Gut 1996; 38: 623-628

A4-3-Oxosteroid 51-reductase deficiency: failureof ursodeoxycholic acid treatment and response tochenodeoxycholic acid plus cholic acid

P T Clayton, K A Mills, AW Johnson, A Barabino, M G Marazzi

AbstractBackground-In some infants with liverdisease, 3-oxo-A4 bile acids are the majorbile acids in urine, a phenomenonattributed to reduced activity of the A4-3-oxosteroid 5p-reductase required for syn-thesis ofchenodeoxycholic acid and cholicacid. These patients form a heterogeneousgroup. Many have a known cause ofhepatic dysfunction and plasma concen-trations of chenodeoxycholic acid andcholic acid that are actually greater thanthose of the 3-oxo-A4 bile acids. It isunlikely that these patients have aprimary genetic deficiency of the 51-reductase enzyme.Aims-To document the bile acid profile,clinical phenotype, and response to treat-ment of an infant with cholestasis,increased plasma concentrations of 3-oxo-A4 bile acids, low plasma concentra-tions of chenodeoxycholic acid and cholicacid, and no other identifiable cause ofliver disease.Patients-This infant was compared withnormal infants and infants with cholesta-sis ofknown cause.Methods-Analysis of bile acids by liquidsecondary ionisation mass spectrometryand gas chromatography - mass spec-trometry,Results-The plasma bile acid profile ofthe patient was unique. She had chroniccholestatic liver disease associated withmalabsorption of vitamins D and E and anormal y-glutamyltranspeptidase whenthe transaminases were increased. Theliver disease failed to improve withursodeoxycholic acid but responded to acombination of chenodeoxycholic acidand cholic acid.Conclusion-Treatment of primary 513-reductase deficiency requires the use ofbile acids that inhibit cholesterol 7a-hydroxylase.(Gut 1996; 38: 623-628)

Keywords: inborn error, bile acid synthesis, giant cellhepatitis, cholestasis.

Diagnosis of inborn errors of bile acid synthe-sis is important because specific treatment isavailable in the form of oral supplements ofbile acids and such treatment is often dra-matically effective.1 2 Rapid diagnosis is madepossible by the use of fast atom bombardment

mass spectrometry (FAB-MS) or liquid sec-ondary ion mass spectrometry (LSI-MS).These techniques can be used to analyse thecholanoids (bile acids and bile alcohols) in aurine sample.2-6 Most children with cholesta-tic liver disease excrete increased amounts ofnormal bile acids (principally the taurine andglycine conjugates of cholic acid and chen-odeoxycholic acid) but in a few, one of fourunusual patterns of cholanoid excretion canbe recognised.6 One of these is characterisedby the presence of ions attributable to theglycine and taurine conjugates of 7ot-hydroxy-3-oxo-4-cholenoic acid and 7oa,122o-dihydroxy-3-oxo-4-cholenoic acid.5-10 It isgenerally accepted that this pattern arises as aresult of reduced activity of the enzymethat converts 7ot-hydroxy-cholest-4-en-3-oneto 7ot-hydroxy-5,B-cholestan-3-one and 7a.,1 2a-dihydroxy-cholest-4-en-3-one to 7, 1 2aL-hydroxy-5,B-cholestan-3-one (Fig 1). As aresult of the &4-3-oxosteroid 5 -reductasedeficiency, 3-oxo-A4 intermediates undergoside chain oxidation to produce the corres-ponding 3-oxo-A4 bile acids, which are con-jugated with glycine and taurine but cannotbe secreted into the bile and appear in theurine.6 The activity of A4-3-oxosteroid 5ot-reductase is preserved, leading to the appear-ance of increased amounts of 5ot [H] -orallo-bile acids in plasma. The difficultyin an individual patient is deciding whetherthe 51-reductase deficiency is the primarycause of the liver disease or whetherreduced activity of the enzyme is a conse-quence of hepatocyte damage. This has alsomade it difficult to define the phenotype,natural history, and response to treatmentof primary A4-3-oxosteroid 5,B-reductasedeficiency.

Patients with 5p-reductase deficiency havebeen treated with various combinations ofbile acids."1 There is little information on theuse of ursodeoxycholic acid alone. This isimportant because this is the bile acid thatis most widely used in the treatment ofcholestatic disorders of childhood. In thispaper we present a child with chroniccholestatic liver disease who failed to respondto ursodeoxycholic acid treatment butimproved rapidly with a combination ofchenodeoxycholic acid and cholic acid. Thelow concentrations of primary bile acids inthe patient's plasma, urine, and duodenaljuice before treatment suggested that thispatient probably had primary A4-3-oxosteroid5,B-reductase deficiency.

