g ar
e no
t ne
cess
arily
reas
ons f
or
excl
usio
n:
- Dep
ress
ive
illne
ss
- Anx
iety
dis
orde
rs
- Hyp
erve
ntila
tion
synd
rom
e.
- Any
con
ditio
n de
fined
pr
imar
ily b
y sy
mpt
oms t
hat
cann
ot b
e co
nfirm
ed b
y di
agno
stic
labo
rato
ry te
stin
g,
e.g.
Fib
rom
yalg
ia, a
nxie
ty,
som
atof
orm
diso
rder
s, no
n-ps
ycho
tic o
r non
-mel
anch
olic
de
pres
sion
, neu
rast
heni
a,
mul
tiple
che
mic
al se
nsiti
vity
- W
ell-c
ontro
lled
med
ical
co
nditi
ons,
such
as:
H
ypot
hyro
idism
or a
sthm
a
- Con
ditio
ns su
ch a
s Lym
e di
seas
e or
syph
ilis t
hat h
as b
een
treat
ed w
ith d
efin
itive
ther
apy
befo
re d
evel
opm
ent o
f CFS
/ME
- I
sola
ted
unex
plai
ned
phys
ical
or
labo
rato
ry e
xam
inat
ion
findi
ngs t
hat a
re in
suffi
cien
t to
stro
ngly
sugg
est t
he e
xist
ence
of
an
excl
usio
nary
con
ditio
n (e
.g. e
leva
ted
antin
ucle
ar
antib
ody
titer
with
out o
ther
fe
atur
es to
supp
ort a
dia
gnos
is).
Fibr
omya
lgia
Syn
drom
e, M
yofa
scia
l Pai
n Sy
ndro
me,
Tem
poro
man
dibu
lar J
oint
Sy
ndro
me,
Irrit
able
Bow
el/B
ladd
er
Synd
rom
e, In
ters
titia
l Cys
titis
, Ray
naud
’s
phen
omen
on, P
rola
psed
Mitr
al V
alve
, D
epre
ssio
n, M
igra
ine,
Alle
rgie
s, M
ultip
le
Che
mic
al S
ensit
iviti
es, H
ashi
mot
o’s
thyr
oidi
tis, S
icca
Syn
drom
e et
c. S
uch
co-
mor
bidi
ties m
ay o
ccur
in th
e se
tting
of
ME/
CFS
. Oth
ers s
uch
as ir
ritab
le b
owel
sy
ndro
me
may
pre
cede
ME/
CFS
by
man
y ye
ars,
but t
hen
beco
me
asso
ciat
ed w
ith it
. Th
e sa
me
hold
s tru
e fo
r mig
rain
es a
nd
depr
essi
on.
Thei
r ass
ocia
tion
is th
us lo
oser
than
bet
wee
n th
e sy
mpt
oms w
ithin
the
synd
rom
e.
ME/
CFS
and
Fib
rom
yalg
ia o
ften
clos
ely
conn
ect a
nd sh
ould
be
cons
ider
ed
“ove
rlapp
ing
synd
rom
es.”
The
com
mitt
ee d
ecid
ed
agai
nst d
evel
opin
g a
com
preh
ensi
ve li
st of
po
tent
ial c
omor
bid
cond
ition
s. Th
ey a
dvis
e th
at
clin
icia
ns m
ay w
ish
to
cons
ider
: - F
ibro
mya
lgia
- M
yofa
scia
l pai
n sy
ndro
me
- I
nter
stiti
al c
ystit
is
- Irri
tabl
e bl
adde
r syn
drom
e
- Ray
naud
’s p
heno
men
on
- Pro
laps
ed m
itral
val
ve
- Dep
ress
ion
- M
igra
ine
- A
llerg
ies
- Mul
tiple
che
mic
al
sens
itivi
ties
- Sic
ca sy
ndro
me
- O
bstru
ctiv
e or
cen
tral
slee
p ap
nea.
PIFS
- po
st-in
fect
ious
fatig
ue sy
ndro
me
SEID
- sy
stem
ic e
xerti
on in
tole
ranc
e di
seas
e C
DC
- C
ente
rs o
f Dis
ease
Con
trol a
nd P
reve
ntio
n PO
TS -
Post
ural
orth
osta
tic ta
chyc
ardi
a sy
ndro
me
15
There is large overlap between the criteria, but also important differences (Table 1). First, the
way symptoms are defined vary substantially, including whether a given symptom or feature
is mandatory or additional. Next, exclusionary conditions and comorbidities vary among the
criteria, e.g. a primary psychiatric diagnosis. Official guidelines provide further
recommendations for the exclusion of differential diagnoses, e.g. by laboratory tests. Table 2
(in Norwegian) shows the recommended supplementary diagnostic investigations according
to the Norwegian ME/CFS Guidelines. Consequently, both aspects i) and ii) above vary
between criteria. There are no objective diagnostic markers in ME/CFS, and this complicate a
precise definition.
Table 2 Recommended supplementary diagnostic investigations according to the Norwegian ME/CFS Guidelines [9] Anbefalt supplerende utredning
Klinisk-kjemiske blodprøver Hb, SR, hvite m diff. telling, trombocytter, jern, transferrin, transferrinmetning og transferrinreseptor, ferritin, Na, K, Ca, P, Mg, glukose, albumin, CRP, ALAT, ASAT, GT, Bilirubin ALP, LD, kreatinin, CK, vitamin B12, folat, Vit. D, fritt T4, TSH, kortisol.
Immunologiske prøver Immunglobuliner, IgG, IgM, IgA, total IgE, ANA-screening, revmatoid faktor, anti-transglutaminase antistoff.
Mikrobiologiske undersøkelser Serologi: EBV (Epstein-Barr virus), CMV(Cytomegalovirus), VZV (Varicella Zoster virus), HSV (Herpes simplex virus), HIV (voksne), Toxoplasma, Borrelia, Mykoplasma pneumonia*, Chlamydophila pneumonia*, Hepatitt B og C, Parvovirus B19. PCR: Humant herpesvirus 6 og ved positiv serologi: EBV, CMV, parvovirus B19.
*Evt bør det gjøres PCR i nasopharynxaspirat på disse mikrobene
Urinprøve Stiks
Billeddiagnostikk Hos barn og unge er røntgen thorax, ultralyd abdomen og MR caput obligatorisk, hos voksne må det utvises skjønn.
Psykiatrisk Semistrukturert intervju og/eller spørreskjema mhp komorbiditet og differensialdiagnostikk
Psykososialt Kartlegging av belastningsfaktorer mhp sårbarhet og mhp psykosomatiske tilstander, særlig viktig hos barn og unge.
Elektrofysiologiske undersøkelser EEG hos barn og unge
Følgende undersøkelser tas kun på spesielle indikasjoner: Spinalpunksjon, fecaltest, vippetest, søvnregistrering, 24-timers EKG, EEG (voksne), vurdering av psykolog/psykiater samt annen nevropsykologisk testing.
16
1.1.4 Epidemiology and burden for patients and society
The prevalence of ME/CFS is uncertain. Many prevalence estimates exceed 1% [26], but the
use of stricter diagnostic criteria such as CCC provides estimates of 0.1 - 0.2% [27]. The
direct and indirect economic cost is estimated to 18-24 billion dollars annually in the US [28],
using a prevalence estimate of 0.42%. In Norway, this corresponds to about 20,000 patients
and an annual cost of 2.5 billion NOK. Consequently, ME/CFS is now widely accepted as a
serious disease with a substantial burden for patients and society [7, 29, 30].
This is also the case in Norway, with the establishment of a national competence service at
Oslo University Hospital in 2012 and the attention of The Ministry of Health, publishing
national guidelines [9] in 2014. The work of patient organizations, as well as public debate
related to treatment, epidemiology and etiology, have increased the awareness of ME/CFS.
Finally, recent public grants support the increasing awareness that this condition needs to be
better understood (e.g. "BehovME" from The Norwegian Research Council [31]). At present,
many patients suffer under the lack of efficient treatment, the lack of consensus among
professionals and experts, and the lack of knowledge, resources and respect in many parts of
the health service.
1.1.5 Etiology and different disease models
Etiology and pathogenesis in ME/CFS are largely unknown, and are expected to be both
complex and heterogenous [32-35]. Different comprehensive mechanistic models for
ME/CFS have been published [36-38]. Two important underlying disease models are
implicated in this discourse, the biopsychosocial model (BPSM) and the bio-medical model
(BMM).
The BPSM states than no complex disease can be viewed as either biological or
psychological, but can only be understood in an integrated model with biological,
psychological and social factors [39]. Different predisposing ("risk"), precipitating
("triggering"), and perpetuating ("prolonging") factors can be identified [34]. Many
researchers point out that BPSM must apply also for ME/CFS [34, 40], but the usage of this
model has been controversial [41]. Specific models within the BPSM is for example the
cognitive-behavioral model [42], which assumes that fatigue is perpetuated by fatigue-related
behavior and beliefs, or the sustained arousal theory [38].
17
In BMM, the causes of ME/CFS are considered to be primarily biologic in origin. Research
into different organ systems has grown steadily during a few decades. Many specific biologic
disease models have substantial scientific support, some of these summarized briefly below.
Yet, there is no firm evidence supporting a single or comprehensive biologic explanation to
ME/CFS [43, 44].
Specific disease models
Infectious agents
Post-infectious fatigue is a documented sequela for many infectious diseases, e.g. glandular
fever (Epstein-Barr virus, EBV), Q-fever and viral meningitis [45]. Reported surges in
ME/CFS after infectious epidemics [11, 34] point to a post-infectious etiology also for
ME/CFS. Several pathogens have been suspected as the main cause of ME/CFS, e.g. EBV,
giardia lamblia, chlamydia pneumonia, enterovirus, different herpes vira and XMRV virus
[34, 40]. These suspicions have not been supported in larger studies [26, 33]. A recent
example is the virus XMRV a few years ago [46-48]. However, there is evidence for ME/CFS
being triggered by infectious agents such as Giardia Lamblia or EBV [49, 50]. Glandular
fever/EBV is established as a cause of chronic fatigue [45, 51], and has been studied
repeatedly as an important factor in ME/CFS pathogenesis [50, 52, 53]. Elevated levels of
antibodies towards certain parts of EBV antigens [54], and increased detection of EBV DNA
in plasma [55] have recently been reported in ME/CFS. Valacyclovir treatment directed
against EBV have shown promising effect for subsets of CFS patients [56], but seemingly
lack replication.
Endocrinology and metabolism
Endocrinological changes have been reported in ME/CFS, e.g. in the hypothalamic-
pituitaryadrenal axis or in thyroid function, and suspected to be involved in the
pathophysiology [57]. Several studies report metabolic changes in ME/CFS (reviewed in
[58]), e.g. Naviaux et al., who found 20 out of 600 metabolites significantly altered, and
concluded with the presence of cellular hypometabolism [59]. Mitochondrial dysfunction is
repeatedly reported as a central disease mechanism, e.g. in publications investigating cellular
respiration and anaerobic thresholds in ME/CFS patients [24, 60-65]. However, these issues
remain controversial, exemplified by a discourse regarding pyruvate dehydrogenase
deficiency [66].
18
Psychology/psychoneurology
Unexplained Chronic fatigue (e.g. neurasthenia) has been associated with different personality
traits, such as conscientiousness or perfectionism [67, 68]. However, these observations have
not been supported for CFS [69, 70]. There is a higher frequency of stressing events prior to
debut of ME/CFS, according to many studies. Anxiety is common in ME/CFS, and health
anxiety was reported in more than 40% of patients in a recent study [71]. As fatigue is
common in depression and chronic fatigue can lead to depression, the overlapping
characteristics of depression and ME/CFS continue to be discussed [71]. Many researchers
explain associations between ME/CFS and psychological factors by the use of wide criteria
not defining primary psychiatric disease as an exclusion for ME/CFS [40]. The potential role
of psychiatry in the development of ME/CFS is highly controversial [33, 40, 44]. A recent
study found that presence of depressive symptoms are not related to general symptom burden
in ME/CFS [72].
Immunology
Immunological dysregulation is suspected to be involved in ME/CFS pathogenesis, and many
aspects of the immune system have been studied. Changes in several different immune cells
have been reported. However, many of these results have failed to reproduce in independent
cohorts [73]. Among the most persistent findings are changes in Natural Killer (NK)-cell
number and function [74, 75]. Differences in subsets and function of B-cells and T-cells have
also been reported [76-78], e.g. significantly lower numbers of regulatory T-cells [74]. There
are also several reports of significantly altered serum cytokine levels in ME/CFS patients [79-
81]. In the largest study, patients were stratified to short or long duration, and a few cytokines,
such as IL12p40, IL-8 and resistin, were associated with both groups, but with opposite effect
size [79]. This finding, together with the naturally fluctuating and tissue-specific cytokine
levels, illustrate the difficulties with such studies. Nevertheless, some cytokines have been
associated with ME/CFS across at least two of these three studies, such as the pro-
inflammatory IFN-g, IL12p40, CCL11 and CSF1, and the pro- or anti-inflammatory IL-4 and
IL-13. Recently, DNA methylation patterns indicating immune dysregulation have been
reported [82].
Autoimmunity
A related hypothesis states that autoimmunity is important in the pathogenesis of ME/CFS,
and this is recently reviewed by different research groups [36, 37, 83, 84]. Central to this
19
hypothesis is that genetically predisposed individuals, triggered by infectious agents (such as
EBV) acting in concert with a variety of environmental factors, could develop an immune
response towards self-antigens, affecting several organ systems and explaining the wide
variety of symptoms [36]. This could be consistent with the reported aberrations from diverse
systems, such as immunological or metabolic changes as referred above, or changes in brain
physiology [85, 86]. Notably, increased levels of auto-antibodies among ME/CFS patients
have previously been reported, e.g. against cellular anchorage molecules [87] or adrenergic
and cholinergic neurotransmitter receptors [88-90], and the role of anti-nuclear antibodies
have also been discussed [91, 92].
1.1.6 Genetic contribution to etiology
Heritability
Heritability is the fraction of phenotypic variation explained by genetic variation [93]. There
are several methods for assessing heritability, e.g. twin studies, and estimates may vary
substantially for any given trait or disease [94]. Thanks to recent technological development,
large genetic association studies have been able to identify specific genetic factors underlying
heritability. However, genetic association studies have revealed smaller, often much smaller,
explanations to phenotypic variation compared to classic heritability-estimates, often referred
to as "missing heritability" [93].
Heritability in ME/CFS
There are many reports of familial aggregation of ME/CFS [95]. However, as close relatives
share both genetic and environmental factors, familial aggregation does not necessarily imply
high heritability. Results from twin studies in chronic fatigue and CFS vary, but collectively
point towards a mixed cause of inherited and environmental factors [95-97]. In two large
Swedish twin-cohorts, the heritability in females with self-reported CFS-like illness and CFS,
was estimated to 31% and 51% respectively, but with wide confidence intervals including 0%
[97, 98]. Albright et al [99] assessed the heritability in CFS with an alternative approach,
where 811 CFS patients and several sets of matched controls were drawn from a large
population database in Utah, containing detailed relationship information and health records.
Average relatedness among CFS patients was significantly increased compared to controls.
This increase remained significant also after excluding close relationships, which are prone to
introducing bias because of substantially shared environment.
20
Genetic associations
Several genetic associations have been reported in ME/CFS [100-104], e.g. with genes
involved in immunological, endocrinological or neurological function, suggesting specific
disease mechanisms. Two small genome-wide association studies (GWAS, see 1.2.3), with
approximately 40 ME/CFS patients and 40 controls each, identified 65 and 442 associated
single nucleotide polymorphisms (SNPs, see 1.2.3), respectively [100, 101]. Main results
were associations to the glutamatergic neurotransmission gene GRIK2 [101], and missense
SNPs in CLEC4M and CCDC157 as well as four SNPs within T-cell receptor loci [100]. Due
to methodological differences, only 28 SNPs were investigated in both studies, and there were
no common findings. In 2017, data from 16 genetic association studies on ME/CFS were
systematically reviewed, but there was little overlap between the investigated loci, and except
an interesting association to the glucocorticoid receptor gene NR3C1, no repeated results
appeared [102]. Finally, several papers have investigated associations between ME/CFS and
the immunologically important human leukocyte antigen (HLA) genes (see 1.2.6-7) [105-
113]. Table 3 provides an overview. All these studies are small and hence of limited power,
and only selected HLA genes are investigated with relatively low resolution. A few
significant associations are reported, but no common results have appeared. A recent review
concluded that no consistently significant genetic association has so far been identified in
ME/CFS [114].
1.1.7 Treatment
There is no established efficient or curative treatment in ME/CFS [9, 43]. Still, numerous
hypothetical treatment regimens are carried out by health care providers, researchers,
caregivers and patients. Generally, these regimens reflect varying beliefs about underlying
etiology, either they focus on psychological and psychosocial aspects, metabolism, nutrition,
neurology, immunology or antibacterial treatment. Despite repeated support for many of these
treatments, consistent scientific support does generally not exist [43]. On the other hand,
ME/CFS patients' need for supportive treatment is well recognized. Pharmacological
treatment (e.g. for pain and sleep disorders), activity management (also known as pacing),
nursing, nutrition, ergotherapy and physiotherapy should be considered [9, 23, 115-117]. Such
treatment are provided both by personal caregivers, private and public organizations or health
services, although the resources, experience and knowledge has been too scarce and still is
unable to fulfill the patient’s needs in many cases.
21
Table 3 Published papers investigating classic HLA genes in ME/CFS.
Publication Patients / controls
Criteria Loci investigated
Selected results (significant results
in bold) Middleton 1991 [118]
59 / 286 Acute onset post viral fatigue syndrome [2]
HLA-DRB, -DQB, -DQA genotyping
-
Fitzgibbon 1996a - - - Referred to as "negative association with HLA-DR4"
Keller 1994 [105] 110 / 616 Chronic fatigue immune dys-function syn-drome (CFIDS)
HLA DR/DQ serotyping
DR4 RR 1,6 (1,0-2,5) DR5 RR 1,7 (1,1-2,8) DQ3 RR 1.8 (1,2-2.8)
Smith 2005 [108] 49 /102 CFS (Fukuda) HLA-DRB1, -DQA1, -DQB1 genotyping
DQA1*01 OR 1.93 DRB1*03 0,122(pts) vs 0,155(ctrs); DQB1*03:03 0.052(pts) vs 0.050(ctrs); DRB1*07:01 0.12(pts) vs 0.13(ctrs)
Underhill 2001 [106]
58 / 134 CFS (Sharpe 1991 syndrome [2]) "in line with" Fukuda
HLA-DRB, -DQB, -DPB (genotyping). HLA-A, -B and -DR/DQ (serotyping)
DRB1*0301/2 0.17(pts) vs 0.26(ctrs); DQB1*03:03 0.12(pts) vs 0.10(ctrs); DRB1*07:01 0.28(pts and ctrs)
Ortega-Hernandez 2009 [111]
44 / 0 CFS (Fukuda) HLA -DR (serotyping) and -DRB1 (genotyping)
-
Helbig 2003 [107] 23 / 162 Post Q-fever fatigue syndrome
HLA-B and -DR genotyping
DR11 allele frequency 19,6% (pts) vs 4,3% (ctrs)
Carlo-Stella 2009 [110]
46 (31) / 141 (99) CFS (Fukuda) HLA-DRB1 genotyping
DRB1*11:04 OR=0.39 DRB1*13:01 OR=2.79
Pasi 2011 [113] 44 / 50 "Certified diagnosis of CFS"
KIR-genes and HLA-KIR-ligands
-
Spitzer 2010 [112] 74 / aggregated population data
CFS (Fukuda) or fibromyalgia
HLA-DQB1*06:02 DQB1*06:02 43% (pts) vs 7,98% (ctrs)
Ledina 2007 [109] 3 / 0 Post-Q-fever-CFS (Fukuda)
HLA-DR -
a = No access to original paper; pts = patients; ctrs = controls.
22
Cognitive behavioral therapy and graded exercise therapy
Cognitive behavioral therapy and graded exercise therapy have been referred to as the only
proven efficient treatments for ME/CFS by several official authorities [9, 23]. A much cited
reference is the PACE trial, published in 2011 [119]. Patients were ascertained with the
Oxford criteria, where primary depression is not exclusionary [2]. In addition, the PACE trial
have been criticized for a number of methodological flaws [120, 121]. Meta-analyses and
reviews have concluded that cognitive behavioral therapy has only moderate effect in
ME/CFS [115] [122], and the subject is still controversial [43]. Many patients report relapses
and in some cases severe deterioration after such treatment [123]. Central authorities in the
USA and some other countries have therefore withdrawn cognitive behavioral therapy from
proven treatment recommendations for patients defined by stricter criteria [124].
Collaboration between patient and caregiver in the management of pacing and comorbid
symptoms such as sleep-disorders, pain, anxiety and depression is currently a more widely
accepted approach [8, 9, 125].
Immunological treatment
Different immuno-active drugs have been investigated as potential therapy in ME/CFS. Many
patients have tried immunoglobulins or corticosteroids, and several studies have investigated
the effect systematically, but without convincing overall results [115]. Since 2004,
oncologists at Haukeland university Hospital (Bergen, Norway) have treated several patients
with long-standing ME/CFS, who concomitantly developed cancer. After conventional cancer
treatment (including methotrexate, cyclophosphamide or rituximab), some of these patients
reported alleviations of core ME/CFS symptoms [126, 127]. Subsequent research trials
addressed the potential effect of immune-modulating and cytotoxic medication in ME/CFS.
The specific anti-CD20 B-cell depletion agent rituximab showed promising results in pilot
studies [126-129], but in a randomized, double-blinded and placebo-controlled trial with 152
Norwegian patients there was no overall effect [130].
23
1.2 GENETICS AND AUTOIMMUNE DISEASES
1.2.1 Brief introduction to genetic variation
The human genome is carried within each nucleated cell in the human body, in the form of
large DNA-molecules structured into 23 pairs of chromosomes. The information in the
genome, the genetic sequence, consists of the succession of 3 billion nucleotide bases. Each
position in this sequence is held by one of four possible nucleotide bases, creating an
enormous potential for variation. Genes are parts of this genetic sequence that encode
biologically active molecules, typically proteins. There are about 22,000 human protein
coding genes [131], and the protein coding part of the genome constitutes only about 1%. A
genetic locus is a specified position on the human genome, i.e. one or several nucleotide bases
or one or several genes. An allele is one alternative genetic variant of a locus. A chromosomal
pair consists of two homologous chromosomes, one inherited from each parent, which
generally contain the same genes (the major exceptions are the sex chromosomes X and Y).
This means that for every locus (except for the sex chromosomes), the sequence is normally
present in two copies. An individual's genotype is the two genetic sequences of a certain
locus. If the information on both homologous loci is identical, the individual is homozygous;
if it is different, the individual is heterozygous. A haplotype is a succession of alleles on one
single chromosomal strand.
The genome is inherited from parents to offspring mainly unchanged. Importantly, certain
changes to the genome - mutation and recombination, occur at a low frequency in each
generation, giving rise to genetic variation, i.e. differences in the genome between individuals
of the same species. Throughout evolution, genetic variation has been distributed within and
between populations. Independent processes of selection, genetic drift and inbreeding have
led to systematic genetic variation between isolated subpopulations [132].
Human genetic variation is estimated and averaged to 0.1% of the genome, which means that
two randomly picked, unrelated individuals are expected to be 99.9% genetically similar
[133]. This variation is of great importance, however, because it explains a large part of the
different characteristics seen in humans, in a complex interplay with environment. Such
observable characteristics, often called traits, could be for instance hair color, height,
personality or disease status. An individual's phenotype describes the status of one or several
traits.
24
The science of genetics studies the genes, genetic variation and heredity between individuals
and within and between populations. The field of disease etiology studies the cause of
diseases. Genetic epidemiology examines the role of inherited factors in disease development
and etiology [134]. For decades, genetic epidemiologists have investigated the role of genetic
factors in determining health and disease, in families and in populations.
1.2.2 Genetics of complex disease
All diseases depend to a certain extent on both genetics and environment. The well-
established model of monogenic versus complex disease is one way to characterize the
genetic contribution [134]. This model actually simplifies a complicated spectrum, but many
diseases fit well within one of these categories. Monogenic diseases constitute a large group
of conditions where genetic changes in one single gene are the major determinants of disease
status. Thousands of monogenic diseases are known, and most of them are extremely rare
[135]. Most often, monogenic, heritable diseases follow the patterns of classic Mendelian
inheritance, for instance autosomal dominant inheritance.
In contrast, the majority of diseases affecting large number of people (e.g. autoimmune
diseases), are complex diseases. Complex diseases are also called multifactorial diseases
because a multitude of factors, including genetic and environmental, determine disease status
[136]. The entire panel of genetic and environmental factors that reduce or increase the risk
for a given disease is generally unknown. In some cases, important environmental and genetic
factors are well known, e.g. coeliac disease, but we still don't know exactly why and how the
disease develops [137]. The genetic contribution is often substantial also in complex disease,
but the inheritance does not follow the distinct patterns typical for monogenic disease [138].
1.2.3 Studying genetics in complex disease: Case - Control association studies
One important tool for investigating the genetic contribution in complex disease is genetic
association studies. These studies investigate associations between genotype and phenotype.
The concept is based on the idea that if a genetic variant increases disease risk, it should be
more frequent among cases than healthy controls [139]. Association studies for a given
phenotype can be directed against one, a few or many genetic loci, e.g. HLA association
studies focusing on variants in the HLA complex, or genome-wide association studies
25
(GWAS), including often hundreds of thousands of common genetic variants called single
nucleotide polymorphisms (SNPs). Association studies have uncovered genetic risk factors
for numerous complex diseases, and in many cases illuminated novel biologic pathways
underlying disease etiology [134].
1.2.4 Immunogenetics and autoimmune disease
Immunogenetics is the study of how genetic factors are involved in the regulation of the
immune system, whose main task is to protect the organism from invading pathogens or other
disease causing processes [140]. The immune system is of enormous complexity, and
immunogenetics is involved with an impressive number of different diseases and conditions
[141].
Autoimmune disease (AID) is a group of disorders where disease activity is caused by an
immune reaction towards the individual's own organs or tissue. There are more than 50
different established AID [142], and autoimmunity is suspected to play an important role for
several other conditions or disease processes, such as atherosclerosis [143]. Different AID
show great variation in epidemiology, symptomatology, chronicity and severity. Some are
organ specific, while others are systemic. Among the best known AID are rheumatoid
arthritis, celiac disease, type 1 diabetes and psoriasis. The prevalence is estimated to 7 - 9% in
the Western World [142, 144]. Notably, fatigue is a common symptom in AID, sometimes,
but not always, correlated to objective disease activity measurements [145-148].
1.2.5 Autoimmune diseases are complex
Although there are exceptional cases of monogenic AID (such as maturity onset diabetes in
the young or autoimmune polyendocrine syndrome), the most common AID all have complex
etiology, with important contribution from both genes and environment. Many genetic
variants associated with AID are located in or near genes with known immunologic function.
Others are located far from known immune-genes. For the large majority of AID, the
strongest genetic association signals map to a genetic region called the Human Leukocyte
Antigen (HLA) complex, which harbors many immunologically important genes. An example
is coeliac disease, with a well-established and very strong association with genes in the HLA
complex. In addition, the recent 10 years of coeliac disease research have established around
26
50 associated variants in other genetic regions; each of these with very small effect size
compared to the HLA association [149].
1.2.6 The HLA complex and classical HLA genes
The HLA complex, spanning 3.6 Mb on the short arm of chromosome 6 (figure 2), is the most
gene-dense region in the human genome [150, 151], harboring more than 200 genes, with
more than 30% involved in immune regulation.
Figure 2 The HLA complex on chromosome 6, illustrating the physical distance between selected genes; classical HLA genes are pointed out in red. The whole complex contains more than 200 genes (figure from hla.alleles.org) [152, 153].
The most studied genes in the HLA complex are the classical HLA genes: HLA-A, -B, -C,
-DRB1, -DQB1 and -DPB1. These genes encode proteins constituting the immunologically
important HLA molecules, central in regulation of the adaptive immune response. The HLA
molecules' main task is to present peptides to T lymphocytes (or T-cells), which are important
27
regulatory and effector cells of the adaptive immune system [154]. Peptides are small parts of
proteins of self or foreign origin, often parts of invading microbes such as virus or bacteria.
The HLA class I genes HLA-A, -B, and -C are expressed in all nucleated cells, and encode the
HLA class I molecules that present intracellular peptides to CD8 positive T-cells. The HLA
class II genes HLA-DRB1, -DQB1, and -DPB1 are expressed primarily in specialized antigen-
presenting cells, and present exogenous peptides to CD4 positive T-cells (figure 3).
Figure 3 Schematic illustration of HLA-peptide presentation to the T-cell receptor for HLA class I and class II molecules. Figure inspired by [140].
The classic HLA genes are among the most polymorphic in the human genome, i.e. they
present the largest amount of genetic variation, both between individuals and between
populations [151, 154]. The characterization of this extreme polymorphism, which required a
special system for nomenclature (figure 4), has accentuated dramatically in recent years due
to the revolution in genetic HLA typing. To date, more than 27,000 different HLA-alleles
28
have been identified and subsequently catalogued in the official WHO nomenclature
committee for factors of the HLA system [155].
Figure 4 The genetic nomenclature system for specifying an HLA allele (http://hla.alleles.org/nomenclature/naming.html) [152, 153].
1.2.7 HLA associations
The HLA complex is the most disease-associated region in the human genome [150, 151], and
HLA-associations have been reported in a number of diseases in addition to AID, especially
infectious disease, but also in neurologic conditions such as Alzheimer's and Parkinson's
disease, as well as in autism, bipolar disorder and schizophrenia [156-158]. Because of the
extreme polymorphism of the HLA genes, as well as the high gene-density and complex
linkage disequilibrium (LD) patterns in the region, it has remained a great challenge to
pinpoint the primary HLA association signal for different diseases [159]. In some cases, this
has been resolved thanks to common efforts with large cohorts of patients and controls,
comparison between ethnicities and technological development such as high-resolution
genetic HLA typing and high-density SNP-arrays covering the HLA complex [154].
Association signals in the HLA region have been shown to originate from either single alleles
in classic HLA genes, specific genotypes, short or longer haplotype stretches, the combination
of alleles of separate HLA loci, or from non-HLA genes within the region.
29
Notably, HLA associations are particularly common in AID, and have been documented for
most, if not all, AID [141]. Table 10 (page 74) provides examples for certain well established
AID. The effect size for a given HLA risk allele and a specific AID can vary from Odds Ratio
(OR) of several hundreds to less than 2. For some AID, negative HLA associations are also
present, i.e. when OR is significantly less than 1 and the allele confers a protective effect. The
population frequency of the associated allele also varies substantially.
The presence of an HLA risk allele is never a sufficient factor for developing an AID [160].
However, the strong HLA associations seen in many established AID clearly points to a
functional role of HLA molecules in disease mechanism. A potential adverse effect of the
complex peptide-presenting function of HLA molecules (described briefly above), is
inappropriate T-cell recognition of self-antigens leading to autoimmunity, or occurrence of
uncontrolled immune response to exogenous antigens [159]. Because of the great HLA
polymorphism, thousands of different HLA molecules exist, but for each classic HLA gene,
one individual encodes only up to a few different molecules. As different HLA molecules
confer ability to bind different peptides [154], specific alleles may facilitate immune
responses leading to autoimmunity. This mechanism is well documented in coeliac disease,
perhaps the best studied AID, where the responsible antigenic peptide gliadin is identified.
Close molecular match between gliadin and a few HLA-DQ molecules encoded by the
associated DQB1 and DQA1 alleles have been established as fundamental and necessary for
development of the autoimmune reaction [137]. Another example is Goodpasture's diseases,
where negatively associated alleles encoding specific HLA-DR molecules cause disease-
protective changes in CD4+ T-cell repertoire and function [161]. For most AID, however,
detailed molecular mechanisms underlying HLA associations have not been identified [159].
The role of HLA variation in the development of autoimmunity is unquestioned, although the
exact mechanisms are incompletely understood.
1.2.8 Immune modulating treatment in AID
The pathophysiological processes in AID are generally very complex, and only partially
understood. In some AID, like rheumatoid arthritis, systemic lupus erythematosus and
psoriasis, immune modulating treatment has proven therapeutic effect. Immune modulation
can reduce disease activity by breaking some part of the succession of signals maintaining the
inflammatory response. Rheumatoid arthritis is a good example, where several different
30
groups of therapeutic immune modulatory drugs can be used, e.g. cortico-steroids, non-steroid
anti-inflammatory drugs, cytostatic or cytotoxic drugs (such as cyclophosphamide) or disease
modifying anti-rheumatic drugs (specific or non-specific immune modulating drugs, such as
rituximab) [162, 163]. Therapeutic approaches often include drugs from these groups
successively or in combination.
31
2 AIM OF THE PROJECT
Our project is based on the hypothesis that autoimmunity or immune dysregulation is
involved in the pathogenesis of ME/CFS.
• Our first aim was to establish a Norwegian HLA reference panel, with high resolution
genetic HLA data from a representative group of healthy Norwegians (Paper I).
• Our second aim was to perform a high resolution HLA association study in ME/CFS,
with a large Norwegian ME/CFS patient group, diagnosed according to the Canadian
Consensus Criteria (Paper II). A thorough investigation of HLA associations is an
important part of the evaluation of the hypothesis of immune dysregulation in
ME/CFS.
• We also aimed to assess the feasibility, safety and potential effect of intravenous
cyclophosphamide infusions in ME/CFS patients, in an open label intervention study
(Paper III). The immunosuppressive agent cyclophosphamide is an established
alternative treatment of certain autoimmune diseases.
• Finally, we wanted to see whether any identified HLA risk alleles were associated to
clinical subgroups or to therapeutic effect (Paper II and Paper III).
32
3 SUMMARY OF RESULTS
HLA -A, -C, -B, -DRB1, -DQB1 and -DPB1 allele and haplotype
frequencies in 4,514 healthy Norwegians (Paper I)
The HLA genes are highly polymorphic, both within and across populations, and a large
population-specific reference panel is optimal when conducting HLA association studies, one
of the main aims of our projects. Therefore, in Paper I we wanted to sample HLA data from
thousands of healthy Norwegians. The recruitment source was The Norwegian Bone Marrow
Donor Registry (NBMDR [164]), containing the HLA type of 40,318 volunteers (July 2017)
and representing all geographic regions in Norway. In HLA association studies, high
resolution genotyping is desirable to be able to reveal biologically relevant associations.
Therefore, we utilized NGS HLA data recently generated in a subset of 4514 healthy adults
from NBMDR. We assessed allele frequencies at 1st field and G group resolutions for the loci
HLA-A, -C, -B, -DRB1, -DQB1 and -DPB1, and performed haplotype estimation. The most
common G group allele per locus were HLA-A*02:01:01G, HLA-C*07:02:01G, HLA-
B*07:02:01G, HLA-DRB1*15:01:01G, HLA-DQB1*02:01:01G and HLA-DPB1*04:01:01G,
and the number of G group alleles with frequency exceeding 1% was 13, 14, 17, 14, 10 and
11, for the same loci, respectively. The most frequent HLA-A~C~B~DRB1~ DQB1 haplotype
was A*01:01:01G~C*07:01:01G~B*08:01:01G~ DRB1*03:01:01G~DQB1*02:01:01G with
an estimated frequency of 7.88%. Prior to our projects, a large Norwegian HLA-reference
material had not been published. By providing geographically representative, high resolution
HLA data from healthy, ethnic Norwegians, Paper I constitute such a reference panel,
enabling thorough HLA association studies in Norwegian population.
33
Human Leukocyte Antigen alleles associated with Myalgic
Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) (Paper II)
ME/CFS is of unknown etiology, and little is known about the pathogenesis. A central
hypothesis is that immune dysregulation or autoimmunity is involved in the pathogenesis.
Thorough investigation of the highly polymorphic and immunologically important HLA
genes is a central part of the evaluation of this hypothesis, as HLA associations are hallmarks
of autoimmune diseases in general. HLA association studies have been performed in
ME/CFS, but the studies are small, and no common results have appeared. Therefore, the
main objective in Paper II was to perform a large and high-resolution HLA association study
in ME/CFS, including several hundred adult Norwegian patients. Different ME/CFS criteria is
shown to identify different subsets of patients with chronic fatigue, and we sought to
minimize such variation by including patients through the 2003 Canadian Consensus Criteria.
We included 426 ME/CFS patients. The mean age at diagnosis for included ME/CFS patients
was 34.7 years, and the most common (45.5% of patients) disease duration was 5-10 years.
82.8% of the patients were female. 41.1% of patients were bed-or housebound, and 86.8%
were unable to work full or part time the previous 6 months before inclusion. High resolution
NGS HLA genotypes were obtained for all patients and compared with data from healthy,
ethnically matched controls from Paper I. HLA alleles were converted to 2nd field resolution
for all individuals prior to association analyses. We evaluated patterns of linkage
disequilibrium and performed haplotype estimation to evaluate the most significant
association signals. We discovered two independent and positive HLA risk associations,
tagged by the alleles HLA-C*07:04 (OR=2.1, 95%CI[1.4-3.1], pnc=0.0001) and HLA-
DQB1*03:03 (OR=1.5, 95%CI[1.1-2.0], pnc=0.005). These alleles were carried by 7.7% and
12.7% of ME/CFS patients, respectively. Both alleles were significantly associated also after
multiple test correction. Two alleles were negatively associated with ME/CFS, namely
B*08:01 (OR=0.7, pnc=0.01) and DPB1*02:01 (OR=0.7, pnc=0.02), but they did not remain
significant after multiple test correction. ME/CFS patients carrying one or both of the tag
alleles had a significantly higher proportion of comorbid AID (OR=2.3, pnc=0.01). No other
clinical variable was significantly associated with the two risk alleles, including gender,
disease severity and reported incidence of infection or vaccination in advance of symptom
debut. In summary, we detected novel HLA associations pointing toward the involvement of
the immune system in ME/CFS pathogenesis.
34
Intravenous cyclophosphamide in Myalgic Encephalomyelitis/
Chronic Fatigue Syndrome. An open-label phase II study (Paper III)
In ME/CFS, immunosuppressive treatment have been conducted both sporadically and in
clinical trials. So far, clinical efficacy from such treatment has not been scientifically
documented in ME/CFS. Immunosuppression is an established alternative in the treatment of
several autoimmune diseases. Importantly, autoimmunity is not established for ME/CFS, but
the hypothesis of immune dysregulation being involved in the pathogenesis, makes
immunosuppression a potentially important therapeutic option. In patients with long standing
ME/CFS and concurrent malignant disease, there has been case-wise observations of clinical
improvement of core ME/CFS symptoms after cancer treatment including the
immunosuppressant cyclophosphamide. Therefore, in this single center, open-label phase II
trial, our objective was to study the feasibility, safety and potential effect of intravenous
cyclophosphamide infusions in ME/CFS patients. We included 40 adult ME/CFS patients
diagnosed according to the CCC. Each patient received six intravenous infusions of
cyclophosphamide at four-week intervals. Most patients received symptomatic medication
during the study period, but none received alternative treatment aimed at ME/CFS. Disease
activity was regularly monitored with SF-36 and other patient questionnaires (including daily
function, PEM and fatigue) and pedometers for 3 to 4 years. The response variable Fatigue
score was constructed as mean change in four self-reported key symptoms. The overall
response rate was 22 of 40 patients (55.0%), and responders had significant improvement in
Fatigue score (p<0.001). There was also a significant improvement in physical function. 75%
of responders still had clinical response after three years, and responders had further
improvement in physical function compared to 18 months. Four years after inclusion, 68% of
responders still had clinical response. Many patients suffered adverse events such as nausea
and constipation, but there were no hematological toxicity, and very few severe adverse
events. Among these 40 patients, the presence of an HLA risk allele from Paper II correlated
to clinical response. 10 of 12 patients (83.3%) carrying DQB1*03:03 and/or C*07:04 had
clinical response to cyclophosphamide, compared to 12 of 28 patients (42.9%) negative for
these alleles (OR=6.67; p=0.028). Our results suggest that intravenous cyclophosphamide is
tolerable in ME/CFS patients and open label treatment induces prolonged clinical
improvement for many patients, and particularly for carriers of previously identified HLA risk
alleles.
35
4 METHODOLOGICAL CONSIDERATIONS 4.1 STUDY POPULATION Figure 5 shows the study populations in the papers of this thesis. The same healthy controls
were included in Paper I and Paper II. All patients are Norwegian, adult ME/CFS patients,
included in three separate groups. 40 patients from inclusion group i) in Paper II constitute the
study population in Paper III.
Figure 5 Study populations of Paper I, Paper II and Paper III. 4.1.1 Choice of diagnostic criteria
A fundamental criterion in genetic epidemiology is precise and standardized phenotyping
[165]. ME/CFS has a complex definition, is clinically heterogenous, and no objective or
biologic diagnostic marker is established [44, 166]. The use of many different diagnostic
criteria in ME/CFS research have probably increased the heterogeneity among patients, and
made it more difficult to identify underlying pathophysiology [167].
In an attempt to control patient heterogeneity, we chose to apply the 2003 Canadian
Consensus Criteria (CCC), which has a more standardized diagnostic protocol than many
36
other criteria [168]. CCC are relatively detailed and strict, identifying patients with more
severe symptoms than other central criteria [5, 169, 170]. The choice of CCC, recommended
in Norwegian guidelines [9] and therefore widely accepted, also allowed for the inclusion of
several hundred patients, which was important for statistical power. Four patients in Paper II
were diagnosed according to the ICC, which are similarly detailed and strict, and also
associated with more severe symptoms and functional impairment [169].
