Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF...
Transcript of Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF...
![Page 1: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/1.jpg)
Leeka KheifetsProfessor
Workshop on Sensitivity of Children to EMFIstanbul, TurkeyJune 2004
Childhood leukemia and EMFChildhood leukemia and EMF
![Page 2: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/2.jpg)
Age-specific childhood leukemia incidence rates in the U.S.
0
1
2
3
4
5
6
7
8
9
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Age (years)
Inci
denc
e ra
te p
er 1
00,0
00 p
er y
ear
ALL MaleALL FemaleALL WhiteALL BlackAML MaleAML FemaleAML WhiteAML Black
![Page 3: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/3.jpg)
Putative Causes of Childhood Leukemia
• Underlying triggers for molecular damage may be inherited at conception, may occur during fetal development or during infancy
• Most likely, a series of detrimental genetic changes accumulate over time
• Confirmed associations explain only 10% of childhood leukemia incidence
• Leukemia results from chromosomal alterations and mutations that disrupt the normal process by which lymphoid or myeloid progenitor cells differentiate
![Page 4: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/4.jpg)
EMF and childhood leukemiaStudyStudy >.3>.3µµTT >.4>.4µµTT
Wertheimer & Leeper* 3.1 1.1-8.5
Fulton 0.5 0.2-1.4
Tomenius 1.5 0.4-5.7
RRRR 95% CI95% CI RRRR 95% CI95% CI
Coghill 1/0
Savitz 3.5 0.8-15.4
Coleman 1.5 0.7-3.5
Myers* 0.8 0.1-9.6
London 1.6 0.7-3.5
Feychting 4.5 1.7-12.0 3.7 1.2-11.4
Olsen 2.0 0.4-10.0 2/0
Verkasalo 2.0 0.2-18.0 6.2 0.7-56.9
Tynes 0/0 0/0
Michaelis 2.4 0.8-7.6 2.0 0.3-15.2
Linet 1.5 0.9-2.4 3.4 1.2-9.6
Dockerty 3/0 0/0
McBride 1.4 0.6-3.2 1.6 0.7-3.7
Green* 4.5 1.3-15.9
UKCCS 1.7 0.4-7.0 1.0 0.3-3.4
Kabuto* 1.7 0.7-3.8
![Page 5: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/5.jpg)
Pooled Analyses of Childhood Leukemia
• Greenland et al., Epidemiology, 2000
• 12 studies with fields; 8 with wire codes
• 2,656 cases; 7,084 controls
• Metric of choice: time-weighted average
• Ahlbom et al., British J. Cancer, 2000
• 9 studies with fields; 2 with wire codes
• 3,247 cases; 10,400 controls
• Metric of choice: geometric mean
ELF
![Page 6: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/6.jpg)
Pooled analysis. Greenland, 2000
![Page 7: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/7.jpg)
Greenland et al., 2000
![Page 8: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/8.jpg)
1
1.5
2.0
< 0.1 0.1 - < 0.2 0.2 - < 0.4 ≤ 0.4 µT
RR
95% C.I..
Ahlbom et al., 2000
![Page 9: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/9.jpg)
Consistent association between childhood
leukemia and exposure > 0.3-0.4 µT
Possible explanations:
•Causal relationship?
•Chance????•Misclassification???•Confounding??•Selection bias?•Other?
Confounding
Leukemia
Selection
Biologically Relevant Exposure
Measured EMF
?
![Page 10: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/10.jpg)
Chance????
Not a chance…
p-value of 0.0001
![Page 11: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/11.jpg)
Misclassification???
• Probably non-differential – Usually dilutes the effect– Misclassification in mid categories can lead to the distortion
of dose-response
• Small reduction in specificity dilutes the effect– Classification of unexposed as unexposed– Big problem, for rare exposure
1,00,99
,98
,97
,96
,95
,94
,93
,92
,91
,90
Odd
s R
atio
Est
imat
e of
Rel
ativ
e R
isk 5,0
4,5
4,0
3,5
3,0
2,5
2,0
1,51,0
Specificity
![Page 12: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/12.jpg)
Confounding??• Numerous factors examined:
–SES–Traffic Density–Chemical exposures–Environmental tobacco smoke–Dietary agents –Viral or other infectious agents
•Not one confounds the association
Simultaneous effect?
![Page 13: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/13.jpg)
Prevalence of the High Risk Level of
the Confounder
Combination of Selected Values Sufficient for a Confounder to Account for an Observed Relative Risk of 2
Ass
ocia
tion
betw
een
Exp
osur
e an
d C
onfo
unde
r (R
elat
ive
Pre
vale
nce)
Association between Confounder and Disease(Relative Risk)
![Page 14: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/14.jpg)
Selection bias?