Institute of ChildHealth and GreatOrmond StreetHospital for Children,LondonP T ClaytonK A MillsAW Johnson

Istituto G Gaslini,Universita di Genova,ItalyA BarabinoM G Marazzi

Correspondence to:Dr P T Clayton, Division ofBiochemistry and Genetics,Institute of Child Health, 30Guilford Street, LondonWC1N 1EH.Accepted for publication31 October 1995

623

on 5 June 2018 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.38.4.623 on 1 April 1996. D

ownloaded from

Clayton, Mills, J7ohnson, Barabino, Marazzi

Figure 1:

of conjugcconjugate.(b) allo-c

Ursodec

ii

m

100- 8060

1- 40n 20

Figure 2:first with 2chenodeox

(OH) (OH) cholestasis with steatorrhoea, failure to thrive,Sidechain oxidation - COOH and clinical rickets. At the age of 3 months,Sidechain oxidation COOH investigations in the Gaslini Institute of Genoa

showed a normal serumalantitrypsin concen-I J,,J, "OH tration, red cell galactose-l-phosphate uridyl

"'OH transferase, and UDP-galactose-4-epimerase..ductase 5a-Reductase 3-oxo-A4bile acids Plasma tyrosine was pu (normal

Sidechain oxidation 40-130), plasma threonine 450F M (normal

(OH) (OH) 65-185), and plasma methionine 10puMCOOH (normal 15-31). The urine organic acid analy-

sis showed increased excretion of 4-hydrox-yphenyl-pyruvic and -lactic acids but no

"'OH H ,H succinyl acetone. Plasma analysis showed a

HO H H low vitamin E concentration (1.4,u M; normalAllo-chenodeoxycholi acid 11.5-35) but a normal vitamin A concentra-

J(OH) A11°acand allo-cholic acid tion. Serum ferritin was increased. Ultrasound(OH)- showed a normal gall bladder and no dilatation

COOH of extrahepatic bile ducts. A liver biopsy at 41 1 months showed lobular disarray resulting from

Chenodeoxycholic acid extensive giant cell transformation and"OH'CheOHand cholic acid necrotic foci with granulocyte accumulation.

H The portal spaces were of normal size andThe effect of reduced activityof A4-3-oxosteroid5,8-reductase: reduced synthesis shape and the interlobular bile duct could

ates of chenodeoxycholic acid and cholic acid and increased synthesis of (a)of 7a-hydroxy-3-oxo-4-cholenoic acid and 7a,12a-dihydroxy-4-cholenoic acid, always be visualised. Some portal tracts were

:henodeoxycholic acid and allo-cholic acid. infiltrated by lymphocytes and some bile ductepithelial cells were vacuolated. The paren-

Case report chyma showed considerable macrovesicularOur patient was the second child of unrelated steatosis and bile pigment accumulation.Sardinian parents. Her mother took carba- Granules of haemosiderin were localisedmazepine throughout the pregnancy. The almost exclusively to Kupffer cells.infant was born at 37 weeks gestation and was Between the ages of 3 and 8 months ourbreast fed. She developed mild jaundice but patient had persistent jaundice with failure toremained well until the third week of life when thrive and steatorrhoea when fed a normalshe developed fever, pronounced jaundice, and formula. She continued to fail to thrive whendrowsiness. Investigations showed: bilirubin fed with Pregestemil, partly because of poor316,uM (conjugated 145,uM), aspartate intake. Treatment with ursodeoxycholic acidaminotransferase (AST) 2279 IU/1 (normal at a dose of 12 mgfkg/d and then at a dose of20-60), alanine aminotransferase (ALT) 1123 20 mgfkg/d did not lead to any clinicalIU/1 (normal 5-45), y-glutamyltranspeptidase improvement or to significant improvement in(y-GT) 102IU/1 (normal 5-51), prothrombin liver function tests (Fig 2). The alkaline phos-time 15.4 seconds (control 12 seconds), ox feto- phatase fell in response to parenteral vitaminprotein 680 000 ,ug/l. Screening tests for D. Vitamins A, E, and K were also givenhepatitis viruses were negative. The plasma parenterally. The urine bile acids were firstcarbamazepine concentration was 1.4,ug/l and analysed when the patient was 0.3 years, theconsequently breast feeding was stopped. The ursodeoxycholic acid treatment having beenpatient's liver function tests showed some stopped for two weeks.improvement but there was persistent At the age of 0.69 years the patient was seen

at Great Ormond Street. She had been receiv-Chenodeoxycholic ing ursodeoxycholic acid treatment at a dose of

Cholic 20 mglkg/d for 0.4 years. Examination showeda weight well below the third percentile, length