The 1994 Fukuda criteria [3] are arguably the most widely used criteria in research the recent
decades [169]. Fukuda and CCC are comparable in many aspects, but an important difference
is that PEM is mandatory only in CCC. PEM, which is referred to as a cardinal symptom of
ME/CFS and a suspected manifestation of the pathophysiology [22, 36], was regarded a
crucial inclusion symptom in our studies, and Fukuda was therefore not applicable. The
Oxford criteria [2] is also commonly used in literature, e.g. in papers investigating the effect
of cognitive behavioral treatment [119, 171]. Notably, primary psychiatric disease is an
accepted comorbidity in the Oxford criteria, and an exclusionary condition Fukuda, CCC and
ICC. This is a fundamentally important distinction, as the introduction of patients with
chronic fatigue due to psychiatric disease will increase the heterogeneity among patients and
likely distort the identification of biologic factors. The IOM criteria from 2015 have less
detailed description of mandatory symptoms, aiming only at "core" symptoms of ME/CFS
after a thorough review of relevant literature [7]. However, these criteria do not provide
exclusionary entities, which makes them vulnerable to increased patient heterogeneity.
Our projects concern immunogenetics and potential effect of immune modulation, and
therefore relate to underlying biological factors. Whether CCC identifies an etiologically
more homogenous population of ME/CFS patients than other criteria is controversial [169].
However, clearly defined patient categories is necessary to identify relevant biologic variables
and treatment efficacy in ME/CFS [172], and we therefore considered CCC the most suitable
criteria.
4.1.2 Patient inclusion criteria
Patient heterogeneity may be substantial in ME/CFS, also when applying a single set of
criteria. Despite detailed descriptions in the criteria, the diagnostic process will vary between
patients [22, 35, 172, 173]. The evaluations of mandatory symptoms and exclusionary
37
diagnoses are partly dependent on professional and personal experience, for instance when
assessing the duration and severity of fatigue, the pattern of PEM, the severity of pain, or
when evaluating whether primary psychiatric disease is present. Prior beliefs of both doctor
and patient could also influence the way symptoms are evaluated.
To get sufficient statistical power in Paper II, we aimed at including more than 400 ME/CFS
patients. To achieve this, we needed to include patients from different sources, i.e. three
inclusion groups (figure 5, page 35). We sought to minimize variations in the diagnostic
procedure between these groups, primarily through the use of clinical criteria, as discussed
above. In inclusion group i) all diagnoses were verified clinically by a small group of
physicians and researchers experienced with ME/CFS [126, 130]. In inclusion group ii) all
patients had uniform medical and psychological evaluation, and the diagnoses were made by
physicians and psychologists employed at the outpatient CFS/ME diagnostic center at Oslo
University Hospital. In inclusion group iii) patients contacted us after announcements in
patient networks, and eligible patients reported their diagnostic evaluation through
questionnaires (attachment b), which were reviewed by a single person experienced with the
diagnostic criteria. Table 4 shows the patient inclusion criteria in Paper II and Paper III.
Table 4 Patient inclusion criteria in Paper II and Paper III.
Paper / Inclusion group Inclusion Criteria
Paper II / i) Norwegian, adult ME/CFS patients diagnosed according to the CCC. Age 18–65 years; disease duration > 2 years; and disease severity mild (>5y duration), mild-to-moderate, moderate, moderate-to-severe, or severe. Patients with very severe disease (completely bedbound and in need of help for all basic activities of daily living) were not included.
Paper II / ii) Norwegian, adult ME/CFS patients diagnosed according to the CCC.
Paper II / iii) Norwegian, adult ME/CFS patients diagnosed according to the CCC/ICC.
Paper III Norwegian, adult ME/CFS patients diagnosed according to the CCC. Age 18–66 years; disease duration > 2 years; and disease severity mild-to-moderate, moderate, moderate-to-severe, or severe. Patients with either mild or very severe disease (completely bedbound and in need of help for all basic activities of daily living) were not included.
38
Symptom severity
We wanted to include ME/CFS patients with all grades of severity from very severe
(completely bed-bound) to mild/moderate (a minimum of 50% reduction in activity level).
Table 1, Paper II and Table 1, Paper III shows the severity of included patients. In inclusion
group i) and ii) in Paper II, there were only 4 and 7 patients, respectively, with very severe
disease. This was anticipated in advance, due to the inclusion criteria in inclusion group i),
and the fact that many severely affected patients do not make their way to the outpatient
diagnostic clinic at the CFS/ME Center at Oslo University Hospital (inclusion group ii). Many
severely affected patients are in fact excluded from participation in research. To secure
inclusion of these patients in Paper II, the advertisement to inclusion group iii) were directed
against bedbound and housebound patients (attachment a), and we arranged home visits with
venipuncture for 23 patients. A total of 48 patients in Paper II were bed-bound at the time of
inclusion.
We could have chosen more stringent inclusion criteria, e.g. by focusing on more extreme
phenotypes and excluding patients with mild to moderate disease, or by focusing on
subgroups with more severe objective symptoms, such as POTS or abnormalities in
laboratory tests. However, there are no validated methods for defining subgroups in ME/CFS
with different etiology [167, 169]. In addition, for both Paper II and Paper III, more strict
inclusion criteria would result in the important drawback of a smaller patient group,
weakening the statistical power. Instead, for all patients within each study, we collected a
number of variables such as symptom severity, initiating event, comorbidities, family history
and prior treatment, and observed whether any of these were correlated to the outcome. This
strategy allowed us to identify potential subgroups with stronger association with HLA risk
alleles or better response to treatment.
4.1.3 Comorbidities in patients
Discerning between comorbid and exclusionary diagnoses is an important and difficult task
when diagnosing ME/CFS. CCC and ICC include lists of disorders that must be excluded
(Table 1), and practical recommendations for evaluating differential diagnoses are given in
public guidelines ([9, 23] and Table 2). Both criteria demand explicitly that if an underlying
diagnosis - active or chronic - can explain a substantial part of the symptoms, a diagnosis of
ME/CFS is excluded. Conversely, ME/CFS can be diagnosed if a patient fulfilling the
39
symptom criteria has a potentially confounding medical condition that is under control [4]. An
illustrating example: Hypothyroidism can cause fatigue. Nevertheless, a patient with well-
regulated medical treatment for hypothyroidism can develop new symptoms and fulfill the
criteria for ME/CFS, since hypothyroidism can be excluded as the main cause of new
symptoms.
In our studies, comorbidities among patients were registered through questionnaires based on
DePaul Symptom Questionnaire [5, 174]. Patients in inclusion group i) and ii) (Figure 5, page
35), had a more uniform diagnostic procedure, including evaluation of comorbidities, than
patients in inclusion group iii. Therefore, in inclusion group iii), where 15 patients reported
potentially exclusionary diagnoses, we got the patients' permission to contact the family
doctor or a specialist to get further details. In total, two patients were excluded due to
diagnoses we regarded as incompatible with ME/CFS, namely uncontrolled thyroid disease
and depressive symptoms that had not been evaluated properly. In all remaining cases, we
found no exclusionary comorbidities. ME/CFS patients in group ii) reported less
comorbidities than the two other groups, e.g. clearly less comorbid AID. This may be due to
natural variation between subgroups of patients, but may also represent a systematic bias due
to differences in diagnostic procedures or in questionnaires.
4.1.4 Patient representativeness
Sampling bias, also called selection bias or ascertainment bias, result from studying subjects
that are not representative for the group intended [175]. Based on our sampling strategy, we
aimed at a patient group in Paper II that represents Norwegian, adult ME/CFS patients
identified through CCC. In Paper III, the low number of participants (N=40) makes (random)
sampling bias an important issue. Patients may not be representative for the ME/CFS group in
underlying disease mechanisms, severity, comorbidities etc. In addition, the most severely
and most mildly affected patients were not included in Paper III (table 4). Even if our study
groups do not represent the entire patient population, results regarding biological factors or
treatment effect could be of general value [176].
4.1.5 Representativeness for healthy controls
The 4514 study subjects in Paper I should ideally represent the entire Norwegian healthy
population. These subjects constitute an arbitrarily chosen subgroup of the complete
40
NBMDR. All individuals in NBMDR had filled the blood banks' general inclusion criteria
[177], and were generally healthy individuals at inclusion. Many diagnoses exclude
participation in NBMDR, e.g. cardiovascular disease, cancer and a range of AID such as type
1 diabetes, lupus, Bechterew's, rheumatoid arthritis and inflammatory bowel disease. Some
medical conditions are not exclusionary, however, such as well-controlled eczema, psoriasis,
asthma or allergy, former or current mild to moderate psychiatric disease, symptom-free
coeliac disease following gluten-free diet or well-regulated hypothyroidism. Subjects were
recruited from all Norwegian blood banks nationwide, and the different regions were well
represented, although not exactly proportionally, as discussed in Paper I. Thus, we consider
our control group representative for healthy Norwegians.
A large sample protects against bias from random variation, but systematic bias may still be
present. NBMDR consists of volunteers, who may not represent the general population on a
number of factors, for instance general health status or socioeconomic factors. Notably, the
aim was representativeness in HLA allele frequencies. The HLA complex is the most disease
associated region in the human genome and is associated with an impressive number of
individual diseases but is not known to be associated with gender, age, general health status or
socioeconomic status [178]. Next, NBMDR volunteers may have been motivated by relatives
with specific diagnoses where allogenic bone marrow transplantation constitute a therapeutic
option. This may cause bias, as genetic factors associated with certain diseases may be over-
represented among the controls. However, a large group of different diagnoses could
potentially be treated with bone marrow replacement, from several forms of cancer to
hemoglobinopathies, metabolic diseases and primary immunodeficiencies [179]. Particular
HLA associations may be present for some of these rare conditions, but for the whole
NBMDR, we do not expect systematic bias.
4.2 STUDY DESIGN
4.2.1 Population Stratification
In genetic association studies it is crucial to consider population stratification, the existence of
systematic genetic differences between populations of different ancestry [132]. This is
particularly important in HLA studies, as many HLA allele frequencies vary dramatically
41
between different populations [180]. If genotype is not causally related to phenotype, but
genotype and phenotype are both associated with ancestry (due to unknown or known factors
of genetic or environmental origin), false conclusions may be drawn due to confounding [181,
182]. Even without any association between phenotype and ancestry, selection bias may lead
to population stratification among cases and controls [183]. Without correction for population
stratification, there is a high risk of both false positive and false negative results [132].
Several methods exist to address this problem. For instance, genetic markers known to
identify different ethnicities, can be used to construct comparable strata within an ethnically
mixed study population [184]. Such strategies are costly and time-consuming, and they are
vulnerable to inaccuracies of the previously established population markers [132, 184].
Alternatively, in GWAS studies, with thousands of genetic markers, one can use principle
component analysis to identify genetically diverse subgroups, and adjust for population
stratification [181]. These strategies were not relevant in Paper II, as we chose to thoroughly
investigate classic HLA genes before applying a genome-wide perspective.
Another method to control population stratification is ethnic matching in the recruitment
phase. Since Norwegian population is relatively genetically homogenous [185], case-control
matching by Norwegian ethnicity is both robust and feasible [132], and was therefore the
strategy we chose in Paper II. The control group was drawn from the NBMDR. To avoid bias
from individuals of non-Norwegian ethnicity, we evaluated all the names, and included only
individuals with a Norwegian-sounding name. This evaluation was performed by one person.
The proportion of individuals excluded with this method (5.5%), corresponded well with the
expected amount of non-Norwegian ethnicities in the registry. Some bias could still be
present. First, because individuals with a non-Norwegian ethnicity may have a Norwegian
name, and second, because ethnic Norwegians may have a non-Norwegian name. This
method was the only possibility, however, as i) the registry holds no information about
ethnicity, and we did not have capacity nor consent to verify ethnicities, and ii) no other
genotype data than HLA was available for the controls. In the patient group, all included
individuals were of Norwegian ethnicity, ensured through self-reported country of birth of
patients, parents and grandparents, as well as patients' self-perceived ethnicity. This is one of
the recommended methods for controlling population stratification [132].
42
4.2.2 Relatedness among subjects
Close relatedness between subjects could lead to bias in genetic association studies such as
Paper II [186]. Relatedness between patients could also lead to false associations in Paper III,
between clinical response and the investigated risk loci. Therefore, we sought to avoid close
relatedness in our studies. All eligible participants reported first, second and third degree
relatives with ME/CFS, and only one patient from each extended family was included. This
procedure was performed within and between each of the three inclusion groups in Paper II,
and for all patients in Paper III. In addition, there were no deviations from HWE among
neither patients nor controls (Supplementary table S1, Paper II), indicating no disturbances
due to relatedness.
4.2.3 Case - control matching
Sampling bias may be overcome by matching patients and controls on different variables, and
matching can protect against false positive associations [175]. In Paper II, we performed
matching for the important variables ethnicity and geography. We did not perform age
matching, and a control person in Paper II was in average nine years younger than a patient at
time of inclusion (mean age 30.6 years and 39.6 years, respectively). The fact that genotypes
do not change throughout life renders this difference less problematic in genetic association
studies, but it can still be a relevant bias. Importantly, individuals in the control group could
develop ME/CFS after the point of inclusion, reducing the study's power to detect true
differences. The proportion of future ME/CFS patients in the control group, however, is
probably considerably smaller than the total ME/CFS prevalence of less than 0.5%, as the
disease in many cases presents itself before the age of 30 years [187]. We therefore consider
the effect of this bias minimal.
In Paper II, the proportion of females was higher in the patient group (83%) than in the
control group (59%). Female overrepresentation is an established feature in ME/CFS, and our
figures correspond well with previous literature [43]. Due to the low number of male patients
(N=73) in our material and hence limited power, we did not conduct gender stratification
when initially assessing HLA associations. Generally, HLA association studies do not report
gender differences in association patterns, and there are no known overall associations
between HLA-type and gender. Still, we assessed whether the identified associations could be
caused by gender bias. There were no significant differences in allele frequencies between
43
male and female ME/CFS patients for the two reported risk alleles (C*07:04 p = 0.22,
DQB1*03:03 p = 0.81). Furthermore, there were no significant gender differences in carrier
frequencies of these alleles, neither among cases (Paper II, Table 3) nor controls (C*07:04
had a carrier frequency of 3.7% in females and 4.0% in males; p = 0.6; DQB1*03:03 had a
carrier frequency of 9.0% in females and 8.3% in males; p = 0.4). Finally, we calculated
gender stratified OR's and performed homogeneity testing specifically for these two
associated alleles. Heterogeneity was rejected (p > 0.5), indicating no gender differences for
the HLA associations we observed. Thus, gender differences between patients and controls
should not represent a major bias in Paper II.
We had no access to additional individual data for the control group, and did not conduct
matching for other variables. This may have caused spurious results, for instance if a certain
HLA associated condition was overrepresented among patients (6.3). Ascertainment of a
control group with more individual data could have enabled closer matching, but at the same
time reduced the size of the control group considerably, weakening statistical power.
Importantly, too close matching could also cause "over-matching" and false negative results
[175].
4.2.4 Collecting clinical data and measuring treatment effect
Information bias is present when measurement errors reveal false differences between study
groups. Most of the baseline clinical data were collected through questionnaires.
Questionnaire-based, self-reported data may be affected by information bias, and recall bias in
particular. Patients' memory may be influenced by their expectations or beliefs or by disease
progress, for instance when reporting infections and other precipitating events before the
debut of ME/CFS, the presence of comorbidities and autoimmune diseases in the family.
Notably, this information was not used to compare with controls (for whom we did not have
equivalent data), but only retrospectively within the patient group of each paper, upon
stratification by disease-associated HLA alleles (Paper II) or by treatment response (Paper
III).
In Paper III, there is great potential for information bias, in the recording of disease severity,
treatment effect and adverse events. The primary outcome clinical response in Paper III was
based on the constructed variable Fatigue Score, calculated as the mean of the four fatigue-
44
related symptoms: “Fatigue”, "PEM", "Need for rest" and “Daily functioning”. These data
were based on questionnaires completed by the patients every second week during follow-up
(attachment c). Secondary outcomes were response duration (also based on Fatigue Score),
changes in daily number of steps, self-reported SF-36 score and self-reported function level.
Number of steps registered with pedometer is an objective, but not necessarily optimal
measurement of treatment effect, as clinical improvement can manifest also as increased
ability to read, talk, listen to music or other meaningful activities that ME/CFS patients are
forced to limit due to symptom burden. As there is no objective method for monitoring
disease activity in ME/CFS, all other outcomes were based on subjective data, challenging the
internal validity of Paper III. By using several measurements in parallel, we intended to
improve the validity. The different objective and subjective variables for clinical response
generally correlated well during the follow-up period, indicating that a clinical improvement
really took place among the responders. Another possible bias of intervention trials like Paper
III, is the selection of patients with extreme values of naturally varying variables, such as
severity. If the same variables are used to measure treatment effect, bias by "regression-to-
the-mean" could be introduced [188]. Notably, the results in Paper III do not show a larger
effect for the more severely affected patients, neither for the whole group, nor for the
subgroups with the identified risk alleles from Paper II.
4.2.5 Placebo effect
Paper III is an uncontrolled, open-label intervention study designed to evaluate the potential
therapeutic effect, feasibility and safety of cyclophosphamide in ME/CFS. The gold standard
in intervention studies is a double-blind, randomized, controlled trial [189]. This cannot
always be achieved, however, such as in the initial phases of assessing treatment safety.
Without a control group, the benefits of randomization and blinding is unachievable.
Knowledge of allocation to the treatment group has a great potential for interfering with both
psychological and physical processes, and thereby affecting the outcome [190]. This
fundamentally important phenomenon, often referred to as the placebo effect, is especially
important for subjective outcomes [190], like the majority of measurements of therapeutic
effect in Paper III. The beneficial treatment effect of cyclophosphamide for ME/CFS patients
in our study may, partly or fully, be due to the placebo effect.
45
4.3 HLA-TYPING
4.3.1 Brief history and our choice of HLA typing
Because of the great importance of HLA in transplantation medicine and immunology,
laboratory techniques for identifying different variants of HLA-molecules or HLA-alleles
have a long history. From the early 1970s, serologic HLA typing revealed specific HLA
associations in autoimmune diseases such as ankylosing spondylitis, psoriasis and coeliac
disease [151]. Serologic typing, using antibodies to identify different molecules [191], can
identify several different alleles of each locus, but has low resolution compared to genetic
typing [192]. Genetic HLA-typing evolved from the 1990s, with sequence specific
oligonucleotide (SSO) and sequence specific primer (SSP) technologies [192] as well as
Sanger sequencing [193]. These techniques can provide low to high resolution HLA typing,
but are relatively low-throughput, time consuming and expensive [192, 193]. In many cases,
these procedures result in unphased data and ambiguities [194]. Due to the rapid development
in genetic sequencing the last decades, modern sequencing techniques, often called Next
Generation Sequencing (NGS), allow for dramatically faster and more accurate HLA-typing
[150, 194, 195]. NGS provides the highest resolution typing available, and is considered the
gold standard in high resolution HLA-typing [195].
We made several choices necessitating the use of HLA typing by NGS. First, we wanted
statistical power to detect HLA-associations of moderate to low effect size, and therefore had
to include several hundreds of patients. Second, we wanted to include all the classic HLA-
genes, both class I and class II. Third, we wanted high resolution to accurately identify alleles
encoding specific HLA molecules, and to improve allele identification and phasing. Only
NGS could offer affordable, high resolution, multi-locus HLA typing for hundreds of patients
in an acceptable amount of time.
An alternative could have been imputed HLA typing through GWAS arrays. Although the
total number of markers may exceed 1 million in different GWAS kits, most kits have too few
markers in the HLA region to capture the extensive polymorphism of the region [159].
Targeted arrays, like Immunochip (Illumina, San Diego, US), with more than 7000 markers in
the HLA region, can be used with designated software and ethnicity-relevant reference panels
to impute HLA alleles of fairly high resolution [154, 159]. The imputation may often be
biased toward certain alleles, which would not be optimal for ME/CFS, where HLA
46
associations have been scarcely investigated. Another alternative could have been so-called
"third generation" NGS HLA typing [150]. These procedures offer exceptionally long reads,
and phasing and haplotyping could be improved. However, these methods have a higher error
rate than our chosen methods, and are also more expensive [150, 196].
4.3.2 Next Generation Sequencing (NGS) procedures in HLA typing
Genetic sequencing of the HLA region is particularly challenging, compared to other genomic
regions, because of the extreme genetic variation [150, 195]. Several different methods and
platforms for HLA sequencing by NGS have been developed. In brief, most current
procedures consist of three steps: i) targeting the desired genetic region, either the complete
HLA region, selected full-length genes or specific exons, ii) the preparation and sequencing
of these target regions, and iii) the assignment of alleles [194].
In our projects, high resolution HLA genotyping by NGS was performed at the Department of
Medical Genetics, Oslo University Hospital for all included ME/CFS patients, and at
Histogenetics (New York, United States) for the control group. At Oslo University Hospital,
HLA genes were amplified (i) with long range PCR (LR-PCR), covering exonic and intronic
sequence of each gene. This procedure requires high DNA quality, and is time consuming as
the ensuing preparation steps are elaborate. Alternatively, amplification of exons only is
relatively quicker and easier [150]. LR-PCR is often the preferred method, however, because
it conveys far more sequencing information, enabling better phasing and allele identification
[195].
The sequencing (ii) was performed on Illumina's MiSeq platform with 2 x 150 basepairs (bp)
paired-end sequencing. There are several alternative platforms, but MiSeq is considered a
very high quality and cost-efficient alternative in high resolution NGS HLA typing [150,
197]. Notably, paired-end sequencing has the advantage of making phasing easier [150], as
two terminal reads from one long fragment can effectively "anchor" two distant
polymorphisms [195]. Allele assignment (iii) was performed in NGSengine (GenDX, Utrecht,
The Netherlands) by mapping sequencing reads to the appropriate position in the appropriate
gene according to the IMGT/HLA Database [198], containing all catalogued HLA alleles. The
resolution was generally 3rd field or higher.
47
At Histogenetics, exon-specific amplification (i) was carried out for at least exon 2 and 3 in
HLA Class I, and at least exon 2 in HLA Class II [196]. Sequencing (ii) was performed with
Sanger technology and NGS applying Illumina MiSeq. Overlapping amplicons and
supplementary sequencing techniques were applied to solve phasing issues [196]. Alleles
were identified (iii) at G group resolution, discriminating alleles with nucleotide sequence
differences in the aforementioned exons [152, 153]. Histogenetics has performed these
procedures in thousands of potential bone marrow donors [196], and exon-specific
amplification may have been the only affordable procedure [199]. By using HLA data already
provided by Histogenetics, we could include a large control cohort and get increased
statistical power. Alternatively, due to limitations of both time and money, performing full-
length HLA typing also for controls would have led to a considerably smaller reference
group.
Thus, the control group HLA data had lower resolution than the patient data, and this may
represent a drawback. Nevertheless, HLA typing by NGS at G group resolution is certainly
high resolution [196]. The exons targeted at G group resolution typing (exon 2 and 3 in HLA
Class I; exon 2 in HLA Class II) harbor most of the genetic variation in HLA Class I and II
genes, and they encode the peptide-binding groove on the HLA molecules [160, 200] (Figure
6). Importantly, most detected HLA associations are based on amino acid differences in the
peptide-binding groove [160, 201]. Such differences correspond to the P group nomenclature
[202], where alleles encoding the same amino acid sequence in the peptide-binding groove
are grouped together. In G group nomenclature, alleles with identical nucleotide sequence in
the exons encoding the peptide-binding groove are grouped together. Consequently, alleles in
the same G group also belong in the same P group (except for null-alleles). Thus, in both
patients and controls, the resolution level was greater than P group, and enabled us to identify
HLA alleles of biologic relevance.
Because of these differences in typing resolution, HLA-data from patients and controls were
not immediately comparable. HLA genotypes in the patient group needed conversion to G
group level to be comparable to the control group. First, alleles were trimmed to second field
resolution, e.g. HLA-DQB1*02:02:03 were trimmed to HLA-DQB1*02:02. Second, all
identified (and trimmed) alleles were compared to a database for G group nomenclature
(http://hla.alleles.org/alleles/text_index.html) [152, 153], and alleles belonging to a G group
with different second field digits (i.e. missense variants outside the exons encoding the
48
Figure 6 Peptide-binding groove of HLA class I and class II molecules. The figures are copied from https://commons.wikimedia.org according to CC BY 2.5 (https://creativecommons.org/licenses/by/2.5/)
peptide-binding groove), were converted to match the second field name of that G group. For
example, HLA-DQB1*02:02 were converted to HLA-DQB1*02:01, because it belongs to the
G group HLA-DQB1*02:01:01G. In total, 91 such conversions were performed for 11
different alleles (Table S5, Paper II). Without these conversions, false associations would
likely have occurred. In summary, the investigation of HLA associations in Paper II should
not be influenced by information bias since both patients and controls were subjected to
similar high resolution HLA typing by NGS.
4.3.3 Ambiguities in allele assignment
In HLA typing, an ambiguity is present when more than two unique alleles per locus is
compatible with the genetic data. Ambiguities can take any of two forms [203]: i) Allelic
ambiguity, when a sequence is compatible with at least two different published alleles, often
For HLA class I, the peptide-binding groove is formed by protein domain !1 and !2, encoded by exon 2 and 3.
For HLA class II, the peptide-binding groove is formed by protein domain !1 and "1,encoded by exon 2 in !- and "-chain genes, respectively.
Figures by User atropos235 on en.wikipedia - Own work, CC BY 2.5,https://commons.wikimedia.org/w/index.php?curid=1805424, https://commons.wikimedia.org/w/index.php?curid=1805483
49
because of uncovered regions, and ii) Genotype ambiguity, when phasing is uncertain due to
polymorphisms on separate amplicons and/or lack of unique overlap between mapped reads.
In the patient group, we accepted no exonic mismatches, and required full coverage of exons
encoding the antigen-binding groove. Hence there should be no allelic ambiguities on G or P
group resolution. Table 5 lists all second and third field ambiguities encountered in the patient
group. In each case, we verified that ambiguous alleles belonged to the same G group, before
performing the conversion described above. There were very few genotype ambiguities, and
when encountered we assigned the most frequent allele or genotype in the population. In the
control group, ambiguities were resolved either with SSP or with NGS sequencing on PacBio,
a "third generation" NGS HLA sequencing procedure that generates very long reads [196].
Table S1, Paper I shows the handling of genotype ambiguities in the control group. In brief,
we accomplished confident allele identification for all individuals for all of the genes HLA -
A, -B, -C, -DRB1, -DQB1 and -DPB1.
4.3.4 Quality control
Before investigating genetic associations, quality control (QC) steps are essential to ensure
high quality genetic data. Some of the important parameters are DNA concentration and
quality (pre- and post-amplification), NGS library quality measurements, sequencing
coverage and depth, genotyping success and ambiguities in allele assignment. At Oslo
University Hospital, DNA concentration and quality were measured at several steps to ensure
accordance with the manufacturer's recommendations. Following NGS, bioinformatic QC
steps were performed at the sequencing facility (NSC, Oslo), to ensure the quality of the data
files containing the sequenced reads. Next, in the mapping procedure, we required
mappability > 80% (the proportion of reads that can be uniquely aligned). Median read depth
was above 150 reads per base for all patients, and we required the heterozygosity balance to
be within 70/30. In total, 2nd field resolution genotyping success exceeded 99% in the patient
group for HLA -A, -B, -C, -DRB1, -DQB1 and -DPB1. At Histogenetics (sequencing of the
control group) QC steps were performed according to their published procedures [196].
Genotyping success exceeded 99.9% for the six aforementioned loci in the control group.
HLA-DRB3, -DRB4 and -DRB5 were not sequenced in the control group and only partially
covered in the patient group.
50
Table 5 Ambiguities encountered in the patient group Allelic ambiguities encountered in the patient group Used in analyses B*13:02:01/B*13:02:05/B*13:02:09/B*13:02:16/B*13:02:20/ B*13:02:21/B*13:38/B*13:69/B*13:96/B*13:99 B*13:02 B*15:01:01/B*15:01:06/B*15:01:07/B*15:01:20/B*15:01:22/ B*15:102/B*15:104/B*15:140/B*15:146/B*15:201/ B*15:227/B*15:228/B*15:247/B*15:320/B*15:321Q B*15:01
DRB1*08:01:01/DRB1*08:77 DRB1*08:01
DRB1*07:01:01/DRB1*07:79 DRB1*07:01
DRB1*01:02:01/DRB1*01:83 DRB1*01:02
DRB1*04:07:01/DRB1*04:92 DRB1*04:07
DRB1*12:01:01/DRB1*12:10 DRB1*12:01
DRB1*09:01:02/DRB1*09:21/DRB1*09:31 DRB1*09:01
DQB1*02:02:01/DQB1*02:97 DQB1*02:01
DQB1*05:03:01/DQB1*05:149 DQB1*05:03
DPB1*13:01:01/DPB1*107:01 DPB1*13:01
DPB1*05:01:01/DPB1*135:01 DPB1*05:01
DPB1*13:01:01/DPB1*107:01/DPB1*519:01 DPB1*13:01
DPB1*04:01:01/DPB1*126:01:01 DPB1*04:01
DPB1*04:02:01/DPB1*105:01:01 DPB1*04:02
DPB1*03:01:01/DPB1*351:01 DPB1*03:01
DPB1*04:02:01/DPB1*463:01:01 DPB1*04:02
DPB1*02:01:02/DPB1*416:01:01 DPB1*02:01
DPB1*04:02:01/DPB1*665:01 DPB1*04:02
DPB1*01:01:01/DPB1*417:01 DPB1*01:01 DPB1*03:01:01/DPB1*03:01:08/DPB1*104:01:01/DPB1*124:01:01/DPB1*351:01/DPB1*669:01/DPB1*675:01/DPB1*676:01 DPB1*03:01 DPB1*04:02:01/DPB1*105:01:01/DPB1*04:02:09/DPB1*04:02:10/DPB1*463:01:01/DPB1*571:01/DPB1*647:01/DPB1*665:01 DPB1*04:02
DPB1*11:01:01/DPB1*654:01 DPB1*11:01
DPB1*17:01:01/DPB1*460:01 DPB1*17:01
DPB1*105:01:01/DPB1*665:01 DPB1*04:02
DPB1*107:01/DPB1*13:01:01 DPB1*13:01
Genotype ambiguities encountered in the patient group Used in analyses DPB1*03:01:01/DPB1*104:01:01/DPB1*14:01:01 + DPB1*10:01:01/DPB1*650:01/DPB1*37:01 DPB1*03:01 + DPB1*10:01 DPB1*03:01:01/DPB1*351:01/DPB1*57:01 + DPB1*04:02:01/DPB1*463:01:01/DPB1*271:01 DPB1*03:01 + DPB1*04:02
HLA sequencing may fail in detecting null alleles (alleles that are not expressed), because
nonsense or frameshift variants may reside in exons that are not covered. This can be avoided
by combining techniques [192], alternatively by applying high resolution genotyping of whole
genes, typically accomplished by long-range PCR based NGS protocols [195]. Such a
protocol was only performed in the patient group in our studies, where we detected one single
null allele (HLA-C*04:09N), not far from the expected number [204]. In Histogenetics'
51
procedure, when encountering alleles potentially representing Common and Well Defined
(CWD) null alleles, additional exons were sequenced to resolve the issue [196]. Other null
alleles are believed to be very uncommon [205] and therefore should not represent a major
concern in our associations studies. Apparent homozygosity for a certain allele, could in some
cases actually be due to "allelic drop out", i.e. when sequence from one of the alleles are not
amplified, often due to sequence incompatibility with the primers. When encountering
homozygosity, we evaluated the coverage to detect clear drops in sequencing depth,
indicating allelic dropout, but found no obvious drops.
4.4 HAPLOTYPES AND LINKAGE DISEQUILIBRIUM
4.4.1 Haplotype frequency estimation
Extended haplotype information is generally not achieved in genetic sequencing. NGS of
classic HLA genes by use of LR-PCR usually enables good phasing and little ambiguity in
allele assignment for each gene. But it does not resolve phasing between the genes [160]. For
instance, having identified heterozygous alleles (genotypes) for each of the three loci HLA-A,
-C, and -B, the haplotypes could exist as any of 4 combinations (figure 7).
Figure 7 Schematic illustration of haplotypes. With the identified heterozygous genotypes over three loci, there are 4 alternative haplotype combinations, and a total of eight possible haplotypes.
Locus:HLA-A HLA-CHLA-B
Genotype01:01 / 02:0102:01 / 07:0107:01 / 08:01
Possible haplotypes: HLA-A HLA-C HLA-B
01:01
02:01
02:01
07:01
07:01
08:01Chromosomal
strands
01:01
02:01
07:01
02:01
07:01
08:01
01:01
02:01
02:01
07:01
08:01
07:01
01:01
02:01
07:01
02:01
08:01
07:01
Alternative 1
Alternative 4
Alternative 3
Alternative 2
52
Genetic associations on allelic level can be addressed without haplotype information. In
genetic association studies, GWAS in particular, the associated genetic markers are often not
biologically significant themselves, but rather markers for causative variants in adjacent loci,
sometimes called a "hitch-hiking" effect [160]. In HLA association studies, haplotype
information is important to decide whether adjacent association signals are independent or
"hitch-hiking", or whether the effect is present only for a combination of alleles on the same
haplotype [159]. Notably, haplotype information is necessary when calculating the strength of
linkage disequilibrium (see 4.4.2). Finally, it is possible that the biologic effect of alleles at
different loci are dependent on the cis or trans position of the alleles, that is, if they are on the
same haplotype or not [141].
Haplotype estimation refers to the process of statistical estimation of haplotypes from
genotype data. There are several methods and computer programs offering HLA haplotype
estimation. We performed haplotype estimation in Pypop 0.7.0 [206] and Arlequin 3.5 [207].
The procedures in both programs are based on a Maximum-likelihood and Expectation-
Maximization (EM) algorithm [208, 209]. The input for these methods is a population with
unphased HLA genotypes. In the first step, initial arbitrary haplotype frequencies are defined.
The second step provides estimated frequencies of multilocus genotypes (the particular
combination of two haplotypes) in the population, under the assumption of HWE. These
estimations are iteratively compared to the population genotype data, and in each step the
haplotype assignments are changed. Finally, when the estimated multilocus genotypes reach a
local maximum in its compatibility with the data, the haplotype estimation is complete.
Settings can be used to define a local maximum, and to increase the number of independent
starting points. The latter is important, to avoid local maxima that deviate from the best
possible fit. HLA haplotype estimation by EM algorithms are widely accepted as robust
methods, as long as the population is large enough, and the individuals are unrelated and
ethnically homogenous [209-211].
4.4.2 Linkage disequilibrium (LD) - concept and methods for calculation
As stated by Slatkin [211], "Linkage disequilibrium is one of those unfortunate terms that
does not reveal its meaning". The most common definition of LD is the nonrandom
association of alleles at different loci. The basic mathematical definition of LD is:
DAB = pAB - pApB
53
where DAB quantifies the LD between allele A and B of two different loci, pAB is the observed
haplotype frequency and pApB is the expected haplotype frequency. In other words, there is
LD between the alleles at the two loci when the observed haplotype frequency differ from
what is expected given independence of the alleles. Independence of the two alleles means
that they are distributed in the population independent of each other, and this is called linkage
equilibrium. The expected haplotype frequency, under the assumption of linkage equilibrium,
is simply the product of the allele frequencies. In many cases, there is no precise information
about the observed haplotype frequency, and estimated frequencies are used instead (see
4.4.1).
LD exists because our genome is inherited in large chromosomal chunks; recombination by
chromosomal crossover is a rare event - only 2-3 in average per chromosome per generation.
Hence, LD is often defined as an association or correlation between two or more linked loci.
However LD may not be due to linkage [151], but to an evolutionary benefit of keeping
alleles of separate loci together [160].
Other measures are derived from D to express LD relative to its maximum value (D' and r2)
and overall LD between multiallelic loci (multiallelic D' and Wn) (figure 8).
D’= %
%&'(D)*+=,
-./[1213, (1 − 12)(1 − 13)]whenD < 0-./[12(1 − 13), (1 − 12)13]whenD > 0
r2= CD
EF(GHEF)EI(GHEI)
D’*=∑ ∑ 1LMNOP′LNO
RNSG
TLSG
Wn*=V∑ ∑ CWX
D /EWEXZX[\
]W[\
^L_(THG,RHG)`G/a
Figure 8 Different measures for Linkage Disequilibrium. * OverallLDbetweentwolociIandJwithmultiplealleles[206,212].
D' is a unidirectional LD measure, and will reach its maximum value 1 if an allele a of locus i
always occurs on haplotype with allele b of locus j, even if the converse is not true (allele b
may be present also on other haplotypes). r2 is a bidirectional LD measure, and will reach its
54
maximum value 1 only if there is perfect LD, meaning that allele a of locus i and allele b of
locus j has the same allele frequency and always occur on the same haplotype. In Paper I, we
chose to express multiallelic LD as Wn, derived from r2, as this is considered a better measure
for highly polymorphic loci [210]. D' is suitable when assessing LD between alleles of
unequal frequencies, and was therefore used to evaluate the independence of associated
alleles in Paper II. Calculation of different LD-measures were performed in the programs
Arlequin, Unphased and PYPOP. As an illustrative overview, table 6 shows LD calculations
for selected allele-pairs.
Table 6 Manually calculated LD measures for selected pair of alleles. (pX = frequency of allele or haplotype X; HFest = estimated haplotype frequency)
Study group
N Allele "A"
Allele "B" pA pB pAB
(HFest) pApB
(1-pA) *(1-pB)
pA* (1-pB)
(1-pA) *pB
D= pAB - pApB
D´ r 2
Patients 426 C*07:04 B*44:02 0.039 0.123 0.035 0.005 0.843 0.034 0.118 0.030 0.90 0.23
Controls 4511 C*07:04 B*44:02 0.019 0.100 0.017 0.002 0.883 0.017 0.098 0.015 0.89 0.14
Patients 426 C*07:04 DQB1*03:03 0.039 0.066 0.005 0.003 0.898 0.036 0.063 0.002 0.06 0.00
Controls 4509 C*07:04 DQB1*03:03 0.019 0.045 0.001 0.001 0.937 0.018 0.044 0.000 -0.35 0.00
Patients 426 B*44:02 DQB1*03:03 0.123 0.066 0.005 0.008 0.819 0.115 0.058 -0.003 -0.34 0.00
Controls 4509 B*44:02 DQB1*03:03 0.100 0.045 0.001 0.005 0.859 0.096 0.041 -0.004 -0.85 0.00
Patients 426 B*57:01 DQB1*03:03 0.046 0.066 0.033 0.003 0.892 0.043 0.063 0.030 0.69 0.33
Controls 4509 B*57:01 DQB1*03:03 0.029 0.045 0.020 0.001 0.928 0.027 0.044 0.019 0.68 0.29
4.5 STATISTICAL ISSUES
4.5.1 Power calculations and significance thresholds - type I and type II errors
The prevailing method in quantitative research is the Null-Hypothesis-Based Significance
Testing (NHST) [213, 214]. The strategy is as follows:
1) We assume a null-hypothesis (H0) and define an opposite, alternative hypothesis (Halt)
2) We conduct our experiments
3) We evaluate the probability of the results under the assumption of H0
4) We reject H0 in favor of Halt if this probability is below a certain threshold
55
Statistical power is a study's ability to detect true effects. Based on NHST, power is defined
as the probability to reject H0 if Halt is true. In association studies, power depend on sample
size, effect size, alpha level and frequency of the investigated factor. With insufficient
statistical power, there is a high risk of false negative results, so called type II errors. This
relationship is given by the formula:
P Type II Errors = 1 - Power
Given the heterogeneity in ME/CFS combined with the likely multifactorial etiology, genetic
risk variants may be assumed to have moderate or small effect size. Therefore, sufficient
power is important. In Paper II, with an uncorrected alpha level of 0.05, we had 80% power to
discover HLA-associations with OR down to 1.75 (given AF > 0.05) and 1.5 (given AF > 0.1)
(Figure 9). For larger effect sizes power will be greater, but we cannot rule out the possibility
of false negative results.
Figure 9 The Power to detect associations as a function of the true OR for given allele frequencies in the study population of 426 patients and 4511 controls in Paper II, provided an alpha level of 0.05. In the NHST strategy summarized above, the threshold in 4) is the fundamentally important
alpha level (or significance level), and is equal to the risk of false positive findings, so called
type I errors. In all papers within this thesis, we have utilized established statistical procedures
within the fields of immunogenetics and clinical intervention. Generally, alpha levels were set
to 0.05, by far the most common level [214].
4.5.2 Multiple test correction
Multiple test correction is an important issue in modern genetic research. As a result of the
rapidly increasing capacity of different methods in molecular biology, multiple hypotheses
56
can be investigated in a single research project. With increasing number of comparisons, the
likelihood of a significant result appearing by chance alone also increases, even if none of the
null hypotheses were true. For this reason, multiple test correction adjusts the p-values or the
alpha levels according to the number of comparisons made in the project to avoid increasing
the rate of false positive findings. There are several different methods, e.g. the Bonferroni
method [215] and the Benjamini-Hochberg method [216].
In Paper II, we conduct multiple comparisons as we compare the frequencies of many
different alleles over several loci, and we choose to apply locus-wise Bonferroni multiple test
correction. Bonferroni is a strict multiple test correction, considering all independent
comparisons being made. Locus-wise correction is often used in HLA association studies
since alleles of neighboring loci typically occur together on specific haplotypes (high LD).
Alleles at different loci are therefore not completely independent, and therefore do not
represent independent statistical tests. Since an allele-wise (multi-locus) Bonferroni multiple
test correction does not take into account the strong LD between the classic HLA loci, the
result could be over-correction leading to type II errors - false negative results. However, with
locus-wise correction, there is the contrasting risk of type I errors, because LD - although
particularly strong - is not complete even in the HLA region. The correction was based on the
number of alleles per locus with a frequency above 0.03 in either patients or controls, to avoid
over-correction due to the many infrequent alleles.