• Participation:– 94-100% in registry based studies– 37-68% among eligible participants interviewed– 9-31% with measurements in matched analysis
• Much worse for controls in most studies
• Rarely the complete selection/participation picture is presented
• Reporting minimizes the problem
![Page 15: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/15.jpg)
Selection bias?
• Evidence for :– Inclusion of partial participants lowered risk in one study
• Previous efforts to quantify the magnitude of selection bias mostly ecological or based on wire codes
• Selection might occur through SES or mobility
Evidence against :– Similar risk in studies with and w/out the high potential
for selection bias– Lack of association w/childhood brain tumors?
![Page 16: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/16.jpg)
Selection bias?
≥0.4 µTAll studies 2.0 (1.3-3.1)Nordic countries 2.1 (0.9-4.9)Rest of the world 1.9 (1.1-3.2)
Ahlbom et al., 2000
![Page 17: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/17.jpg)
Other?
• Contact Current?• Melatonin???
• Susceptible subpopulation (e.g. children with fusion genes generated by
chromosome translocation TEL-AML1)
![Page 18: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/18.jpg)
Causal relationship?
• Pro– Consistent Epidemiology– Specificity
• Con– Lack of supportive laboratory evidence– No known biophysical mechanism for
carcinogenesis
![Page 19: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/19.jpg)
Multiple-Bias Modeling
• Selection bias likely present, but unlikely to fully explain the association
• Confounding is probably less important• Misclassification greatly increases uncertainty,
making both no association and a strong association more plausible
• Probability that the combination of misclassification, selection bias, confounding and random error explain the association 2-4%
Greenland, 2005
![Page 20: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/20.jpg)
Some Observations….
• Role of pooled and meta-analysis
• Study is only as large as its smallest cell
• Childhood Leukemia and SES
• Selection bias – implications for case-control studies
![Page 21: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/21.jpg)
IARC Criteria
• For each type of cancer, classify human and animal data separately as:
– Sufficient– Limited– Inadequate– Lack of effect
Group 1: Is carcinogenic to humansGroup 2A: Probably is carcinogenicGroup 2B: Possibly is carcinogenicGroup 3: Not classifiableGroup 4: Is probably not carcinogenic
N.B. Greatest weight given to epidemiology
![Page 22: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/22.jpg)
Limited evidence of carcinogenicity
….usually based on evidence in humans which is considered credible, but chance, bias or confounding could not be ruled out with reasonable confidence
![Page 23: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/23.jpg)
• ELF magnetic fields classified as Group 2B “Possibly Carcinogenic” based on – limited human data based on
childhood leukemia studies– inadequate animal data
• Other exposures and outcomes considered “inadequate to classify”
IARCIARC EvaluationExtremely Low Frequencies (ELF) 2002
![Page 24: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/24.jpg)
![Page 25: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/25.jpg)
Animal studies
• Good animal model for childhood leukemia? • Relevant exposure?• Intensity within 1 order of magnitude of human
exposure? • Power to detect small risk?
![Page 26: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/26.jpg)
Scientific Issues in ELF Risk Assessment
• Refutation of alternative hypotheses
Human
Animal
+
+
??
?
• Uncertainty and misclassification in epidemiologic studies
• Exposure distribution, risk function and how to combine them
• Reconciling epidemiologic and toxicological data
• Relevance and weight of biophysical arguments
![Page 27: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/27.jpg)
RF Towers and Childhood Leukemia
Study N RR 95%CI
Selvin 52 0.7 NSMaskarinec 12 2.0 0.1-8.3Hocking 1.6 1.1-2.3Dolk 79 1.1 0.6-2.0McKenzie 0.9 0.6-1.4Cooper 1 1.1 0.0-6.3Michelozzi 8 2.2 1.0-41
![Page 28: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/28.jpg)
RF and Childhood Leukemia
• Small numbers
• Previously identified clusters
• Exposure assessment based on distance
N.B. Studies completely uninformative
![Page 29: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/29.jpg)
What can new epidemiologic studies of ELF childhood
leukemia contribute?
• Unlikely to substantially change the observed association
Need to be designed to test specific hypotheses:– other aspects of exposure
– selection bias
– susceptible subgroups
![Page 30: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/30.jpg)
Childhood Leukemia and RF
• No evidence to date- numerous case reports
• ELF exposure from mobile phones- need data
• Low environmental exposures- virtually no relevant information
• Feasibility of well designed studies unknown- can do a lot to improve existing attempts
![Page 31: Leeka Kheifets - WHO · Leeka Kheifets Professor Workshop on Sensitivity of Children to EMF Istanbul, Turkey June 2004 Childhood leukemia and EMF](https://reader036.fdocuments.net/reader036/viewer/2022070111/605001d2a493fe6d743b5c8c/html5/thumbnails/31.jpg)