)xycholic just below the third percentile, and head cir-cumference just below the second percentile.

s0- She was wasted and jaundiced but there wereno stigmata of chronic liver disease. The liver

)o _ \edge was palpable 5-5 cm from the costalmargin in the mid-clavicular line and 5 cm

so- \ from the xiphisternum in the midline.Investigations showed: haemoglobin 96 g/l

0 with normallindices, ferritin 245 gg/l (normal

0o 7-150), serum iron 13.7 jiM (normal 14-22),)o _ A total iron binding capacity 54-9 ,uM (normal)o _ \42-66), copper 17.3 ,uM (normal 12.6-26.8),0o - manganese 254 nM (normal 73-210). Clotting)o l studies normal plasma calcium,

0.1 0.3 0.5 0.7 0.9 1.1 phosphate, alkaline phosphatase, and albumin.Age (y) Plasma amino acids were normal with the

The response of the total bilirubin and aspartate aminotransferase to treatment exception of threonine (284 ,uM; normalursodeoxycholic acid (12 mg/kg/dfollowed by 20 mg/kg/d) and then with 70-220). The total bilirubin was 88 ,uMycholic acid plus cholic acid (8 mg/kg/d of each). (conjugated 35 jiM); the AST 511 IU/1

624

on 5 June 2018 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.38.4.623 on 1 April 1996. D

ownloaded from

A4-3-Oxosteroid 5,8-reductase deficiency

U)

a)c

a)

100 510

460 511406 49444 479

480 497O E ;,463rrrrrrrrrrrrrrrrrrrrrrrrr4548

425 528

566

420 440 460 480 50-0 520 540 560

1 1 years the patient was asymptomatic withnormal liver function tests (apart from aminimally increased alkaline phosphatase). Bythe age of 1.4 years she was above the thirdcentile for height and weight.

599 615

587 607t

580., . 6 3(20 631580 600 620 640

Mass/charge ratio (m/z)Figure 3: Negative ion liquid secondary ionisation mass spectrum obtainedfrom the urineofour patient before commencement of bile acid therapy. The identities of the major ionswere shown to be: m/z 444, glycine-conjugated 7a-hydroxy-3-oxo-4-cholenoic acid;m/z 460 glycine-conjugated 7a12a-dihydroxy-3-oxo-4-cholenoic acid; m/z 494,taurine-conjugated 7a-hydroxy-3-oxo-4-cholenoic acid; m/z 510 glycine-conjugated7a12a-dihydroxy-3-oxo-4-cholenoic acid.

(normal 20-60); the ALT 252 IU/l (normal5-45); -y-GT 36 IU/1 (normal 5-51); choles-terol 4-6 mM (normal 2-5-4 9). Because oftheparenteral therapy the plasma vitamin D was

slightly increased but the vitamin E was stilllow at 4-8 ,uM (normal 1 1-5-35). An abdomi-nal ultrasound showed hepatomegaly only.The liver biopsy showed preservation of thearchitecture but lobular disarray with numer-ous giant hepatocytes and infiltration bypolymorphs, particularly around areas ofcholestasis. The portal tracts were infiltratedby inflammatory cells including numerousneutrophils and showed periportal fibrosis.There was extensive fatty change but no ironaccumulation. The appearances were those ofa giant cell hepatitis. Electron microscopy(Professor B Lake) showed that the hepato-cytes contained abundant bile pigment con-

taining some lamellar bodies. Mitochondriawere generally ofnormal structure with a densematrix. Peroxisomes were plentiful and slightlylarger than normal (diameter >1 ,uM). Thecanalicular microvilli did not appear stunted.Overall, the liver biopsy showed no evidence ofimprovement during the period of ursodeoxy-cholic acid treatment and some progression offibrosis.

At 0.7 years our patient started treatmentwith chenodeoxycholic acid and cholic acid at adose of 8 mg/kg/d of both. There was an initialrise in AST followed by a fall, and a steady fallin bilirubin (Fig 2) to normal values over a

period of three months. The steatorrhoearesolved and the patient showed dramatic catchup weight gain so that she was on the third per-centile for weight at one year. In contrast withthe transaminases, the y-GT continued to riseduring the first two months of treatment to a

maximum value of 225 IU/1 but fell thereafterto within the normal range. At the age of