4.5.3 Statistical methods in stratification analyses
In Paper II, we stratified the patients based on presence of risk alleles, and investigated eight
dichotomous clinical variables (related to initiating events, comorbidities, family history and
gender) with OR and chi-square significance testing (Paper II, table 3). After data collection
in Paper III, treatment effect was stratified by gender, severity, previous treatment and
genetically by the presence of either of the two risk alleles from Paper II, and associations
were compared using OR (Paper III, figures 2, 4 and 6). For genetic stratification, P-values
were calculated with Fisher's exact tests, due to the low number of individuals in some of the
contingency table cells [217]. In both papers, since stratification analyses were only
performed for a limited number of pre-selected variables, we did not perform multiple test
correction.
57
5 ETHICAL CONSIDERATIONS In this PhD-project, we follow all relevant regulations regarding scientific and ethical
standards, including patient and data security. Our studies are appreciated by Regional
Committees for Medical and Health Research Ethics (REK numbers 2015/1547 and
2014/1672), the privacy department ("Personvernombudet") at Oslo University Hospital for
Paper II, and the National Medicines Agency in Norway for Paper III. Paper III was pre-
registered as a clinical trial (EudraCT no. 2014-004029-41, www.clinicaltrials.gov
NCT02444091). Patient representatives were involved in the planning phase, e.g. related to
inclusion criteria and patient questionnaires.
The research's potential benefits must be balanced against any burden the research may inflict
on patients and families. For instance, since PEM is a central feature in ME/CFS, many
patients will suffer clinical deterioration after completing detailed questionnaires or after
venipuncture. All patients must be properly informed before consenting to participation.
When collecting blood samples from severely sick patients in their homes or at institutions,
we carefully attempted to disturb and provoke as little as possible, to minimize severe
deterioration.
Especially, treatment with potentially harmful drugs, such as cyclophosphamide in Paper III,
must be carefully considered. Cyclophosphamide treatment commonly induces adverse
effects such as nausea, constipation, hematuria or infections. After the initial and unintended
observations by Fluge, Mella and colleagues, of a possible clinical effect of
cyclophosphamide in ME/CFS, a thorough investigation of the safety and effects of such
treatment was truly scientifically warranted. Such investigation may only be performed with
full informed consent, a carefully planned and approved protocol, close surveillance of
patients during and after administration, as well as supervision by professionals familiar with
cyclophosphamide treatment and the handling of adverse effects. All these factors being
present for Paper III, combined with the clinical severity for many ME/CFS patients, and the
lack of knowledge, support and treatment, indicate an ethical justification of our research.
Researchers should also be aware that patients' expectations for improved treatment or a
better understanding of disease mechanisms may in some cases be unrealistic. Patients may
58
have wanted to receive individual HLA-data, and we informed them that such data will not be
made available, and we included this point in the written consent form.
ME/CFS is a disease of unknown cause, with various and partially conflicting disease models,
and with no established treatment [44]. The diastasis between biologic and mental causes is
fundamental in this discourse, and has possibly led to more suffering for patients [218, 219].
Disease burden is substantial in ME/CFS, and many patients are suffering from severe
symptoms. Patients often experience lack of knowledge and understanding from health care
providers. Research into biological aspects of CFS/ME is essential for a better understanding
of this complex disease, better and more precise diagnostics, and potentially more effective
treatment or prevention. On the other hand, increasing focus on biologic etiology could also
be said to decrease the focus on other issues, such as psychosocial etiology or cognitive and
psychological treatment, which may also be important. Therefore, in ME/CFS research, we
are particularly obliged to scientific thoroughness, including a cautious and sober
interpretation of results. In our experience, most patients and their relatives are well aware of
the complexity of ME/CFS, and they can handle the uncertainty resulting from different
hypotheses.
59
6 DISCUSSION 6.1 ARE HLA ASSOCIATIONS PRESENT IN ME/CFS?
6.1.1 Positive associations from Paper II compared to previous literature
Previous HLA association studies in ME/CFS are small (N≤110 patients), with great
variation in patient inclusion criteria and limited genotyping resolution for selected HLA loci,
and none have performed multiple test correction (table 3, page 21). Hence, these studies are
of limited power, and are prone to both false positive and false negative results. In our
association study (Paper II) we wanted to overcome some of these challenges, and we
performed high resolution genetic HLA typing of 426 ME/CFS patients and 4511 ethnically
matched controls and reported multiple test corrected p-values. We identified two novel HLA
associations, tagged by the alleles HLA-C*07:04 and DQB1*03:03.
The association with HLA-C*07:04 (Paper II) cannot be directly compared to the literature,
since no previous studies in ME/CFS have investigated HLA-C alleles, and there is only a few
reported results for HLA class I alleles. HLA-B*44:02 is generally occurring with C*07:04 in
our data (see 6.2), and two previous studies have performed low resolution serologic HLA-B
typing [106, 107]. The broad antigen HLA-B12, which is not specific for B*44:02 as it
detects both B*44 and B*45 alleles, was not associated with CFS in these studies.
In contrast, the DQB1 locus and the DQ molecule have been investigated in several studies
[105, 106, 108], enabling comparison with the other association from Paper II, with
DQB1*03:03. In the largest previous study, Keller et al compared HLA-DR and -DQ
serotypes in 110 Caucasian patients with Chronic fatigue immune dysfunction syndrome
(CFIDS) and 616 healthy Caucasian controls [105]. The patients were diagnosed according to
the Holmes Criteria [1], and CFIDS was defined as a subgroup with positive findings in viral
reactivation patterns and B- and T-cell tests, indicating post-infectious debut and a certain
degree of immune dysfunction. The strongest significant association was with the serotype
HLA-DQ3 with an OR of 1.8 (95%CI: 1.2 - 2.8). This serotype corresponds to HLA-
DQB1*03 in 1st field genetic nomenclature. We did not originally investigate 1st field
associations, but by combining the five different identified 2nd field alleles that constitute
DQB1*03 (namely DQB1*03:01 to DQB1*03:05), we could compare our results with Keller
et al (table 8). The sum of these five alleles are slightly, but not significantly, more common
60
among patients (AF 0.34 versus 0.32). In some diseases (e.g. multiple sclerosis) initially
identified, low resolution HLA associations (DR2) are subsequently refined to specific second
field alleles (DRB1*15:01) [220]. Similarly, it is possible that the DQ3 association in Keller
et al (low resolution), is actually driven by a primary association with DQB1*03:03 (high
resolution). In CFS, higher resolution HLA-DQB1 typing have previously only been
performed in two smaller cohorts [106, 108] without revealing significant associations.
However, DQB1*03:03 was the only DQB1 allele with a higher frequency among patients
(table 3, page 21). Thus, there is some support in existing literature for an association with
DQB1*03:03.
Table 8 Number and frequency of DQB1*03 alleles among 426 ME/CFS patients and 4511 controls. Total number of identified alleles: 852 in patients and 9018 in controls
HLA-DQB1 allele
No of alleles in patients
Allele frequency in patients
No of alleles in controls
Allele frequency in controls
03:01 128 0.1502 1215 0.1347 03:02 102 0.1197 1262 0.1399 03:03 56 0.0657 406 0.0450 03:04 1 0.0012 15 0.0017 03:05 0 0.0000 3 0.0003 All combined 287 0.34 2901 0.32
DQB1*03:03, DQB1*03:02 and DQB1*03:01, the three most common second field alleles
constituting DQB1*03, are in separate G and P groups and therefore encode proteins with
amino acid differences in the peptide binding groove (Figure 10). Unique amino acid
sequences may convey unique peptide binding properties [154, 201]. Thus, DQB1*03:03
being the only DQB1*03 allele associated with ME/CFS fits well within the hypothesis of an
immune-driven disease mechanism, potentially involving specific HLA-peptide interaction
and presentation. DQB1*03:01 was also more prevalent among patients, although only
DQB1*03:03 was significantly associated in our material. Alternatively, different alleles
encoding identical amino acid sequence in parts of the peptide binding groove may constitute
the HLA association within subgroups of patients. This is highly speculative for ME/CFS, but
in line with the established theory of "shared epitope" in rheumatoid arthritis [221].
61
Allele/AA Position 20 30 40 50 60 70
DQB1*03:03 FKGMCYFTNG TERVRLVTRY IYNREEYARF DSDVGVYRAV TPLGPPDAEY WNSQKEVLER
DQB1*03:02 ---------- ---------- ---------- ---------- ------A--- ----------
DQB1*03:01 --A------- -----Y---- ---------- ----E----- ---------- ---------- Figure 10 Illustration of amino acid (AA) sequence differences in part of the HLA-DQ β chain proteins encoded by most of exon 2 of HLA-DQB1*03:03 compared with DQB1*03:02 and DQB1*03:01. Exon 2 in HLA-DQB1 encodes the β chain part of the peptide binding groove of the HLA-DQ molecule.
A few ME/CFS association studies based on SNP genotyping report HLA associations as
well. Recently, genome wide SNP-data from 383 ME/CFS patients, identified by the Fukuda
criteria, was compared to a large SNP database, and the reported associations included SNPs
in HLA-DRB1 and -DQA1 [103]. However, through filtering algorithms the authors focused
only on SNPs with estimated high deleteriousness, and neither standard significance testing
nor population stratification were addressed. Thus, the reported results are not directly
relevant for associations in the highly complex and polymorphic HLA region. Results from a
large GWAS in the UK Biobank, with many phenotypes, including a self-reported diagnosis
of CFS, were published online, although not in peer-reviewed scientific journal [222]
(https://biobankengine.stanford.edu, accessed September 2018). There were no genome-wide
significant results for CFS, but a SNP in the BTNL2-gene (200Kb from DRB1) was more
prevalent among patients, with OR of 1.28 and minor allele frequency at 5%, representing an
interesting similarity in frequency with our DQB1*03:03 risk allele. Whether this SNP could
be a tag for DQB1*03:03, or potentially the B*57:01-DQB1*03:03 haplotype, is an
interesting question that has not been addressed in this thesis, but will be further investigated
in future projects in our research group.
6.1.2 Previously reported significant associations compared to our results
Smith et al [108] reported an association between CFS and DQA1*01 (OR 1.93, p=0.008)
that is often cited in the literature. This result cannot be directly evaluated in our material
since the HLA-DQA1 locus was not genotyped in the control group. DQA1*01 was prevalent
62
in our patient group (Paper II) with AF of 45% and carrier frequency of 70.6%. By haplotype
estimation in our patient group, we evaluated which second field HLA-DQB1 alleles that
most often occurred with DQA1*01. None of these DQB1 alleles were associated with
ME/CFS when comparing with the control group (table 9), except for the very infrequent
allele DQB1*05:04 (0.6% in cases vs 0.09% in controls). These alleles combined were also
evenly distributed, with an OR of 1.1. Hence, our data do not support the main result from
Smith et al.
Table 9 HLA-DQB1 alleles estimated to be on haplotype with HLA-DQA1*01 in the patient group of Paper II (N=420). Comparison of the frequency of these DQB1 alleles in patients versus controls (N=4509).
DQB1-allele on haplotype with DQA1*01
Alleles in patient group
Alleles in control group
OR p= OR lower CI
OR upper CI
06:02 142 1472 1.1 0.6 0.9 1.3 05:01 80 965 0.9 0.3 0.7 1.1 06:03 66 725 1.0 0.9 0.8 1.3 06:04 54 462 1.3 0.09 1.0 1.7 05:03 24 227 1.2 0.5 0.8 1.8 05:04 5 8 7.0 0.0002 2.5 19.7
All combined 371 3859 1.1 0.4 0.9 1.2
Carlo-Stella et al [110] reported an association between CFS and DRB1*13:01 (OR 2.79,
p=0.006). In our material, this allele had almost equal AF in patients and controls, with an OR
of 1.0, leaving no support for this result. Alleles corresponding to the reported associations
DR4 and DR5 in Keller et al also had ORs very close to 1.0 in our data.
Spitzer et al [112] reported that 43% of CFS patients carried DQB1*06:02, an allele strongly
associated with narcolepsy and reported in up to 98% of narcolepsy patients [223]. However,
the great majority of patients in Spitzer et al were not referred for evaluation of chronic
fatigue, but were retrospectively evaluated to fulfill a CFS diagnosis. Many patients also
fulfilled the diagnosis of narcolepsy, truly a devastating bias. Two other studies report higher
frequencies of DQB1*06 and DQB1*06:02 among CFS patients, but without significant
associations [106, 108]. In our material, the association with DQB1*06:02 had no support
(OR 1.1, table 9).
63
The discrepancies between previous literature and our results may reflect the heterogeneity in
ME/CFS, and the fact that different inclusion criteria have been applied. Furthermore, there is
a non-negligible risk of false positive results in these previous studies, due to limited size and
the lack of multiple test correction.
6.1.3 Negative HLA associations
We also observed alleles that were more frequent in controls than in ME/CFS patients, i.e.
negative associations (Paper II). The HLA-B*08:01 allele is part of the common Northern
European ancestral haplotype AH8.1, comprising C*07:01-B*08:01-DRB1*03:01-
DQB1*02:01 [224]. In our material, all these alleles, as well as the entire AH8.1 haplotype,
were somewhat less common among ME/CFS patients, but only B*08:01 showed significant
association with Pnc<0.05. In addition, significant negative association between ME/CFS and
DPB1*02:01 was observed. This allele is not a part of AH8.1, shown by negative LD with
HLA-B*08:01 (D’=-0.41). In line with this, the haplotype A*01:01-C*07:01-B*08:01-
DRB1*03:01-DQB1*02:01-DPB1*02:01 had low frequency compared to the same haplotype
with any DPB1 allele (0.4% vs 7.9%, Paper I). Thus, the haplotype AH8.1 and the allele
DPB1*02:01 seem to be two independent associations. DPB1 has not been investigated in
CFS/ME previously. In contrast, HLA-DRB1 is the most frequently studied HLA locus in
CFS. In four out of five studies, DR3/DRB1*03 was less prevalent in the patient group [106-
108, 111], while in the fifth study the frequency was similar in both groups [105]. DRB1*03
corresponds almost exactly to DRB1*03:01, as their frequencies in the control group are
identical (table S2 and S3 Paper I). The AH8.1 haplotype is positively associated with a wide
variety of AID, including myasthenia gravis, systemic lupus erythematosus and coeliac
disease [224], but also negatively associated to rheumatoid arthritis [94, 225] and
inflammatory bowel disease [226]. Therefore, negative association to the general
"autoimmunity risk haplotype" AH8.1 could potentially indicate a protection against ME/CFS
based on immune regulating functions of HLA. Alternatively, the negative association may
reflect that patients with chronic fatigue and also AH8.1-associated autoimmune diagnoses
have been excluded from the ME/CFS diagnosis. Importantly, although AH8.1 seems truly
less prevalent among ME/CFS patients based on previous literature and our results, a multiple
test corrected significant association has not been reported.
64
6.2 DETERMINING THE PRIMARY ASSOCIATION SIGNALS IN THE HLA COMPLEX
We identified four potential HLA risk alleles: C*07:04, B*57:01, B*44:02 and DQB1*03:03
(Paper II, table 2). Importantly, in HLA-association studies, an associated allele may not be
the primary risk allele, but "hitch-hiking" due to strong LD with the causal one [141]. The
LD-patterns in our Norwegian dataset told us that most patients carrying C*07:04 also carry
B*44:02 (D' 0.90), and most patients carrying DQB1*03:03 also carry B*57:01 (D' 0.69),
while C*07:04 and DQB1*03:03 displayed linkage equilibrium (D' 0.06). Hence, the
associated alleles represent two allele-pairs mainly occurring on two distinct haplotypes,
namely C*07:04 - B*44:02 and B*57:01 - DQB1*03:03. Both haplotypes had frequencies
below 5%, but were among the 10 most frequent haplotypes in our patient group. The latter
haplotype occurred almost exclusively with HLA-DRB1*07:01, but this allele also occurred
on other haplotypes and was not associated with ME/CFS (OR=1.2, p=0.1). It is often seen
that haplotypes, rather than single alleles are the mediators of disease risk [151]. In other
cases, detailed mapping in large enough populations, have broken down associated haplotypes
and discovered that individual alleles are the mediators of disease risk [227]. In our data, the
incomplete (although high) LD, and the differences in allele frequency show that the
associated alleles also occur on other haplotypes. Therefore, we aimed to identify the apparent
primary association of each allele-pair (C*07:04 - B*44:02 and B*57:01 - DQB1*03:03)
(Paper II). However, none of the alleles on each haplotype showed association in absence of
the other, possibly due to the limited size of our patient samples, combined with the above
mentioned high LD.
We decided to report only one of the alleles on each of these risk haplotypes, even though we
could not statistically pinpoint the primary associations. C*07:04 was reported as the first tag
allele, as it remained significant after multiple test correction, and had a higher OR. HLA-C
was also associated on the global level, while HLA-B was not (Paper II). Regarding the
second allele-pair, B*57:01 and DQB1*03:03, there were only small differences in p-value
and effect size. DQB1*03:03 was about one and a half times as prevalent as B*57:01 in both
patients and controls. HLA-DQB1 was associated on the global level, while HLA-B was not
(Paper II), and DQB1 is also the locus being reported in previous literature, as discussed
below. Resultingly, DQB1*03:03 was chosen as the second tag allele. Whether these alleles
are primarily associated with ME/CFS, or hitch-hiking due to other genetic variants in or near
the HLA complex, remains to be settled. Several other genes within the HLA complex could
65
also be candidate genes for ME/CFS. An examples of an association within the HLA-region
that is not driven by a classic HLA gene is the HFE gene (close to HLA-A) in
hemochromatosis [228]. Furthermore, given the immunological importance of the HLA
complex, causal genetic variants in several genes may be involved. In an ongoing project in
our research group, we assess the genetic associations across the HLA complex in the same
patient material using the dense SNP panel Immunochip (Illumina, San Diego, US).
6.3 VALIDITY OF THE REPORTED HLA ASSOCIATIONS Any HLA-associated variable differentially distributed among patients and controls may
confound the results in Paper II. First, the distribution of psychiatric comorbidity may be
different between the groups. Previous or current moderate to severe depressive symptoms
were present for nearly 10% of patients, corresponding well to figures from literature [26].
We do not have equivalent data for controls, but since former or current mild to moderate
psychiatric disease do not exclude from registration in NBMDR, we can safely assume their
presence among controls. Estimated 12-month and life-time prevalence of moderate
depression in Oslo is 7.3% and 17.8% respectively [229], and large differences between
patients and controls seem improbable. Furthermore, depression is not a plausible confounder
as it is not associated with classic HLA alleles [230]. Next, fibromyalgia has considerable
clinical overlap with ME/CFS [168], and it is accepted as a comorbidity in CCC [4]. It is
probably much less common in the control group, but since there are no established HLA
associations for fibromyalgia, and specifically not for our reported risk alleles [231] [232], it
is not a plausible confounder. Finally, age and gender differences, discussed in 4.2.3, are not
plausible sources for bias or confounding.
Different distribution of HLA-associated AID is an important potential confounder. In our
material, 14.3% of the patients reported comorbidity of autoimmune diseases. The great
majority of these patients (>80%) had thyroiditis/hypothyroidism or psoriasis, and there were
only sporadic cases of Graves disease, coeliac disease, type 1 diabetes, inflammatory bowel
disease, alopecia areata, vitiligo, rheumatoid arthritis and pernicious anemia. We can safely
assume that hypothyroidism and psoriasis are present also in the control group, being
relatively common diseases compatible with inclusion in NBMDR if in a stable phase [177]
(and 4.1.3). The same goes for a few of the other registered AID, such as coeliac disease. We
66
had no comorbidity data for the controls, but the estimated general prevalence of AID is 7-9%
[142, 144], and AID is therefore probably more prevalent among patients and may constitute
a confounder. Importantly, the two positively associated alleles C*07:04 and DQB1*03:03
are not known to be associated with any AID, specifically not to thyroiditis and psoriasis. In
two studies with South-east Asian populations, DQB1*03:03 has been associated with
pityriasis rosea and the rare psoriasis-subtype pustular psoriasis, respectively [233, 234], but
both studies were small (55 and 26 patients), and results have not been replicated.
DQB1*03:03 does sometimes occur on haplotype with HLA-C*06:02, and the latter is indeed
associated with psoriasis [235, 236]. However, the LD between these two alleles in our
patient population was only D'= 0.39, not strong enough to drive the association. Notably,
HLA-C*06:02 was equally distributed in both groups (6.6% in patients, 6.1% in controls, p=
n.s.). Hence, these comorbidities are not plausible confounders.
Ethnicity matching was performed in Paper II to control population stratification, but genetic
differences may be present between Norwegian regions. An example is the higher prevalence
of founder mutations in certain regions, e.g. distinct BRCA and ATM mutations [237, 238].
Therefore, we ensured that all Norwegian regions were well represented among both patients
and controls, but not necessarily in exactly similar proportions. Both reported risk alleles were
associated with ME/CFS in all three patient inclusion groups in Paper II (figure 5, page 35).
However, the DQB1*03:03 frequency varied, and was more prevalent in the group with the
majority of patients from Western Norway (allele frequency 7.2%, 5.2% and 6.8% in the three
groups). Importantly, this association was not driven by regional population stratification, as
there were no differences in DQB1*03:03 allele frequency between the whole control group
(4.5%) and control individuals recruited through blood banks in Western Norway (4.4%).
The carrier frequency of DQB1*03:03 was substantially higher in Paper III, with patients
mainly from Western Norway (25%), than among all patients in Paper II (12.5%). Notably,
the patients' risk-allele status was unknown to investigators when patients were selected and
when clinical results were registered. The increased DQB1*03:03 frequency likely represents
a coincidence because of the relatively small sample size of 40 patients. It is also possible that
the risk allele DQB1*03:03 increases the liability to develop ME/CFS more in Western
Norway than elsewhere, because of potential differences in HLA-environment interactions,
such as for instance different antigenic repertoire. This has been suggested for established
AID, such as multiple sclerosis [239].
67
6.4 IMMUNE MODULATION IN ME/CFS
6.4.1 Experiences from rituximab intervention
One of the inclusion groups in Paper II comprised the 152 ME/CFS patients from the double-
blinded RCT published in 2019, where infusions with rituximab, a specific and reversible B-
cell depletion agent, had no overall therapeutic effect compared to placebo [130]. Specific
HLA alleles may influence the effect of immune modulating treatment [240, 241], and despite
no overall effect, HLA associations may have been present in the intervention group in the
rituximab trial. Therefore, we retrospectively evaluated whether the HLA risk alleles
identified in Paper II were more prevalent among patients with clinical improvement after
rituximab infusion. There were no associations between rituximab response and C*07:04 and
DQB1*03:03 (unpublished data). The initial therapeutic effect of rituximab, in case-reports,
open label studies and in one smaller randomized study [126-129], may be due to different
etiology among subgroups of patients. An alternative explanation is that the placebo effect
had a significant impact in the initial rituximab studies.
6.4.2 Effect and safety of cyclophosphamide administration
In addition to rituximab, some ME/CFS patients at the oncology unit at Haukeland university
hospital have received cyclophosphamide or iphosphamide for malignant lymphoma or breast
cancer, and seven patients have reported substantial improvement in core ME/CFS symptoms.
Cyclophosphamide is a cytotoxic agent with a less specific immunosuppressant effect than
rituximab. Through the main mechanism of DNA alkylation [242], cyclophosphamide
induces cell death among several subsets of lymphocytes, particularly rapidly proliferating
cells [243]. Like rituxmab, cyclophosphamide is also an established alternative in the
treatment of certain AID [244, 245], and therefore further investigation of cyclophosphamide
tretament in ME/CFS was relevant to the hypothesis of an immune-driven etiology.
Adverse effects
One of the main aims in Paper III was to evaluate the safety of cyclophosphamide
administration to ME/CFS patients. There were no signs of myelosuppression or severe
toxicity among the 40 patients. Nausea, constipation, haematuria, irregular menstruation and
other mild to moderate adverse effects were common (approximately 90% of patients). A
serious possible adverse reaction occurred for one patient who was hospitalized for two weeks
68
due to gradually aggravated POTS. It is uncertain whether the aggravation, which
subsequently declined, was caused by the intervention. There were 10 other serious adverse
events such as infections, rash, headache and palpitations. Two female patients in mid-forties
experienced menopause during follow-up, possibly a side-effect of cyclophosphamide. These
adverse effects, compatible with reports from cyclophosphamide treatment in established AID
[245, 246] are not negligible, although temporary. Importantly, they should be interpreted in
the light of the potential long lasting clinical improvement for many patients.
Clinical response
Different variables all point to a clear therapeutic effect of cyclophosphamide in our ME/CFS
patient cohort in Paper III. Clinical response, registered for 22/40 patients (55%), was delayed
with 3-6 months from the first infusion (figure 2, panel C, Paper III), and lasted for at least 3
years for 75%. Experiences from intravenous pulse-treatment with cyclophosphamide in
established AID is in line with the timing of this effect [247]. It is plausible that placebo effect
and registration bias due to patients' expectations (4.2.4-5) have had some impact on these
results. Compared to the effect of rituximab in the initial trials [128], however, the clinical
improvement of core ME/CFS symptoms in Paper III was more homogenous and long-
lasting, suggesting an immunosuppressive therapeutic potential in ME/CFS.
Cyclophosphamide doses given to patients in Paper III (monthly infusion of 600-700 mg/m2)
correspond to treatment for systemic lupus erythematosus or lupus nephritis. Dosages used in
cancer treatment, typically much higher, can be used for the complete eradication of
hematopoietic cell, while lower doses are relatively selective for T-cells (especially regulatory
T-cells), but also affects B-cells and other immune cells [243]. Changed numbers of
regulatory T-cells have been reported in ME/CFS, although there is a lack of replicated results
[74]. Cyclophosphamide reduces the cytokines IFN-g and IL12 [248], which both
interestingly showed increasing levels according to severity in ME/CFS patients [80]. In
Paper III, however, clinical improvement showed no association to disease severity, and was
in fact smaller for ME/CFS patients with severe diasese. The limited size of Paper III and the
lack of a placebo group are fundamental challenges, and clear conclusions cannot drawn. Still,
our results support the hypothesis of immunological involvement in ME/CFS pathogenesis,
and warrants a future double-blinded randomized controlled cyclophosphamide trial.
69
6.5 CLINICALLY OR ETIOLOGICALLY DISTINCT SUBGROUPS OF ME/CFS PATIENTS Some studies identify subgroups of ME/CFS patients, and present associations between
biologic variables (such as genetic data) and different subgroups based on long or short
duration [79], age of onset [111], immune signatures [105], specific symptoms [249] or
severity of symptoms [250]. Gene expression patterns are correlated to clinical subtypes
[251], and specific methylation patterns are associated with subgroups based on differences in
physical functioning and post-exertional malaise [252]. Associations between acute infection-
triggered ME/CFS and SNPs in PTPN22 and CTLA4 have recently been reported; both SNPs
have established associations with multiple AID [104]. However, no biomarkers have proven
to consistently identify ME/CFS subgroups, despite continuous efforts to investigate
differences between patients [251, 252]. The heterogeneity in ME/CFS may be substantial,
challenging the validity of the diagnosis [44, 166, 170]. The application of many different
diagnostic criteria in ME/CFS research has probably increased heterogeneity among patients,
and made it more difficult to identify underlying pathophysiology [22, 35, 167, 172, 173].
Therefore, identification of biologic variables in subgroups of patients is fundamental when
further investigating ME/CFS etiology and potential treatment [102, 114].
In Paper II, the subgroup of ME/CFS patients with either of the two associated alleles had a
significantly higher comorbidity of established autoimmune disease (Paper II, table 3). The
following autoimmune diseases were present in this subgroup, ordered by frequency:
Hashimoto's thyroiditis/hypothyroidism (10 patients), psoriasis (8), rheumatoid arthritis (2),
alopecia areata (1) and inflammatory bowel disease (1). In line with diagnostic criteria [4],
these conditions have been evaluated as comorbidities for the patients in question, and not
exclusionary diagnoses. AID often cluster in families, and there are common genetic risk-
factors for several AID [144, 253]. In individuals with an autoimmune condition, as well as in
their families, there is an increased frequency of other AID as well [254]. Therefore, higher
frequency of autoimmunity among risk-alleles carriers may indicate a subgroup in ME/CFS
with an autoimmune disease mechanism.
In Paper III, carriers of either of the risk alleles had a significantly higher response to
cyclophosphamide than non-carriers, and may also constitute an etiologially distinct
subgroup. There are several reports of associations between specific HLA alleles and
response to immune modulatory treatment in AID [240, 255, 256]. To our knowledge, this is
70
not reported specifically for cyclophosphamide nor for the alleles C*07:04 and DQB1*03:03.
The subgroup of patients carrying either of the two risk alleles, was not significantly
distinguishable by any other patient characteristic investigated in Paper II, including age,
gender, infectious onset or present severity (examplified in figure 11). It can be argued that
the lack of specificity of our self-reported variables makes it harder to detect such possible
associations. Similarly, in Paper III no other patient characteristic was significantly associated
with response, including gender, severity, disease duration, infectious onset or previous
rituximab treatment or response (figure 4, Paper III). Response rates were higher among
patients with moderate or moderate-to-severe disease, compared to the four patients with
severe disease.
Figure 11 Output from SPSS with disease severity boxplots by risk allele status, stratified by inclusion group (figure 5, page 35) among 426 ME/CFS patients in Paper II. There were no significant correlations (neither with nor without stratification). Disease severity was measured as activity level in DePaul Symptom Questionnaire, question 79 (DSQ79) [5, 174]. Legend, DSQ79: 1: I am not able to work or do anything, and I am bedridden. 2: I can walk around the house, but I cannot do light housework. 3: I can do light housework, but I cannot work part-time. 4: I can only work part time at work or on some family responsibilities. 5: I can work full time, but I have no energy left for anything else. 6: I can work full time and finish some family responsibilities but I have no energy left for anything else. 7: I can do all work or family responsibilities without any problems with my energy.
PaperII_InclusionGroup
iii)ii)i)
Disease_Severity_DSQ
79
6
5
4
3
2
1
10
Risk_allele_status
Page 1
71
Our projects were not designed to identify clear subgroups of patients. However, the
combined results point to the following interesting hypothesis: The overall HLA association
described in Paper II may be driven by subgroups of ME/CFS patients with a higher
tendency to autoimmunity (Paper II), and a substantially better response to immune-
modulating treatment (Paper III).
6.6 IS AUTOIMMUNITY PART OF ME/CFS PATHOGENESIS?
6.6.1 Definition of an autoimmune disease
Although the term autoimmunity is not precisely defined, a fundamental criterion for
autoimmune diseases is the presence of an auto-inflammatory process in the affected tissue or
organ in the form of self-reactive antibodies or T-cells [257]. In established AID, target tissue
inflammation can generally be proven, such as intestinal villous atrophy in celiac disease,
joint inflammation in rheumatoid arthritis and ankylosing spondylitis and skin lesions in
psoriasis. In some cases, specific auto-antibodies can be detected, demonstrating humoral
auto-immunity, e.g. in coeliac disease or seropositive rheumatoid arthritis. However, many
established AID lack strong and uniform associations to specific auto-antibodies, such as
multiple sclerosis [258] and inflammatory bowel disease [259]. Some clinically affected
patients with coeliac disease, with typical intestinal inflammation, classic HLA-type and later
proven effect of gluten-free diet lack detectable auto-antibodies [260]. Identification of the
causative antigen is required by some [257], although precise knowledge about the antigens
involved in autoimmunity is rare [141]. Examples are coeliac disease (gluten [261]) and
Goodpasture's disease (specific parts of type IV collagen in renal basal membrane [262]).
Often, it is also required that immunization in experimental models can induce an analogous
immune response and a similar disease phenotype [257], demonstrated in e.g. Goodpasture's
disease [161]. The types of treatment with proven efficacy can also indirectly indicate
whether autoimmunity is involved in disease etiology.
6.6.2 Evidence for immune dysregulation in ME/CFS
Humoral or cell mediated autoimmune reactions in ME/CFS has not been documented.
However, several quantitative and qualitative immune cell perturbations have been reported
for NK-cells and subsets of both B-cells and T-cells [74, 75] [76-78], including impaired
72
metabolism in CD4+ and CD8+ T-cells [263]. Associations between infectious onset
ME/CFS and SNPs in PTPN22 and CTLA4 were recently reported; both these genes play a
key role in regulating B and T-cell activation [104]. Furthermore, several studies have found
elevated autoantibodies in ME/CFS, e.g. towards adrenergic and cholinergic neurotransmitter
receptors [88-90]. In order for an autoimmune response to progress to a pathologic outcome, a
number of inflammatory modulators are required [142], and changed levels of cytokines have
been reported in ME/CFS [79-81], albeit with little overlap across studies.
The many reports of immune dysregulation in ME/CFS fit within disease models where
autoimmune reactions are thought to explain the wide variety of symptoms seen in ME/CFS
[36, 37]. However, these theories remain hypothetical, as none of the above listed criteria for
autoimmunity have been documented for ME/CFS. For example, the reported autoantibodies
in ME/CFS are not consistent between studies. This may be due to methodological
differences, false negative results or different etiologies between subsets of patients, but false
positive or non-generalizable results are also plausible.
We intended to perform a thorough screen for auto-antibodies in our patient cohort, and did
preliminary investigations in collaboration with researchers at the Department of
Immunology, Oslo University Hospital. Sera from 87 ME/CFS patients and 87 healthy
controls were screened for autoantibodies against an array of 10,000 recombinantly
manufactured human proteins. Results were analyzed with hierarchical clustering and
principal component analysis in R (RStudio, PBC, Boston, US), but no evident associations
appeared, and there were no significant associations (unpublished data). A few proteins
yielded high signals in subgroups of patients, indicating high levels of auto-antibodies, but
these proteins did not correspond to previously published auto-antibodies among ME/CFS
patients.
In light of the repeated findings of altered NK-cell function in ME/CFS patients, investigation
of genes encoding the NK-cell receptor are relevant. These killer-immunoglobulin-like
receptor genes (KIR genes) are highly polymorphic and complex [264], and intricate
interaction with HLA class I alleles are central to the activity of NK-cells. The HLA class I
risk allele C*07:04 (Paper II), could potentially be exerting its biologic relevance through this
interaction. Pasi et al investigated KIR genes and HLA-KIR ligands in CFS patients and
controls [113]. One of the reported findings was that a certain genotype lacking a specific
73
HLA-C interacting KIR gene was clearly more prevalent in patients, and the authors
speculated that CFS may be caused by a partially deficient interaction between innate (KIR)
and adaptive (HLA) immunity. However, the study was small and the significance did not
survive multiple test correction. In this thesis, KIR genes were not investigated, but we
characterized the KIR ligand subtypes of HLA class I alleles and compared these between
patients and controls without revealing any significant associations (unpublished data). We
plan further work with KIR genes in ME/CFS in our research group, including investigations
of correlations such as in Pasi et al.
6.6.3 The impact of HLA associations
HLA associations are documented for virtually all AID. Generally, if HLA associations are
present for a given condition, the pivotal role of HLA-molecule-dependent antigen
presentation to T-cells are thought to be involved [262]. However, HLA associations are not
sufficient for determining the etiology of a disease (subgroup) [160]. Therefore, our results
cannot be taken as evidence for an autoimmune disease mechanism in ME/CFS.
Some AID present very strong HLA associations, like HLA-B27 in ankylosing spondylitis or
DQB1*06:02 in narcolepsy, while others are moderate or weak, and both common alleles and
rare alleles are involved [141]. This is illustrated in Table 10, listing established AID
associated HLA alleles and their frequencies in our Norwegian control cohort from Paper I.
Importantly, the HLA associations with ME/CFS in our data (Paper II) are comparable with
some of these established associations, such as the most cited HLA association for Crohn's
disease represented by the relatively rare allele DRB1*01:03.
HLA associations are also reported for many conditions without established autoimmune
etiology, such as Alzheimer's disease, Parkinson's disease, bipolar disorder or schizophrenia
[156-158], although in some cases with conflicting evidence [265] [230]. These associations
influence the field of research, and for many conditions novel hypotheses of autoimmune
involvement in the pathogenesis have emerged [143, 266]. Similarly, the HLA associations in
our material add to the many reports of immune dysregulation in ME/CFS, and may increase
the focus on the hypothesis of an autoimmune disease mechanism.
74
Tabl
e 10
Ex
ampl
es o
f HLA
ass
ocia
tions
in d
iffer
ent e
stab
lishe
d A
ID.
Dise
ase
HLA
alle
le (o
r hap
loty
pe w
hen
indi
cate
d)
OR
Refe
renc
e Fr
eque
ncy
in 4
514
heal
thy
Norw
egia
ns (P
aper
I)
Croh
n's d
iseas
e DR
B1*0
1:03
2.
5 (6
.9 in
chi
ldre
n)
[226
] 0.
014
Mul
tiple
scle
rosis
DR
B1*1
5:01
3.
6 [1
97]
0.16
5
DRB1
*15:
01-D
QA1
*01:
02-D
QB1
*06:
02 (h
aplo
type
) 2.
6 [2
67]
0.16
2 Ad
diso
n's d
iseas
e DR
B1*0
3:01
2.
9 [2
68]
0.13
6
DRB1
*04:
04
3.3
[268
] 0.
060
Psor
iasis
C*
06:0
2 4.
2a [2
69]
0.06
1 Sy
stem
ic lu
pus e
ryth
emat
osus
DR
B1*1
5:01
1.
3 [2
70]
0.16
5
DRB1
*03:
01
1.9
- 2.3
[2
70, 2
71]
0.13
6 Ty
pe1
diab
etes
DQ
B1*0
6:02
0.
17
[272
] 0.
163
DQ
B1*0
3:02
3.
1 [2
72]
0.14
0
DQB1
*02:
01
4.4
[272
] 0.
193b
Nar
cole
psy
DQB1
*06:
02
251
[223
] 0.
163
DQ
B1*0
6:03
0.
19
[223
] 0.
080
Rheu
mat
oid
arth
ritis
DRB1
*04:
01
4.4
[94]
0.
115
Juve
nile
mya
sthe
nia
grav
is DR
B1*0
3:01
2.
6 [2
73]
0.13
6 Ju
veni
le m
yast
heni
a gr
avis
DRB1
*04:
04
2.7
[273
] 0.
060
Late
-ons
et m
yast
heni
a gr
avis
DRB1
*15:
01
2.4
[274
] 0.
165
Coel
iac
dise
ase
DQB1
*02:
01-D
QA1
*05:
01 (h
aplo
type
) 21
.6
[275
] - c
Anky
losin
g sp
ondy
litis
B27
>100
(RR)
[1
59]
0.06
0 G
oodp
astu
re's
dise
ase
DRB1
*15:
01
8.5
[262
] 0.
167
a: C
alcu
late
d fr
om fi
gure
s in
the
refe
renc
e. b
: Fre
quen
cy o
f DQ
B1*
02:0
1:01
G, w
hich
incl
udes
e.g
. DQ
B1*
02:0
2. c
: HLA-DQA1
was
not
in
vest
igat
ed in
Pap
er I,
and
pre
cise
hap
loty
pe fr
eque
ncie
s cou
ld th
eref
ore
not b
e es
timat
ed.
75
7 CONCLUSIONS AND FUTURE PERSPECTIVES In this thesis, we present results that hopefully contribute to a better understanding of
ME/CFS. Our overriding hypotheses is that the immune system is involved in the
pathogenesis. We focus on the importance of biologic factors in ME/CFS, and simultaneously
do not refute the co-existence of e.g. immunogenetic and psychosocial contributing factors
[40, 43], well compatible with a sensible interpretation of the biopsychosocial model. Our
findings fit with the notion that ME/CFS is a complex disease, with heterogeneity both
clinically and etiologically.
We have established a Norwegian HLA reference data set (Paper I), published the first large
and thorough HLA association study in ME/CFS (Paper II), and evaluated the safety and
potential therapeutic effect of cyclophosphamide in ME/CFS patients (Paper III). Our main
finding is two novel HLA associations for adult, Norwegian ME/CFS patients, represented by
a significantly higher prevalence of HLA-C*07:04 and HLA-DQB1*03:03 among patients.
These risk-alleles were further significantly associated to comorbid autoimmune disease
among patients, and to long lasting clinical improvement after open-label treatment with the
immune modulating drug cyclophosphamide. Our results are in favor of immunological
involvement in the predisposing, precipitating and/or perpetuating factors in ME/CFS
pathogenesis.
Immunological or autoimmune involvement in ME/CFS pathogenesis is still uncertain. Future
studies should continue to address immune dysregulation and immunogenetics in ME/CFS.
Relevant strategies are studies of auto-antibodies and cytokines as well as NK-, T- and B-cell
levels and function. Characterization of KIR, T-cell and B-cell receptor genes, and SNP
analyses focused on multiple immune genes (such as the Immunochip), should also be
performed. Gene expression and epigenetics would also be relevant to address, as well as
continued efforts to identify biomarkers, e.g. micro-RNAs [114]. Many studies reporting
immune disturbances in ME/CFS are relatively small, and others employ complex processing
of large data sets. In both cases, we should expect a certain amount of false positive results,
and replication in larger and independent patient cohorts is indispensable. Notably, the HLA
associations reported herein also need replication before considered established for ME/CFS.
76
Further studies with fine-mapping of the HLA region is also necessary to pinpoint the primary
associations. Other interesting strategies are analyses of HLA and KIR genes in combination.
ME/CFS is a heterogenous disease, and further thorough patient categorization and sub-
phenotyping are important to reveal different etiologies and potentially improve treatment
[102, 114]. Genome-wide association studies with thousands of ME/CFS patients are
important in this respect, both to identify candidate genes and pathways, and for clinical
subgrouping. Ideally, multiple large cohorts of clinically well-characterized patients should be
available, e.g. with more precise data about infectious agents and immunological parameters
in patients with post-infectious ME/CFS.
No established efficient treatment exists for ME/CFS, and potential treatment should continue
to be scientifically investigated. Our results from Paper III indicate that cyclophosphamide
treatment is relatively tolerable for ME/CFS patients, and induce a long lasting clinical
improvement for a large subgroup of patients. However, since there was no control group, the
placebo effect was not adressed. Until the clinical effect has been investigated in a larger,
randomized and controlled trial, cyclophosphamide intervention should not be performed in
ME/CFS patients.