MethodsThe sample preparation used in the analysis ofurinary cholanoids by negative ion FAB-MS orLSI-MS has been described previously.3 Inthis study we used LSI-MS (using a caesiumion gun operated at 10 kV) rather than FAB-MS; the spectra obtained with the two ionisa-tion techniques do not differ significantly.'2The methods used for the analysis of bile acidsin plasma, urine, and duodenal juice by GC-MS2-4 13 14 and for the analysis of plasmasterols (including cholestanol) by GC-MS15have been described previously. The results ofplasma and urine analyses from our patientwere compared with samples from infants andchildren who were being investigated in ourlaboratories for a possible bile acid synthesisdefect or peroxisomal disorder. Thus the'normal' controls came largely from childrenwith neurological symptoms such as hypoto-nia, developmental delay, fits, deafness, etc;this group had no clinical or biochemical evi-dence of liver disease. The 'cholestatic con-trols' consisted of infants with clear evidence ofliver disease; the patient was assigned to thecontrol group when the cause of the liverdisease was discovered. This group includedchildren with biliary atresia, galactosaemia, a1antitrypsin deficiency, severe coarctation of theaorta, etc.

Results

Samples obtained before ursodeoxycholic acidtreatmentNegative ion LSI-MS analysis of the urinesample obtained from our patient at 0.3 yearsproduced the result shown in Fig 3. The majorions in the range mass/charge (m/z) ratio 400 tom/z 640 were m/z 444, 460, 494, . and510, which correspond to the quasimolecular(M-H)-ions for the glycine and taurine con-jugates of 7ox-hydroxy-3-oxo-4-cholenoic acidand 7a, 1 2oL-dihydroxy-3-oxo-4-cholenoic acid.The glycine and taurine conjugates of cheno-deoxycholic acid and cholic acid (mlz 448, 464,498, and 514) were not detectable above thebackground. In the normal controls, bile acidpeaks were just detectable in some cases; the 3-oxo-,A4 bile acid peaks were never bigger thanthe normal bile acid peaks. In the cholestaticcontrols, 3-oxo-A4 bile acids were sometimesdetected, particularly in children with severeliver dysfunction but the normal saturated bileacids were always present and accounted for30-970/o of the total. The urine sample fromour patient showed a number of additionalpeaks in the LSI mass spectrum (Fig 3). Theidentities of these are being investigated.

Table I shows results of the plasma sampleobtained at 0.3 years. The 3-oxo-A4 bileacids, 7ao-hydroxy-3-oxo-4-cholenoic acid, and

625

on 5 June 2018 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.38.4.623 on 1 April 1996. D

ownloaded from

Clayton, Mills, J7ohnson, Barabino, Marazzi

TABLE I Concentrations of 3-oxo-A14 bile acids, allo bile acids, and normal bile acids(chenodeoxycholic acid and cholic acid) in the plasma ofour patient and in plasmasamples from infants with normal liverfunction and infants with cholestasis ofknown cause

Plasma concentration (p£M)

Normal infants Infants with cholestasisBile acid Patient (n=38) (n=30)

7ot-hydroxy-3-oxo-4-cholenoic acid 1-94 ND 0-6.527a,12ct-dihydroxy-3-oxo-4-cholenoic acid 2-07 ND 0-4 36Allochenodeoxycholic acid ND 0-0 1 0-5-2Allocholic acid 0-76 0-0.15 0-3.1Chenodeoxycholic acid ND 0-2-12-7 13-4-181Cholic acid ND 04-6-68 4 7-403

ND=not detectable (<0 05 uM).

TABLE II Percentage composition of the mixture of bile acids in plasma ofour patient,infants with normal liverfunction, and infants with cholestasis ofknown cause

% Total plasma bile acid concentration

Normal infants Infants with cholestasisBile acid Patient (n=38) (n=30)

7ox-hydroxy-3-oxo-4-cholenoic acid 41 0 0-6 57a, 1 2oc-dihydroxy-3-oxo-4-cholenoic acid 43 0 0-4.4Allochenodeoxycholic acid 0 0-2.3 0-2.0Allocholic acid 16 0-2.3 0-14Chenodeoxycholic acid 0 31-91 11-88Cholic acid 0 8-69 10-88

7x, 12a-dihydroxy-3-oxo-4-cholenoic acid, werethe major bile acids and allo-cholic acid(3o,7o, 1 2a-trihydroxy-5ac-cholanic acid) wasreadily detectable whereas the concentrationsof cholic acid and chenodeoxycholic acid werebelow the limit of detection of the method(<005 FM). The two 3-oxo-A4 bile acids wereundetectable in the plasma of normal infantsbut they were occasionally detected in theplasma of infants with severe liver dysfunction.Concentrations as high as 6-5 FLM and 4-4 FLM(Table I) were seen in two infants who hadsevere coarctation and, as a result, both liverfailure and a reduced urine output. In allpatients apart from the case described in detailabove, however, an increase in the plasma con-centration of 3-oXo-1A4 bile acids was associ-ated with an increase in the concentration ofchenodeoxycholic acid and cholic acid. Thus,as Table II shows, the 3-oXo-A4 bile acidsaccounted for <15% of the total plasma bileacid mixture in all but the patient described inthis report for whom the figure was 84%.The plasma bile acid chromatogram from

our patient showed some additional minorpeaks, in particular one with a mass spectrumcontaining the ions m/z 414, 378, 363, 294,267, 145, 129, suggestive of the trimethylsilyleither of a side chain hydroxylated derivative