The complexity of ME/CFS, combined with the severity and the lack of knowledge, warrant
broad scientific investigation of multiple possible etiological hypotheses. The combination of
central nervous system symptoms and immunological alterations are the core features of
ME/CFS, according to some researchers, who propose terms like psychoneurobiology to
further address and study such interactions between different organ systems [33, 43]. In fact,
the explanatory autoimmune ME/CFS models are good examples of such holistic approaches
[36, 37]. Continued thorough research is necessary to expand the knowledge about this
disease, which in turn will hopefully benefit the millions of patients suffering from ME/CFS.
77
References 1. Holmes, G.P., et al., Chronic fatigue syndrome: a working case definition. Ann Intern
Med, 1988. 108(3): p. 387-9. 2. Sharpe, M.C., et al., A report--chronic fatigue syndrome: guidelines for research. J R
Soc Med, 1991. 84(2): p. 118-21. 3. Fukuda, K., et al., The chronic fatigue syndrome: a comprehensive approach to its
definition and study. International Chronic Fatigue Syndrome Study Group. Ann Intern Med, 1994. 121(12): p. 953-9.
4. Carruthers, B.M., et al., Myalgic Encephalomyelitis/Chronic Fatigue Syndrome. Journal of Chronic Fatigue Syndrome, 2003. 11(1): p. 7-115.
5. Jason LA, E.M., Porter N, Brown M, Brown AA, Hunnell J, et al. , The development of a revised Canadian myalgic encephalitis chronic fatigue syndrome case definition. Am J Biochem Biotechnol, 2010. 6: p. 120-35.
6. Carruthers, B.M., et al., Myalgic encephalomyelitis: International Consensus Criteria. J Intern Med, 2011. 270(4): p. 327-38.
7. Committee on the Diagnostic Criteria for Myalgic Encephalomyelitis/Chronic Fatigue, S., P. Board on the Health of Select, and M. Institute of, The National Academies Collection: Reports funded by National Institutes of Health, in Beyond Myalgic Encephalomyelitis/Chronic Fatigue Syndrome: Redefining an Illness. 2015, National Academies Press (US). Copyright 2015 by the National Academy of Sciences. All rights reserved.: Washington (DC).
8. (NICE), N.I.f.H.a.C.E. ME/CFS guidelines in development [GID-NG10091]. 2007; Available from: https://www.nice.org.uk/guidance/indevelopment/gid-ng10091.
9. Helsedirektoratet, Nasjonal veileder. Pasienter med CFS/ME: Utredning, diagnostikk, behandling, rehabilitering, pleie og omsorg. 2015.
10. Prevention, U.C.f.D.C.a., What is ME/CFS. 2018. 11. Straus, S.E., History of chronic fatigue syndrome. Rev Infect Dis, 1991. 13 Suppl 1: p.
S2-7. 12. Hashimoto, N., [History of chronic fatigue syndrome]. Nihon Rinsho, 2007. 65(6): p.
975-82. 13. Dowsett, E.G., et al., Myalgic encephalomyelitis--a persistent enteroviral infection?
Postgrad Med J, 1990. 66(777): p. 526-30. 14. Ramsay, M., Myalgic encephalomyelitis: a baffling syndrome. Nurs Mirror, 1981.
153(15): p. 40-1. 15. Aaronson, L.S., et al., Defining and measuring fatigue. Image J Nurs Sch, 1999. 31(1):
p. 45-50. 16. Shen, J., J. Barbera, and C.M. Shapiro, Distinguishing sleepiness and fatigue: focus on
definition and measurement. Sleep Med Rev, 2006. 10(1): p. 63-76. 17. Bates, D.W., et al., Prevalence of fatigue and chronic fatigue syndrome in a primary
care practice. Arch Intern Med, 1993. 153(24): p. 2759-65. 18. Landmark-Hoyvik, H., et al., The genetics and epigenetics of fatigue. Pm r, 2010. 2(5):
p. 456-65. 19. van't Leven, M., et al., Fatigue and chronic fatigue syndrome-like complaints in the
general population. Eur J Public Health, 2010. 20(3): p. 251-7. 20. Loge, J.H., O. Ekeberg, and S. Kaasa, Fatigue in the general Norwegian population:
normative data and associations. J Psychosom Res, 1998. 45(1): p. 53-65.
78
21. Carmel, L., et al., Gene expression profile of empirically delineated classes of unexplained chronic fatigue. Pharmacogenomics, 2006. 7(3): p. 375-86.
22. Jason, L.A., et al., Problems in defining post-exertional malaise. J Prev Interv Community, 2015. 43(1): p. 20-31.
23. (NICE), N.I.f.H.a.C.E. ME/CFS guidelines - diagnosis and management. 2007; Available from: https://www.nice.org.uk/guidance/cg53.
24. Lien, K., et al., Abnormal blood lactate accumulation during repeated exercise testing in myalgic encephalomyelitis/chronic fatigue syndrome. Physiol Rep, 2019. 7(11): p. e14138.
25. Bestpractice.bmj, Chronic fatigue syndrome/Myalgic encephalomyelitis. 2019. 26. Rimbaut, S., et al., Chronic fatigue syndrome - an update. Acta Clin Belg, 2016: p. 1-8. 27. Nacul, L.C., et al., The functional status and well being of people with myalgic
encephalomyelitis/chronic fatigue syndrome and their carers. BMC Public Health, 2011. 11: p. 402.
28. Jason, L.A., et al., The economic impact of ME/CFS: individual and societal costs. Dyn Med, 2008. 7: p. 6.
29. NIH - National Institutes of Health. 30. NICE - National Institute for Health and Care Excellence. 31. Forskningsrådet. Behovsidentifisert forskning om CFS/ME. 2019 [cited 2020;
Available from: https://www.forskningsradet.no/om-forskningsradet/publikasjoner/2019/behovsidentifisert-forskning-om-cfsme--behov-me/.
32. Devanur, L.D. and J.R. Kerr, Chronic fatigue syndrome. J Clin Virol, 2006. 37(3): p. 139-50.
33. Ortega-Hernandez, O.D. and Y. Shoenfeld, Infection, vaccination, and autoantibodies in chronic fatigue syndrome, cause or coincidence? Ann N Y Acad Sci, 2009. 1173: p. 600-9.
34. Prins, J.B., J.W. van der Meer, and G. Bleijenberg, Chronic fatigue syndrome. Lancet, 2006. 367(9507): p. 346-55.
35. Stadje, R., et al., The differential diagnosis of tiredness: a systematic review. BMC Fam Pract, 2016. 17(1): p. 147.
36. Blomberg, J., et al., Infection Elicited Autoimmunity and Myalgic Encephalomyelitis/Chronic Fatigue Syndrome: An Explanatory Model. Front Immunol, 2018. 9: p. 229.
37. Sotzny, F., et al., Myalgic Encephalomyelitis/Chronic Fatigue Syndrome - Evidence for an autoimmune disease. Autoimmun Rev, 2018. 17(6): p. 601-609.
38. Wyller, V.B., H.R. Eriksen, and K. Malterud, Can sustained arousal explain the Chronic Fatigue Syndrome? Behav Brain Funct, 2009. 5: p. 10.
39. Engel, G.L., The need for a new medical model: a challenge for biomedicine. Science, 1977. 196(4286): p. 129-36.
40. Griffith, J.P. and F.A. Zarrouf, A systematic review of chronic fatigue syndrome: don't assume it's depression. Prim Care Companion J Clin Psychiatry, 2008. 10(2): p. 120-8.
41. Geraghty, K.J. and A. Esmail, Chronic fatigue syndrome: is the biopsychosocial model responsible for patient dissatisfaction and harm? Br J Gen Pract, 2016. 66(649): p. 437-8.
79
42. Janse, A., G. Bleijenberg, and H. Knoop, Prediction of long-term outcome after cognitive behavioral therapy for chronic fatigue syndrome. J Psychosom Res, 2019. 121: p. 93-99.
43. Cortes Rivera, M., et al., Myalgic Encephalomyelitis/Chronic Fatigue Syndrome: A Comprehensive Review. Diagnostics (Basel), 2019. 9(3).
44. Sandler, C.X. and A.R. Lloyd, Chronic fatigue syndrome: progress and possibilities. Med J Aust, 2020.
45. Hickie, I., et al., Post-infective and chronic fatigue syndromes precipitated by viral and non-viral pathogens: prospective cohort study. Bmj, 2006. 333(7568): p. 575.
46. Alberts, B., Retraction. Science, 2011. 334(6063): p. 1636. 47. Rasa, S., et al., No evidence of XMRV provirus sequences in patients with myalgic
encephalomyelitis/chronic fatigue syndrome and individuals with unspecified encephalopathy. New Microbiol, 2014. 37(1): p. 17-24.
48. Simmons, G., et al., Failure to confirm XMRV/MLVs in the blood of patients with chronic fatigue syndrome: a multi-laboratory study. Science, 2011. 334(6057): p. 814-7.
49. Naess, H., et al., Chronic fatigue syndrome after Giardia enteritis: clinical characteristics, disability and long-term sickness absence. BMC Gastroenterol, 2012. 12: p. 13.
50. Katz, B.Z., et al., Chronic fatigue syndrome after infectious mononucleosis in adolescents. Pediatrics, 2009. 124(1): p. 189-93.
51. Pedersen, M., et al., Predictors of chronic fatigue in adolescents six months after acute Epstein-Barr virus infection: A prospective cohort study. Brain Behav Immun, 2019. 75: p. 94-100.
52. Kristiansen, M.S., et al., Clinical symptoms and markers of disease mechanisms in adolescent chronic fatigue following Epstein-Barr virus infection: An exploratory cross-sectional study. Brain Behav Immun, 2019. 80: p. 551-563.
53. Loebel, M., et al., Deficient EBV-specific B- and T-cell response in patients with chronic fatigue syndrome. PLoS One, 2014. 9(1): p. e85387.
54. Loebel, M., et al., Serological profiling of the EBV immune response in Chronic Fatigue Syndrome using a peptide microarray. PLoS One, 2017. 12(6): p. e0179124.
55. Shikova, E., et al., Cytomegalovirus, Epstein-Barr virus, and human herpesvirus-6 infections in patients with myalgic small ie, Cyrillicncephalomyelitis/chronic fatigue syndrome. J Med Virol, 2020.
56. Lerner, A.M., et al., Valacyclovir treatment in Epstein-Barr virus subset chronic fatigue syndrome: thirty-six months follow-up. In Vivo, 2007. 21(5): p. 707-13.
57. Tanriverdi, F., et al., The hypothalamo-pituitary-adrenal axis in chronic fatigue syndrome and fibromyalgia syndrome. Stress, 2007. 10(1): p. 13-25.
58. Tomas, C. and J. Newton, Metabolic abnormalities in chronic fatigue syndrome/myalgic encephalomyelitis: a mini-review. Biochem Soc Trans, 2018. 46(3): p. 547-553.
59. Naviaux, R.K., et al., Metabolic features of chronic fatigue syndrome. Proc Natl Acad Sci U S A, 2016. 113(37): p. E5472-80.
60. Bohne, V.J.B. and O. Bohne, Suggested pathology of systemic exertion intolerance disease: Impairment of the E3 subunit or crossover of swinging arms of the E2 subunit of the pyruvate dehydrogenase complex decreases regeneration of cofactor dihydrolipoic acid of the E2 subunit. Med Hypotheses, 2019. 130: p. 109260.
80
61. Fluge, O., et al., Metabolic profiling indicates impaired pyruvate dehydrogenase function in myalgic encephalopathy/chronic fatigue syndrome. JCI Insight, 2016. 1(21): p. e89376.
62. Missailidis, D., et al., An Isolated Complex V Inefficiency and Dysregulated Mitochondrial Function in Immortalized Lymphocytes from ME/CFS Patients. Int J Mol Sci, 2020. 21(3).
63. Morris, G., et al., Myalgic encephalomyelitis/chronic fatigue syndrome: From pathophysiological insights to novel therapeutic opportunities. Pharmacol Res, 2019. 148: p. 104450.
64. Ohba, T., et al., Myalgic Encephalomyelitis/Chronic Fatigue Syndrome Induced by Repeated Forced Swimming in Mice. Biol Pharm Bull, 2019. 42(7): p. 1140-1145.
65. Snell, C.R., et al., Discriminative validity of metabolic and workload measurements for identifying people with chronic fatigue syndrome. Phys Ther, 2013. 93(11): p. 1484-92.
66. YT, B., Svak kobling mellom kronisk utmattelsessyndrom og pyruvat dehydrogenasemangel. Tidsskr Nor Legeforen 2017. 137: p. 1870-1.
67. Coetzee, N., D.J.F. Maree, and B.N. Smit, The relationship between chronic fatigue syndrome, burnout, job satisfaction, social support and age among academics at a tertiary institution. Int J Occup Med Environ Health, 2019. 32(1): p. 75-85.
68. Leone, S.S., et al., Two sides of the same coin? On the history and phenomenology of chronic fatigue and burnout. Psychol Health, 2011. 26(4): p. 449-64.
69. Loades, M.E., et al., Perfectionism and beliefs about emotions in adolescents with chronic fatigue syndrome and their parents: a preliminary investigation in a case control study nested within a cohort. Psychol Health, 2019. 34(7): p. 850-866.
70. Wood, B. and S. Wessely, Personality and social attitudes in chronic fatigue syndrome. J Psychosom Res, 1999. 47(4): p. 385-97.
71. Daniels, J., H. Parker, and P.M. Salkovskis, Prevalence and treatment of Chronic Fatigue Syndrome/Myalgic Encephalomyelitis and co-morbid severe health anxiety. Int J Clin Health Psychol, 2020. 20(1): p. 10-19.
72. Natelson, B.H., et al., The effect of comorbid medical and psychiatric diagnoses on chronic fatigue syndrome. Ann Med, 2019. 51(7-8): p. 371-378.
73. Theorell, J., et al., Unperturbed Cytotoxic Lymphocyte Phenotype and Function in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome Patients. Front Immunol, 2017. 8: p. 723.
74. Rivas, J.L., et al., Association of T and NK Cell Phenotype With the Diagnosis of Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS). Front Immunol, 2018. 9: p. 1028.
75. Curriu, M., et al., Screening NK-, B- and T-cell phenotype and function in patients suffering from Chronic Fatigue Syndrome. J Transl Med, 2013. 11: p. 68.
76. Brenu, E.W., et al., A Preliminary Comparative Assessment of the Role of CD8+ T Cells in Chronic Fatigue Syndrome/Myalgic Encephalomyelitis and Multiple Sclerosis. J Immunol Res, 2016. 2016: p. 9064529.
77. Mensah, F.F.K., et al., CD24 Expression and B Cell Maturation Shows a Novel Link With Energy Metabolism: Potential Implications for Patients With Myalgic Encephalomyelitis/Chronic Fatigue Syndrome. Front Immunol, 2018. 9: p. 2421.
78. Mensah, F.K.F., et al., Chronic fatigue syndrome and the immune system: Where are we now? Neurophysiol Clin, 2017. 47(2): p. 131-138.
81
79. Hornig, M., et al., Distinct plasma immune signatures in ME/CFS are present early in the course of illness. Sci Adv, 2015. 1(1).
80. Montoya, J.G., et al., Cytokine signature associated with disease severity in chronic fatigue syndrome patients. Proc Natl Acad Sci U S A, 2017.
81. Roerink, M.E., et al., Cytokine signatures in chronic fatigue syndrome patients: a Case Control Study and the effect of anakinra treatment. J Transl Med, 2017. 15(1): p. 267.
82. Trivedi, M.S., et al., Identification of Myalgic Encephalomyelitis/Chronic Fatigue Syndrome-associated DNA methylation patterns. PLoS One, 2018. 13(7): p. e0201066.
83. Morris, G., et al., The emerging role of autoimmunity in myalgic encephalomyelitis/chronic fatigue syndrome (ME/cfs). Mol Neurobiol, 2014. 49(2): p. 741-56.
84. Staines, D.R., Is chronic fatigue syndrome an autoimmune disorder of endogenous neuropeptides, exogenous infection and molecular mimicry? Med Hypotheses, 2004. 62(5): p. 646-52.
85. Biswal, B., P. Kunwar, and B.H. Natelson, Cerebral blood flow is reduced in chronic fatigue syndrome as assessed by arterial spin labeling. J Neurol Sci, 2011. 301(1-2): p. 9-11.
86. Duffy, F.H., et al., EEG spectral coherence data distinguish chronic fatigue syndrome patients from healthy controls and depressed patients--a case control study. BMC Neurol, 2011. 11: p. 82.
87. Maes, M., et al., IgM-mediated autoimmune responses directed against anchorage epitopes are greater in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) than in major depression. Metab Brain Dis, 2012. 27(4): p. 415-23.
88. Loebel, M., et al., Antibodies to beta adrenergic and muscarinic cholinergic receptors in patients with Chronic Fatigue Syndrome. Brain Behav Immun, 2016. 52: p. 32-39.
89. Tanaka, S., et al., Autoantibodies against muscarinic cholinergic receptor in chronic fatigue syndrome. Int J Mol Med, 2003. 12(2): p. 225-30.
90. Wirth, K. and C. Scheibenbogen, A Unifying Hypothesis of the Pathophysiology of Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS): Recognitions from the finding of autoantibodies against ss2-adrenergic receptors. Autoimmun Rev, 2020: p. 102527.
91. Nishikai, M., [Antinuclear antibodies in patients with chronic fatigue syndrome]. Nihon Rinsho, 2007. 65(6): p. 1067-70.
92. von Mikecz, A., et al., High frequency of autoantibodies to insoluble cellular antigens in patients with chronic fatigue syndrome. Arthritis Rheum, 1997. 40(2): p. 295-305.
93. Zaitlen, N. and P. Kraft, Heritability in the genome-wide association era. Hum Genet, 2012. 131(10): p. 1655-64.
94. Viatte, S. and A. Barton, Genetics of rheumatoid arthritis susceptibility, severity, and treatment response. Semin Immunopathol, 2017. 39(4): p. 395-408.
95. Crawley, E. and G. Davey Smith, Is chronic fatigue syndrome (CFS/ME) heritable in children, and if so, why does it matter? Arch Dis Child, 2007. 92(12): p. 1058-61.
96. Buchwald, D., et al., A twin study of chronic fatigue. Psychosom Med, 2001. 63(6): p. 936-43.
97. Schur, E., et al., Twin analyses of fatigue. Twin Res Hum Genet, 2007. 10(5): p. 729-33.
98. Sullivan, P.F., et al., Twin analyses of chronic fatigue in a Swedish national sample. Psychol Med, 2005. 35(9): p. 1327-36.
82
99. Albright, F., et al., Evidence for a heritable predisposition to Chronic Fatigue Syndrome. BMC Neurol, 2011. 11: p. 62.
100. Schlauch, K.A., et al., Genome-wide association analysis identifies genetic variations in subjects with myalgic encephalomyelitis/chronic fatigue syndrome. Transl Psychiatry, 2016. 6: p. e730.
101. Smith, A.K., et al., Convergent genomic studies identify association of GRIK2 and NPAS2 with chronic fatigue syndrome. Neuropsychobiology, 2011. 64(4): p. 183-94.
102. Wang, T., et al., A systematic review of the association between fatigue and genetic polymorphisms. Brain Behav Immun, 2017. 62: p. 230-244.
103. Perez, M., et al., Genetic Predisposition for Immune System, Hormone, and Metabolic Dysfunction in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome: A Pilot Study. Frontiers in pediatrics, 2019. 7: p. 206-206.
104. Steiner, S., et al., Autoimmunity-Related Risk Variants in PTPN22 and CTLA4 Are Associated With ME/CFS With Infectious Onset. Frontiers in immunology, 2020. 11: p. 578-578.
105. Keller, R.H., et al., Association between HLA class II antigens and the chronic fatigue immune dysfunction syndrome. Clin Infect Dis, 1994. 18 Suppl 1: p. S154-6.
106. Underhill, J.A., et al., Lack of association between HLA genotype and chronic fatigue syndrome. Eur J Immunogenet, 2001. 28(3): p. 425-8.
107. Helbig, K.J., et al., Variation in immune response genes and chronic Q fever. Concepts: preliminary test with post-Q fever fatigue syndrome. Genes Immun, 2003. 4(1): p. 82-5.
108. Smith, J., et al., Association of chronic fatigue syndrome with human leucocyte antigen class II alleles. J Clin Pathol, 2005. 58(8): p. 860-3.
109. Ledina, D., et al., Chronic fatigue syndrome after Q fever. Med Sci Monit, 2007. 13(7): p. Cs88-92.
110. Carlo-Stella, N., et al., Molecular study of receptor for advanced glycation endproduct gene promoter and identification of specific HLA haplotypes possibly involved in chronic fatigue syndrome. Int J Immunopathol Pharmacol, 2009. 22(3): p. 745-54.
111. Ortega-Hernandez, O.D., et al., Autoantibodies, polymorphisms in the serotonin pathway, and human leukocyte antigen class II alleles in chronic fatigue syndrome: are they associated with age at onset and specific symptoms? Ann N Y Acad Sci, 2009. 1173: p. 589-99.
112. Spitzer, A.R. and M. Broadman, A retrospective review of the sleep characteristics in patients with chronic fatigue syndrome and fibromyalgia. Pain Pract, 2010. 10(4): p. 294-300.
113. Pasi, A., et al., Excess of activating killer cell immunoglobulinlike receptors and lack of HLA-Bw4 ligands: a twoedged weapon in chronic fatigue syndrome. Mol Med Rep, 2011. 4(3): p. 535-40.
114. Dibble, J.J., S.J. McGrath, and C.P. Ponting, Genetic Risk Factors of ME/CFS: A Critical Review. Hum Mol Genet, 2020.
115. Castro-Marrero, J., et al., Treatment and management of chronic fatigue syndrome/myalgic encephalomyelitis: all roads lead to Rome. Br J Pharmacol, 2017. 174(5): p. 345-369.
116. Galeoto, G., et al., The effect of physiotherapy on fatigue and physical functioning in chronic fatigue syndrome patients: A systematic review. Clin Ter, 2018. 169(4): p. e184-e188.
83
117. Goudsmit, E.M., et al., Pacing as a strategy to improve energy management in myalgic encephalomyelitis/chronic fatigue syndrome: a consensus document. Disabil Rehabil, 2012. 34(13): p. 1140-7.
118. Middleton, D., D.A. Savage, and D.G. Smith, No association of HLA class II antigens in chronic fatigue syndrome. Dis Markers, 1991. 9(1): p. 47-9.
119. White, P.D., et al., Comparison of adaptive pacing therapy, cognitive behaviour therapy, graded exercise therapy, and specialist medical care for chronic fatigue syndrome (PACE): a randomised trial. Lancet, 2011. 377(9768): p. 823-36.
120. Vink, M. and A. Vink-Niese, Multidisciplinary rehabilitation treatment is not effective for myalgic encephalomyelitis/chronic fatigue syndrome: A review of the FatiGo trial. Health Psychol Open, 2018. 5(2): p. 2055102918792648.
121. Vink, M. and A. Vink-Niese, Graded exercise therapy for myalgic encephalomyelitis/chronic fatigue syndrome is not effective and unsafe. Re-analysis of a Cochrane review. Health Psychol Open, 2018. 5(2): p. 2055102918805187.
122. Castell, B.D., N. Kazantzis, and R.E. Moss-Morris, Cognitive behavioral therapy and graded exercise for chronic fatigue syndrome: A meta-analysis. Clinical Psychology: Science and Practice, 2011. 18: p. 311–324.
123. Forening, N.M.E., Norges ME-forenings brukerundersøkelse. 2011. 124. Services, U.S.D.o.H.H., Effective Health Care Program - Diagnosis and Treatment of
Myalgic Encephalomyelitis/ Chronic Fatigue Syndrome. 2016. 125. Bjorkum, T., C.E. Wang, and K. Waterloo, [Patients' experience with treatment of
chronic fatigue syndrome]. Tidsskr Nor Laegeforen, 2009. 129(12): p. 1214-6. 126. Fluge, O., et al., Benefit from B-lymphocyte depletion using the anti-CD20 antibody
rituximab in chronic fatigue syndrome. A double-blind and placebo-controlled study. PLoS One, 2011. 6(10): p. e26358.
127. Fluge, O. and O. Mella, Clinical impact of B-cell depletion with the anti-CD20 antibody rituximab in chronic fatigue syndrome: a preliminary case series. BMC Neurol, 2009. 9: p. 28.
128. Fluge, O., et al., B-Lymphocyte Depletion in Myalgic Encephalopathy/ Chronic Fatigue Syndrome. An Open-Label Phase II Study with Rituximab Maintenance Treatment. PLoS One, 2015. 10(7): p. e0129898.
129. Lunde, S., et al., Serum BAFF and APRIL Levels, T-Lymphocyte Subsets, and Immunoglobulins after B-Cell Depletion Using the Monoclonal Anti-CD20 Antibody Rituximab in Myalgic Encephalopathy/Chronic Fatigue Syndrome. PLoS One, 2016. 11(8): p. e0161226.
130. Fluge, O., et al., B-Lymphocyte Depletion in Patients With Myalgic Encephalomyelitis/Chronic Fatigue Syndrome: A Randomized, Double-Blind, Placebo-Controlled Trial. Ann Intern Med, 2019. 170(9): p. 585-593.
131. Salzberg, S.L., Open questions: How many genes do we have? BMC Biol, 2018. 16(1): p. 94.
132. Hellwege, J.N., et al., Population Stratification in Genetic Association Studies. Curr Protoc Hum Genet, 2017. 95: p. 1.22.1-1.22.23.
133. Auton, A., et al., A global reference for human genetic variation. Nature, 2015. 526(7571): p. 68-74.
134. Seyerle, A.A. and C.L. Avery, Genetic epidemiology: the potential benefits and challenges of using genetic information to improve human health. N C Med J, 2013. 74(6): p. 505-8.
84
135. OMIM.org. Online Mendelian Inheritance in Man. 2019 [cited 2019 November 15th]; Available from: https://www.omim.org/statistics/geneMap.
136. Lohmueller, K.E., et al., Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nat Genet, 2003. 33(2): p. 177-82.
137. De Re, V., R. Magris, and R. Cannizzaro, New Insights into the Pathogenesis of Celiac Disease. Front Med (Lausanne), 2017. 4: p. 137.
138. Khera, A.V., et al., Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat Genet, 2018. 50(9): p. 1219-1224.
139. Broad_Institute. Association Studies. 2017; Available from: https://www.broadinstitute.org/news/after-decade-genome-wide-association-studies-new-phase-discovery-pushes.
140. Nicholson, L.B., The immune system. Essays Biochem, 2016. 60(3): p. 275-301. 141. Caillat-Zucman, S., Molecular mechanisms of HLA association with autoimmune
diseases. Tissue Antigens, 2009. 73(1): p. 1-8. 142. Rose, N.R., Prediction and Prevention of Autoimmune Disease in the 21st Century: A
Review and Preview. Am J Epidemiol, 2016. 183(5): p. 403-6. 143. Cinoku, II, C.P. Mavragani, and H.M. Moutsopoulos, Atherosclerosis: Beyond the lipid
storage hypothesis. The role of autoimmunity. Eur J Clin Invest, 2020. 50(2): p. e13195.
144. Cooper, G.S., M.L. Bynum, and E.C. Somers, Recent insights in the epidemiology of autoimmune diseases: improved prevalence estimates and understanding of clustering of diseases. J Autoimmun, 2009. 33(3-4): p. 197-207.
145. Chalah, M.A. and S.S. Ayache, Is there a link between inflammation and fatigue in multiple sclerosis? J Inflamm Res, 2018. 11: p. 253-264.
146. Griggs, S. and N.S. Morris, Fatigue Among Adults With Type 1 Diabetes Mellitus and Implications for Self-Management: An Integrative Review. Diabetes Educ, 2018. 44(4): p. 325-339.
147. Nguyen, M.H., K. Bryant, and S.G. O'Neill, Vitamin D in SLE: a role in pathogenesis and fatigue? A review of the literature. Lupus, 2018. 27(13): p. 2003-2011.
148. Torres, J., et al., Crohn's disease. Lancet, 2017. 389(10080): p. 1741-1755. 149. Coleman, C., et al., Common polygenic variation in coeliac disease and confirmation
of ZNF335 and NIFA as disease susceptibility loci. Eur J Hum Genet, 2016. 24(2): p. 291-7.
150. Carapito, R., M. Radosavljevic, and S. Bahram, Next-Generation Sequencing of the HLA locus: Methods and impacts on HLA typing, population genetics and disease association studies. Hum Immunol, 2016. 77(11): p. 1016-1023.
151. Vandiedonck, C. and J.C. Knight, The human Major Histocompatibility Complex as a paradigm in genomics research. Brief Funct Genomic Proteomic, 2009. 8(5): p. 379-94.
152. Robinson, J., et al., The IPD and IMGT/HLA database: allele variant databases. Nucleic Acids Res, 2015. 43(Database issue): p. D423-31.
153. Robinson, J., et al., IMGT/HLA database--a sequence database for the human major histocompatibility complex. Tissue Antigens, 2000. 55(3): p. 280-7.
154. Matzaraki, V., et al., The MHC locus and genetic susceptibility to autoimmune and infectious diseases. Genome Biol, 2017. 18(1): p. 76.
85
155. System, T.W.N.C.f.F.o.t.H. Nomenclature for Factors of the HLA System. 2020 [cited 2020; Available from: http://hla.alleles.org/alleles/index.html.
156. Cohly, H.H. and A. Panja, Immunological findings in autism. Int Rev Neurobiol, 2005. 71: p. 317-41.
157. Raha-Chowdhury, R., S.R. Andrews, and J.R. Gruen, CAT 53: a protein phosphatase 1 nuclear targeting subunit encoded in the MHC Class I region strongly expressed in regions of the brain involved in memory, learning, and Alzheimer's disease. Brain Res Mol Brain Res, 2005. 138(1): p. 70-83.
158. Tamouza, R., et al., HLA genetics in bipolar disorder. Acta Psychiatr Scand, 2018. 138(5): p. 464-471.
159. Caillat-Zucman, S., New insights into the understanding of MHC associations with immune-mediated disorders. Hla, 2017. 89(1): p. 3-13.
160. Lokki, M.L. and R. Paakkanen, The complexity and diversity of major histocompatibility complex challenge disease association studies. Hla, 2019. 93(1): p. 3-15.
161. Ooi, J.D., et al., Dominant protection from HLA-linked autoimmunity by antigen-specific regulatory T cells. Nature, 2017.
162. Wang, Y., et al., Effects of glucocorticoids on osteoporosis in rheumatoid arthritis: a systematic review and meta-analysis. Osteoporos Int, 2020.
163. Sussman, M., et al., Cost-utility analyses of targeted immunomodulators in rheumatoid arthritis: systematic review. J Med Econ, 2020: p. 1.
164. The Norwegian bone marrow donor registry. Available from: https://oslo-universitetssykehus.no/fag-og-forskning/nasjonale-og-regionale-tjenester/det-norske-benmargsgiverregisteret.
165. Founti, P., et al., Biobanks and the importance of detailed phenotyping: a case study--the European Glaucoma Society GlaucoGENE project. Br J Ophthalmol, 2009. 93(5): p. 577-81.
166. Sunnquist, M., et al., A Comparison of Case Definitions for Myalgic Encephalomyelitis and Chronic Fatigue Syndrome. J Chronic Dis Manag, 2017. 2(2).
167. Maclachlan, L., et al., Are current chronic fatigue syndrome criteria diagnosing different disease phenotypes? PLoS One, 2017. 12(10): p. e0186885.
168. Meeus, M., et al., What is in a name? Comparing diagnostic criteria for chronic fatigue syndrome with or without fibromyalgia. Clin Rheumatol, 2016. 35(1): p. 191-203.
169. Brurberg, K.G., et al., Case definitions for chronic fatigue syndrome/myalgic encephalomyelitis (CFS/ME): a systematic review. BMJ Open, 2014. 4(2): p. e003973.
170. Jason, L.A., et al., Examining the Institute of Medicine's Recommendations Regarding Chronic Fatigue Syndrome: Clinical Versus Research Criteria. J Neurol Psychol, 2015. 2015(Suppl 2).
171. Stubhaug, B., et al., A 4-Day Mindfulness-Based Cognitive Behavioral Intervention Program for CFS/ME. An Open Study, With 1-Year Follow-Up. Front Psychiatry, 2018. 9: p. 720.
172. Jason, L.A., et al., Clinical Criteria Versus a Possible Research Case Definition in Chronic Fatigue Syndrome/Myalgic Encephalomyelitis. Fatigue, 2017. 5(2): p. 89-102.
173. Sharif, K., et al., On chronic fatigue syndrome and nosological categories. Clin Rheumatol, 2018. 37(5): p. 1161-1170.
86
174. Strand EB, L.K., Jason LA, Tveito K, Diep L M, Valla SS, Sunnquist M, Helland IB, Herder I, Dammen T. , Comparing the DePaul Symptom Questionnaire with physician assessments: A preliminary study. Fatigue: Biomedicine, Health & Behavior, 2016. 4: p. 52-62.
175. Mann, C.J., Observational research methods. Research design II: cohort, cross sectional, and case-control studies. Emerg Med J, 2003. 20(1): p. 54-60.
176. Skarpsno, E.S., Representative patient samples are unnecessary in clinical research. Tidsskr Nor Laegeforen, 2019. 139(4).
177. Helsedirektoratet. Veileder for transfusjonstjenesten i Norge utgave 7.3 2017. 2017 [cited 2020; Available from: https://www.helsedirektoratet.no/veiledere/transfusjonstjenesten-i-norge-utgave-73/Transfusjonstjenesten%20i%20Norge%20utgave%207.3%20–%20Veileder.pdf.
178. Kennedy, A.E., U. Ozbek, and M.T. Dorak, What has GWAS done for HLA and disease associations? Int J Immunogenet, 2017. 44(5): p. 195-211.
179. Thomas, E.D., Bone marrow transplantation: a review. Semin Hematol, 1999. 36(4 Suppl 7): p. 95-103.
180. Hollenbach, J.A., et al., Analytical methods for disease association studies with immunogenetic data. Methods Mol Biol, 2012. 882: p. 245-66.
181. Price, A.L., et al., New approaches to population stratification in genome-wide association studies. Nat Rev Genet, 2010. 11(7): p. 459-63.
182. Wacholder, S., N. Rothman, and N. Caporaso, Population stratification in epidemiologic studies of common genetic variants and cancer: quantification of bias. J Natl Cancer Inst, 2000. 92(14): p. 1151-8.
183. Cardon, L.R. and L.J. Palmer, Population stratification and spurious allelic association. Lancet, 2003. 361(9357): p. 598-604.
184. Campbell, C.D., et al., Demonstrating stratification in a European American population. Nat Genet, 2005. 37(8): p. 868-72.
185. Harbo, H.F., et al., Two genome-wide linkage disequilibrium screens in Scandinavian multiple sclerosis patients. J Neuroimmunol, 2003. 143(1-2): p. 101-6.
186. Balding, W.A.a.D.J., Population Structure and Cryptic Relatedness in Genetic Association Studies. Statistical Science, 2009. 24(4): p. 451-471.
187. Bakken, I.J., et al., Two age peaks in the incidence of chronic fatigue syndrome/myalgic encephalomyelitis: a population-based registry study from Norway 2008-2012. BMC Med, 2014. 12: p. 167.
188. Ioannidis, J.P. and F.B. Karassa, Bias in uncontrolled therapeutic trials in rheumatology due to selection of populations with extreme characteristics. J Rheumatol, 2001. 28(8): p. 1881-7.
189. Backmann, M., What's in a gold standard? In defence of randomised controlled trials. Med Health Care Philos, 2017. 20(4): p. 513-523.
190. Schulz, K.F. and D.A. Grimes, Blinding in randomised trials: hiding who got what. Lancet, 2002. 359(9307): p. 696-700.
191. Berger, A., HLA typing. Bmj, 2001. 322(7280): p. 218. 192. Dunckley, H., HLA typing by SSO and SSP methods. Methods Mol Biol, 2012. 882: p. 9-
25. 193. Erlich, R.L., et al., Next-generation sequencing for HLA typing of class I loci. BMC
Genomics, 2011. 12: p. 42.
87
194. Kishore, A. and M. Petrek, Next-Generation Sequencing Based HLA Typing: Deciphering Immunogenetic Aspects of Sarcoidosis. Front Genet, 2018. 9: p. 503.
195. Profaizer, T. and A. Kumanovics, Human Leukocyte Antigen Typing by Next-Generation Sequencing. Clin Lab Med, 2018. 38(4): p. 565-578.
196. Cereb, N., et al., Advances in DNA sequencing technologies for high resolution HLA typing. Hum Immunol, 2015. 76(12): p. 923-7.
197. Michalik, J., et al., The association of HLA-DRB1 and HLA-DQB1 alleles with genetic susceptibility to multiple sclerosis in the Slovak population. Neurol Res, 2015. 37(12): p. 1060-7.
198. Robinson, J., et al., IPD-IMGT/HLA Database. Nucleic Acids Res, 2020. 48(D1): p. D948-d955.
199. Lange, V., et al., Cost-efficient high-throughput HLA typing by MiSeq amplicon sequencing. BMC Genomics, 2014. 15: p. 63.
200. Wieczorek, M., et al., Major Histocompatibility Complex (MHC) Class I and MHC Class II Proteins: Conformational Plasticity in Antigen Presentation. Front Immunol, 2017. 8: p. 292.
201. van Deutekom, H.W. and C. Kesmir, Zooming into the binding groove of HLA molecules: which positions and which substitutions change peptide binding most? Immunogenetics, 2015. 67(8): p. 425-36.
202. hla.alleles.org. Nomenclature for Factors of the HLA System. [cited 2020; Available from: http://hla.alleles.org/alleles/g_groups.html.
203. Adams, S.D., et al., Ambiguous allele combinations in HLA Class I and Class II sequence-based typing: when precise nucleotide sequencing leads to imprecise allele identification. J Transl Med, 2004. 2(1): p. 30.
204. Elsner, H.A. and R. Blasczyk, Immunogenetics of HLA null alleles: implications for blood stem cell transplantation. Tissue Antigens, 2004. 64(6): p. 687-95.
205. Henry, J., et al., Frequency of Class I common or well documented null alleles in National Marrow Donor Program high resolution typing programs. Human Immunology, 2015. 76: p. 133.
206. Lancaster, A.K., et al., PyPop update--a software pipeline for large-scale multilocus population genomics. Tissue Antigens, 2007. 69 Suppl 1: p. 192-7.
207. Excoffier, L. and H.E. Lischer, Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour, 2010. 10(3): p. 564-7.
208. Dudbridge, F., Pedigree disequilibrium tests for multilocus haplotypes. Genet Epidemiol, 2003. 25(2): p. 115-21.
209. Excoffier, L. and M. Slatkin, Maximum-likelihood estimation of molecular haplotype frequencies in a diploid population. Mol Biol Evol, 1995. 12(5): p. 921-7.
210. Single RM, G.P.-A., Lancaster AK, Briggs A, Barcellos LF, Hollenbach JA, Mack SJ, Thomson G, Haplotype Estimation and Linkage Disequilibrium Methods Manual: Version 0.1.8 (June 3, 2011) Estimating Haplotype Frequencies and Linkage Disequilibrium Parameters in the HLA and KIR Regions, in Available online at: www.ImmPort.org. 2011.
211. Slatkin, M., Linkage disequilibrium--understanding the evolutionary past and mapping the medical future. Nat Rev Genet, 2008. 9(6): p. 477-85.
212. Hedrick, P.W., Gametic disequilibrium measures: proceed with caution. Genetics, 1987. 117(2): p. 331-41.
88
213. Whelan, B.M.M., A.B., The “Null Hypothesis” of Precision Agriculture Management. Precision Agriculture - Kluwer Academic Publishers, 2000. 2(3): p. 265–279.
214. Dumas-Mallet, E., et al., Low statistical power in biomedical science: a review of three human research domains. R Soc Open Sci, 2017. 4(2): p. 160254.
215. Shi, Q., E.S. Pavey, and R.E. Carter, Bonferroni-based correction factor for multiple, correlated endpoints. Pharm Stat, 2012. 11(4): p. 300-9.
216. Benjamini, Y.H., Y., Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society. Series B (Methodological), 1995. 57(1): p. 289-300.
217. Jung, S.H., Stratified Fisher's exact test and its sample size calculation. Biom J, 2014. 56(1): p. 129-40.
218. Karfakis, N., The biopolitics of CFS/ME. Stud Hist Philos Biol Biomed Sci, 2018. 70: p. 20-28.
219. Ulvestad, E., Chronic fatigue syndrome defies the mind-body-schism of medicine. Medicine, Health Care and Philosophy, 2008. 11(3): p. 285.
220. Hollenbach, J.A. and J.R. Oksenberg, The immunogenetics of multiple sclerosis: A comprehensive review. J Autoimmun, 2015. 64: p. 13-25.
221. Raychaudhuri, S., et al., Five amino acids in three HLA proteins explain most of the association between MHC and seropositive rheumatoid arthritis. Nat Genet, 2012. 44(3): p. 291-6.
222. UK_Biobank. 2018 [cited 2018 September]; Available from: https://biobankengine.stanford.edu.
223. Tafti, M., et al., DQB1 locus alone explains most of the risk and protection in narcolepsy with cataplexy in Europe. Sleep, 2014. 37(1): p. 19-25.
224. Candore, G., et al., Pathogenesis of autoimmune diseases associated with 8.1 ancestral haplotype: effect of multiple gene interactions. Autoimmun Rev, 2002. 1(1-2): p. 29-35.
225. Sandoughi, M., et al., Frequency of HLA-DRB1 alleles in rheumatoid arthritis patients in Zahedan, southeast Iran. Ann Saudi Med, 2011. 31(2): p. 171-3.
226. Ashton, J.J., et al., Review article: the genetics of the human leucocyte antigen region in inflammatory bowel disease. Aliment Pharmacol Ther, 2019. 50(8): p. 885-900.
227. Patsopoulos, N.A., et al., Fine-mapping the genetic association of the major histocompatibility complex in multiple sclerosis: HLA and non-HLA effects. PLoS Genet, 2013. 9(11): p. e1003926.
228. Eijkelkamp, E.J., T.R. Yapp, and L.W. Powell, HFE-associated hereditary hemochromatosis. Can J Gastroenterol, 2000. 14(2): p. 121-5.
229. Health, N.I.o.P. Report: “Psykisk helse i Norge”. 2018 [cited 2020; Available from: https://www.fhi.no/publ/2018/psykisk-helse-i-norge/.
230. Glanville, K.P., et al., Classical Human Leukocyte Antigen Alleles and C4 Haplotypes Are Not Significantly Associated With Depression. Biol Psychiatry, 2019.
231. Branco, J.C., et al., HLA Studies in Fibromyalgia. Journal of Musculoskeletal Pain, 1996. 4(3): p. 21-27.
232. Yunus, M.B., Genetic factors in fibromyalgia syndrome. Zeitschrift für Rheumatologie, 1998. 57(2): p. S61-S62.
233. Fu, L.Y., et al., Association of HLA-DQB1*03:03 with pityriasis rosea in Chinese patients. Clin Exp Dermatol, 2018. 43(4): p. 389-392.
89
234. Ozawa, A., et al., HLA class I and II alleles and susceptibility to generalized pustular psoriasis: significant associations with HLA-Cw1 and HLA-DQB1*0303. J Dermatol, 1998. 25(9): p. 573-81.
235. Haapasalo, K., et al., The Psoriasis Risk Allele HLA-C*06:02 Shows Evidence of Association with Chronic or Recurrent Streptococcal Tonsillitis. Infect Immun, 2018. 86(10).
236. Stawczyk-Macieja, M., et al., ERAP1 and HLA-C*06 are strongly associated with the risk of psoriasis in the population of northern Poland. Postepy Dermatol Alergol, 2018. 35(3): p. 286-292.
237. Laake, K., et al., Characterization of ATM mutations in 41 Nordic families with ataxia telangiectasia. Hum Mutat, 2000. 16(3): p. 232-46.
238. Heramb, C., et al., BRCA1 and BRCA2 mutation spectrum - an update on mutation distribution in a large cancer genetics clinic in Norway. Hered Cancer Clin Pract, 2018. 16: p. 3.
239. Bihrmann, K., et al., Small-scale geographical variation in multiple sclerosis: A case-control study using Danish register data 1971-2013. Mult Scler Relat Disord, 2018. 23: p. 40-45.
240. Iwaszko, M., et al., Polymorphisms within the human leucocyte antigen-E gene and their associations with susceptibility to rheumatoid arthritis as well as clinical outcome of anti-tumour necrosis factor therapy. Clin Exp Immunol, 2015. 182(3): p. 270-7.
241. Zhu, X., et al., Human leukocyte antigen class I and class II genes polymorphisms might be associated with interferon alpha therapy efficiency of chronic hepatitis B. Antiviral Res, 2011. 89(3): p. 189-92.
242. Hall, A.G. and M.J. Tilby, Mechanisms of action of, and modes of resistance to, alkylating agents used in the treatment of haematological malignancies. Blood Rev, 1992. 6(3): p. 163-73.
243. Ahlmann, M. and G. Hempel, The effect of cyclophosphamide on the immune system: implications for clinical cancer therapy. Cancer Chemother Pharmacol, 2016. 78(4): p. 661-71.
244. Bar, F., et al., Cyclophosphamide Pulse Therapy in Severe Refractory Crohn's Disease: A Retrospective Multicenter Case Series. Inflamm Intest Dis, 2018. 2(3): p. 139-146.
245. Teles, K.A., et al., Cyclophosphamide administration routine in autoimmune rheumatic diseases: a review. Rev Bras Reumatol Engl Ed, 2017. 57(6): p. 596-604.
246. Bär, F., et al., Cyclophosphamide Pulse Therapy in Severe Refractory Crohn's Disease: A Retrospective Multicenter Case Series. Inflamm Intest Dis, 2018. 2(3): p. 139-146.
247. Barta, Z., L. Tóth, and M. Zeher, Pulse cyclophosphamide therapy for inflammatory bowel disease. World J Gastroenterol, 2006. 12(8): p. 1278-80.
248. Ogino, M.H. and P. Tadi, Cyclophosphamide, in StatPearls. 2020, StatPearls Publishing LLC.: Treasure Island (FL).
249. Nishikai, M., et al., Autoantibodies to a 68/48 kDa protein in chronic fatigue syndrome and primary fibromyalgia: a possible marker for hypersomnia and cognitive disorders. Rheumatology (Oxford), 2001. 40(7): p. 806-10.
250. de Vega, W.C., et al., Epigenetic modifications and glucocorticoid sensitivity in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS). BMC Med Genomics, 2017. 10(1): p. 11.
90
251. Kerr, J.R., et al., Seven genomic subtypes of chronic fatigue syndrome/myalgic encephalomyelitis: a detailed analysis of gene networks and clinical phenotypes. J Clin Pathol, 2008. 61(6): p. 730-9.
252. de Vega, W.C., et al., Integration of DNA methylation & health scores identifies subtypes in myalgic encephalomyelitis/chronic fatigue syndrome. Epigenomics, 2018. 10(5): p. 539-557.
253. Marquez, A., et al., Meta-analysis of Immunochip data of four autoimmune diseases reveals novel single-disease and cross-phenotype associations. Genome Med, 2018. 10(1): p. 97.
254. Cárdenas-Roldán, J., A. Rojas-Villarraga, and J.M. Anaya, How do autoimmune diseases cluster in families? A systematic review and meta-analysis. BMC Med, 2013. 11: p. 73.
255. Chang, D.-y., et al., Association of HLA Genes with Clinical Outcomes of ANCA-Associated Vasculitis. Clinical Journal of the American Society of Nephrology, 2012. 7(8): p. 1293-1299.
256. Dand, N., et al., HLA-C*06:02 genotype is a predictive biomarker of biologic treatment response in psoriasis. J Allergy Clin Immunol, 2019. 143(6): p. 2120-2130.
257. Rose, N.R. and C. Bona, Defining criteria for autoimmune diseases (Witebsky's postulates revisited). Immunol Today, 1993. 14(9): p. 426-30.
258. Wanleenuwat, P. and P. Iwanowski, Role of B cells and antibodies in multiple sclerosis. Mult Scler Relat Disord, 2019. 36: p. 101416.
259. Seow, C.H., et al., Novel anti-glycan antibodies related to inflammatory bowel disease diagnosis and phenotype. Am J Gastroenterol, 2009. 104(6): p. 1426-34.
260. Dore, M.P., et al., Clinical and genetic profile of patients with seronegative coeliac disease: the natural history and response to gluten-free diet. BMJ Open Gastroenterol, 2017. 4(1): p. e000159.
261. Hardy, M.Y. and J.A. Tye-Din, Coeliac disease: a unique model for investigating broken tolerance in autoimmunity. Clin Transl Immunology, 2016. 5(11): p. e112.
262. Phelps, R.G. and A.J. Rees, The HLA complex in Goodpasture's disease: a model for analyzing susceptibility to autoimmunity. Kidney Int, 1999. 56(5): p. 1638-53.
263. Mandarano, A.H., et al., Myalgic encephalomyelitis/chronic fatigue syndrome patients exhibit altered T cell metabolism and cytokine associations. J Clin Invest, 2020. 130(3): p. 1491-1505.
264. Middleton, D. and F. Gonzelez, The extensive polymorphism of KIR genes. Immunology, 2010. 129(1): p. 8-19.
265. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature, 2007. 447(7145): p. 661-78.
266. De Virgilio, A., et al., Parkinson's disease: Autoimmunity and neuroinflammation. Autoimmun Rev, 2016. 15(10): p. 1005-11.
267. Lincoln, M.R., et al., Epistasis among HLA-DRB1, HLA-DQA1, and HLA-DQB1 loci determines multiple sclerosis susceptibility. Proc Natl Acad Sci U S A, 2009. 106(18): p. 7542-7.
268. Skinningsrud, B., et al., Multiple loci in the HLA complex are associated with Addison's disease. J Clin Endocrinol Metab, 2011. 96(10): p. E1703-8.
269. Mallon, E., et al., HLA-CW*0602 is a susceptibility factor in type I psoriasis, and evidence Ala-73 is increased in male type I psoriatics. J Invest Dermatol, 1997. 109(2): p. 183-6.
91
270. Morris, D.L., et al., Unraveling multiple MHC gene associations with systemic lupus erythematosus: model choice indicates a role for HLA alleles and non-HLA genes in Europeans. Am J Hum Genet, 2012. 91(5): p. 778-93.
271. Fernando, M.M., et al., Identification of two independent risk factors for lupus within the MHC in United Kingdom families. PLoS Genet, 2007. 3(11): p. e192.
272. Hamzeh, A.R., et al., Association of HLA-DQA1 and -DQB1 alleles with type I diabetes in Arabs: a meta-analyses. Tissue Antigens, 2015. 86(1): p. 21-7.
273. Popperud, T.H., et al., Juvenile myasthenia gravis in Norway: HLA-DRB1*04:04 is positively associated with prepubertal onset. PLoS One, 2017. 12(10): p. e0186383.
274. Maniaol, A., et al., Late Onset Myasthenia Gravis Is Associated with HLA DRB1*15:01 in the Norwegian Population. PLoS One, 2012. 7(5).
275. Gutierrez-Achury, J., et al., Fine mapping in the MHC region accounts for 18% additional genetic risk for celiac disease. Nat Genet, 2015. 47(6): p. 577-8.
II
PAPER II
1Scientific RepoRtS | (2020) 10:5267 | https://doi.org/10.1038/s41598-020-62157-x
www.nature.com/scientificreports
Human Leukocyte Antigen alleles associated with Myalgic encephalomyelitis/chronic fatigue Syndrome (Me/cfS)Asgeir Lande1,2*, Øystein fluge3, Elin B. Strand4,5, Siri T. flåm1, Daysi D. Sosa6, Olav Mella3, torstein egeland2,7, Ola D. Saugstad8, Benedicte A. Lie1,2,7 & Marte K. Viken1,7
the etiology and pathogenesis of Myalgic encephalomyelitis/chronic fatigue Syndrome (Me/cfS) are unknown, and autoimmunity is one of many proposed underlying mechanisms. Human Leukocyte Antigen (HLA) associations are hallmarks of autoimmune disease, and have not been thoroughly investigated in a large ME/CFS patient cohort. We performed high resolution HLA -A, -B, -C, -DRB1, -DQB1 and -DPB1 genotyping by next generation sequencing in 426 adult, Norwegian ME/CFS patients, diagnosed according to the Canadian Consensus Criteria. HLA associations were assessed by comparing to 4511 healthy and ethnically matched controls. Clinical information was collected through questionnaires completed by patients or relatives. We discovered two independent HLA associations, tagged by the alleles HLA-C*07:04 (OR 2.1 [95% CI 1.4–3.1]) and HLA-DQB1*03:03 (OR 1.5 [95% CI 1.1–2.0]). These alleles were carried by 7.7% and 12.7% of ME/CFS patients, respectively. The proportion of individuals carrying one or both of these alleles was 19.2% in the patient group and 12.2% in the control group (OR 1.7 [95% CI 1.3–2.2], pnc = 0.00003). ME/CFS is a complex disease, potentially with a substantial heterogeneity. We report novel HLA associations pointing toward the involvement of the immune system in ME/CFS pathogenesis.
Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is a disabling disorder characterized by med-ically unexplained fatigue, post-exertional malaise and a variety of additional symptoms, such as chronic pain, sleep disturbances and cognitive difficulties. ME/CFS is diagnosed on clinical grounds alone, and different sets of criteria specify the mandatory symptoms as well as recommendations for the exclusion of differential diagno-ses1–3. The specificity and validity of different diagnostic criteria have been questioned, yet there is no agreement on the level of heterogeneity in ME/CFS, and there is no consensus on how to categorize different subgroups4–8.
The pathogenesis and etiology of ME/CFS are unknown, with several models having been proposed9. One central hypothesis states that autoimmunity is part of the pathophysiology10,11. ME/CFS has been reported to be partly heritable12,13, consistent with a multifactorial etiology dependent on both genetic and environmental fac-tors. This is the prevailing model for a vast number of diseases, including established autoimmune diseases (AID). Several publications report immunological alterations among ME/CFS patients, e.g. changes in natural killer (NK) cell function14,15, cytokine levels16,17, and DNA methylation patterns consistent with immune dysregula-tion18. Some of these findings have failed to reproduce in other studies, which could be due to differences in meth-odology, the complexity and heterogeneity of ME/CFS, and lack of power due to small sample sizes19. Resultingly, the autoimmunity hypothesis warrants further evaluation. A characteristic feature of AID is genetic association with certain human leukocyte antigen (HLA) alleles20. Thus, a thorough investigation of HLA associations in ME/CFS is relevant, although HLA associations per se cannot be used as evidence regarding disease etiology21. Studies
1Department of Medical Genetics, University of Oslo and Oslo University Hospital, Oslo, Norway. 2Institute of Clinical Medicine, University of Oslo, Oslo, Norway. 3Department of Oncology and Medical Physics, Haukeland University Hospital and Department of Clinical Science, University of Bergen, Bergen, Norway. 4National Advisory Unit on CFS/ME, Oslo University Hospital, Oslo, Norway. 5Faculty of Health Science, VID Specialized University, Stavanger, Norway. 6CFS/ME Center, Oslo University Hospital, Oslo, Norway. 7Department of Immunology, Oslo University Hospital, Oslo, Norway. 8Department of Pediatric Research, Oslo University Hospital, University of Oslo, Oslo, Norway. *email: [email protected]
open
2Scientific RepoRtS | (2020) 10:5267 | https://doi.org/10.1038/s41598-020-62157-x
www.nature.com/scientificreportswww.nature.com/scientificreports/
of HLA associations in CFS have been published, but with great variation in patient inclusion criteria and HLA typing methodology22–30. No reproducible, significant associations are evident across these studies. In the largest study, including 110 patients, the strongest significant association was with HLA-DQ3 with an odds ratio (OR) of 1.8 (95% CI 1.2–2.8)22. Associations with HLA alleles DQA1*01, DRB1*13:01 and DQB1*06:02 have also been reported25,27,29. The great majority of these studies include less than 50 patients, and are underpowered for the detection of moderate to weak associations. Hence, in this study we aimed to conduct a comprehensive investiga-tion of HLA associations in a large ME/CFS cohort, applying modern, high resolution HLA typing.
ResultsCharacterization of patient and control groups. We included 426 ME/CFS patients and 4511 healthy, ethnically matched controls. All patients had been diagnosed in Norway according to the 2003 Canadian Consensus Criteria2, except for four patients where the similarly strict 2010 International Consensus Criteria3 were applied. Demographic and clinical characteristics of patients and controls are shown in Table 1. The mean age at diagnosis for ME/CFS patients was 34.7 years, 82.8% were female, and most patients (45.5%) had a disease duration of 5–10 years, from symptom debut to inclusion. 12.5% of the patients had severe or very severe disease (bedridden). An additional 28.6% had moderate to severe disease (strictly housebound). A total of 41.1% of ME/CFS patients were bed- or housebound, and 86.8% of patients were unable to work full or part time the previous 6 months.
HLA alleles associated with ME/CFS. In all patients and controls, we obtained 2nd field resolution geno-types of HLA class I genes HLA -A, -B and -C and class II genes HLA -DRB1, -DQB1 and -DPB1. This resolution distinguishes HLA alleles that encode specific HLA proteins. No significant deviations from Hardy-Weinberg equilibrium were noted at any HLA loci, neither in the patient group nor in the control group (Supplementary Table S1). Allele frequencies for all observed HLA Class I and Class II alleles are presented in Supplementary Table S2. Global association tests for each HLA locus (Supplementary Table S3) were significant for HLA-C (p = 0.04) and HLA-DQB1 (p = 0.04). When comparing individual allele frequencies between patients and con-trols, four HLA risk alleles emerged (Table 2): C*07:04 (OR = 2.1 [95% CI 1.4–3.1], pnc = 0.0001, pc = 0.001), B*57:01 (OR = 1.6 [95% CI 1.2–2.3], pnc = 0.004, pc < 0.05), DQB1*03:03 (OR = 1.5 [95% CI 1.1–2.0], pnc = 0.005, pc < 0.05) and B*44:02 (OR = 1.3 [95% CI 1.0–1.6], pnc = 0.03, pc = n.s.). In order to evaluate the dependency of these associations, we measured the degree of linkage disequilibrium (LD) between the four alleles within the patient group (Fig. 1). Strong LD was observed between C*07:04 and B*44:02 (D’ = 0.90) as well as between B*57:01 and DQB1*03:03 (D’ = 0.69), indicating that these alleles may occur on two distinct haplotypes. The first haplotype, C*07:04 - B*44:02, had an estimated frequency of 3.5% in the patient group and 1.7% in the control group, resulting in an OR of 2.1 (95% CI 1.4–3.1, pnc = 0.0002). The second haplotype, B*57:01 - DQB1*03:03, had an estimated frequency of 3.3% in the patient group and 2.0% in the control group, resulting in an OR of 1.7 (95% CI 1.1–2.5 pnc = 0.01). To further evaluate which allele on each of the two haplotypes represents the
Patients Controls
Mean age, years (min, max)a,b 39.5 (17, 79) 30.6 (19, 52)
Percentage femalesa,c 82.8 59.8
Mean age at diagnosis, years (min, max)d 34.7 (8, 65) —
Disease duration, percentage of patientse
1–5 years 18.1 —
5–10 years 45.5 —
10–15 years 26.4 —
>15 years 10.1 —
Disease severity, measured by activity assessment in DSQ79*, percentage of patientsf
Cat.1: Bedridden 12.5 —
Cat.2: Strictly housebound 28.6 —
Cat.3: Light housework 45.7 —
Cat.4: Able to work part time 12.2 —
Cat.5: Able to work full time 0.8 —
Cat.6: Handling some family obligations 0.3 —
Cat.7: Handling work and family obligations 0 —
Table 1. Demographic and clinical characteristics of ME/CFS patients and healthy controls. aValid number of controls: 4510. Valid number of patients: b426, c424, d345, e288, f385. *DePaul Symptom Questionnaire, question no. 7949,50: Cat.1: I am not able to work or do anything, and I am bedridden. Cat.2: I can walk around the house, but I cannot do light housework. Cat.3: I can do light housework, but I cannot work part-time. Cat.4: I can only work part time at work or on some family responsibilities. Cat.5: I can work full time, but I have no energy left for anything else. Cat.6: I can work full time and finish some family responsibilities but I have no energy left for anything else. Cat.7: I can do all work or family responsibilities without any problems with my energy.
3Scientific RepoRtS | (2020) 10:5267 | https://doi.org/10.1038/s41598-020-62157-x
www.nature.com/scientificreportswww.nature.com/scientificreports/
most significant association, we performed Svejgaard analyses between C*07:04 and B*44:02 as well as between B*57:01 and DQB1*03:03 (Supplementary Data Sheet S4). None of the two alleles on either of the two haplotypes reached significance when testing their independent association, which is not surprising due to the strong LD mentioned above. We report C*07:04 and DQB1*03:03 as tag alleles for the ME/CFS associations, since these alleles occur at the two loci initially showing global association.
We next wanted to make sure that these HLA associations were not due to gender differences between the cases and controls (82.8% vs 59.8% females, respectively). No significant gender differences were observed between the carrier frequencies of these alleles in either cases (Table 3) or controls (C*07:04 had a carrier fre-quency of 3.7% in females and 4.0% in males; p = 0.6; DQB1*03:03 had a carrier frequency of 9.0% in females and 8.3% in males; p = 0.4). Furthermore, after stratifying cases and controls according to gender, heterogeneity was rejected (p > 0.5) between the OR values obtained for females only and males only, indicating no gender differences between the HLA associations observed in ME/CFS.
The two loci HLA-DRB1 and HLA-DQB1 are physically close, and exhibit particularly strong LD. In our data-set, DQB1*03:03 occurred most frequently with DRB1*07:01. The haplotype DRB1*07:01 - DQB1*03:03 had estimated frequencies of 4.7% and 2.9% in the patient and control group, respectively (OR 1.7 [95% CI 1.2–2.3], pnc = 0.003). Among patients, this haplotype was without exception estimated to carry DQA1*02:01. Genotype data for the HLA-DQA1 locus was only obtained for patients, and was not available for the controls.
There were two alleles with a negative association with ME/CFS, suggesting a potential protection, namely B*08:01 (OR = 0.7 [95% CI 0.6–0.9], pnc = 0.01, pc = n.s.) and DPB1*02:01 (OR = 0.7 [95% CI 0.6–0.9], pnc = 0.02, pc = n.s.) (Table 2). These alleles were not in LD (D’ = −0.29), indicating that the associations are independent. The most frequent B*08:01 haplotype in Norway is the highly conserved so-called autoimmune and ancestral AH8.1 haplotype31 (C*07:01-B*08:01-DRB1*03:01-DQB1*02:01). This haplotype had reduced estimated fre-quency in the patient group compared to the control group (8.2% vs. 10.3%, OR = 0.8, pnc = 0.06), albeit not significantly.
HLA risk allele carriers and clinical characteristics. The proportion of individuals carrying the allele C*07:04 was 7.7% in the patient group and 3.8% in the control group, while 12.7% of the patients and 8.7% of the controls carried DQB1*03:03 (Supplementary Data Sheet S4). The proportion of individuals carrying one or both of the two alleles was 19.2% in the patient group and 12.2% in the control group (OR 1.7, pnc = 0.00003, 95% CI[1.3–2.2]). Table 3 shows the distribution of clinical characteristics in the patient group, including stratification for C*07:04 and/or DQB1*03:03. Neither gender, initiating events, comorbidity of depression or fibromyalgia, nor AID or ME/CFS among 1st degree relatives were associated with the risk alleles. However, ME/CFS patients carrying one or both of the risk alleles had a significantly higher proportion of comorbid AID (OR = 2.3 [95% CI 1.2–4.3], pnc = 0.01). The frequency of comorbid AID was significantly increased also when stratifying for
HLA allele ME/CFS, n (%) Controls, n (%) OR (95% CI) pnc pc*C*07:04 33 (3.9) 172 (1.9) 2.1 (1.4–3.1) 0.0001 0.001
B*57:01 39 (4.6) 259 (2.9) 1.6 (1.2–2.3) 0.004 <0.05
DQB1*03:03 56 (6.6) 406 (4.5) 1.5 (1.1–2.0) 0.005 <0.05
B*44:02 105 (12.3) 904 (10.0) 1.3 (1.0–1.6) 0.03 0.3
B*08:01 83 (9.7) 1146 (12.7) 0.7 (0.6–0.9) 0.01 0.1
DPB1*02:01 78 (9.2) 1067 (11.8) 0.7 (0.6–1.0) 0.02 0.1
Table 2. HLA alleles showing association (pnc < 0.05) with ME/CFS in 426 patients and 4511 healthy controls. n = number of alleles. *Calculated by locus-wise Bonferroni multiple test correction.
Figure 1. Illustration of Linkage disequilibrium (LD) calculated as D’ values in the patient group between the four alleles with a significant association to ME/CFS. D’ measures the strength of LD between two alleles, and ranges from −1 to 0 for negative LD, and from 0 to 1 for positive LD. The number in each square is the D’ value between the two alleles listed diagonally above the square. Blue colors indicate the strength of LD, with darker colors for stronger LD.
4Scientific RepoRtS | (2020) 10:5267 | https://doi.org/10.1038/s41598-020-62157-x
www.nature.com/scientificreportswww.nature.com/scientificreports/
C*07:04 alone (OR = 2.9 [95% CI 1.2–6.6], pnc = 0.01), but not when stratifying for DQB1*03:03 alone (OR = 1.6 [95% CI 0.8–3.4], pnc = n.s.). These patients, carrying HLA risk alleles, had the following AID, ordered by fre-quency: Hashimoto’s thyreoiditis/hypothyreosis, psoriasis, rheumatoid arthritis, alopecia areata and Crohn’s dis-ease or ulcerative colitis.
DiscussionIn this project, we performed high resolution HLA genotyping by next generation sequencing (NGS) in 426 adult, Norwegian ME/CFS patients, diagnosed according to the Canadian Consensus Criteria2. There are no previous publications with comprehensive HLA genotyping by NGS in this patient group. We discovered two independent HLA associations, tagged by the alleles HLA-C*07:04 and HLA-DQB1*03:03.
To our knowledge, associations with HLA-C alleles have not previously been studied in ME/CFS. In 1994, Keller et al. performed serologic HLA-DR and DQ typing in 110 patients with Chronic fatigue immune dysfunc-tion syndrome (CFIDS)22. The patients were diagnosed with the Holmes Criteria32, and CFIDS was defined as a subgroup with positive findings in viral reactivation patterns and B- and T-cell tests, indicating post-infectious debut and a certain degree of immune dysfunction. The authors found a significant association (OR = 1.8) with the serotype HLA-DQ3. Serologic HLA typing is of low resolution compared to genetic typing33. HLA-DQ3 cor-responds to HLA-DQB1*03 in genetic nomenclature, where DQB1*03:03 is one of the three largest subgroups. Higher resolution HLA-DQB1 typing have been performed in two smaller cohorts (<58 patients), and even though statistically not significant, DQB1*03:03 was observed slightly more frequent among CFS patients, diag-nosed with the Fukuda criteria, than among controls23,25. Hence, the findings in existing literature is compatible with the association between ME/CFS and DQB1*03:03 in our material.
HLA-B*08:01 showed reduced frequency in ME/CFS compared to controls in our material. This allele most often occur on the haplotype C*07:01-B*08:01-DRB1*03:01-DQB1*02:01, which was also less prevalent among ME/CFS patients in our material. This ancestral haplotype, AH8.1, is a risk factor for a wide variety of AID, including myasthenia gravis, systemic lupus erythematosus and coeliac disease31, but protective against rheuma-toid arthritis34,35. In the existing literature on HLA and CFS, HLA-DRB1 is the locus most frequently studied. In four out of five studies, the frequency of DR3/DRB1*03 was lower in the patient group23–25,28, while in the fifth study the frequency was similar in both groups22. Hence, this haplotype seems truly less prevalent among ME/CFS patients.
Some HLA associations previously reported in CFS are not supported by our results27,29. The often cited association with DQA1*01 reported by Smith et al.25 cannot be evaluated in our material since the HLA-DQA1 locus was not genotyped in the control group. In our patient group, DQA1*01 occurred on haplotype with the following DQB1 alleles, ordered by frequency: DQB1*06:02, DQB1*05:01, DQB1*06:03, DQB1*06:04 and DQB1*05:03, and neither of these (pnc > 0.1, Supplementary Table S2), nor all combined (OR = 1.0, pnc = 0.6) were associated with ME/CFS.
The present HLA study in ME/CFS is to our knowledge the largest performed to date (other studies comprise ≤ 110 CFS patients). Our study had 80% power to discover HLA-associations with OR ≥ 1.5 given an allele fre-quency > 0.05. Interestingly, both C*07:04 and DQB1*03:03 remained significant after Bonferroni multiple test correction. Notably, we performed locus-wise multiple test correction, i.e. correcting for the number of alleles tested at each locus, since alleles at different HLA loci are in strong LD, and therefore do not represent independ-ent tests. The Bonferroni method is considered a strict multiple test correction36, but on the other hand locus-wise correction does not take into account the lack of complete LD between the investigated loci. Taken together, our
N (valid) TotalDQB1* 03:03 +
DQB1* 03:03 - pnc
OR (95% CI) C* 07:04 +
C* 07:04 - pnc
OR (95% CI)
C*07:04 + and/or DQB1* 03:03 +
C*07:04 - and DQB1* 03:03 - pnc
OR (95% CI)
Gender: Female 424 351 (82.8%)
44 (81.5%)
307 (83.0%) 0.7 0.9
(0.4–1.8) 25 (75.8%) 326 (83.0%) 0.2 0.6
(0.3–1.3)65 (79.3%)
286 (83.1%) 0.3 0.7
(0.4–1.3)
Symptom start following infectious episode
408 297 (72.8%)
41 (80.4%)
256 (71.7%) 0.2 1.6
(0.8–3.4) 18 (60.0%) 279 (73.8%) 0.1 0.5
(0.2–1.1)56 (73.7%)
241 (72.6%) 0.8 1.0
(0.6–1.9)
Symptom start following vaccination 407 64
(15.7%) 5 (9.8%) 59 (16.6%) 0.2 0.6
(0.2–1.4) 5 (16.7%) 59 (15.6%) 0.9 1.1
(0.4–3.0) 9 (11.8%) 55 (16.6%) 0.3 0.7
(0.3–1.4)
Previous or current moderate to severe depressive symptoms
407 39 (9.6%) 5 (9.8%) 34 (9.6%) 1.0 1.0 (0.4–2.8) 2 (6.7%) 37 (9.8%) 0.6 0.7
(0.2–2.9) 7 (9.2%) 32 (9.7%) 0.9 0.9 (0.4–2.2)
Comorbidity: Fibromyalgia 408 34 (8.3%) 2 (3.9%) 32 (9.0%) 0.2 0.4
(0.1–1.8) 3 (10.0%) 31 (8.2%) 0.7 1.2 (0.4–4.3) 5 (6.6%) 29 (8.7%) 0.5 0.7
(0.3–2.0)
Comorbidity: Autoimmune disease 405 58
(14.3%)10 (20.0%)
48 (13.5%) 0.2 1.6
(0.8–3.4) 9 (30.0%) 49 (13.1%) 0.01 2.9
(1.2–6.6)18 (24.0%)
40 (12.1%) 0.01 2.3
(1.2–4.3)
1st-degree relative with autoimmune disease
408 154 (37.7%)
24 (47.1%)
130 (36.4%) 0.1 1.5
(0.9–2.8) 7 (23.3%) 147 (38.9%) 0.1 0.5
(0.2–1.1)30 (39.5%)
124 (37.3%) 0.7 1.1
(0.7–1.8)
1st-degree relative with CFS/ME 403 59
(14.6%)11 (22.0%)
48 (13.6%) 0.1 1.8
(0.9–3.7) 3 (10.0%) 56 (15.0%) 0.5 0.6
(0.2–2.1)13 (17.3%)
46 (14.0%) 0.5 1.3
(0.7–2.5)
Table 3. Clinical characteristics of 426 ME/CFS patients: in total and stratified according to the presence of either DQB1*03:03 or C*07:04 or both.
5Scientific RepoRtS | (2020) 10:5267 | https://doi.org/10.1038/s41598-020-62157-x
www.nature.com/scientificreportswww.nature.com/scientificreports/
results need verification in independent cohorts. In general, established HLA associations are reproducible across different populations, but susceptibility loci can also vary between populations20,21,37,38. Therefore, HLA associa-tions in ME/CFS should also be investigated in populations of different ancestry.
Both of the ME/CFS associations observed in our data set were evident at 2nd field resolution (i.e. C*07:04 and DQB1*03:03), which distinguishes alleles encoding amino acid differences. Interestingly, the other C*07 and DQB1*03 alleles were not associated, emphasizing the importance of high resolution HLA genotyping.
We report C*07:04 and DQB1*03:03 as tag alleles for two independent HLA risk associations in ME/CFS, as these alleles are in linkage equilibrium (D’ = 0.06). However, they could still be markers for either one common, or two independently associated, variants outside the loci tested in this study. Alternatively, the associated alleles reported herein could constitute a functional relevance themselves. HLA class I alleles, like C*07:04, could influ-ence disease risk through their interactions with CD8 positive cytotoxic T lymphocytes20,39. Disturbances in CD8 positive T lymphocytes have been reported in ME/CFS, although the results are somewhat conflicting10. Another important function of HLA class I alleles is to serve as ligands for NK cell receptors. Altered numbers and func-tion of NK cells have been reported by several independent researchers in ME/CFS14,15,40. The other associated allele, DQB1*03:03, is an HLA class II allele, which is also interesting in regard to the hypothesis of autoimmun-ity in ME/CFS. In certain well studied AID, associated HLA class II alleles have been shown to exhibit unique peptide binding properties, as well as HLA-TCR restriction, directly influencing the acquired immune response, e.g. with the production of specific auto-antibodies41. A dysregulated activity of CD4 positive T lymphocytes, the principal cell type interfering with HLA class II alleles, have been discussed as a central mechanism in ME/CFS42. Several studies report increased levels of specific auto-antibodies in ME/CFS patients43, e.g. to neurotransmitter receptors, although most of these lack verification in additional cohorts.
Another interesting question is whether these HLA associations are driven by subgroups of patients, and thereby representing stronger risk alleles. This is a relevant aspect in a complex disease like ME/CFS, where different causal mechanisms may be at play in different subgroups. In our study, ME/CFS patients carrying HLA risk alleles had a significantly higher comorbidity of established AID. We are not aware of any publications reporting associations between C*07:04 or DQB1*03:03 and the AID affecting some of our patients. Therefore, it is unlikely that the associations in our material are driven by HLA associations with already established AID. Familial aggregation is observed for many specific AID, as well as for autoimmunity in general44. 1st degree relatives of ME/CFS patients in our study have a high prevalence of AID (Table 3). These observations could potentially be due to an element of autoimmunity in ME/CFS, or within a subgroup. No other patient characteristics were dominant among the ME/CFS patients with HLA risk alleles, specifically neither self-reported infectious onset nor current disease severity. It can be argued that the lack of validity of self-reported data precludes the detection of possible subgroup identifiers.
In conclusion, we report novel HLA associations in a large cohort of ME/CFS patients fulfilling the Canadian Consensus Criteria, thereby supporting the involvement of the immune system in the ME/CFS pathogenesis.
Materials and MethodsThis study is approved by the Norwegian Regional Committees for Medical and Health Research Ethics45. All methods and data handling were performed according to relevant national and institutional regulations and guidelines. All patients gave informed consent. In three cases, written consent was given by a close relative due to the patient being severely ill and unable to sign. A total of 426 adult, Norwegian ME/CFS patients were included. All had been diagnosed in Norway according to the 2003 Canadian Consensus Criteria2, except for four patients where the similarly strict 2010 International Consensus Criteria3 were applied. There were three separate recruit-ment groups for ME/CFS patients: 214 patients were recruited from recent and ongoing trials with Rituximab46,47 and Cyclophosphamide (Rekeland IG et al., submitted, NCT02444091); 116 patients were recruited from the CFS/ME biobank at Oslo University Hospital; 96 patients were recruited via announcements in patient networks, including patient organizations. Patients from the latter two groups were not included in clinical trials. Duplicates within or between the three recruitment groups were excluded. All patients provided the identity of any 1st, 2nd or 3rd degree relatives with ME/CFS, and we excluded close relatives to ensure that only one patient per extended family was included. Norwegian ethnicity was ensured by evaluation of sur- and family names of all patients, country of birth of parents and grandparents as well as self-perceived ethnicity. The control group consisted of 4511 ethnically matched, healthy individuals drawn from the Norwegian Bone Marrow Donor Registry48. Clinical information was collected for the ME/CFS patients through questionnaires completed by patients or close relatives. The categories applied in this study were gender, age at diagnosis, initiating events, disease dura-tion and severity, comorbidities and family history. Most of the questions were based on the DePaul Symptom Questionnaire49,50. Infection or vaccination as initiating event was self-reported, and in many cases, the time from event to symptom debut was not specified. The disease severity was assessed with self-reported activity level dur-ing the previous 6 months, as stated through the DePaul Symptom Questionnaire, question no. 79.
HLA genotyping by next generation sequencing. In 426 ME/CFS patients, we performed high resolu-tion, targeted, next generation sequencing (NGS) of HLA class I genes HLA -A, -B and -C and class II genes HLA -DRB1, -DQB1, -DPB1, -DQA1. Amplification and library preparation were performed with kits from GenDx (Utrecht, The Netherlands) and Illumina (San Diego, USA), 2 ×150 bp paired-end sequencing was performed by The Norwegian Sequencing Centre with Illumina MiSeq Reagent Kit v2 (300-cycles), and HLA genotypes were obtained by analyzing sequencing reads with NGSengine from GenDx, using the IMGT/HLA Database51. The median sequencing depth was above 150 reads per called base. The 4511 healthy Norwegian controls had pre-viously been HLA typed by NGS52. Both patient and control genotypes were analyzed at 2nd field resolution for HLA -A, -B, -C, -DRB1, -DQB1 and -DPB1. HLA alleles can be genotyped at resolution level from 1st field to 4th field. 2nd field resolution distinguishes alleles that encode amino acid differences, i.e. specific HLA proteins, and is therefore of great biological relevance. The genotyping success exceeded 99% in the patient group and 99.9%
6Scientific RepoRtS | (2020) 10:5267 | https://doi.org/10.1038/s41598-020-62157-x
www.nature.com/scientificreportswww.nature.com/scientificreports/
in the control group for all loci. In the control group, alleles were originally identified at a G group resolution, and certain alleles from the patient group were therefore converted to avoid typing method bias (Supplementary Table S5).
Data analyses were performed in Unphased 3.0.10 and Pypop 0.7.053,54. Assessment of Hardy-Weinberg equilibrium was performed with a chi-square test with a significance level of 0.05. Haplotype frequencies were estimated with an expectation-maximization method for unknown gametic phase. Global associations for each locus were calculated with a likelihood ratio test, with a rare allele frequency threshold of 0.01. Genetic associ-ations were investigated on allelic and haplotypic levels, and ORs with 95% confidence intervals (95% CI) were calculated with Woolf ’s formula comprising Haldane’s correction. Risk allele ORs were calculated also with gen-der stratification, and homogeneity tests were performed with the logit-based estimator. LD calculations and Svejgaard tests were carried out to examine the degree of independence between the associated alleles55. The LD measure D’ was calculated according to the formula D’ = D/Dmax, where Dmax = min [pA(1-pB), (1-pA)pB] for D > 0, Dmax = min [pApB, (1-pA)(1-pB)] for D < 0, D is the standard mathematical definition of LD between alleles A and B56, and p is the frequency of the stated allele. Because of the comparison of multiple allele frequencies we performed locus-wise Bonferroni correction. For each locus, non-corrected p-values were multiplied by the total number of alleles detected at that locus, excluding alleles with a frequency less than 3% in both the control group and the patient group. The significance level was 0.05. Only haplotypes consisting of associated alleles were inves-tigated, and multiple test correction was therefore not applied on the haplotypic level.
Investigation of clinical data. The clinical information was gathered separately for each of the three recruitment groups, controlled in one common database, and exported to SPSS57 for statistical analyses. The patient group was stratified according to the presence of specific HLA alleles, and eight dichotomized clinical variables were assessed with OR calculation by binary logistic regression, and chi-square significance testing.
Restrictions on the availability of material. Individual genotypes of patients are not made available due to Norwegian privacy regulations and laws.
Received: 2 September 2019; Accepted: 3 March 2020;Published: xx xx xxxx
References1. Fukuda, K. et al. The chronic fatigue syndrome: a comprehensive approach to its definition and study. International Chronic Fatigue
Syndrome Study Group. Annals of internal medicine 121, 953–959 (1994).2. Carruthers, B. M. et al. Myalgic Encephalomyelitis/Chronic Fatigue Syndrome. Journal of Chronic Fatigue Syndrome 11, 7–115,
https://doi.org/10.1300/J092v11n01_02 (2003).3. Carruthers, B. M. et al. Myalgic encephalomyelitis: International Consensus Criteria. Journal of internal medicine 270, 327–338,
https://doi.org/10.1111/j.1365-2796.2011.02428.x (2011).4. de Vega, W. C., Erdman, L., Vernon, S. D., Goldenberg, A. & McGowan, P. O. Integration of DNA methylation & health scores
identifies subtypes in myalgic encephalomyelitis/chronic fatigue syndrome. Epigenomics 10, 539–557, https://doi.org/10.2217/epi-2017-0150 (2018).
5. Williams, T. E., Chalder, T., Sharpe, M. & White, P. D. Heterogeneity in chronic fatigue syndrome - empirically defined subgroups from the PACE trial. Psychological medicine 47, 1454–1465, https://doi.org/10.1017/s0033291716003615 (2017).
6. Sunnquist, M., Jason, L. A., Nehrke, P. & Goudsmit, E. M. A Comparison of Case Definitions for Myalgic Encephalomyelitis and Chronic Fatigue Syndrome. Journal of chronic diseases and management 2 (2017).
7. Maclachlan, L. et al. Are current chronic fatigue syndrome criteria diagnosing different disease phenotypes? PloS one 12, e0186885, https://doi.org/10.1371/journal.pone.0186885 (2017).
8. Jason, L. A., McManimen, S., Sunnquist, M., Newton, J. L. & Strand, E. B. Clinical Criteria Versus a Possible Research CaseDefinition in Chronic Fatigue Syndrome/Myalgic Encephalomyelitis. Fatigue: biomedicine, health & behavior 5, 89–102, https://doi.org/10.1080/21641846.2017.1299077 (2017).
9. Griffith, J. P. & Zarrouf, F. A. A systematic review of chronic fatigue syndrome: don’t assume it’s depression. Primary care companion to the Journal of clinical psychiatry 10, 120–128 (2008).
10. Sotzny, F. et al. Myalgic Encephalomyelitis/Chronic Fatigue Syndrome - Evidence for an autoimmune disease. Autoimmunity reviews 17, 601–609, https://doi.org/10.1016/j.autrev.2018.01.009 (2018).
11. Blomberg, J., Gottfries, C. G., Elfaitouri, A., Rizwan, M. & Rosen, A. Infection Elicited Autoimmunity and MyalgicEncephalomyelitis/Chronic Fatigue Syndrome: An Explanatory Model. Frontiers in immunology 9, 229, https://doi.org/10.3389/fimmu.2018.00229 (2018).
12. Albright, F., Light, K., Light, A., Bateman, L. & Cannon-Albright, L. A. Evidence for a heritable predisposition to Chronic FatigueSyndrome. BMC neurology 11, 62, https://doi.org/10.1186/1471-2377-11-62 (2011).