TABLE in Plasma bile acid concentrations in our patient (a) before bile acid treatment;(b) during treatment with ursodeoxycholic acid, and (c) during treatment withchenodeoxycholic acid plus cholic acid

Plasma concentration (,uM)

Age 03y Age 069y Age 081yNo Ursodeoxycholic Chenodeoxycholic

Bile acid treatment acid and cholic acid

7a-hydroxy-3-oxo-4-cholenoic acid 1-94 2-74 ND7ct,12n-dihydroxy-3-oxo-4-cholenoic acid 2-07 3.05 NDAllochenodeoxycholic acid ND ND NDAllocholic acid 0-76 0-43 NDChenodeoxycholic acid ND 0-06 24-84Cholic acid ND 0-26 7-72Ursodeoxycholic acid ND 14-02 2-03

ND=not detectable (<0 05 gM).

of 7ac, 1 2a-dihydroxy-cholest-4-en-3-one. Theurine LSI mass spectrum also suggested that a3-oxo-4 cholestenetriol may be present both asthe sulphate (m/z 51 1) and as the glucuronide(m/z 623). Sjovall has also reported the urinaryexcretion of bile alcohols with a 3-oxo-A4structure in patients with 5f-reductase defi-ciency. 16

Samples obtained during ursodeoxycholic acidtreatmentA plasma sample obtained when the patientwas 069 years and receiving ursodeoxycholicacid treatment showed that there had been nofall in the plasma concentration of 7ot-hydroxy-3-oxo-4-cholenoic acid and 7a, 1 2a-dihy-droxy-3-oxo-4-cholenoic acid (Table III).Ursodeoxycholic acid was the major plasmabile acid and metabolites of ursodeoxycholicacid were also readily detectable (for example,the compound tentatively identified byKoopman et al as 2 1-hydroxy-ursodeoxycholicacid'7). The plasma concentration ofcholestanol was normal. The LSI-MS analysisof urine showed that 7a-hydroxy-3-oxo-4-cholenoic acid and 7o, 122aL-dihydroxy-3-oxo-4-cholenoic acid were still present in urinealthough the major peak was m/z 471 (possiblysulphated ursodeoxycholic acid). Table IVshows the results of analysis of the urineby GC-MS (after enzymatic deconjugationof bile acids). The major bile acids in the urinewere 7ot, 1 2o-dihydroxy-3-oxo-4-cholenoicacid, 7ox-hydroxy-3-oxo-4-cholenoic acid andursodeoxycholic acid. Several metabolites ofursodeoxycholic acid were present; the majorones were identified as 21-, 6P-, 1p-, and 2p-hydroxy-ursodeoxycholic acid. Chenodeoxy-cholic acid and cholic acid were undetectable.

Bile samples obtained after stoppingursodeoxycholic acid treatmentDuodenal bile from our patient contained avery low cholanoid concentration of 19 KM(normal 2.5-23 mM). The major cholanoidspresent were cholic acid (13 FiM), 7a, 12ot-dihydroxy-4-cholenoic acid (1. 1 KM), and313,7a-dihydroxy-4-cholenoic acid (073 KM).Several unidentified compounds were present.

Samples obtained during treatment withchenodeoxycholic acid and cholic acidA urine sample obtained when the patient hadbeen receiving chenodeoxycholic acid andcholic acid treatment for five days producedthe result shown in Table IV, column 2. Therehad been a small fall in the urinary excretion of7c,12oc-dihydroxy-3-oxo-4-cholenoic acid anda larger fall in the urinary excretion of 7ot-hydroxy-3-oxo-4-cholenoic acid. A plasmasample obtained when the patient was 0x8years and receiving chenodeoxycholic acidand cholic acid treatment showed plasmaconcentrations of chenodeoxycholic acidand cholic acid that were slightly above thenormal ranges and undetectable concentra-tions of 7a-hydroxy-3-oxo-4-cholenoic acid

626

on 5 June 2018 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.38.4.623 on 1 April 1996. D

ownloaded from

A4-3-Oxosteroid 5/3-reductase deficiency

TABLE IV Urinary excretion of bile acids by the patient while taking ursodeoxycholic acidand while taking chenodeoxycholic acid and cholic acid