13. Schur, E., Afari, N., Goldberg, J., Buchwald, D. & Sullivan, P. F. Twin analyses of fatigue. Twin research and human genetics: the officialjournal of the International Society for Twin Studies 10, 729–733, https://doi.org/10.1375/twin.10.5.729 (2007).
14. Rivas, J. L., Palencia, T., Fernandez, G. & Garcia, M. Association of T and NK Cell Phenotype With the Diagnosis of Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS). Frontiers in immunology 9, 1028, https://doi.org/10.3389/fimmu.2018.01028 (2018).
15. Curriu, M. et al. Screening NK-, B- and T-cell phenotype and function in patients suffering from Chronic Fatigue Syndrome. Journalof translational medicine 11, 68, https://doi.org/10.1186/1479-5876-11-68 (2013).
16. Montoya, J. G. et al. Cytokine signature associated with disease severity in chronic fatigue syndrome patients. Proceedings of the National Academy of Sciences of the United States of America, https://doi.org/10.1073/pnas.1710519114 (2017).
17. Hornig, M. et al. Distinct plasma immune signatures in ME/CFS are present early in the course of illness. Science advances 1, https://doi.org/10.1126/sciadv.1400121 (2015).
18. Trivedi, M. S. et al. Identification of Myalgic Encephalomyelitis/Chronic Fatigue Syndrome-associated DNA methylation patterns.PloS one 13, e0201066, https://doi.org/10.1371/journal.pone.0201066 (2018).
19. Mensah, F. K. F., Bansal, A. S., Ford, B. & Cambridge, G. Chronic fatigue syndrome and the immune system: Where are we now? Neurophysiologie clinique = Clinical neurophysiology 47, 131–138, https://doi.org/10.1016/j.neucli.2017.02.002 (2017).
20. Matzaraki, V., Kumar, V., Wijmenga, C. & Zhernakova, A. The MHC locus and genetic susceptibility to autoimmune and infectious diseases. Genome biology 18, 76, https://doi.org/10.1186/s13059-017-1207-1 (2017).
7Scientific RepoRtS | (2020) 10:5267 | https://doi.org/10.1038/s41598-020-62157-x
www.nature.com/scientificreportswww.nature.com/scientificreports/
21. Lokki, M. L. & Paakkanen, R. The complexity and diversity of major histocompatibility complex challenge disease association studies. Hla 93, 3–15, https://doi.org/10.1111/tan.13429 (2019).
22. Keller, R. H. et al. Association between HLA class II antigens and the chronic fatigue immune dysfunction syndrome. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America 18(Suppl 1), S154–156 (1994).
23. Underhill, J. A., Mahalingam, M., Peakman, M. & Wessely, S. Lack of association between HLA genotype and chronic fatigue syndrome. European journal of immunogenetics: official journal of the British Society for Histocompatibility and Immunogenetics 28, 425–428 (2001).
24. Helbig, K. J. et al. Variation in immune response genes and chronic Q fever. Concepts: preliminary test with post-Q fever fatiguesyndrome. Genes and immunity 4, 82–85, https://doi.org/10.1038/sj.gene.6363912 (2003).
25. Smith, J. et al. Association of chronic fatigue syndrome with human leucocyte antigen class II alleles. Journal of clinical pathology 58, 860–863, https://doi.org/10.1136/jcp.2004.022681 (2005).
26. Ledina, D. et al. Chronic fatigue syndrome after Q fever. Medical science monitor: international medical journal of experimental and clinical research 13, Cs88–92 (2007).
27. Carlo-Stella, N. et al. Molecular study of receptor for advanced glycation endproduct gene promoter and identification of specificHLA haplotypes possibly involved in chronic fatigue syndrome. International journal of immunopathology and pharmacology 22, 745–754, https://doi.org/10.1177/039463200902200320 (2009).
28. Ortega-Hernandez, O. D. et al. Autoantibodies, polymorphisms in the serotonin pathway, and human leukocyte antigen class IIalleles in chronic fatigue syndrome: are they associated with age at onset and specific symptoms? Annals of the New York Academy of Sciences 1173, 589–599, https://doi.org/10.1111/j.1749-6632.2009.04802.x (2009).
29. Spitzer, A. R. & Broadman, M. A retrospective review of the sleep characteristics in patients with chronic fatigue syndrome and fibromyalgia. Pain practice: the official journal of World Institute of Pain 10, 294–300, https://doi.org/10.1111/j.1533-2500.2009.00352.x (2010).
30. Pasi, A. et al. Excess of activating killer cell immunoglobulinlike receptors and lack of HLA-Bw4 ligands: a twoedged weapon in chronic fatigue syndrome. Molecular medicine reports 4, 535–540, https://doi.org/10.3892/mmr.2011.447 (2011).
31. Candore, G., Lio, D., Colonna Romano, G. & Caruso, C. Pathogenesis of autoimmune diseases associated with 8.1 ancestralhaplotype: effect of multiple gene interactions. Autoimmunity reviews 1, 29–35 (2002).
32. Holmes, G. P. et al. Chronic fatigue syndrome: a working case definition. Annals of internal medicine 108, 387–389 (1988). 33. Dunckley, H. HLA typing by SSO and SSP methods. Methods in molecular biology (Clifton, N.J.) 882, 9–25, https://doi.
org/10.1007/978-1-61779-842-9_2 (2012). 34. Sandoughi, M., Fazaeli, A., Bardestani, G. & Hashemi, M. Frequency of HLA-DRB1 alleles in rheumatoid arthritis patients in
Zahedan, southeast Iran. Annals of Saudi medicine 31, 171–173, https://doi.org/10.4103/0256-4947.78205 (2011). 35. Viatte, S. & Barton, A. Genetics of rheumatoid arthritis susceptibility, severity, and treatment response. Seminars in immunopathology
39, 395–408, https://doi.org/10.1007/s00281-017-0630-4 (2017). 36. Colquhoun, D. An investigation of the false discovery rate and the misinterpretation of p-values. Royal Society open science 1,
140216, https://doi.org/10.1098/rsos.140216 (2014). 37. Caillat-Zucman, S. Molecular mechanisms of HLA association with autoimmune diseases. Tissue antigens 73, 1–8, https://doi.
org/10.1111/j.1399-0039.2008.01167.x (2009). 38. Zhou, F. et al. Deep sequencing of the MHC region in the Chinese population contributes to studies of complex disease. Nature
genetics 48, 740–746, https://doi.org/10.1038/ng.3576 (2016). 39. Gough, S. C. & Simmonds, M. J. The HLA Region and Autoimmune Disease: Associations and Mechanisms of Action. Current
genomics 8, 453–465, https://doi.org/10.2174/138920207783591690 (2007). 40. Brenu, E. W. et al. Longitudinal investigation of natural killer cells and cytokines in chronic fatigue syndrome/myalgic
encephalomyelitis. Journal of translational medicine 10, 88, https://doi.org/10.1186/1479-5876-10-88 (2012). 41. Dendrou, C. A., Petersen, J., Rossjohn, J. & Fugger, L. HLA variation and disease. Nature reviews. Immunology 18, 325–339, https://
doi.org/10.1038/nri.2017.143 (2018). 42. Brenu, E. W. et al. Immunological abnormalities as potential biomarkers in Chronic Fatigue Syndrome/Myalgic Encephalomyelitis.
Journal of translational medicine 9, 81, https://doi.org/10.1186/1479-5876-9-81 (2011). 43. Ortega-Hernandez, O. D. & Shoenfeld, Y. Infection, vaccination, and autoantibodies in chronic fatigue syndrome, cause or
coincidence? Annals of the New York Academy of Sciences 1173, 600–609, https://doi.org/10.1111/j.1749-6632.2009.04799.x (2009). 44. Cardenas-Roldan, J., Rojas-Villarraga, A. & Anaya, J. M. How do autoimmune diseases cluster in families? A systematic review and
meta-analysis. BMC medicine 11, 73, https://doi.org/10.1186/1741-7015-11-73 (2013). 45. Regional committees for medical and health related ethics, https://helseforskning.etikkom.no/prosjekterirek/prosjektregister?p_
dim=34977&_ikbLanguageCode=us&p_operation=new (Folder no: 2015/1547) 46. Fluge, O. et al. Benefit from B-lymphocyte depletion using the anti-CD20 antibody rituximab in chronic fatigue syndrome. A
double-blind and placebo-controlled study. PloS one 6, e26358, https://doi.org/10.1371/journal.pone.0026358 (2011). 47. Fluge, O. et al. B-Lymphocyte Depletion in Patients With Myalgic Encephalomyelitis/Chronic Fatigue Syndrome: A Randomized,
Double-Blind, Placebo-Controlled Trial. Annals of internal medicine, https://doi.org/10.7326/m18-1451 (2019). 48. The Norwegian bone marrow donor registry, https://oslo-universitetssykehus.no/fag-og-forskning/nasjonale-og-regionale-
tjenester/det-norske-benmargsgiverregisteret (2018) 49. Jason, L. A. et al. The development of a revised Canadian myalgic encephalitis chronic fatigue syndrome case definition. Am. J.
Biochem. Biotechnol. 6, 120–135 (2010). 50. Strand, E. B. et al. Comparing the DePaul Symptom Questionnaire with physician assessments: A preliminary study. Fatigue:
Biomedicine, Health & Behavior 4, 52–62 (2016). 51. IPD-IMGT/HLA Database, https://www.ebi.ac.uk/ipd/imgt/hla/ (2018) 52. Lande, A., Andersen, I., Egeland, T., Lie, B. A. & Viken, M. K. HLA -A, -C, -B, -DRB1, -DQB1 and -DPB1 allele and haplotype
frequencies in 4514 healthy Norwegians. Human immunology 79, 527–529, https://doi.org/10.1016/j.humimm.2018.04.012 (2018). 53. Dudbridge, F. Pedigree disequilibrium tests for multilocus haplotypes. Genetic epidemiology 25, 115–121, https://doi.org/10.1002/
gepi.10252 (2003). 54. Lancaster, A. K., Single, R. M., Solberg, O. D., Nelson, M. P. & Thomson, G. PyPop update–a software pipeline for large-scale
multilocus population genomics. Tissue antigens 69(Suppl 1), 192–197, https://doi.org/10.1111/j.1399-0039.2006.00769.x (2007). 55. Svejgaard, A. & Ryder, L. P. HLA and disease associations: detecting the strongest association. Tissue antigens 43, 18–27 (1994). 56. Slatkin, M. Linkage disequilibrium–understanding the evolutionary past and mapping the medical future. Nature reviews. Genetics
9, 477–485, https://doi.org/10.1038/nrg2361 (2008). 57. IBM Corp. Released 2017. IBM SPSS Statistics for Macintosh, Version 25.0. Armonk, NY: IBM Corp.
AcknowledgementsWe are very grateful to the patients who participated in the study, and we thank the Norwegian Bone Marrow Donor Registry for access to controls. We would also like to express our gratitude to Kari Sørland and Kristin Risa at Department of Oncology and Medical Physics, Haukeland University Hospital, Wenche Kristiansen and Hilde Haukeland at the CFS/ME Biobank, Oslo University Hospital, and employees at the Norwegian Sequencing
8Scientific RepoRtS | (2020) 10:5267 | https://doi.org/10.1038/s41598-020-62157-x
www.nature.com/scientificreportswww.nature.com/scientificreports/
Centre, University in Oslo and Oslo University Hospital. This study was funded by the Kavli Trust and the Research Council of Norway. We express our deepest gratitude. The funders did not influence the research or the manuscript in any way.
Author contributionsØ.F., O.M., T.E., O.S., B.L. and M.V. initiated the project and applied for funding. A.L., Ø.F., E.S., D.S., O.M. and M.V. recruited and included patients. A.L., S.F. and M.V. performed sample handling, laboratory work and HLA genotyping. A.L. performed analyses and wrote the first draft of the manuscript. All authors participated in editing the manuscript and approved the final draft.
competing interestsA.L., Ø.F., E.S., S.F., D.S., O.M., T.E., B.L. and M.V. all declare that they have no conflict of interest. O.S. is member of EMECC - European M.E. Clinicians Council; otherwise he has no conflict of interest.
Additional informationSupplementary information is available for this paper at https://doi.org/10.1038/s41598-020-62157-x.Correspondence and requests for materials should be addressed to A.L.Reprints and permissions information is available at www.nature.com/reprints.Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre-ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per-mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© The Author(s) 2020
a) Locus Obs Het Exp Het P-HWE HLA-A 346 349.9 0.83HLA-C 390 381.2 0.65HLA-B 393 389.6 0.86
HLA-DRB1 379 382.4 0.86HLA-DQB1 377 375.0 0.92HLA-DPB1 317 312.0 0.78
b) Locus Obs Het Exp Het P-HWE HLA-A 3724 3751.7 0.65HLA-C 4043 4034.0 0.89HLA-B 4114 4126.8 0.84
HLA-DRB1 4121 4096.9 0.71HLA-DQB1 3966 3936.7 0.64HLA-DPB1 3394 3409.9 0.79
Table S1Hardy-Weinberg Equilibrium parameters for 6 HLA loci at 2nd field resolution among a) 426 ME/CFS patients, and b) 4511 healthy controls
Obs Het: Observed Heterozygosity; Exp Het: Expected Heterozygosity; P-HWE: p-value for Hardy-Weinberg Equilibrium deviation.
Table S2
Locus Allele Allele countAllele frequency Allele count
Allele frequency
HLA-A 01:01 117 0.1373 1433 0.158801:02 0 0.0000 2 0.000202:01 293 0.3439 2889 0.320202:02 0 0.0000 3 0.000302:05 6 0.0070 61 0.006802:06 5 0.0059 12 0.001302:07 0 0.0000 4 0.000402:08 0 0.0000 1 0.000102:17 0 0.0000 1 0.000102:20 0 0.0000 1 0.000102:118 0 0.0000 1 0.000102:474 0 0.0000 1 0.000102:581 0 0.0000 2 0.000203:01 137 0.1608 1401 0.155303:02 0 0.0000 7 0.000803:17 0 0.0000 1 0.000103:56 0 0.0000 1 0.000103:62 0 0.0000 1 0.000103:201 0 0.0000 1 0.000111:01 46 0.0540 547 0.060623:01 12 0.0141 121 0.013423:06 0 0.0000 1 0.000124:02 74 0.0869 772 0.085624:03 4 0.0047 13 0.001424:07 0 0.0000 3 0.000325:01 26 0.0305 222 0.024625:11 0 0.0000 1 0.000126:01 14 0.0164 147 0.016326:08 2 0.0024 10 0.001129:01 4 0.0047 7 0.000829:02 21 0.0247 181 0.020130:01 2 0.0024 56 0.006230:02 4 0.0047 27 0.003030:04 0 0.0000 9 0.001031:01 24 0.0282 301 0.033431:08 0 0.0000 1 0.000132:01 18 0.0211 300 0.033332:63 0 0.0000 1 0.000133:01 1 0.0012 31 0.003433:03 1 0.0012 19 0.002134:01 0 0.0000 2 0.000234:02 1 0.0012 1 0.0001
Allele count and frequency for all detected HLA alleles among patients and controls. HLA Class I and Class II loci are listed in separate tabs
ME/CFS patients Controls
66:01 2 0.0024 30 0.003368:01 35 0.0411 365 0.040568:02 2 0.0024 22 0.002468:35 0 0.0000 1 0.000169:01 1 0.0012 4 0.000474:01 0 0.0000 1 0.000174:03 0 0.0000 2 0.000274:05 0 0.0000 1 0.0001Total 852 0.99998 9022 0.9996
HLA-C 01:02 31 0.0364 393 0.043602:02 36 0.0423 457 0.050702:10 0 0.0000 4 0.000403:02 2 0.0024 18 0.002003:03 50 0.0587 559 0.062003:04 126 0.1479 1380 0.153003:05 0 0.0000 2 0.000203:14 2 0.0024 3 0.000303:220 0 0.0000 2 0.000204:01 80 0.0939 782 0.086704:03 0 0.0000 2 0.000204:15 0 0.0000 2 0.000204:09N 1 0.0012 0 0.000005:01 88 0.1033 792 0.087805:09 0 0.0000 1 0.000106:02 56 0.0657 549 0.060906:26 0 0.0000 1 0.000107:01 111 0.1303 1383 0.153307:02 154 0.1808 1540 0.170707:04 33 0.0387 172 0.019107:21 0 0.0000 4 0.000407:24 1 0.0012 3 0.000307:36 0 0.0000 1 0.000107:51 0 0.0000 3 0.000307:124 0 0.0000 1 0.000107:212 0 0.0000 1 0.000107:XX 0 0.0000 1 0.000108:01 2 0.0024 19 0.002108:02 12 0.0141 193 0.021412:02 0 0.0000 20 0.002212:03 25 0.0293 265 0.029414:02 3 0.0035 67 0.007415:02 16 0.0188 147 0.016315:04 0 0.0000 5 0.000615:05 1 0.0012 5 0.000615:06 0 0.0000 2 0.000215:11 0 0.0000 3 0.000316:01 17 0.0200 193 0.0214
16:02 3 0.0035 8 0.000916:04 0 0.0000 1 0.000117:01 2 0.0024 35 0.003918:01 0 0.0000 3 0.0003Total 852 0.99999 9022 1.0000
HLA-B 07:02 138 0.1620 1433 0.158807:04 2 0.0024 3 0.000307:05 1 0.0012 5 0.000607:31 0 0.0000 1 0.000108:01 83 0.0974 1146 0.127008:04 1 0.0012 4 0.000408:26 0 0.0000 1 0.000113:01 0 0.0000 1 0.000113:02 6 0.0070 95 0.010514:01 7 0.0082 94 0.010414:02 6 0.0070 102 0.011315:01 75 0.0880 908 0.100615:02 0 0.0000 3 0.000315:03 1 0.0012 5 0.000615:07 2 0.0024 15 0.001715:13 0 0.0000 1 0.000115:16 0 0.0000 4 0.000415:17 1 0.0012 5 0.000615:18 1 0.0012 9 0.001015:21 0 0.0000 1 0.000115:24 0 0.0000 1 0.000115:25 0 0.0000 1 0.000115:39 0 0.0000 1 0.000115:110 0 0.0000 1 0.000115:363 0 0.0000 1 0.000118:01 34 0.0399 330 0.036618:03 0 0.0000 1 0.000127:02 1 0.0012 18 0.002027:05 42 0.0493 525 0.058235:01 53 0.0622 543 0.060235:02 1 0.0012 11 0.001235:03 7 0.0082 70 0.007835:08 3 0.0035 13 0.001435:10 0 0.0000 1 0.000135:17 0 0.0000 1 0.000135:XX 0 0.0000 1 0.000137:01 9 0.0106 122 0.013538:01 5 0.0059 50 0.005539:01 4 0.0047 55 0.006139:05 0 0.0000 2 0.000239:06 7 0.0082 65 0.007239:24 0 0.0000 1 0.0001
40:01 97 0.1139 940 0.104240:02 11 0.0129 161 0.017940:06 0 0.0000 3 0.000340:94 0 0.0000 1 0.000141:01 0 0.0000 9 0.001041:02 1 0.0012 23 0.002642:02 0 0.0000 3 0.000344:02 105 0.1232 904 0.100244:03 27 0.0317 289 0.032044:04 0 0.0000 1 0.000144:05 1 0.0012 8 0.000944:08 0 0.0000 1 0.000144:XX 0 0.0000 1 0.000145:01 6 0.0070 41 0.004546:01 0 0.0000 5 0.000647:01 3 0.0035 24 0.002748:01 2 0.0024 13 0.001449:01 9 0.0106 110 0.012250:01 3 0.0035 29 0.003250:02 0 0.0000 4 0.000451:01 36 0.0423 309 0.034351:05 0 0.0000 2 0.000251:08 1 0.0012 3 0.000351:09 0 0.0000 1 0.000152:01 0 0.0000 22 0.002453:01 1 0.0012 12 0.001354:01 0 0.0000 2 0.000255:01 12 0.0141 105 0.011655:02 0 0.0000 1 0.000156:01 3 0.0035 47 0.005257:01 39 0.0458 259 0.028757:02 0 0.0000 4 0.000458:01 5 0.0059 31 0.003458:02 0 0.0000 1 0.000173:01 0 0.0000 3 0.0003Total 852 0.99996 9022 0.9993
Locus Allele Allele countAllele frequency Allele count
Allele frequency
HLA-DRB1 01:01 77 0.0912 761 0.084401:02 0 0.0000 41 0.004501:03 6 0.0071 123 0.013603:01 96 0.1137 1231 0.136503:14 1 0.0012 0 0.000004:01 101 0.1197 1035 0.114804:02 0 0.0000 18 0.002004:03 5 0.0059 73 0.0081
ME/CFS patients Controls
04:04 42 0.0498 537 0.059604:05 1 0.0012 30 0.003304:06 1 0.0012 4 0.000404:07 10 0.0119 68 0.007504:08 3 0.0036 36 0.004004:10 0 0.0000 1 0.000104:38 0 0.0000 1 0.000107:01 86 0.1019 776 0.086108:01 36 0.0427 366 0.040608:02 0 0.0000 14 0.001608:03 0 0.0000 16 0.001808:04 3 0.0036 4 0.000408:10 0 0.0000 1 0.000109:01 14 0.0166 142 0.015810:01 4 0.0047 68 0.007511:01 28 0.0332 310 0.034411:02 0 0.0000 14 0.001611:03 2 0.0024 35 0.003911:04 7 0.0083 59 0.006511:14 1 0.0012 5 0.000611:28 0 0.0000 1 0.000112:01 23 0.0273 195 0.021612:02 0 0.0000 3 0.000313:01 64 0.0758 703 0.078013:02 60 0.0711 527 0.058413:03 1 0.0012 30 0.003313:04 0 0.0000 1 0.000113:06 0 0.0000 1 0.000113:15 0 0.0000 1 0.000114:01 24 0.0284 226 0.025114:02 2 0.0024 11 0.001214:03 0 0.0000 1 0.000114:04 0 0.0000 3 0.000314:05 0 0.0000 1 0.000114:07 0 0.0000 1 0.000114:22 0 0.0000 1 0.000115:01 145 0.1718 1491 0.165315:02 0 0.0000 17 0.001916:01 1 0.0012 32 0.003516:02 0 0.0000 3 0.0003Total 844 1.0000 9018 1.0000
HLA-DQB1 02:01 142 0.1667 1744 0.193402:14 0 0.0000 1 0.000103:01 128 0.1502 1215 0.134703:02 102 0.1197 1262 0.139903:03 56 0.0657 406 0.045003:04 1 0.0012 15 0.0017
03:05 0 0.0000 3 0.000304:01 0 0.0000 1 0.000104:02 41 0.0481 391 0.043405:01 82 0.0962 965 0.107005:02 1 0.0012 42 0.004705:03 24 0.0282 227 0.025205:04 5 0.0059 8 0.000906:01 0 0.0000 21 0.002306:02 142 0.1667 1472 0.163206:03 67 0.0786 725 0.080406:04 56 0.0657 462 0.051206:09 4 0.0047 55 0.006106:16 0 0.0000 2 0.000206:164 0 0.0000 1 0.000106:XX 1 0.0012 0 0.0000Total 852 1.0000 9018 0.9999
HLA-DPB1 01:01 43 0.0506 611 0.067802:01 78 0.0918 1067 0.118302:02 7 0.0082 39 0.004303:01 108 0.1271 1003 0.111204:01 402 0.4729 4010 0.444704:02 97 0.1141 1033 0.114505:01 28 0.0329 243 0.027006:01 8 0.0094 144 0.016009:01 3 0.0035 46 0.005110:01 6 0.0071 104 0.011511:01 15 0.0177 147 0.016313:01 10 0.0118 99 0.011014:01 9 0.0106 84 0.009315:01 6 0.0071 52 0.005816:01 3 0.0035 85 0.009417:01 6 0.0071 60 0.006719:01 8 0.0094 95 0.010520:01 7 0.0082 52 0.005823:01 4 0.0047 34 0.003824:01 0 0.0000 2 0.000226:01 0 0.0000 2 0.000231:01 0 0.0000 1 0.000135:01 0 0.0000 1 0.000145:01 1 0.0012 0 0.000051:01 0 0.0000 1 0.000171:01 1 0.0012 0 0.0000115:01 0 0.0000 1 0.0001129:01 0 0.0000 1 0.0001424:01 0 0.0000 1 0.0001Total 850 1.0000 9018 0.9999
Table S3Tests of global association
LocusLikelihood ratio chisq No. of alleles df p-value
HLA-A 11.01 12 11 0.443HLA-C 22.86 14 13 0.043HLA-B 26.21 18 17 0.071HLA-DRB1 15.08 15 14 0.373HLA-DQB1 17.69 10 9 0.039HLA-DPB1 17.2 12 11 0.102df: Degrees of freedom
Table S5
Locus Allele No. of alleles Belonging to G group
Used in analysis
HLA-C 01:127 6 01:02:01G 01:02HLA-C 07:18 2 07:01:01G 07:01HLA-C 17:03 2 17:01:01G 17:01HLA-B 15:220 1 15:03:01G 15:03HLA-B 44:27 1 44:02:01G 44:02HLA-DRB1 01:77 1 01:01:01G 01:01HLA-DRB1 14:54 24 14:01:01G 14:01HLA-DQB1 02:02 49 02:01:01G 02:01HLA-DQB1 03:19 1 03:01:01G 03:01HLA-DPB1 104:01 3 03:01:01G 03:01HLA-DPB1 463:01 1 04:02:01G 04:02
Conversion of alleles in the patient group (N = 426) belonging to G groups differing on 2nd field resolution
Supp
lem
enta
ry D
ata
Shee
t S4
- Pag
e1
C*07
:04
B*44
:02
Pati
ents
Cont
rols
++
3115
5+
-2
16-
+69
704
--
324
3636
426
4511
Que
stio
nCa
lcul
atio
nsA
nsw
er
OR
log
X2p-
valu
e
OR
lo
wer
C
I
OR
up
per
CI
Alle
le C
*07:
04 a
ssoc
iate
d?33
171
2.13
10.
7625
.69
14.7
10.
0001
1.44
83.
137
Yes
393
4340
Alle
le B
*44:
02 a
ssoc
iate
d?10
085
91.
304
0.27
68.9
44.
860.
027
1.03
1.65
1Ye
s32
636
52
C*07
:04
asso
ciat
ed in
B*4
4:02
+?31
155
2.04
10.
7118
.31
9.31
0.00
21.
291
3.22
6Ye
s69
704
C*07
:04
asso
ciat
ed in
B*4
4:02
-?2
161.
403
0.34
1.77
0.2
0.65
30.
321
6.12
8N
ot si
gnifi
cant
lyIn
depe
nden
t ass
ocia
tion
of C
*07:
04?
324
3636
Not
sign
ifica
ntly
B*44
:02
asso
ciat
ed in
C*0
7:04
+?31
155
1.6
0.47
1.66
0.37
0.54
40.
357.
313
Not
sign
ifica
ntly
216
B*44
:02
asso
ciat
ed in
C*0
7:04
-?69
704
1.1
0.1
51.8
80.
470.
493
0.83
81.
444
No
Inde
pend
ent a
ssoc
iati
on o
f B*4
4:02
?32
436
36N
o
Eval
uati
ons o
f ind
epen
dent
ass
ocia
tion
for e
ach
of th
e tw
o al
lele
s on
each
of t
he
hapl
otyp
es C
*07:
04 -
B*44
:02
and
B*57
:01
- DQ
B1*0
3:03
(in
sepa
rate
tabs
). M
etho
ds
base
d on
Sve
jgaa
rd, A
. & R
yder
, L. P
. HLA
and
dis
ease
ass
ocia
tion
s: d
etec
ting
the
stro
nges
t as
soci
atio
n. T
issu
e an
tige
ns 4
3, 1
8-27
(199
4)
Supp
lem
enta
ry D
ata
Shee
t S4
- Pag
e2B*
57:0
1D
QB1
*03:
03Pa
tien
tsCo
ntro
ls+
+27
183
+-
1073
-+
2721
1-
-36
240
4242
645
09Q
uest
ion
Calc
ulat
ions
Ans
wer
OR
log
X2p-
valu
e
OR
lo
wer
C
I
OR
up
per
CI
Alle
le B
*57:
01 a
ssoc
iate
d?37
256
1.58
0.46
29.6
46.
20.
013
1.10
22.
265
Yes
389
4253
Alle
le D
QB1
*03:
03 a
ssoc
iate
d?54
394
1.51
60.
4241
.69
7.22
0.00
71.
119
2.05
4Ye
s, m
ore
sign
ifica
ntly
372
4115
B*57
:01
asso
ciat
ed in
DQ
B1*0
3:03
+?27
183
1.15
30.
1411
.87
0.24
0.62
40.
653
2.03
7N
o27
211
B*57
:01
asso
ciat
ed in
DQ
B1*0
3:03
-?10
731.
530.
428.
571.
550.
214
0.78
32.
988
Not
sign
ifica
ntly
Inde
pend
ent a
ssoc
iati
on o
f B*5
7:01
?36
240
42N
ot si
gnifi
cant
ly
DQ
B1*0
3:03
ass
ocia
ted
in B
*57:
01+?
2718
31.
077
0.07
6.4
0.04
0.85
10.
496
2.33
7N
o10
73
DQ
B1*0
3:03
ass
ocia
ted
in B
*57:
01-?
2721
11.
429
0.36
22.3
32.
840.
092
0.94
42.
163
Not
sign
ifica
ntly
Inde
pend
ent a
ssoc
iati
on o
f DQ
B1*0
3:03
?36
240
42N
ot si
gnifi
cant
ly
III
PAPER III
CLINICAL TRIALpublished: 29 April 2020
doi: 10.3389/fmed.2020.00162
Frontiers in Medicine | www.frontiersin.org 1 April 2020 | Volume 7 | Article 162
Edited by:
James N. Baraniuk,
Georgetown University Medical
Center, United States
Reviewed by:
Lucinda Bateman,
Bateman Horne Center, United States
Indre Bileviciute-Ljungar,
Karolinska Institutet (KI), Sweden
*Correspondence:
Øystein Fluge
Specialty section:
This article was submitted to
Family Medicine and Primary Care,
a section of the journal
Frontiers in Medicine
Received: 14 October 2019
Accepted: 09 April 2020
Published: 29 April 2020
Citation:
Rekeland IG, Fosså A, Lande A,
Ktoridou-Valen I, Sørland K,
Holsen M, Tronstad KJ, Risa K,
Alme K, Viken MK, Lie BA, Dahl O,
Mella O and Fluge Ø (2020)
Intravenous Cyclophosphamide in
Myalgic Encephalomyelitis/Chronic
Fatigue Syndrome. An Open-Label
Phase II Study. Front. Med. 7:162.
doi: 10.3389/fmed.2020.00162
Intravenous Cyclophosphamide inMyalgic Encephalomyelitis/ChronicFatigue Syndrome. An Open-LabelPhase II StudyIngrid G. Rekeland 1, Alexander Fosså 2, Asgeir Lande 3, Irini Ktoridou-Valen 1,Kari Sørland 1, Mari Holsen 4, Karl J. Tronstad 5, Kristin Risa 1, Kine Alme 1,Marte K. Viken 3,6, Benedicte A. Lie 3,6, Olav Dahl 5, Olav Mella 1,5 and Øystein Fluge 1,5*
1Department of Oncology and Medical Physics, Haukeland University Hospital, Bergen, Norway, 2Department of Oncology,
Norwegian Radium Hospital, Oslo University Hospital, Oslo, Norway, 3Department of Medical Genetics, Oslo University
Hospital and Faculty of Medicine, University of Oslo, Oslo, Norway , 4Clinical Research Unit, Haukeland University Hospital,
Bergen, Norway, 5Department of Biomedicine, University of Bergen, Bergen, Norway, 6Department of Immunology,
University of Oslo and Oslo University Hospital-Rikshospitalet, Oslo, Norway
Introduction: Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is a
disease with high symptom burden, of unknown etiology, with no established treatment.
We observed patients with long-standing ME/CFS who got cancer, and who reported
improvement of ME/CFS symptoms after chemotherapy including cyclophosphamide,
forming the basis for this prospective trial.
Materials and methods: This open-label phase II trial included 40 patients with
ME/CFS diagnosed by Canadian criteria. Treatment consisted of six intravenous infusions
of cyclophosphamide, 600–700 mg/m2, given at four-week intervals with follow-up for
18 months, extended to 4 years. Response was defined by self-reported improvements
in symptoms by Fatigue score, supported by Short Form 36 (SF-36) scores, physical
activity measures and other instruments. Repeated measures of outcome variables were
assessed by General linear models. Responses were correlated with specific Human
Leukocyte Antigen (HLA) alleles.
Results: The overall response rate by Fatigue score was 55.0% (22 of 40 patients).
Fatigue score and other outcome variables showed significant improvements compared
to baseline. The SF-36 Physical Function score increased from mean 33.0 at baseline
to 51.5 at 18 months (all patients), and from mean 35.0 to 69.5 among responders.
Mean steps per 24 h increased from mean 3,199 at baseline to 4,347 at 18 months (all
patients), and from 3,622 to 5,589 among responders. At extended follow-up to 4 years
68% (15 of 22 responders) were still in remission. Patients positive for HLA-DQB1∗03:03
and/or HLA-C∗07:04 (n= 12) had significantly higher response rate compared to patients
negative for these alleles (n = 28), 83 vs. 43%, respectively. Nausea and constipation
were common grade 1–2 adverse events. There were one suspected unexpected serious
adverse reaction (aggravated POTS) and 11 serious adverse events in eight patients.
Conclusion: Intravenous cyclophosphamide treatment was feasible for ME/CFS
patients and associated with an acceptable toxicity profile. More than half of the patients
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
responded and with prolonged follow-up, a considerable proportion of patients reported
ongoing remission. Without a placebo group, clinical response data must be interpreted
with caution. We nevertheless believe a future randomized trial is warranted.
Clinical Trial Registration: www.ClinicalTrials.gov, identifier: NCT02444091.
Keywords: myalgic encephalomyelitis, chronic fatigue syndrome, ME, CFS, cyclophosphamide, clinical trial,
medical treatment, HLA
INTRODUCTION
Myalgic Encephalomyelitis/Chronic Fatigue Syndrome(ME/CFS) is a disease of unknown etiology characterizedby post-exertional malaise (PEM) (1, 2), sleep disturbanceswith inadequate restitution (3), fatigue, pain and sensoryhypersensitivity, cognitive and several other symptoms. Thediagnosis relies on exclusion of other disorders associated withfatigue, and there are no confirmatory diagnostic tests. Usingthe Canadian consensus criteria (4), an estimated 0.1% of thepopulation suffer from ME/CFS (5), affecting women 3–4 timesmore often than men. ME/CFS has profound impact on qualityof life for patients and their caretakers (6, 7). The socio-economiccosts are high, and there is an urgent need for elucidation of thedisease mechanisms, for improved diagnostic approaches, andfor rational treatment (8).
We hypothesized that ME/CFS could be a variant of anautoimmune disease, with a role for B-cells and possiblyautoantibodies. Several observations suggest that immunedysregulation and low-grade inflammation may be involvedin the pathogenesis of ME/CFS (9–11). A review (12)summarizes data indicating autoimmunity as a possibleetiological factor. Mechanisms may include dysregulationsof cytokines (13), alterations in lymphocyte subsets (14) andpresence of autoantibodies (15–17). A study with peptide arraysdemonstrated an immunosignature based on serum antibodiesthat separated ME/CFS cases from healthy controls (18). Also,elderly patients with ME/CFS have an increased risk of B-celllymphomas, especially marginal zone lymphomas known tobe associated with autoimmunity or chronic infections (19).Recent research suggests disturbed turnover of complex lipids,fatty acids and amino acids and impaired energy metabolismas possible features of ME/CFS (20–23), possibly linked tolow-grade inflammation (24).
There is evidence for a genetic predisposition in ME/CFS(25, 26). The immunologically important Human LeukocyteAntigen (HLA) genes were previously investigated in smallME/CFS cohorts, and certain class II alleles have been foundmore prevalent among patients (27–29). A recent study of alarger Norwegian cohort of patients and controls, identifiedtwo potential HLA risk alleles, namely HLA-C∗07:04 and HLA-DQB1∗03:03 (30).
At present, there is no established treatment for ME/CFS. Inour oncology unit, we have observed seven patients with long-standingME/CFS, who reported significant improvement of theirME/CFS symptoms after chemotherapy for either malignant
lymphoma or breast cancer. These seven patients all receivedchemotherapy including the cytotoxic drugs cyclophosphamideor ifosfamide, and one patient also received rituximab. Wedecided to pursue these observations in separate clinical trials.
Rituximab is a monoclonal antibody that targets CD20 onthe surface of B-cells, resulting in reversible B-cell depletion(31). Initial small studies testing rituximab in ME/CFS (32–34)indicated that a subgroup could benefit from B-cell depletion.However, in a recent Norwegian multicenter, randomized,double-blind and placebo-controlled trial, we reported nosignificant outcome differences between the rituximab andplacebo groups (35).
Cyclophosphamide, an alkylating agent widely used in cancertreatment (36), induces immunosuppression and is also usedto treat immune-mediated diseases like systemic lupus (SLE),rheumatoid arthritis, vasculitis, and multiple sclerosis (37–40).Based on the assumed immune disturbance in ME/CFS, theobserved improvement in ME/CFS symptoms could be due tothe immunosuppressive effect of cyclophosphamide (41).
In 2014, we treated four ME/CFS patients with six infusions ofcyclophosphamide every 4 weeks. Two of the patients reportedsubstantial improvement of their ME/CFS symptoms, lastingmore than 4 years for one of them. In these pilot experiences,there were no infections, neutropenia, thrombocytopenia orunexpected adverse events. We decided to conduct a prospectivetrial to further investigate feasibility, efficacy and safety ofcyclophosphamide treatment in ME/CFS patients.
METHODS
Trial DesignThe CycloME study (EudraCT no. 2014-004029-41,ClinicalTrials.gov NCT02444091) was designed as an open-label phase II trial comprising 40 patients with ME/CFS. Thestudy was approved by the Regional Committees for Medicaland Health Research Ethics (2014/1672) and by the NationalMedicines Agency in Norway. Originally planned for 18 monthsfollow-up, the protocol was amended for prolonged observationof patients up to 4 years after start of treatment. The protocol isavailable as supporting information (Data Sheet 1).
Setting and Patient InclusionSince 2011 patients with a likely diagnosis of ME/CFS have beenreferred to the Department of Oncology and Medical Physics,Haukeland University Hospital (HUH), for possible inclusion inclinical trials. Based on available information and proximity to
Frontiers in Medicine | www.frontiersin.org 2 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
the treating hospital, patients previously included in trials withrituximab and newly referred patients were invited to receiveinformation about the trial. Following signed informed consent,the patients were screened for eligibility.
Inclusion criteria were: a diagnosis of ME/CFS according tothe Canadian criteria (4); age 18–66 years; disease duration morethan 2 years; and disease severity mild-to-moderate, moderate,moderate-to-severe, or severe. Patients with either mild or verysevere disease (completely bedbound and in need of help for allbasic activities of daily living) were not included. The exclusioncriteria and pre-treatment evaluation are detailed in the trialprotocol (Data Sheet 1).
Recruitment lasted from March 2015 until December 2015.All 40 patients were included at the Department of Oncologyand Medical Physics, HUH. Seven patients had parts of theirtreatment and follow-up at the Department of Oncology, OsloUniversity Hospital (OUH).
Follow-up was originally completed in August 2017, withassessments for prolonged follow-up performed in January 2018and April 2019.
Patient RegistrationsAt baseline, patients recorded severity of a range of commonME/CFS symptoms including PEM, fatigue, cognitive symptomsand pain, using a numerical rating scale of 1–10. During18 months follow-up, patients were asked to complete asymptom questionnaire every 2 weeks, recording change or nochange to the same range of symptoms. The relative scale forsymptom change ranged from 0 to 6, in which three denotedno change from baseline; 4, 5, and 6 slight, moderate, andmajor improvement; and 2, 1, and 0 slight, moderate, andmajor worsening, respectively. This scale was adapted fromthe validated Clinical Global Impression Scale, which has beenused previously in ME/CFS (42). The primary outcome variableFatigue score, which has not been validated, was calculated everysecond week during follow-up as the mean change score forthe four fatigue-related items: “Fatigue,” “PEM,” “Need for rest,”and “Daily functioning.” At baseline and every 2 weeks, patientsalso recorded their percent function level on a scale from 1 to100%, where 100% denoted a completely healthy state. A set ofexamples was provided to facilitate this assessment. Samples ofall questionnaires are enclosed under Supporting Information(Data Sheet 1). Outcome measures also included the Short Form36 Health Survey (SF-36) ver. 1.2 in Norwegian translation(43, 44), at baseline, every 3 months during follow-up and atextended follow-up assessments at 24–30 and 38–48 months.Fatigue Severity Scale was recorded at 3-months intervals until18 months (45, 46). Physical activity level was recorded usingan electronic SenseWear armband continuously for 5 to 7 daysin a home setting (47, 48), at baseline and repeated in the timeintervals 7–9, 11–12, 17–18, 24–30, and 38–48 months after startof treatment.
Intervention and Follow-UpSix 30-minute intravenous infusions of cyclophosphamidewere administered at 4-week intervals with 600 mg/m2 atthe first and 700 mg/m2 at further cycles. Patients received
premedication with ondansetron 8mg and dexamethasone 4mg,when necessary enforced by aprepitant 125mg day 1, and80mg days 2 and 3. Patients with hematuria or dysuria inprevious cycles were given oral uromitexan (Data Sheet 1).Patients used cold-caps (Elasto-Gel R©, Southwest technologies,North Kansas City, USA) during infusions to reduce hairthinning. Each infusion was preceded by routine blood tests,including hematology, and a visit with a physician or studynurse. After the first and second infusions, a nadir bloodsample was collected between days 10 and 14 after infusion.If there were no signs of neutropenia or thrombocytopeniaafter the first two treatments, no further blood tests betweentreatments were required. Throughout the 18 months follow-up, patients attended consultations with an investigator every3 months. Adverse events were registered continuously at eachtreatment visit and at follow-up every 3 months and summarizedaccording to Common Toxicity Criteria for Adverse Events(CTCAE) ver. 4.03. The Viedoc R© electronic CRF system (PCGSolutions) was used for data collection and management in thestudy. There were no interim analyses. The trial was externallymonitored by the Department for Research and Developmentat HUH.
OutcomesResponse to treatment was defined as Fatigue score ≥4.5 for aminimum of 6 consecutive weeks, occurring at any time pointduring treatment or within 18 months follow-up. The trialhad two primary endpoints based on this definition: (i) overallresponse rate and (ii) changes in Fatigue score compared tobaseline through 18 months follow up. These endpoints werealso analyzed separately for the treatment-naïve patients (with noprevious rituximab exposure).
Secondary endpoints included: (i) response durationcalculated as the sum of response periods each of at least sixconsecutive weeks with mean Fatigue score ≥4.5; and changesfrom baseline to specific timepoints of (ii) SF-36 scores forPhysical Function subscale (SF-36-PF) and Physical componentsummary score (SF-36-PCS); (iii) self-reported percent functionlevel; (iv) mean number of steps per 24 h. Adverse events duringthe 18 months of follow-up from start of treatment were anadditional secondary endpoint.
HLA TypingHigh-resolution HLA genotyping was conducted as part of alarger study (30). In short, HLA-A, -B, -C, -DRB1, -DQB1, -DQA1, and -DPB1 alleles were genotyped using NGSgo kits andNGSengine software from GenDX (Utrecht, the Netherlands),and 2 × 150 bp paired-end sequencing on a Miseq instrument(Illumina, San Diego, USA) at the Norwegian Sequencing Centre,Oslo. The association analysis between HLA risk alleles andclinical response was not specified in the protocol, and wasperformed retrospectively in the data analysis phase. Only thepotential HLA risk alleles identified by Lande et al. (30), i.e.,HLA-C∗07:04 and HLA-DQB1∗03:03, were investigated.
Frontiers in Medicine | www.frontiersin.org 3 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
Statistical AnalysisDescriptive methods were used to characterize the sample, withmean and standard deviation (SD) for normally distributed data,and median with range [min-max, or interquartile range (IQR)]for skewed data. Primary and secondary outcome measureswere analyzed by the intention-to-treat principle. Changes frombaseline through 18 months follow-up were assessed by GeneralLinear Model for repeated measures (GLM), including time asa predictor. Greenhouse-Geisser corrections were used for allGLM analyses because Mauchly’s tests were significant (p <
0.001), indicating violations of the sphericity assumption. Thechanges through follow-up, compared to baseline, were assessedby the within-subjects effects for time. Simple contrasts in thetime domain were used to assess the changes from baselineto each specific time interval or time point during follow-up,with the effect sizes from the parameter estimates [means and95% confidence intervals (CI)]. To assess differences betweengroups GLM repeated measures were performed with p-value(Greenhouse-Geisser corrected) from the interaction time-by-group. Groups analyzed were sex, ME/CFS severity, ME/CFSduration, previous rituximab treatment, infection prior to debutof ME/CFS symptoms, and specific HLA alleles. The distributionof sex, ME/CFS severity and the proportion of respondersamong carriers and non-carriers of the two aforementionedHLA-alleles, were compared using Odds Ratio (OR) and Fisher’sexact tests.
All tests were two-sided with a significance level of 0.05.Missing data were replaced using the last value carried forward(LVCF) method. All analyses were performed using IBM SPSSStatistics ver.25 (IBM Corp., Armonk, USA), and GraphpadPrism ver.8 (GraphPad Software, La Jolla, USA).
Role of the Funding SourcesThe research group for ME/CFS at Department of Oncology andMedical Physics (HUH) has received funding from the KavliTrust and the Norwegian Ministry of Health and Care Services.The HLA sequencing has received funding from the Kavli Trustand Norwegian Research Council. The funders had no role intrial design, data collection, analysis, decision to publish, orpreparation of the manuscript.
RESULTS
Study PopulationThe flow chart for patient screening, inclusion, treatment andfollow-up is shown in Figure 1. Among available referralswith adequate medical information, we randomly selected 50patients for eligibility screening. Ten patients were excludeddue to violation of eligibility criteria, or declined to participate.We included 25 rituximab-naïve patients and 15 patients withprevious rituximab intervention.
Table 1 shows baseline characteristics for all included patients(n= 40), the rituximab-naïve patients (n= 25), and patients with(n = 22) or without (n = 18) a response to cyclophosphamideaccording to the definition of the primary endpoint of the study.
Medical history and concomitant diseases atbaseline, and concomitant medication during study
follow-up, are summarized in Supplementary Tables 1, 3.Supplementary Table 2 shows previous treatment by trialparticipants. Some kind of cognitive therapy had been triedby 52.5%, graded exercise or other physical therapy by 45.0%,adaptive pacing by 37.5%, vitamin B12 injections by 40.0%, andlow dose naltrexone by 37.5%. None of the patients received anyalternative intervention aimed at ME/CFS during the trial.
Thirty-one patients received all preplanned six infusions,three patients received five infusions, four received fourinfusions, and two received three infusions (Figure 1). Thereasons for omitting infusions were either withdrawal of consent(two cases after cycle 4), or high symptom burden (seven cases).All the decisions to omit infusions were in agreement with thetrial investigators. Thus, nine patients (22.5%) deviated from theplanned treatment protocol.
Missing DataFor the 18 months study period, there were missing data for thetwo patients who withdrew from study after ∼5 months (bothnon-responders at the time of withdrawal), and for one non-responding patient with severe ME/CFS who failed to completeself-reported forms from 4 months onwards. Except for thesethree patients, there were eight missing data items out of 1,560raw data for the variable Fatigue score. SenseWear activityarmband data were complete at baseline, and had missing datafrom the two withdrawals during follow-up.
Primary OutcomeThe overall response rate, i.e., proportion of patients with Fatiguescore ≥4.5 for at least six consecutive weeks, was 22 out of 40patients (55.0%, 95% CI 39.8–69.3%). Among the rituximab-naïve patients, 14 out of 25 patients achieved a clinical response(56.0%, 95% CI 36.9–73.4%).
Changes in Fatigue score during 18 months follow-up, withcomparisons of mean Fatigue score at each 3-month intervalto baseline are shown in Figure 2, for all patients (Figure 2A),rituximab-naïve patients (Figure 2B), patients with a response(Figure 2C), and no response during follow-up (Figure 2D).Repeated measures of Fatigue score showed significant increasesfrom baseline, with similar improvements among the rituximab-naïve patients as observed in all patients. The Fatigue scoreincreased significantly from baseline to 9 months after start oftreatment and further through 18 months follow-up. Amongthe 18 patients with no response, the Fatigue score decreasedsignificantly from baseline to 3 and 6 months, and thereafterreturned to near baseline level. Figure 2 also shows thecourses of mean Fatigue score through 18 months’ follow-up, subgrouped by ME/CFS disease severity (Figure 2E), andby presence/absence of HLA risk alleles (Figure 2F) in whichpatients with HLA-DQB1∗03:03 and/or HLA-C∗07:04 reportedhigher improvements of Fatigue score through follow-up thanthose negative for these alleles (p= 0.05).
Secondary OutcomesChanges of SF-36-PF and percent function level through each3-month interval, and mean steps per 24 h (at baseline, 7–9, 11–12, and 17–18 months), are shown in Figures 3A–L. Outcomes
Frontiers in Medicine | www.frontiersin.org 4 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
FIGURE 1 | Flow diagram of enrollment, treatment and follow-up in the CycloME study.
are shown for all patients and for the rituximab-naïve group,as well as for patients with and without response accordingto the study criteria. There were significant improvements ofall outcome variables from baseline through 18 months follow-up among all 40 patients, with mean SF-36-PF increasing from33.0 at baseline to a maximum 51.5 at 18 months follow-up(p < 0.001). Among 25 rituximab-naïve patients, mean SF-36-PF increased from 34.0 at baseline to 49.8 at 18 months(p = 0.001). Among 22 responders, mean SF-36-PF increasedfrom 35.0 at baseline to 69.5 at 18 months (p < 0.001). For18 non-responders there was only a slight increase of SF-36-PF from 30.6 at baseline to a maximum of 34.4 at 3 months,and with no significant changes through the remaining studyfollow-up. Similar patterns of significant changes were seenthrough follow-up, as compared to baseline, for percent functionlevel and for mean steps per 24 h, and also for SF-36-PCS(not shown).
Figure 4 shows the courses of SF-36-PF by subgroups.There were no significant interactions time-by-group forsex, severity, disease duration, infection prior to ME/CFS,or previous treatment with rituximab, i.e., the changes inSF-36-PF over time were similar in all subgroups, exceptfor HLA risk allele defined subgroups (see below). Thereason for showing SF-36-PF in these plots was to enablecomparison of data to other reported studies, in which SF-36-PF has often been used. There was no significant overallinteraction between time and ME/CFS severity (p = 0.51),although the small group (n = 6) with severe diseasehad no clinically relevant increase in SF-36-PF, from 8.3at baseline to a maximum of 11.7 at 12 months follow-up. The severe ME/CFS group included two patients withmissing data (one withdrawal and one who failed to completeregistration). However, seven patients with moderate-to-severedisease had similar improvements of the outcome measures
Frontiers in Medicine | www.frontiersin.org 5 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
TABLE 1 | Baseline characteristics of the study population are shown for the intention-to-treat population, for rituximab-naïve patients and for patients with or without
clinical response.
Characteristic All patients (n = 40) Rituximab-naïvea (n = 25) Respondersb (n = 22) Non-respondersc (n = 18)
Female, n (%) 31 (77.5) 18 (72.0) 18 (81.8) 13 (72.2)
Male, n (%) 9 (22.5) 7 (28.0) 4 (18.2) 5 (27.8)
Age, female pts, mean (min–max) 43.0 (25.0–61.1) 41.5 (26.6–54.6) 41.8 (25.0–60.3) 44.6 (26.6–61.1)
Age, male pts, mean (min–max) 37.6 (21.5–53.3) 35.1 (21.5–50.8) 39.5 (21.5–53.3) 36.0 (23.4–50.8)
BMI female ptsd, mean (min–max) 24.5 (17.1–33.1) 24.6 (17.1–33.1) 24.1 (17.1–32.7) 24.9 (19.0–33.1)
BMI male ptsd, mean (min–max) 24.5 (17.4–30.6) 23.4 (17.4–29.2) 25.9 (17.4–30.6) 23.4 (21.1–26.9)
Rituximab-naïvea, n (%) 25 (62.5) 25 (100.0) 14 (63.6) 12 (66.7)
Previous rituximab treatmente, n (%) 15 (37.5) 0 9 (40.9) 6 (33.3)
ME/CFS disease duration
2–5 years, n (%) 7 (17.5) 7 (28.0) 5 (22.7) 2 (11.1)
5–10 years, n (%) 13 (32.5) 7 (28.0) 5 (22.7) 8 (44.4)
10–15 years, n (%) 9 (22.5) 4 (16.0) 6 (27.3) 3 (16.7)
>15 years 11 (27.5) 7 (28.0) 6 (27.3) 5 (27.8)
ME/CFS disease severity
Mild/Moderate, n (%) 14 (35.0) 10 (40.0) 9 (40.9) 5 (27.8)
Moderate, n (%) 13 (32.5) 7 (28.0) 9 (40.9) 4 (22.2)
Moderate/severe, n (%) 7 (17.5) 5 (20.0) 4 (18.2) 3 (16.7)
Severef, n (%) 6 (15.0) 3 (12.0) 0 6 (33.3)
Infection prior to ME/CFSg, n (%) 26 (65.0) 17 (68.0) 15 (68.2) 11 (61.1)
SF36 Physical Functionh, mean (min–max) 33.0 (0–65) 34.0 (0–65) 35.0 (10–65) 30.6 (0–65)
SF36 Physical component summary scorei, mean (min–max) 23.3 (13.5–41.6) 24.5 (14.6–41.6) 23.1 (13.5–41.6) 23.5 (14.6–31.0)
Steps, mean per 24 h, mean (min–max) 3,199 (568–9,637) 3,282 (568–9,637) 3,622 (1,083–8,178) 2,681 (568–9,637)
Total function levelj, mean (min–max) 16.9 (5–40) 17.0 (5–30) 19.3 (10–40) 14.1 (5–25)
HLA-DQB1*03:03 pos, n (%)k 10 (25.0) 6 (24.0) 9 (40.9) 1 (5.6)
HLA-C *07:04 pos, n (%) 4 (10.0) 2 (8.0) 3 (13.6) 1 (5.6)
HLA-DQB1*03:03 and/or HLA-C*07:04 pos, n (%) 12 (30.0) 6 (24.0) 10 (45.5) 2 (11.1)
aPatients with no previous rituximab intervention.bClinically significant responders, including 18 patients with long response duration (≥30 weeks), three with moderate response duration (14–28 weeks) and one with marginal response
duration (6–12 weeks).cPatients with no clinically significant response.dBody Mass Index (kg/m2 ).ePatients treated with rituximab in previous trial (KTS-2-2010) n = 14, or outside a clinical trial (n = 1).fTwo of six patients with severe ME/CFS withdrew from the study after four infusions.gSelf-reported infection prior to onset of ME/CFS disease.hShort Form 36 (SF-36) physical function subscale (scale 0–100).iSF-36 Physical Health Summary Score, norm-based with population mean 50.jBaseline self-reported function level (scale 0–100%).kHLA-types determined as part of a larger study (30).
as patients with either moderate or mild-to-moderate disease.Supplementary Figure 1 shows the courses during follow-up, forthe SF-36 subscales Vitality, Social Function, and Bodily Pain(Supplementary Figures 1A–F), and also the Fatigue SeverityScale (Supplementary Figures 1G,H), all showing that theresponders report improvement during follow-up which weinterpret to be of clinical significance.
Out of nine patients included in the trial who had receivedprevious rituximab treatment without reporting improvementof ME/CFS symptoms, four achieved a clinical response aftercyclophosphamide intervention. Patients with HLA allelesHLA-DQB1∗03:03 and/or HLA-C∗07:04 reported higherimprovements of SF-36-PF through follow-up than thosenegative for these alleles (p= 0.05) (Figure 4F).
Clinical Response DurationsAmong the 22 patients with response, the total duration orresponse was median 44 weeks (range 6–70 weeks) within 18months follow-up. Themedian ratio of clinical response durationto follow-up was 0.56 (range 0.08–0.90). Response duration was≥ 30 weeks in 18 patients, 14–28 weeks in three patients, and6–12 weeks in one patient.
The median time to first response was 22 weeks (range 2–42 weeks). There were no significant differences in time to firstresponse by sex, disease severity, disease duration, infectionprior to ME/CFS, or by previous rituximab treatment (datanot shown).
Out of 22 responders, 17 patients (77.3%) reported a sustainedresponse with Fatigue score of least 4.5 at the end of 18 months
Frontiers in Medicine | www.frontiersin.org 6 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
FIGURE 2 | Fatigue score (primary end point), means with 95% CI at time points through 18 months follow-up, from self-reported symptom scores every second
week. The scale is 0–6, where 3 indicates no change from baseline and higher scores indicate less fatigue. (A): All included patients (n = 40). (B): Treatment-naïve
patients (not previously exposed to rituximab, n = 25). (C): Responders during follow-up (n = 22). (D) Non-responders during follow-up (n = 18). P-values from
General Linear Model repeated measures assessing changes in Fatigue score from baseline. (E,F) show mean Fatigue score (with 95% CI) through 18 months’
follow-up, subgrouped by ME/CFS disease severity (E), and presence/absence of HLA risk alleles (F). P-values from General Linear Model for interaction
time-by-group, assessing difference between subgroups in repeated measures of Fatigue score over time compared to baseline. P-values: * <0.05; ** <0.01;
*** <0.001. CI, confidence intervals.
follow-up. Among all 40 included patients, 21 (52.5%) reporteda Fatigue score of at least 4.0 (slight improvement) at endof follow-up.
Prolonged Follow-UpFollowing two approved protocol amendments, patients hadadditional visits or telephone interviews with recordings ofSF-36 and percent function level and SenseWear physicalactivity measurements at 24–30 and 38–48 months follow-up. Due to the risk of recall bias, Fatigue score comparedto baseline was not recorded at these late visits. Instead,patients were asked to self-assess whether their symptoms hadrelapsed, remained unchanged or had improved further sincethe end of trial (18 months). The changes of SF-36-PF, percentfunction level and mean steps, from baseline until extendedfollow-up at 38–48 months, by response status, are shown inFigures 5A–C.
At the 38–48 months visit, 36 out of 38 patients stillin the study completed the interview including assessmentof their percent function level, 35 recorded SF-36 formsand 32 completed SenseWear activity measurements. Out of22 responders, 20 completed the interview; 15 were still inremission, while five reported a complete or partial relapse.
For 20 responders with available SF-36 recordings at 38–48 months, the mean SF-36-PF was 70.8 (range 25–100)compared to mean 69.5 at 18 months. SenseWear activityregistration was available for 19 out of 22 responders at38–48 months with mean 6,415 steps per 24 h (SD 2,764),compared to mean 5,589 (SD 2,017) at 18 months (Figure 5C).Six patients with missing SenseWear data at 38–48 monthsincluded two responders in ongoing remission, one in relapse andthree non-responders.
At baseline, only two of the responders had part-time workparticipation. During follow-up, at least nine out of 22 respondersreturned to either part-time of full-time work or studies.
HLA DataTwelve of the 40 patients (30.0%) carried either of the twospecific HLA risk alleles. Ten of the 12 patients (83.3%)positive for HLA alleles DQB1∗03:03 and/or C∗07:04 had aresponse, compared to 12 out of 28 patients (42.9%) negativefor these HLA alleles (OR = 6.67; p = 0.028; Figure 6). Theallele HLA-C∗07:04 was present in four out of 40 patients(10.0%), and three (75.0%) of these were responders. HLA-DQB1∗03:03 was detected in 10 out of 40 patients (25.0%),and 9 out of 10 (90.0%) were responders, compared to 13
Frontiers in Medicine | www.frontiersin.org 7 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
FIGURE 3 | SF-36 Physical Function (SF-36-PF) (A–D), percent function level (E–H), and mean steps per 24 h (I–L), means with 95% CI, at time points through 18
months follow-up. Outcome data for all included patients (n = 40) (A,E,I); Treatment-naïve (not previously exposed to rituximab, n = 25) (B,F,J); Responders during
follow-up (n = 22) (C,G,K); Non-responders during follow-up (n = 18) (D,H,L). P-values from General Linear Model repeated measures assessing changes in
outcome variable from baseline. P-values: * <0.05; ** <0.01; *** <0.001. SF-36 Physical Function with scale 0–100, higher number indicates better function. CI,
confidence intervals; SF-36, Short Form 36.
out of 30 HLA-DQB1∗03:03 negative patients (43.3%); OR =
11.8, p= 0.013.Among 12 patients with the HLA alleles DQB1∗03:03
and/or C∗07:04, 7 (58.3%) had mild-to-moderate, 3 (25%)had moderate, and 2 (16.7%) had moderate-to-severeME/CFS. Contrary, among 28 patients without these HLAalleles, 7 (25.0%) had mild-to-moderate, 10 (35.7%) hadmoderate, and 11 (39.3%) had moderate-to-severe ME/CFS(p = 0.05). Eleven out of 12 patients (91.7%) with HLAalleles DQB1∗03:03 and/or C∗07:04 were female, comparedto 20 out of 28 patients (71.4%) without these alleles(p= 0.23).
Adverse EventsAdverse events (AE) for the complete period of 18 monthsfollow-up are shown in Table 2. Thirty-three patients (82.5%)reported AEs of CTCAE grade ≥ 2, of which gastrointestinalevents such as nausea and constipation were the mostcommon. Out of 16 grade 3–4 events in 11 patients, 11resulted in hospitalization and were reported as serious adverseevents (SAE, Supplementary Table 4). There was one suspectedunexpected serious adverse reaction (SUSAR), in a femalepatient with moderate-to-severe ME/CFS who was a non-responder in the study. She experienced gradual worseningof postural orthostatic tachycardia syndrome (POTS) after
Frontiers in Medicine | www.frontiersin.org 8 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
FIGURE 4 | SF-36 Physical Function (SF-36-PF), means with 95% CI through 18 months follow-up, by subgroups. P-values from General Linear Model for interaction
time-by-group, assessing differences between the subgroups in repeated measures of SF-36-PF over time, compared to baseline. (A): By men vs. women; (B) By
ME/CFS severity; (C): By ME/CFS disease duration; (D) With or without self-reported infection prior to ME/CFS; (E) With or without previous rituximab treatment; (F):
With or without HLA risk alleles (HLA DQB1*03:03 and/or HLA C*07:04). SF-36, Short Form 36; CI, confidence intervals; HLA, Human Leukocyte Antigen.
cycle 4, resulting in hospital admission for 2 weeks. She hadexperienced periods of similar POTS aggravations regularlysince she became ill with ME/CFS 18 years before studyinclusion. Her POTS symptoms gradually returned to baselinelevel, but study treatment was discontinued. With routineblood sampling before each cycle and after cycle 1 and 2,there was no sign of hematological toxicity. Two womenboth aged ≥41years at inclusion, experienced menopause afterstart of treatment, two others reported irregular menstrualbleeding that persisted to end of follow-up. One patientwithout a clinical response suffered a sudden death ofunknown cause 4 years after inclusion in the study, i.e., 42months after the last infusion, with no probable relation tothe intervention.
DISCUSSION
The present open-label phase II study with cyclophosphamideinfusions was well conducted with little missing data. Morethan half of the patients had clinical response according tothe predefined criteria, many with long-lasting improvement ofsymptoms. At extended follow-up 3–4 years after inclusion 68%of responders were still in remission.
In general, the toxicity to cyclophosphamide infusions inME/CFS patients was moderate, and there were few seriousadverse events and no registered hematological toxicity. Themost common side effects were nausea and general malaiselasting for 1–2 weeks after each infusion. ME/CFS patientsreported more nausea and discomfort after cyclophosphamide
Frontiers in Medicine | www.frontiersin.org 9 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
FIGURE 5 | Outcome variables at extended follow-up, by response status. Means with 95% CI at different time points including extended follow-up at 24–30 and at
38–48 months. (A): SF-36 Physical Function (SF-36-PF); (B) percent function level; (C): mean steps per 24 h. Numbers of patients at the different time points through
follow-up are shown below the graphs. SF-36, Short Form 36; CI, confidence intervals.
FIGURE 6 | Frequency of HLA risk alleles (HLA DQB1*03:03 and/or HLA
C*07:04) in responders and non-responders during follow-up. P-value from
Fischer’s exact test. HLA, Human Leukocyte Antigen.
than cancer patients typically do at similar doses, in linewith the generally low stress tolerance and sensitivity to drugsreported by many patients. We reinforced the anti-emeticregimen with aprepitant during the study in efforts to reducethe nausea experienced by the patients during the first daysafter infusion. Fertility concerns are an important toxicity issuewith chemotherapy. Cyclophosphamide is an alkylating agentassociated with ovarian failure and the risk increases withhigher cumulative doses and with increasing age (49, 50). Onestudy with intravenous infusions, applying similar cumulativedoses (mean 9.1 gram) as in the present study, and meanage 31 years, reported ovarian failure in 13%, and transient
amenorrhea in 20% of the patients (51). In our present trial, twowomen aged 41 and 46 years at inclusion experienced prematuremenopause, and two others reported irregular menstruationprobably induced by the treatment at end of follow-up. Incontrast to spontaneous premature menopause, chemotherapyassociated ovarian dysfunction can resume over time (years) insome patients, even after a prolonged period of amenorrhea andelevated gonadotropin levels (52).
Since the 6-month initial treatment period with repeatedcyclophosphamide infusions in some patients led to increasedsymptom burden and side effects, the extent and duration ofimprovement in ME/CFS symptoms are important aspects tojustify the intervention. We therefore extended the follow-upperiod, and collected additional clinical data from participants,at 2–3 and 3–4 years after inclusion. The response durations weresustained for most of the responders. Out of 22 responders 82%were still in remission at 2–3 years and 68% at 3–4 years extendedfollow-up. Seven even reported further improvement comparedto their status at 18 months follow-up. Also of note, three of thepatients who registered relapse at 3–4 years still reported a 2-foldincrease of their percent function levels as compared to baseline.Thus, responders’ self-reported percent function levels, SF-36Physical Function with increase from mean 35 at baseline tomean >70 at 12 months, and measured levels of physical activity(steps per 24 h), reflect clinically meaningful improvements oftheir abilities and activities of daily life. For comparison, themeanSF-36 Physical Function in the general population is 84.2 (95%CI71.9–96.5) (53).
Compared to the randomized RituxME trial assessingrituximab vs. placebo in ME/CFS patients (35), the patternsof improvement among patients in the present CycloME trialseemed to be more homogeneous. In CycloME the clinicalresponses occurred earlier than in the RituxME trial; at median22 weeks compared to 41 weeks. In the CycloME study theresponse rates were comparable between men and women, as
Frontiers in Medicine | www.frontiersin.org 10 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
TABLE 2 | Patients with adverse events of CTCAE grade 1–4 during 18 months
follow-up.
≥1 ≥2 3–4* Related to
study
treatment†
Patients with ≥ 1
adverse event
39 (97.5%) 33 (82.5%) 11 (27.5%) 29 (72.5%)
Nausea 36 (90%) 15 (37.5%) 0 36 (90%)
Constipation 22 (55%) 9 (22.5%) 1 (2.5%) 19 (47.5%)
Diarrhea 7 (17.5%) 1 (2.5%) 0 6 (15%)
Stomach pain 9 (22.5%) 2 (5%) 1 (2.5%) 7 (17.5%)
Infections 24 (60%) 15 (37.5%) 3 (7.5%) 13 (32.5%)
Irregular menstrual
bleeding
7 (17.5%) 3 (7.5%) 0 7 (17.5%)
Premature
menopause
2 (5%) 1 (2.5%) 0 2 (5%)
Haematuria 6 (15%) 1 (2.5%) 0 6 (15%)
Urinary bladder
symptoms**5 (12.5%) 3 (7.5%) 0 5 (12.5%)
Hair loss 4 (10%) 0 0 4 (10%)
Rash or urticaria 6 (15%) 4 (10%) 1 (2.5%) 5 (12.5%)
Headache 12 (30%) 3 (7.5%) 1 (2.5%) 9 (22.5%)
Dizziness 6 (15%) 3 (7.5%) 0 5 (12.5%)
Edema of face or
limbs
6 (15%) 0 0 5 (12.5%)
Palpitations or
tachycardia
4 (10%) 2 (5%) 1 (2.5%) 2 (5%)
*11 grade 3–4 events for 8 patients were reported as SAE. See Supplementary Table 4
for details.†Possible, probable or very likely relation to study treatment.**Bladder/urinary tract pain or increased urinary frequency.
opposed to higher response in women in the RituxME trial.The response rates were higher among patients with moderateor moderate-to-severe disease, compared to the 4 patients withsevere ME/CFS who completed the intervention. In an ongoingaddition to the trial (part B), feasibility and response rateare investigated in a small number of additional patients withsevere ME/CFS, to gain experience and to decide whether severepatients may be included in a possible future randomized trialassessing cyclophosphamide intervention.
The response rates were similar among patients who wererituximab-naïve and patients who had participated in previoustrials with rituximab intervention (32, 33). Also, four out ofnine patients with no improvement after previous rituximabintervention experienced clinical benefit after cyclophosphamidein the present study.
Interestingly, the presence of either of the two HLA riskalleles, previously shown to be associated with ME/CFS (HLA-DQB1∗03:03 andHLA-C∗07:04) (30), was predictive for responseto cyclophosphamide. In contrast there was no associationbetween presence of these HLA alleles and clinical improvementamong patients included in the RituxME trial (35) (datanot shown).
The carrier frequency of any of these HLA risk alleles was30% among ME/CFS patients in this trial, which is higherthan the 19.1% reported in the recent study of 426 Norwegian
ME/CFS patients (30). Western Norway is well represented inthis large cohort, and the frequency of DQB1∗03:03 and C∗07:04from Western Norway sources did not differ from the nationalfrequency (data not shown). Therefore, geographical bias is not aprobable explanation.
The association between cyclophosphamide response andthe HLA risk alleles could be due to a true treatment effectin individuals carrying these alleles. There are several reportsof associations between specific HLA alleles/haplotypes andresponses to immune modulatory treatments (54–57), but toour knowledge this has not been demonstrated specificallyfor cyclophosphamide. Another possibility is that carriersof these HLA risk alleles constitute a subgroup amongME/CFS patients with an immune-driven pathomechanismgenerally responding better to immune modulating treatment.Finally, the observed association between the HLA risk allelesand response to cyclophosphamide could be coincidental,but warrants further investigation in a possible futurerandomized trial.
There are no biomarkers for ME/CFS or disease activity,and assessments of symptom changes consequently have torely largely on self-recorded subjective variables. To increasethe validity of the measurements, we used several differentvariables to measure symptom changes. These variables generallyshowed the same patterns of improvement and worsening ofME/CFS symptoms during the follow-up period. Self-reportedimprovements in Fatigue score, percent function level and SF-36 Physical Function scores correlated well, and with increasedlevels of physical activity. “Steps per 24 hours” is an objectivemeasure, but not a perfect way to validate symptom improvementbecause individual patients will use their improved energy fordifferent purposes. Some will walk, while some will prefer to reador increase the time for social activity.
The initial patient observations in our cancer clinic, ofpatients with long-standing ME/CFS who developed cancer,and who reported relief of ME/CFS symptoms after cancertreatment, included seven cases treated with cyclophosphamide(or ifosfamide), and in one case the combination ofcyclophosphamide and rituximab. Our hypothesis was thatME/CFS in a subgroup of patients could be caused by animmunological dysfunction, possibly with a variant of anautoimmune pathomechanism. In the present study, thefrequency of self-reported infection prior to ME/CFS debut(65%) was in line with other reports (58). Also, there was ahigh occurrence of autoimmunity among first-degree relatives(55.0%). Both observations may support an immunological basisfor the disease. Initial phase II studies with rituximab (32, 33)suggested that a subgroup of patients could benefit from B-celldepletion therapy. Conversely, in the double-blind, placebo-controlled, multicenter, phase III RituxME trial there were nosignificant differences between the rituximab and placebo groupsfor any of the primary or secondary outcome measures (35).Taking the RituxME results into account, we have to interpretthe data from the present open-label CycloME trial with caution.Patient selection, placebo mechanisms, patient’s expectationsin clinical trials, and natural variation of symptoms over timemay be operative (59, 60). Until a randomized trial has been
Frontiers in Medicine | www.frontiersin.org 11 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
performed, there is not sufficient evidence for a beneficial effectof cyclophosphamide in ME/CFS patients.
Other study limitations are self-referral, use of self-reportedprimary outcome measures with possible recall bias, and theinclusion of patients who had participated in previous studieswith rituximab intervention. Although inclusion relied on strictdiagnostic criteria, the unknown etiology of ME/CFS and lack ofspecific biomarkers could introduce unintended heterogeneity ofthe patient sample.
When comparing the response data from the CycloME andRituxME studies, it is important to consider the completelydifferent modes of action of the two drugs. Rituximabis a monoclonal antibody which selectively depletes B-cells expressing the CD20 protein on their surface, whilecyclophosphamide has broader immunosuppressive effectson several subsets of lymphocytes. The main mechanism ofcyclophosphamide is the ability to covalently bind an alkylgroup, affecting mainly the DNA (61). This interaction isirreversible and leads to inhibition of DNA replication andapoptosis, producing cell death amongst resting and dividingwhite blood cells and leading to impaired humoral and cellularimmune responses (62). Rapidly proliferating cells are mostsensitive to cyclophosphamide (41). This feature is utilized incancer therapy, but also to influence activated immune cellsthat are present in different immune-mediated diseases (37).The effects and side-effects of cyclophosphamide are highlydose dependent. High doses can be used for the completeeradication of hematopoietic cells, but lower doses are relativelyselective for T-cells, especially T-regulatory cells (T-regs).Cyclophosphamide affects T-regs, which have a generallyhigher proliferation rate than other T-cell subsets such as theT-helper (Th) cells, but also affects B-cells and other cellsof the immune system (41). T-regs have an important rolein down-regulating the effects of Th cells, and help preventautoimmune diseases by maintaining self-tolerance (63). Ahigher frequency of T-regs in ME/CFS patients compared tohealthy controls has been reported in some studies (64–66).The T-reg markers are also general T-cell activation markers(63). Thus, cyclophosphamide may interfere with the balancebetween immune cell subsets and possibly counteract adisease-facilitating environment.
Although the double blinded RituxME trial showed nosignificant differences between the rituximab and placebogroups for the outcome measures (35), there may still bea subgroup of ME/CFS patients that have an autoantibody-mediated disease where only few patients have autoantibody-production from early CD20-positive plasmablasts that can betargeted by rituximab. Other patients may still have autoantibodyproduction, but from long-lived CD20-negative plasma cells.This mechanism is active in several rituximab refractoryautoimmune diseases and could be compatible both with the totalexperience from our rituximab trials, and with the data fromthe present cyclophosphamide trial. Cytotoxic chemotherapy,such as cyclophosphamide, may inhibit B-cell activation andproliferation to new antibody-secreting cells, thus reducing theshort-lived plasma cell compartment and recruitment of matureplasma cells (67).
If an autoantibody-mediated mechanism is operative in asubgroup of ME/CFS patients, the nature of possible endogenoustargets for pathogenic immunoglobulins is still elusive. Increasedserum levels of autoantibodies against several G-protein coupledreceptors have been shown in ME/CFS (16). Clinical symptomssuggest inadequate regulation of autonomic functions andblood flow, also demonstrated in a recent study of reducedcerebral blood flow during head-up tilt test with orthostaticstress using Doppler flow imaging of carotid and vertebralarteries (68). Recent observations of patients with unexplainedexertional intolerance and dyspnea demonstrated a subgroupwith low ventricular filling pressure (preload failure) in uprightposition during cardiopulmonary exercise tests, related toreduced venous pressure (69, 70). Also, in patients withunexplained exertional intolerance, a subgroup had impairedsystemic oxygen extraction, which may be associated withmicrocirculatory dysregulation or mitochondrial dysfunction(71). One might speculate on the possibility of an autoimmuneprocess indirectly of directly affecting blood vessels, or againstsmall nerve fibers including autonomic nerves regulating bloodvessel function. Small fiber neuropathy (SFN) is associatedwith fatigue, postural orthostatic tachycardia syndrome (POTS),gastrointestinal disturbances and abnormal sweating (72). SFNhas been demonstrated in 49% of fibromyalgia patients (73), andin up to 43% of patients with preload failure, many of whom hadsymptoms suggestive of ME/CFS (70). This could be associatedwith inadequate autoregulation of blood flow according to thedemands of tissues, with local hypoxia and lactate accumulationon limited exertion, and with metabolic adjustments which couldbe secondary and compensatory in efforts to restore cellularenergy balance (20, 21, 23, 74, 75). Microvasculopathy mayalso be reflected in arterial endothelial dysfunction which hasbeen demonstrated in ME/CFS (76), and also investigated insubstudies to the CycloME and RituxME trials (manuscriptsin preparation).
The growing evidence for immune disturbances in ME/CFS,experience with cyclophosphamide in other autoimmunediseases, with broad immunosuppressive effects on severallymphocyte subsets including B-cells and T-regs, and the hereinreported association between HLA risk alleles and clinicalresponse to cyclophosphamide intervention, support that theobserved relief of ME/CFS symptoms could be a drug effecttargeting the underlying disease mechanisms. We stronglyadvise patients and physicians not to use cyclophosphamide forME/CFS patients outside of clinical trials before a randomizedtrial has been conducted, to evaluate the possible benefits ofthe drug.
CONCLUSION
This study shows that cyclophosphamide intervention isfeasible for ME/CFS patients. The growing evidence forimmune alterations in ME/CFS and the high symptomburden with very low quality of life, we believe canjustify use of an immune modulating drug with possibleside effects. The treatment period was demanding for
Frontiers in Medicine | www.frontiersin.org 12 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
most patients, but in total the toxicity was interpreted asacceptable. The treatment was associated with long-lastingimprovements of ME/CFS symptoms for approximately halfof patients. However, due to the lack of a placebo group,response data must be interpreted with great caution.In the further work to find effective treatment, we willconsider a new multicenter, randomized, double-blind andplacebo-controlled trial with cyclophosphamide. Should thistrial prove cyclophosphamide to be beneficial for ME/CFSpatients, this could also be important in the search for relevantdisease mechanisms.
DATA AVAILABILITY STATEMENT
The datasets generated from this study are available onreasonable request to the corresponding author.
ETHICS STATEMENT
The studies involving human participants were reviewed andapproved by The Regional Committees for Medical and HealthResearch Ethics (2014/1672), and by the National MedicinesAgency in Norway. The patients/participants provided theirwritten informed consent to participate in this study.
AUTHOR CONTRIBUTIONS
IR, ØF, KS, and OM: conception and design. IR, ØF, AF, KT,AL, MV, BL, and OM: analyses and interpretation. IR, ØF, AF,IK-V, MH, KS, and OM: inclusion and follow-up of patients.IR, ØF, AF, IK-V, KS, KR, KA, MH, OM, AL, MV, and BL:collection and assembly of data. IR, ØF, KS, KR, KA, OD, andOM: administrative, technical, biobank and logistic support. IR,ØF, AF, KS, and OM: drafting the article. All authors: criticalrevision of the article and final approval of the manuscript.
FUNDING
The CycloME trial received funding from the Kavli Trust. Thestudy received partial funding from the Norwegian Ministry ofHealth and Care Services.
ACKNOWLEDGMENTS
We acknowledge the staff at the Department of Oncology,Haukeland University Hospital, the Clinical Research Unit,Haukeland University Hospital, and the outpatient clinic atthe Norwegian Radium Hospital (Oslo University Hospital) forproviding facilities to perform the clinical trial.
SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be foundonline at: https://www.frontiersin.org/articles/10.3389/fmed.2020.00162/full#supplementary-material
Supplementary Figure 1 | SF-36 subscales (raw scores) and Fatigue Severity
Scale during follow-up until 18 months, shown for the Intention-to-treat population
(n = 40) in (A,C,E,G), and for responders (n = 22) vs. non-responders (n = 18) in
(B,D,F,H). The SF-36 subscales for Vitality (A,B), Social Function (C,D), Bodily
Pain (E,F), and Fatigue Severity Scale (G,H) are shown. SF-36 subscales with
scale 0–100, higher number indicates better function. Fatigue Severity Scale with
scores 7–63, higher score indicates more fatigue. SF-36, Short Form 36; CI,
confidence intervals; SD, standard deviation.
Supplementary Table 1 | Medical history and concomitant diseases reported at
baseline, shown by System Organ Class (SOC) and CTCAE term.
Supplementary Table 2 | Previous treatments for ME/CFS, reported at baseline.
Supplementary Table 3 | Concomitant medication during 18 months follow-up
(shown by ATC-code).
Supplementary Table 4 | Serious Adverse Events during 18 months follow-up
(System Organ Class, CTCAE term, SAE category and relation to treatment).
Trial protocol.
Data Sheet 1 | Trial protocol.
REFERENCES
1. Brown A, Jason LA. Meta-analysis investigating post-exertional
malaise between patients and controls. J Health Psychol. (2018)
1359105318784161. doi: 10.1177/1359105318784161
2. Chu L, Valencia IJ, Garvert DW, Montoya JG. Deconstructing post-
exertional malaise in myalgic encephalomyelitis/ chronic fatigue
syndrome: a patient-centered, cross-sectional survey. PLoS ONE. (2018)
13:e0197811. doi: 10.1371/journal.pone.0197811
3. Castro-Marrero J, Zaragoza MC, Gonzalez-Garcia S, Aliste L, Saez-Francas
N, Romero O, et al. Poor self-reported sleep quality and health-related
quality of life in patients with chronic fatigue syndrome/myalgic
encephalomyelitis. J Sleep Res. (2018) 27:e12703. doi: 10.1111/jsr.
12703
4. Carruthers BM, Jain AK, De Meirleir KL, Peterson DL, Klimas NG, Lerner
AM, et al. Myalgic encephalomyelitis/ chronic fatigue syndrome: clinical
working case definition, diagnostic and treatment protocols. J Chronic Fatigue
Syndr. (2003) 11:7–36. doi: 10.1300/J092v11n01_02
5. Nacul LC, Lacerda EM, Pheby D, Campion P, Molokhia M, Fayyaz S, et al.
Prevalence of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS)
in three regions of England: a repeated cross-sectional study in primary care.
BMCMed. (2011) 9:91. doi: 10.1186/1741-7015-9-91
6. Falk Hvidberg M, Brinth LS, Olesen AV, Petersen KD, Ehlers
L. The health-related quality of life for patients with myalgic
encephalomyelitis / chronic fatigue syndrome (ME/CFS). PLoS ONE.
(2015) 10:e0132421. doi: 10.1371/journal.pone.0132421
7. Nacul LC, Lacerda EM, Campion P, Pheby D, Drachler Mde L, Leite
JC, et al. The functional status and well being of people with myalgic
encephalomyelitis/chronic fatigue syndrome and their carers. BMC Public
Health. (2011) 11:402. doi: 10.1186/1471-2458-11-402
8. Institute of medicine. The National Academies Collection: Reports funded
by National Institutes of Health. Beyond Myalgic Encephalomyelitis/Chronic
Fatigue Syndrome: Redefining an Illness. Washington, DC: National
Academies Press (2015).
9. Blomberg J, Gottfries CG, Elfaitouri A, Rizwan M, Rosen A.
Infection elicited autoimmunity and myalgic encephalomyelitis/chronic
fatigue syndrome: an explanatory model. Front Immunol. (2018)
9:229. doi: 10.3389/fimmu.2018.00229
10. Sharif K, Watad A, Bragazzi NL, Lichtbroun M, Martini M, Perricone C, et al.
On chronic fatigue syndrome and nosological categories. Clin Rheumatol.