Urinary excretion (p.mol/mmol creatinine)

Age O7y Age O81yAge 0-69y Cheno- Cheno-Ursodeoxycholic deoxycholic deoxycholic

Bile acid acid and cholic acid and cholic acid

7a-hydroxy-3-oxo-4-cholenoic acid 22 9.1 17o, 1 2ot-dihydroxy-3-oxo-4-cholenoic acid 94 82 22Chenodeoxycholic acid ND 2-6 1-5Cholic acid ND 32 4-3Ursodeoxycholic acid 37 ND 0-9

ND=not detectable.

and 7ao,12cx-dihydroxy-3-oxo-4-cholenoic acid(Table III). The urine analysis showed that the3-oxo-/A4 bile acids were still present in urinebut the amounts being excreted (measured inp.mol/mmol creatinine) were much lower thatthey had been when the patient was receivingursodeoxycholic acid treatment (Table IV).

DiscussionIn the future, the gold standard for the diag-nosis of primary 5p-reductase deficiency willbe the finding of a gene mutation that leads toan inactive enzyme. This may not be far off as

the cDNA for the human enzyme has beencloned and the enzyme expressed in COScells.'8 For the time being it is necessary to use

other criteria. We believe that the patientdescribed in this report had primary 5p-reduc-tase deficiency for the following reasons: (a) noother cause of chronic cholestatic liver diseasewas identified; (b) the clinical features were

similar to those caused by deficiency of theenzyme 3,3-hydroxy-A5-C27-steroid dehydro-genase, which catalyses a step in bile acid syn-thesis that just precedes the 5,B-reductasereaction.' That is, firstly, the chronic cholesta-tic liver disease was associated with steator-rhoea, rickets (because of malabsorption ofvitamin D), a low plasma vitamin E concentra-tion, and a -y-GT that was normal, or onlyslightly raised at a time when the transaminaseswere considerably increased.' 19 Secondly, theabnormality of bile acid concentrations in bile,plasma, and urine was similar to that seen

in 3,3-hydroxy-A5-C27-steroid dehydrogenasedeficiency; the concentrations of chenodeoxy-cholic acid and cholic acid in bile were

<50 ,uM, the concentrations in plasma were

<3.5 ,uM (that is, in the low or low-normalrange), and the concentrations of abnormal(unsaturated) bile acids were considerablyincreased. We have analysed plasma bile acidsfrom 30 infants and children who had a knowncause of cholestatic liver disease but who were

also excreting 3-oxo-A4 bile acids in their urine(secondary 5,-reductase deficiency). Thesepatients had increased plasma concentrationsof chenodeoxycholic acid, increased or highnormal plasma concentration of cholic acid,and plasma concentrations of 3-oxo-A4 bileacids, which were lower than those of cheno-deoxycholic acid and cholic acid. (c) Thepatient described in this report showed a

complete response to treatment with cheno-deoxycholic acid and cholic acid. This is

strong evidence that defective bile acid syn-thesis is the cause and not the consequence ofliver disease.The cause(s) of liver cell damage in

1A4-3-oxosteroid 5p-reductase deficiency and33-hydroxy-A5-C27-steroid dehydrogenasedeficiency are not known. Two major hypothe-ses have been proposed. The first suggests thatloss of the bile acid dependent component ofbile flow leads to accumulation of toxiccompounds in the hepatocyte. The secondsuggests that it is the accumulation of bile acidprecursors or unsaturated bile acids, or both,which cannot be secreted into bile that isresponsible for liver cell damage.20 21 Stiegeret al have shown that the accumulation of 3-oxo-A4 bile acids in 5,B-reductase deficiencycan be expected to produce adverse effects.They showed that taurine conjugated 7c-hydroxy-3-oxo-4-cholenoic acid stronglyinhibits the transport of taurocholate by theATP dependent canalicular bile acid trans-porter.22 Ursodeoxycholic acid treatment canbe expected to fuel bile flow but ursodeoxy-cholic acid does not inhibit the first step in bileacid synthesis, cholesterol 7ox-hydroxylase, andtherefore cannot be as effective as chenodeoxy-cholic acid and cholic acid at reducing the syn-thesis of bile acid precursors and unsaturatedbile acids.23 The patient described in thisreport failed to respond to ursodeoxycholicacid treatment but showed a clear response toprimary bile acid (chenodeoxycholic acid andcholic acid) treatment. This suggests that inhi-bition of cholesterol 7oa-hydroxylase is import-ant for successful treatment of A4-3-oxosteroid5f-reductase deficiency. From a previousreport of bile acid treatment for 51-reductasedeficiency'0 it was not possible to say whetherursodeoxycholic acid had a beneficial effect asit was only used alone for four days. The com-bination of chenodeoxycholic acid and cholicacid led to reduced excretion of 3-oxo-lA4 bileacids, however, it was abandoned because itseemed to be producing diarrhoea. Longtermtreatment with a combination of ursodeoxy-cholic acid and cholic acid led to normalisationof liver function tests and resolution of liverbiopsy abnormalities. It took a year for thebilirubin to normalise unlike our patient whohad a normal bilirubin within three months ofstarting treatment with chenodeoxycholic acidand cholic acid. However, the cases are notstrictly comparable; the patients described byDaugherty et al had more severe cholestasis atthe onset of treatment and we know, from ourexperience of treating 3p-hydroxy-1A5-C27-steroid dehydrogenase deficiency, that patientswith severe cholestasis take longer to respond.'Such patients may also show a deterioration inliver function tests when treated with chen-odeoxycholic acid alone but improve quicklywith a combination of chenodeoxycholic acidand cholic acid.While chenodeoxycholic acid plus cholic