(2018) 37:1161–70. doi: 10.1007/s10067-018-4009-2
11. Morris G, Berk M, Galecki P, Maes M. The emerging role of autoimmunity
in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). Mol
Neurobiol. (2014) 49:741–56. doi: 10.1007/s12035-013-8553-0
Frontiers in Medicine | www.frontiersin.org 13 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
12. Sotzny F, Blanco J, Capelli E, Castro-Marrero J, Steiner S, Murovska
M, et al. Myalgic encephalomyelitis/chronic fatigue syndrome -
evidence for an autoimmune disease. Autoimmun Rev. (2018)
17:601–9. doi: 10.1016/j.autrev.2018.01.009
13. Montoya JG, Holmes TH, Anderson JN, Maecker HT, Rosenberg-Hasson
Y, Valencia IJ, et al. Cytokine signature associated with disease severity in
chronic fatigue syndrome patients. Proc Natl Acad Sci USA. (2017) 114:E7150–
8. doi: 10.1073/pnas.1710519114
14. Brenu EW, Huth TK, Hardcastle SL, Fuller K, Kaur M, Johnston
S, et al. Role of adaptive and innate immune cells in chronic
fatigue syndrome/myalgic encephalomyelitis. Int Immunol. (2014)
26:233–42. doi: 10.1093/intimm/dxt068
15. Scheibenbogen C, Loebel M, Freitag H, Krueger A, Bauer S, Antelmann
M, et al. Immunoadsorption to remove beta2 adrenergic receptor
antibodies in Chronic Fatigue Syndrome CFS/ME. PLoS ONE. (2018)
13:e0193672. doi: 10.1371/journal.pone.0193672
16. Loebel M, Grabowski P, Heidecke H, Bauer S, Hanitsch LG, Wittke K,
et al. Antibodies to beta adrenergic and muscarinic cholinergic receptors
in patients with Chronic Fatigue Syndrome. Brain Behav Immun. (2016)
52:32–9. doi: 10.1016/j.bbi.2015.09.013
17. Tanaka S, Kuratsune H, Hidaka Y, Hakariya Y, Tatsumi KI, Takano T, et al.
Autoantibodies against muscarinic cholinergic receptor in chronic fatigue
syndrome. Int J Mol Med. (2003) 12:225−30.
18. Gunther OP, Gardy JL, Stafford P, Fluge O, Mella O, Tang P, et al.
Immunosignature analysis of myalgic encephalomyelitis/chronic
fatigue syndrome (ME/CFS). Mol Neurobiol. (2019) 56:4249–
57. doi: 10.1007/s12035-018-1354-8
19. Chang CM, Warren JL, Engels EA. Chronic fatigue syndrome and
subsequent risk of cancer among elderly US adults. Cancer. (2012) 118:5929–
36. doi: 10.1002/cncr.27612
20. Fluge O, Mella O, Bruland O, Risa K, Dyrstad SE, Alme K, et al.
Metabolic profiling indicates impaired pyruvate dehydrogenase function
in myalgic encephalopathy/chronic fatigue syndrome. JCI Insight. (2016)
1:e89376. doi: 10.1172/jci.insight.89376
21. Germain A, Ruppert D, Levine SM, Hanson MR. Metabolic profiling of a
myalgic encephalomyelitis/chronic fatigue syndrome discovery cohort reveals
disturbances in fatty acid and lipid metabolism. Mol Biosyst. (2017) 13:371–
9. doi: 10.1039/C6MB00600K
22. Yamano E, Sugimoto M, Hirayama A, Kume S, Yamato M, Jin G,
et al. Index markers of chronic fatigue syndrome with dysfunction
of TCA and urea cycles. Sci Rep. (2016) 6:34990. doi: 10.1038/
srep34990
23. Naviaux RK, Naviaux JC, Li K, Bright AT, Alaynick WA, Wang L, et al.
Metabolic features of chronic fatigue syndrome. Proc Natl Acad Sci USA.
(2016) 113:E5472–80. doi: 10.1073/pnas.1607571113
24. Nagy-Szakal D, Barupal DK, Lee B, Che X, Williams BL, Kahn EJR,
et al. Insights into myalgic encephalomyelitis/chronic fatigue syndrome
phenotypes through comprehensive metabolomics. Sci Rep. (2018)
8:10056. doi: 10.1038/s41598-018-28477-9
25. Albright F, Light K, Light A, Bateman L, Cannon-Albright LA. Evidence for
a heritable predisposition to Chronic Fatigue Syndrome. BMC Neurol. (2011)
11:62. doi: 10.1186/1471-2377-11-62
26. Schur E, Afari N, Goldberg J, Buchwald D, Sullivan PF. Twin analyses of
fatigue. Twin Res Hum Genet. (2007) 10:729–33. doi: 10.1375/twin.10.5.729
27. Keller RH, Lane JL, Klimas N, Reiter WM, Fletcher MA, van Riel F,
et al. Association between HLA class II antigens and the chronic fatigue
immune dysfunction syndrome. Clin Infect Dis. (1994) 18 (Suppl. 1):S154–
6. doi: 10.1093/clinids/18.Supplement_1.S154
28. Carlo-Stella N, Bozzini S, De Silvestri A, Sbarsi I, Pizzochero C, Lorusso L,
et al. Molecular study of receptor for advanced glycation endproduct gene
promoter and identification of specific HLA haplotypes possibly involved in
chronic fatigue syndrome. Int J Immunopathol Pharmacol. (2009) 22:745–
54. doi: 10.1177/039463200902200320
29. Smith J, Fritz EL, Kerr JR, Cleare AJ, Wessely S, Mattey DL. Association of
chronic fatigue syndrome with human leucocyte antigen class II alleles. J Clin
Pathol. (2005) 58:860–3. doi: 10.1136/jcp.2004.022681
30. Lande A, Fluge Ø, Strand EB, Flåm ST, Sosa DD, Mella O,
et al. Human leucocyte antigen alleles associated with Myalgic
Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS). Sci Rep. (2020)
10:5267. doi: 10.1038/s41598-020-62157-x
31. Pierpont TM, Limper CB, Richards KL. Past, present, and future of rituximab-
the world’s first oncology monoclonal antibody therapy. Front Oncol. (2018)
8:163. doi: 10.3389/fonc.2018.00163
32. Fluge O, Risa K, Lunde S, Alme K, Rekeland IG, Sapkota D, et al. B-
Lymphocyte depletion in myalgic encephalopathy/ chronic fatigue syndrome.
An open-label phase II study with rituximab maintenance treatment. PLoS
ONE. (2015) 10:e0129898. doi: 10.1371/journal.pone.0129898
33. Fluge O, Bruland O, Risa K, Storstein A, Kristoffersen EK, Sapkota D, et al.
Benefit fromB-lymphocyte depletion using the anti-CD20 antibody rituximab
in chronic fatigue syndrome. A double-blind and placebo-controlled study.
PLoS ONE. (2011) 6:e26358. doi: 10.1371/journal.pone.0026358
34. Fluge O, Mella O. Clinical impact of B-cell depletion with the anti-CD20
antibody rituximab in chronic fatigue syndrome: a preliminary case series.
BMC Neurol. (2009) 9:28. doi: 10.1186/1471-2377-9-28
35. Fluge O, Rekeland IG, Lien K, Thurmer H, Borchgrevink PC, Schafer C, et al.
B-lymphocyte depletion in patients with myalgic encephalomyelitis/chronic
fatigue syndrome: a randomized, double-blind, placebo-controlled trial. Ann
Intern Med. (2019) 170:585–93. doi: 10.7326/M18-1451
36. Emadi A, Jones RJ, Brodsky RA. Cyclophosphamide and
cancer: golden anniversary. Nat Rev Clin Oncol. (2009) 6:638–
47. doi: 10.1038/nrclinonc.2009.146
37. Brummaier T, Pohanka E, Studnicka-Benke A, Pieringer H. Using
cyclophosphamide in inflammatory rheumatic diseases. Eur J Intern Med.
(2013) 24:590–6. doi: 10.1016/j.ejim.2013.02.008
38. Riley P, Maillard SM, Wedderburn LR, Woo P, Murray KJ, Pilkington
CA. Intravenous cyclophosphamide pulse therapy in juvenile
dermatomyositis. A review of efficacy and safety. Rheumatology. (2004)
43:491–6. doi: 10.1093/rheumatology/keh082
39. Patti F, Lo Fermo S. Lights and shadows of cyclophosphamide
in the treatment of multiple sclerosis. Autoimmune Dis. (2011)
2011:961702. doi: 10.4061/2011/961702
40. Fox LP, Pandya AG. Pulse intravenous cyclophosphamide
therapy for dermatologic disorders. Dermatol Clin. (2000)
18:459–73. doi: 10.1016/S0733-8635(05)70194-7
41. Ahlmann M, Hempel G. The effect of cyclophosphamide on the immune
system: implications for clinical cancer therapy. Cancer Chemother
Pharmacol. (2016) 78:661–71. doi: 10.1007/s00280-016-3152-1
42. Blacker CV, Greenwood DT, Wesnes KA, Wilson R, Woodward
C, Howe I, et al. Effect of galantamine hydrobromide in chronic
fatigue syndrome: a randomized controlled trial. JAMA. (2004)
292:1195–204. doi: 10.1001/jama.292.10.1195
43. Loge JH, Kaasa S, Hjermstad MJ, Kvien TK. Translation and performance of
the Norwegian SF-36 Health Survey in patients with rheumatoid arthritis. I.
Data quality, scaling assumptions, reliability, and construct validity. J Clin
Epidemiol. (1998) 51:1069–76. doi: 10.1016/S0895-4356(98)00098-5
44. Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey
(SF-36). I. Conceptual framework and item selection. Med Care. (1992)
30:473–83. doi: 10.1097/00005650-199206000-00002
45. Lerdal A, Wahl A, Rustoen T, Hanestad BR, Moum T. Fatigue in the general
population: a translation and test of the psychometric properties of the
Norwegian version of the fatigue severity scale. Scand J Public Health. (2005)
33:123–30. doi: 10.1080/14034940410028406
46. Krupp LB, LaRocca NG, Muir-Nash J, Steinberg AD. The
fatigue severity scale. Application to patients with multiple
sclerosis and systemic lupus erythematosus. Arch Neurol. (1989)
46:1121–3. doi: 10.1001/archneur.1989.00520460115022
47. Almeida GJ, Wasko MC, Jeong K, Moore CG, Piva SR. Physical activity
measured by the SenseWear Armband in women with rheumatoid arthritis.
Phys Ther. (2011) 91:1367–76. doi: 10.2522/ptj.20100291
48. Scheers T, Philippaerts R, Lefevre J. Variability in physical activity patterns as
measured by the SenseWear Armband: howmany days are needed? Eur J Appl
Physiol. (2012) 112:1653–62. doi: 10.1007/s00421-011-2131-9
49. Sklar CA, Mertens AC, Mitby P, Whitton J, Stovall M, Kasper C, et al.
Premature menopause in survivors of childhood cancer: a report from
the childhood cancer survivor study. J Natl Cancer Inst. (2006) 98:890–
6. doi: 10.1093/jnci/djj243
Frontiers in Medicine | www.frontiersin.org 14 April 2020 | Volume 7 | Article 162
Rekeland et al. Intravenous Cyclophosphamide in ME/CFS
50. Cui W, Phillips KA. Conservative management of reproductive cancers.
Ovarian protection during treatment. Best Pract Res Clin Obstet Gynaecol.
(2019) 55:49–58. doi: 10.1016/j.bpobgyn.2018.06.010
51. Mok CC, Ying KY, Ng WL, Lee KW, To CH, Lau CS, et al. Long-
term outcome of diffuse proliferative lupus glomerulonephritis
treated with cyclophosphamide. Am J Med. (2006) 119:355.e325–
33. doi: 10.1016/j.amjmed.2005.08.045
52. Demeestere I, Brice P, Peccatori FA, Kentos A, Dupuis J, Zachee P, et al.
No Evidence for the benefit of gonadotropin-releasing hormone agonist in
preserving ovarian function and fertility in lymphoma survivors treated with
chemotherapy: final long-term report of a prospective randomized trial. J Clin
Oncol. (2016) 34:2568–74. doi: 10.1200/JCO.2015.65.8864
53. Ware JE Jr, Gandek B. Overview of the SF-36 health survey and the
international quality of life assessment (IQOLA) project. J Clin Epidemiol.
(1998) 51:903–12. doi: 10.1016/S0895-4356(98)00081-X
54. Iwaszko M, Swierkot J, Kolossa K, Jeka S, Wiland P, Bogunia-Kubik K.
Polymorphisms within the human leucocyte antigen-E gene and their
associations with susceptibility to rheumatoid arthritis as well as clinical
outcome of anti-tumour necrosis factor therapy. Clin Exp Immunol. (2015)
182:270–7. doi: 10.1111/cei.12696
55. Dand N, Duckworth M, Baudry D, Russell A, Curtis CJ, Lee SH,
et al. HLA-C∗06:02 genotype is a predictive biomarker of biologic
treatment response in psoriasis. J Allergy Clin Immunol. (2019) 143:2120–
30. doi: 10.1016/j.jaci.2018.11.038
56. Nakao S, Takamatsu H, Chuhjo T, Ueda M, Shiobara S, Matsuda
T, et al. Identification of a specific HLA class II haplotype strongly
associated with susceptibility to cyclosporine-dependent aplastic anemia.
Blood. (1994) 84:4257–61. doi: 10.1182/blood.V84.12.4257.bloodjournal841
24257
57. Chang DY, Luo H, Zhou XJ, Chen M, Zhao MH. Association of HLA genes
with clinical outcomes of ANCA-associated vasculitis. Clin J Am Soc Nephrol.
(2012) 7:1293–9. doi: 10.2215/CJN.13071211
58. Chu L, Valencia IJ, Garvert DW, Montoya JG. Onset patterns and course of
myalgic encephalomyelitis/chronic fatigue syndrome. Front Pediatr. (2019)
7:12. doi: 10.3389/fped.2019.00012
59. Cho HJ, Hotopf M, Wessely S. The placebo response in the treatment of
chronic fatigue syndrome: a systematic review and meta-analysis. Psychosom
Med. (2005) 67:301–13. doi: 10.1097/01.psy.0000156969.76986.e0
60. Frisaldi E, Shaibani A, Benedetti F. Why we should assess
patients’ expectations in clinical trials. Pain Ther. (2017)
6:107–10. doi: 10.1007/s40122-017-0071-8
61. Hall AG, Tilby MJ. Mechanisms of action of, and modes of resistance to,
alkylating agents used in the treatment of haematological malignancies. Blood
Rev. (1992) 6:163–73. doi: 10.1016/0268-960X(92)90028-O
62. Barnes H, Holland AE, Westall GP, Goh NS, Glaspole IN.
Cyclophosphamide for connective tissue disease-associated
interstitial lung disease. Cochrane Database Syst Rev. (2018)
1:Cd010908. doi: 10.1002/14651858.CD010908.pub2
63. Corthay A. How do regulatory T cells work? Scand J Immunol. (2009)
70:326–36. doi: 10.1111/j.1365-3083.2009.02308.x
64. Ramos S, Brenu E, Broadley S, Kwiatek R, Ng J, Nguyen T, et al. Regulatory
T, natural killer T and gammadelta T cells in multiple sclerosis and chronic
fatigue syndrome/myalgic encephalomyelitis: a comparison. Asian Pac J
Allergy Immunol. (2016) 34:300–5. doi: 10.12932/AP0733
65. Curriu M, Carrillo J, Massanella M, Rigau J, Alegre J, Puig J, et al. Screening
NK-, B- and T-cell phenotype and function in patients suffering from Chronic
Fatigue Syndrome. J Transl Med. (2013) 11:68. doi: 10.1186/1479-5876-11-68
66. Brenu EW, van Driel ML, Staines DR, Ashton KJ, Ramos SB, Keane J,
et al. Immunological abnormalities as potential biomarkers in Chronic
Fatigue Syndrome/Myalgic Encephalomyelitis. J Transl Med. (2011)
9:81. doi: 10.1186/1479-5876-9-81
67. Hoyer BF, Moser K, Hauser AE, Peddinghaus A, Voigt C, Eilat D, et al.
Short-lived plasmablasts and long-lived plasma cells contribute to chronic
humoral autoimmunity in NZB/W mice. J Exp Med. (2004) 199:1577–
84. doi: 10.1084/jem.20040168
68. van Campen C, Verheugt FWA, Rowe PC, Visser FC. Cerebral blood
flow is reduced in ME/CFS during head-up tilt testing even in the
absence of hypotension or tachycardia: A quantitative, controlled
study using Doppler echography. Clin Neurophysiol Pract. (2020)
5:50–8. doi: 10.1016/j.cnp.2020.01.003
69. Oldham WM, Lewis GD, Opotowsky AR, Waxman AB, Systrom DM.
Unexplained exertional dyspnea caused by low ventricular filling pressures:
results from clinical invasive cardiopulmonary exercise testing. Pulm Circ.
(2016) 6:55–62. doi: 10.1086/685054
70. Joseph P, Sanders J, Oaklander AL, Arevalo Rodriguez TCA, Oliveira R, Faria
Urbina M, et al. The patophysiology of Chronic Fatigue Syndrome: Results
from an invasive cardiopulmonary exercise laboratory. Am J Respir Crit Care
Med. (2019) 199:A6902.
71. Melamed KH, Santos M, Oliveira RKF, Urbina MF, Felsenstein D,
Opotowsky AR, et al. Unexplained exertional intolerance associated with
impaired systemic oxygen extraction. Eur J Appl Physiol. (2019) 119:2375–
89. doi: 10.1007/s00421-019-04222-6
72. Oaklander AL, Nolano M. Scientific advances in and clinical approaches
to small-fiber polyneuropathy: a review. JAMA Neurol. (2019) 76:1240–
51. doi: 10.1001/jamaneurol.2019.2917
73. Grayston R, Czanner G, Elhadd K, Goebel A, Frank B, Uceyler N,
et al. A systematic review and meta-analysis of the prevalence of small
fiber pathology in fibromyalgia: Implications for a new paradigm in
fibromyalgia etiopathogenesis. Semin Arthritis Rheum. (2019) 48:933–
40. doi: 10.1016/j.semarthrit.2018.08.003
74. Armstrong CW, McGregor NR, Lewis DP, Butt HL, Gooley PR. Metabolic
profiling reveals anomalous energy metabolism and oxidative stress pathways
in chronic fatigue syndrome patients. Metabolomics. (2015) 11:1626–
39. doi: 10.1007/s11306-015-0816-5
75. Lien K, Johansen B, Veierod MB, Haslestad AS, Bohn SK, Melsom MN,
et al. Abnormal blood lactate accumulation during repeated exercise testing
in myalgic encephalomyelitis/chronic fatigue syndrome. Physiol Rep. (2019)
7:e14138. doi: 10.14814/phy2.14138
76. Newton DJ, Kennedy G, Chan KK, Lang CC, Belch JJ, Khan F. Large and small
artery endothelial dysfunction in chronic fatigue syndrome. Int J Cardiol.
(2012) 154:335–6. doi: 10.1016/j.ijcard.2011.10.030
Conflict of Interest: The authors declare that the research was conducted in the
absence of any commercial or financial relationships that could be construed as a
potential conflict of interest.
Copyright © 2020 Rekeland, Fosså, Lande, Ktoridou-Valen, Sørland, Holsen,
Tronstad, Risa, Alme, Viken, Lie, Dahl, Mella and Fluge. This is an open-access
article distributed under the terms of the Creative Commons Attribution License (CC
BY). The use, distribution or reproduction in other forums is permitted, provided
the original author(s) and the copyright owner(s) are credited and that the original
publication in this journal is cited, in accordance with accepted academic practice.
No use, distribution or reproduction is permitted which does not comply with these
terms.
Frontiers in Medicine | www.frontiersin.org 15 April 2020 | Volume 7 | Article 162
Supplementary Table 1. Medical history and concomitant diseases reported at baseline, shown by System Organ Class (SOC) and CTCAE term
System Organ Class (SOC) CTCAE term n %
Endocrine disorders Hypothyroidism 4 10.0
Immune system disorder Allergy 16 40.0
Musculoskeletal and connective tissue disorder Fibromyalgia 3 7.5
Psychiatric disorder Depression 4 10.0
Psychiatric disorder Anxiety 4 10.0
Other Other 14 35.0
Supplementary Table 2. Previous treatments for ME/CFS, reported at baseline.
Type of treatment, n (%) n % Cognitive therapy (CT)
“Lightning Process” (LP) 13 32.5
Mindfulness 11 27.5
Other CT 4 10.0
Any CT (LP, Mindfulness, Other) 21 52.5
Physical therapy
Graded exercise therapy (GET) 6 15.0
Other physical therapy 15 37.5
GET or other physical therapy 18 45.0
Activity management (adaptive pacing) 15 37.5
Not received any of these treatments 8 20.0
Not answered 1 2.5
Medical treatments
Nexavir 2 5.0
Vitamin B12-injections 16 40.0
Long term antibiotics 10 25.0
Low dose naltrexone 15 37.5
Rituximab 15 37.5
Not received any of these treatments 7 17.5
Not answered 1 2.5
Supplementary Table 3. Concomitant medication during 18 months follow-up (shown by ATC-code)
ATC code Description N=40
A02 Antacids, n (%) 9 (22.5)
A03, A04 Antiemetics, n (%) 10 (25.0)
A06 Laxantia, n (%) 2 (5.0)
B01 Antithrombotic agents, n (%) 1 (2.5)
B03A, B03BB Vitamin B12 supplements, n (%) 5 (12.5)
C07 Betablockers, n (%) 3 (7.5)
C08, C09 Antihypertensive agents, n (%) 4 (10.0)
C10 Statins, n (%) 2 (5.0)
G01, J01-05 Antibiotics, n (%) 13 (32.5)
G03A Contraceptives (systemic), n (%) 4 (10.0)
H03 Thyroid hormone replacement, n (%) 4 (10.0)
M01A NSAID, n (%) 17 (42.5)
N02A Opioidsa, n (%) 11 (27.5)
N02B Paracetamol, n (%) 15 (37.5)
N02C Antimigraine agents, n (%) 5 (12.5)
N03A Antiepileptic agents, n (%) 1 (2.5)
N05B Anxiolytica, n (%) 4 (10.0)
N05C Hypnotics and sedativesb, n (%) 20 (50.0)
N06 Antidepressants, n (%) 6 (15.0)
R01, R03,R06A, S01G Allergy and asthma medications, n (%) 20 (50.0)
Other medications, n (%) 30 (75.0)
Dietary supplements (non-ATC) , n (%) 10 (25.0)
a: 11 out of 40 patients received opioids at any time during follow-up. Among these, only 2 used tramadol daily on a regular basis, and 9 used codeine phosphate or tramadol on demand. None of the patients used any stronger opioids. b: 20 out of 40 patients had used hypnotics regularly or sporadically at any time during follow-up, 4 of whom had tried more than one type of hypnotic. Among the 20, 10 had used melatonin, 10 had used zopiclone and 4 patients had used nitrazepam or zolpidem.
IV
a) Information letter (in Norwegian) for recruitment of severely affected ME/CFS patients in Paper II
b) Questionnaire (in Norwegian) for ME/CFS patiens included in Paper II, inclusion group iii)
c) Questionnaires for symptom registry before treatment and during follow-up in Paper III
ATTACHMENTS
Vil du delta i forskning? Rekruttering av pasienter med alvorlig ME til ImmunoME-studien ved Oslo Universitetssykehus.
Hva er ImmunoME-studien? Formålet med ImmunoME er å forstå sykdommen ME bedre fra et genetisk og biologisk perspektiv. Vi ønsker særlig å undersøke om immunsystemet har en rolle i utviklingen av ME. Vi vil gjøre en grundig kartlegging av gener som er sentrale for immunsystemet for å se etter forskjeller mellom ME-pasienter og kontrollpersoner. Vi vil også undersøke forekomst av antistoffer hos ME-pasientene. Vi ønsker nå å rekruttere pasienter med den mest alvorlige formen for ME. Analysene vil bli gjort på gruppenivå, og det kan derfor ikke gis individuell tilbakemelding til deltakerne.
Du kan delta hvis: • Du er over 18 år• Du har ME-diagnosen etter Canada-kriteriene• Du har eller har hatt den mest alvorlige formen for ME
Med den mest alvorlige formen for ME mener vi at du er sengeliggende stort sett hele døgnet, og er ute av stand til å arbeide eller utføre dagligdagse aktiviteter.
Du kan inkluderes om du pr. i dag har denne alvorlige formen for ME eller om du tidligere har hatt den sammenhengende i minst to år.
Deltakelse innebærer: • Å få tilsendt informasjons- og samtykkeskriv og signere dette• Å svare på et spørreskjema for kliniske opplysninger• Å avgi blodprøve
Pårørende kan hjelpe til med å fylle ut skjemaet. For svært syke pasienter vil blodprøvetaking kunne være belastende, og vi vil etterstrebe at dette gjøres så skånsomt som mulig. Hjemmebesøk vil kunne være aktuelt.
Godkjenning Studien er godkjent av Regional forskningsetisk komité (2015/1547 REK sør-øst B). I utformingen av prosjektet har vi samarbeidet med et brukerpanel bestående av pasienter og pårørende.
Kontakt • Asgeir Lande. E-post: [email protected]
Telefon: 23016672 • Benedicte A. Lie. E-post: [email protected] • Marte K. Viken. E-post: [email protected] Vennligst ta kontakt dersom du ønsker mer informasjon eller er interessert i å delta. Studiegruppen ved Oslo Universitetssykehus består av: • Prosjektleder og professor Benedicte A. Lie,
Avdeling for medisinsk genetikk • Professor og overlege Ola Didrik Saugstad,
Pediatrisk forskningsinstitutt • Professor og overlege Torstein Egeland,
Avdeling for immunologi og transfusjonsmedisin • Postdoktor Marte K. Viken,
Avdeling for immunologi og transfusjonsmedisin • PhD-stipendiat og lege Asgeir Lande,
Avdeling for medisinsk genetikk
______________________________________________________________Pasientkode/navn/født
1
SPØRSMÅL SOM SKAL BESVARES FOR INKLUDERING I ImmunoME - FORSKNING PÅ ME OG IMMUNSYSTEMET
Navn
Adresse
Telefon Mobil: Fast:
Epost
Kjønn: c K c M
Fødselsdato (dd.mm.åååå):
Hva er ditt fødeland?
Hva er dine foreldres fødeland?
Mor:
Far:
Hva er dine besteforeldres fødeland?
Mormor:
Morfar:
Farmor:
Farfar:
Hvilken etnisitet vil du si at du tilhører?
Merknad: Opplysninger om etnisitet er viktig for korrekt tolkning av de genetiske undersøkelsene, da det er kjent at genetiske forhold kan variere mellom folkeslag. Opplysningene vil ikke bli brukt i andre sammenhenger.
Hva er din nåværende sivilstatus?
c gift/partnerskap/samboer c separert/skilt
c enke/enkemann c enslig
Hvor mange barn har du?
Hva er dine barns fødselsår?
______________________________________________________________Pasientkode/navn/født
2
Har du fått ME-diagnosen stilt av lege?
c Ja,
årstall:
c Nei
c Vet ikke
Hvis ja, hvilket kriteriesett er brukt?
c Canada (Carruthers et al., 2003)
c ICC (Internasjonale Konsensuskriterier, 2011)
c Fukuda
c Jason (for barn/ungdommer)
c Vet ikke
c Andre kriterier, hvilke:
Når begynte dine symptomer på ME?
Årstall: Kommentar:
Startet din utmattelse etter at du hadde opplevd noe av det følgende?
(Vennligst sett kryss for ett eller flere alternativer.
Angi også tiden, i antall måneder, fra hendelsens start til symptomstart.)
c Infeksjon:
c Mononukleose (kyssesyke)
c Borreliose
c Giardia
c Mage/tarm:
c Annet:
Angi tiden (i mnd):
-
-
-
-
-
c Vaksinering:
c Meningokokk-B
c Svineinfluensa (Pandemrix)
c Annet:
-
-
-
c Ulykke: -
c Operasjon: -
c Kraftig psykisk påkjenning: -
c Annet: -
c Vet ikke
Kommentar:
______________________________________________________________Pasientkode/navn/født
3
Hvilket utsagn beskriver best ditt aktivitetsnivå i løpet av de siste 6 månedene?
(Kryss av for ett alternativ.)
c Jeg kan gjøre alt arbeid og familiære forpliktelser uten noen problemer med min energi. c Jeg kan arbeide fulltid og fullføre noen familiære forpliktelser, men har ikke energi til noe annet. c Jeg kan arbeide fulltid, men har ingen energi til overs til noe annet. c Jeg kan bare arbeide deltid på jobben eller med familiære forpliktelser. c Jeg kan gjøre lett husarbeid, men jeg kan ikke arbeide deltid. c Jeg kan gå rundt i huset, men jeg kan ikke gjøre lett husarbeid. c Jeg er ikke i stand til å arbeide eller å gjøre noen ting, og jeg er sengeliggende. c Jeg er sengebundet og må ha hjelp til basale funksjoner, som for eksempel matinntak. Eventuell kommentar:
Dersom du tidligere har hatt mer alvorlige ME- symptomer enn du har nå, hvilket utsagn beskriver best ditt aktivitetsnivå i den mest alvorlige 6-måneders-fasen?
(Kryss av for ett alternativ og angi når du opplevde bedring fra den mest alvorlige fasen.)
c Jeg kunne gjøre alt arbeid og familiære forpliktelser uten noen problemer med min energi. c Jeg kunne arbeide fulltid og fullføre noen familiære forpliktelser, men hadde ikke energi til noe annet. c Jeg kunne arbeide fulltid, men hadde ingen energi til overs til noe annet. c Jeg kunne bare arbeide deltid på jobben eller med familiære forpliktelser. c Jeg kunne gjøre lett husarbeid, men jeg kunne ikke arbeide deltid. c Jeg kunne gå rundt i huset, men jeg kunne ikke gjøre lett husarbeid. c Jeg var ikke i stand til å arbeide eller å gjøre noen ting, og jeg var sengeliggende. c Jeg var sengebundet og måtte ha hjelp til basale funksjoner, som for eksempel matinntak. Bedring fra den mest alvorlige fasen inntraff i ............................ (årstall)
Eventuell kommentar:
Har du vært sengebundet som følge av ME-sykdommen?
c Ja, antall år (feks 1,5):
c Nei
Kommentar:
______________________________________________________________Pasientkode/navn/født
4
Er du inkludert i ME-biobanken på OUS?
(ME/CFS - Tematisk Biobank tilknyttet CFS/ME- senteret på Aker Sykehus)
c Ja
c Nei
c Vet ikke
Kommentar:
Har du deltatt i Rituximab-studiene for ME-pasienter, administrert ved Haukeland Universitets-sykehus?
c Ja,
årstall:
c Nei
c Vet ikke
Kommentar:
Har du deltatt i CycloME- studien for ME-pasienter, administrert ved Haukeland Universitets-sykehus?
c Ja
c Nei
c Vet ikke
Kommentar:
Har du deltatt i andre studier/biobanker?
(Hvis ja - vennligst spesifisér)
c Ja
c Nei
c Vet ikke
Spesifisér:
Har noen av dine slektninger med ME meldt seg for deltagelse i ImmunoME?
(Hvis ja – vennligst angi hvilke(n) slektning(er))
c Ja
c Nei
c Vet ikke
Slektning:
-
-
Har noen av dine slektninger med ME deltatt i noen av de nevnte eller i andre studier/biobanker?
(Hvis ja - vennligst spesifisér slektning og studie/biobank)
c Ja
c Nei
c Vet ikke
Slektning: Studie/biobank:
- -
- -
- -
- -
______________________________________________________________Pasientkode/navn/født
5
Har du fått følgende medisinsk behandling for din ME-sykdom?
Rituximab-behandling
c Ja c Nei
c Vet ikke
c Deltager i ovennevnte studie
Start/stans, dose (hvis kjent):
Immunglobulin-behandling
(f.eks Octagam, Gammanorm)
c Ja c Nei
c Vet ikke
Start/stans, type, dose (hvis kjent):
Langvarig antibiotika-behandling
c Ja c Nei
c Vet ikke
Start/stans, type, dose (hvis kjent):
Annen behandling for ME
(Vennligst spesifiśer)
Hvilke andre medikamenter bruker du eller har du brukt?
-
-
-
-
-
Startet (år):
Sluttet (år):
Indikasjon (grunnen til medikament-bruken):
Dose (hvis kjent):
______________________________________________________________Pasientkode/navn/født
6
Har du eller noen av dine anførte slektninger hatt noen av de følgende sykdommene?
(Sett kryss hvis du eller en slektning har eller har hatt sykdommen. Kommentér eventuelt nederst.)
Du selv Mor Far Dine barn Dine søsken
ME/CFS
Fibromyalgi
MS (multippel sklerose)
Myasthenia Gravis
Diabetes type I
Lavt stoffskifte (hypotyreose)
Høyt stoffskifte (hypertyreose)
Psoriasis
Sarkoidose
Cøliaki
PBC (primær biliær cirrhose)
PSC (primær skleroserende kolangitt)
Inflammatorisk tarmsykdom (ulcerøs kolitt eller Crohns sykdom)
Sjøgrens sykdom
SLE (systemisk lupus erytematosus)
Leddgikt (reumatoid artritt)
Bekhterevs sykdom
Annen autoimmun sykdom
(vennligst spesifisér)
______________________________________________________________Pasientkode/navn/født
7
Du selv Mor Far Dine barn Dine søsken
Benskjørhet (osteoporose)
Slitasjegikt (artrose)
Kreft
(vennligst angi type)
Hjertesykdom
(vennligst angi type)
Nyresykdom
(vennligst angi type)
Behandlingstrengende psykisk lidelse
(vennligst angi type)
Autisme/Asperger syndrom
Astma
Kronisk bronkitt
KOLS / Emfysem
Fjernet mandler (tonsillektomi)
Fjernet blindtarmen (appendektomi)
Migrene
Annet
(vennligst spesifisér)
______________________________________________________________Pasientkode/navn/født
8
Eventuelle kommentarer:
Tusen takk for innsatsen.
Symptom self-report form before treatment.
CycloME part A/KTS-7-2015.
Appendix D1
Version: 1.0 Document date: 20.09.2014. Page 1 of 4
1
SYMPTOM SELF-REPORT FORM BEFORE TREATMENT
Study ID no:
Date:
This form is to be completed prior to first treatment only.
On the next page, please record your individual symptoms and to which
degree these affect you prior to treatment.
We would like to record your baseline assessment to use as a basis for
comparison for any changes in your symptoms.
Please grade each symptom on a scale from 1 to 10:
1: not affected, as before you were taken ill
2: barely affected
3: slightly affected
4: sligthly to moderately affected
5: moderately affected
6: moderately to considerably affected
7: considerably affected
8: considerably affected
9: intensely affected
10: severely affected, severely ill
In the bottom box, please state your “total function level” expressed as a
percentage of a healthy state, which equals 100% and represents your
function level before you became ill with ME/CFS (please refer to the
separate list of examples).
Symptom self-report form before treatment.
CycloME part A/KTS-7-2015.
Appendix D1
Version: 1.0 Document date: 20.09.2014. Page 2 of 4
2
Please state the number (1 to 10) which best describes your condition
during the last 3 months. 1: not affected,--------- 5: moderately affected,----------- 10: severely affected
(In the bottom box please state a percentage from 1 to 100 %)
Before treatment
Completion date
FATIGUE
Fatigue
Post-exertional malaise
Need for rest
Daily functioning
PAIN
Muscle pain
Headache
Joint pain
Skin pain
COGNITIVE
Memory problems
Concentration problems
Reduced ability to think clearly
Mood swings, melancholy
OTHER SYMPTOMS
Sleep disturbances
Nausea
Diarrhoea
Constipation
Dizziness
Sensitivity to light
Sensitivity to noise
Visual disturbances
Perspiration
Palpitations
Dry mouth
Rash
Tender lymph nodes
Sore throat
Disturbed body temperature
Cold hands and feet
HOW DO YOU RATE YOUR
OVERALL CONDITION THE LAST
2 WEEKS (1 to 10)?
HOW DO YOU RATE YOU
“TOTAL FUNCTION LEVEL”
(PERCENTAGE OF HEALTHY
STATE, 0 to 100 %)?
Symptom self-report form before treatment.
CycloME part A/KTS-7-2015.
Appendix D1
Version: 1.0 Document date: 20.09.2014. Page 3 of 4
3
”Total function level” (0-100%); list of examples. In your self-report folder you must document any changes to your symptoms once every two
weeks throughout follow-up. You must compare the severity of your symptoms to your
symptoms before you started treatment, on a scale from 0 to 6 (where 3= no change from
baseline). On this scale you register subjectively how you perceive any change in each
symptom compared to baseline (no change, slight, moderate or major change).
We also wish to record your assessment of your condition and function level compared to an
entirely healthy state, i.e. your condition before you were taken ill with ME/CFS.
The scale for this assessment is 0 to 100 %, where 100 % = healthy, with no ME/CFS
symptoms. You are required to register your total function level (in %) once every two weeks
on the self-report form.
-A patient who is confined to bed or sofa all day, and who needs assistance for basic tasks,
may have a function level of less than 5% of a healthy state.
-A patient who is mainly bedbound, perhaps capable of pottering about, and barely outside
their home, may have a function level of between 5 and 10% of a healthy state.
-A patient who is mainly inactive, but can engage in light indoor activities for shorter periods,
and may be capable of doing a round of shopping or short errands a couple of times a week,
may have a function level of between 10 and 15% of a healthy state.
-A patient who engages in some activity, or can take short walks and participate in limited
socal activity, may have a function level of between 10 and 15% of a healthy state.
-A patient who can engage in e.g. study, leisure or work-related activities for one or two days
per week, and can take light to moderate walks with no major restrictions, may have a
function level of approx. 40% of a healthy state.
-A patient who engages socially with family most days, who can go on holidays, take walks,
perhaps spend parts of the day studying, reading or working with a computer, may have a
function level of 60 to 70% of a healthy state.
-A patient who experiences slight restrictions in their physical and social function level, but
can otherwise in principle perform activities almost as in healthy state (but at a reduced
rate/duration/pace), may have a function level of 80 to 90% of a healthy state.
-A person who is not affected by ME symptoms and feels entirely healthy will register 90 to
100%.
-A patient who has experienced no improvement or worsening, will register an unchanged
function level throughout follow-up compared to baseline.
-A patient with a reduced function level will register lower percentages during follow-up than
at baseline.
-A patient experiencing improvements will register higher percentages during follow-up than
at baseline.
It is up to you to decide on the approximately right percentage for your total function level,
compared to a healthy state (100%), both at baseline and once every two weeks, and
record this in the bottom box of the self-report forms in your folder.
Symptom self-report form before treatment.
CycloME part A/KTS-7-2015.
Appendix D1
Version: 1.0 Document date: 20.09.2014. Page 4 of 4
4
IF YOU WISH, YOU MAY ALSO DESCRIBE IN MORE DETAIL
HOW YOU PERCEIVE YOUR SYMPTOMS:
Please also make notes of any medications you use
(including herbal medicines and nutritional supplements):
Self-report form every two weeks
CycloME part A/KTS-7-2015.
Appendix E1
Version: 1.0 Document date: 20.09.2014. Page 1 of 9
Remember that every two weeks you must compare your symptoms (on a
scale of 0 to 6) to your symptoms BEFORE the start of trial
1
SELF REPORT FORM FOR SYMPTOM CHANGE
EVERY TWO WEEKS THROUGHOUT TRIAL FOLLOW-UP
Study ID no:……..
This form must be filled in every two weeks from start of treatment throughout
the trial follow-up period (at least 18 months)
On this form, we ask you to register any changes in your symptoms of Chronic
Fatigue Syndrome (ME/CFS), regularly every second week.
Please consider how you have experienced your condition over the last two
weeks, always compared to before you started treatment (which equals the
number 3).
SCALE
0 1 2 3 4 5 6
Major
worsening
Moderate
worsening
Slight
worsening
No change Slight
improvement
Moderate
improvement
Major
improvement
In the bottom box you must also record your experience of your «total function
level», expressed as a percentage of an entirely healthy state, i.e. your condition
before you became ill with ME/CFS (please refer to separate sheet of examples in
your folder).
Self-report form every two weeks
CycloME part A/KTS-7-2015.
Appendix E1
Version: 1.0 Document date: 20.09.2014. Page 2 of 9
Remember that every two weeks you must compare your symptoms (on a
scale of 0 to 6) to your symptoms BEFORE the start of trial
2
0: major worsening, 1: moderate worsening, 2: slight worsening, 3: no change,
4: slight improvement, 5: moderate improvement, 6: major improvement
Time from treatment (WEEKS) 0 2 4 6 8 10 12 14 16 Date
FATIGUE
Fatigue 3
Post-exertional malaise 3
Need for rest 3
Daily functioning 3
PAIN
Muscle pain 3
Headache 3
Joint pain 3
Skin pain 3
COGNITIVE
Memory problems 3
Concentration problems 3
Reduced ability to think clearly 3
Unstable moods, melancholy 3
OTHER SYMPTOMS
Sleep disturbances 3
Nausea 3
Diarrhoea 3
Constipation 3
Dizziness 3
Sensitivity to light 3
Sensitivity to noise 3
Visual disturbances 3
Perspiration 3
Palpitations 3
Dry mouth 3
Rash 3
Tender lymph nodes 3
Sore throat 3
Urinary bladder dysfunction 3
Disturbed body temperature 3
Cold hands and feet 3
HOW DO YOU RATE YOUR
OVERALL CONDITION THE
LAST 2 WEEKS (0 to 6)? 3
HOW DO YOU RATE YOU
“TOTAL FUNCTION LEVEL”
(PERCENTAGE OF HEALTHY
STATE, 0 to 100 %)?