acid is probably the treatment of choice for apatient such as the one described in this report(who had a low plasma concentration ofchenodeoxycholic acid), we would urgecaution in its use in patients who have a high

627

on 5 June 2018 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.38.4.623 on 1 April 1996. D

ownloaded from

628 Clayton, Mills, J7ohnson, Barabino, Marazzi

plasma concentration of chenodeoxycholicacid (and thus we would not use it in a patientwith secondary 5p-reductase deficiency). Thereason for this is that, if the plasma concentra-tion of chenodeoxycholic acid is increased, it isprobable that the hepatocyte concentration isalso increased and hydrophobic dihydroxy bileacids such as chenodeoxycholic acid arethought to contribute to the liver cell damagethat occurs in cholestasis.

Although we have emphasised the similaritiesbetween our patient and infants with 3p-hydroxy-A5-C27-steroid dehydrogenase defi-ciency, the initial presentation with profoundliver dysfunction at two weeks has not beendescribed in the latter inborn error of bile acidsynthesis. An association ofprofound liver dys-function, the histological features of neonatalhaemochromatosis, and excretion of 3-oxo-A4bile acids has been described recently and theauthors suggested that primary 5p-reductasedeficiency can cause accumulation of iron andneonatal liver failure.8 We have argued that it ismore probable that these patients representanother example of secondary 5,-reductasedeficiency.9 We did not find evidence of ironstorage in the liver biopsy specimen of ourpatient, serum iron was at the lower end of thenormal range and iron binding capacity wasnot reduced. Thus on these criteria there wasno evidence of iron overload. The serumferritin concentrations was slightly increasedbut within the range expected from a widerange of hepatic insults and not in the rangeseen in haemochromatosis. Thus this patientprovides further evidence that disorders of bileacid synthesis do not produce haemochro-matosis.

1 Clayton PT. Inborn errors of bile acid synthesis. In:Fernandes J, Saudubray J-M, van den Berghe G, eds.Inborn metabolic diseases. Diagnosis and Treatment. Berlin:Springer-Verlag, 1995: 341-8.

2 Clayton PT, Casteels M, Mieli-Vergani G, Lawson AM.Familial giant cell hepatitis with low bile acid concentra-tions and increased urinary excretion of specific bilealcohols. A new inborn error of bile acid synthesis? PediatrRes 1995; 37: 424-31.

3 Lawson AM, Madigan MJ, Shortland DB, Clayton PT.Rapid diagnosis of Zellweger syndrome and infantileRefsum disease by fast atom bombardment of urine bilesalts. Clin Chim Acta 1986; 161: 221-31.

4 Clayton PT, Leonard JV, Lawson AM, Setchell KDR,Andersson S, Egestad B, et al. Familial giant cell hepatitisassociated with synthesis of 3P,7ac-dihydroxy- and3,,7os,12cs-trihydroxy-5-cholenoic acids. J Clin Invest1987; 79: 1031-8.

5 Clayton PT, Patel E, Lawson AM, Carruthers RA, TannerMS, Strandvik B, et al. 3-Oxo-delta-4 bile acids in liverdisease [Letter]. Lancet 1988; i: 1283-4.

6 Clayton PT. Bile acid metabolism in children with hepato-biliary disease. International Pediatrics 1995; 10: 44-50.

7 Setchell KDR, Suchy FJ, Welsh MB, Zimmer-NechemiasL, Heubi J, Balistreri WF. A4-3-Oxosteroid 5,B-reductasedeficiency described in identical twins with neonatalhepatitis. A new inborn error in bile acid synthesis. Jf ClinInvest 1988; 82: 2148-57.

8 Schneider BL, Setchell KDR, Whitington PF, Neilson KASuchy FJ. A4-3-Oxosteroid 5,B-reductase deficiency caus-ing neonatal liver failure and neonatal hemochromatosis.Jf Pediatr 1994; 124: 234-8.

9 Clayton PT. A4-3-oxosteroid 5,B-reductase deficiency andneonatal hemochromatosis [Letter]. Jf Pediatr 1994; 125:845-6.

10 Setchell KDR, Schneider BL, Suchy FJ, Whitington PF.A4-3-Oxosteroid 5,B-reductase deficiency and neonatalhemochromatosis. Reply [Letter]. Jf Pediatr 1994; 125:846.

11 Daugherty CC, Setchell KDR, Heubi JE, Balistreri W.Resolution of liver biopsy alterations in three siblings withbile acid treatment of an inborn error of bile acid meta-bolism (A4-3-oxosteroid 5,B-reductase deficiency).Hepatology 1993; 18: 1096-101.

12 Setchell KDR, Lawson AM. Bile acids. In: Lawson AM, ed.Mass spectrometry. Berlin: Walter de Gruyter, 1989:55-109.

13 Clayton PT, Muller DPR. A simplified gas-liquid chro-matographic method for the estimation of non-sulphatedplasma bile acids. Clin Chim Acta 1980; 105: 401-5.

14 Clayton PT, Lake BD, Hall NA, Shortland DB, CarruthersRA, Lawson AM. Plasma bile acids in patients with per-oxisomal dysfunction syndromes; analysis by capillary gaschromatography - mass spectrometry. EurJrPediatr 1987;146: 166-73.

15 Clayton PT, Bowron A, Mills KA, Masoud A, Casteels M,Milla PJ. Phytosterolaemia in children with parenteralnutrition-associated cholestatic liver disease. Gastro-enterology 1993; 105: 1806-13.

16 Sjovall J. Mass spectrometry in studies of inherited andacquired diseases of bile acid synthesis and metabolism.In: Matsumoto I, Kuhara T, Mamer OA, Sweetman L,Calderhead RG, eds. Advances in chemical diagnosis andtreatment of metabolic disorders. Vol 2. Kanazawa:Kanazawa Medical University Press, 1994: 107-22.

17 Koopman BJ, Wolthers BG, van der Molen JC, Nagel GT,Kruizinga W. Abnormal urinary bile acids in a patient suf-fering from cerebrotendinous xanthomatosis during oraladministration of ursodeoxycholic acid. Biochim BiophysActa 1987; 917: 238-46.

18 Kondo K-H, Kai M-H, Setoguchi Y, Eggersten G, SjoblomP, Setoguchi T, et al. Cloning and expression of cDNA ofhuman A4-3-Oxosteroid 5,-reductase and substratespecificity of the expressed enzyme. Eur Jf Biochem 1994;219: 357-63.

19 Jacquemin E, Setchell KDR, O'Connell CO, Estrada A,Maggiore G, Schmitz J, et al. A new cause of progres-sive intrahepatic cholestasis: 3,B-hydroxy-C27-steroiddehydrogenase/isomerase deficiency. J Pediatr 1994;125: 379-84.

20 Ichimiya H, Nazer H, Gunasekaran T, Clayton P, Sjovall J.Treatment of chronic liver disease caused by 3,B-hydroxy-A5-C27-steroid dehydrogenase deficiency with cheno-deoxycholic acid. Arch Dis Child 1990; 65: 1121-4.

21 Ichimiya H, Egestad B, Nazer H, Baginski S, Clayton PT,Sjovall J. Bile acids and bile alcohols in a child withhepatic 3p-hydroxy-A5-C27-steroid dehydrogenase defi-ciency: effects of chenodeoxycholic acid treatment. J LipidRes 1991; 32: 829-41.

22 Stieger B, Zhang J, O'Neill B, Sjovall J, Meier PJ. Transportof taurine conjugates of 7a-hydroxy-3-oxo-4-cholenoicacid and 3,B,7a-dihydroxy-5-cholenoic acid in rat liverplasma membrane vesicles. In: van Berge Henegouwen, etal, eds. Cholestatic liver diseases. Dordrecht: KluwerAcademic, 1994: 82-7.

23 Koopman BJ, Wolthers BG, van der Molen JC, Nagel GT,Waterreus RJ, Oosterhuis HGJGH. Capillary gaschromatographic determinations or urinary bile acids andbile alcohols in CTX-patients proving the ineffectivity ofursodeoxycholic acid treatment. Clin Chim Acta 1984;142: 103-11.

on 5 June 2018 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.38.4.623 on 1 April 1996. D

ownloaded from