A

io

cfea

im

m©

K

1

tajo

i3

1d

Digestive and Liver Disease 39 (2007) 351–355

Liver, Pancreas and Biliary Tract

Inheritance of hyperbilirubinemia: Evidence fora major autosomal recessive gene

M. Clementi a, E. Di Gianantonio a, L. Fabris b, P. Forabosco c,M. Strazzabosco d, R. Tenconi a, L. Okolicsanyi b,∗

a Clinical Genetics and Epidemiology, Department of Pediatrics, University of Padua, Italyb Department of Surgical and Gastroenterological Sciences, University of Padua, Italy

c Centro Nazionale Ricerche (CNR), Cagliari, Italyd Department of Internal Medicine, Section of Digestive Diseases, TAC Building,

Room 233, Yale University, New Haven, USA

Received 6 September 2006; accepted 18 December 2006Available online 7 March 2007

bstract

Background and aim. To clarify the precise mode of inheritance of Gilbert syndrome, an unconjugated familial hyperbilirubinemia, wherempaired bilirubin conjugation is caused by reduced UGT1A1 activity determined by a defective function of the A(TA)6TAA promoter regionf the UGT1A1 gene.

Subjects and methods. Serum bilirubin levels were measured in a large, homogeneous resident population from North-Eastern Italy,onsisting of 1.639 males (age 44.5 ± 13.9, range 18–89 years), and 1.420 females (age 45.1 ± 15.0, range 18–85). In 112 nuclear familiesrom hyperbilirubinemic probands living in the same area a complex segregation analysis was then performed. In both samples we carefullyxcluded potentially confounding factors of bilirubin levels (alcohol abuse, excessive cigarette smoking, drug consumption, overt haemolysisnd liver disease).

Results. Mean serum bilirubin concentrations are higher in males than in females, showing fluctuations through the different age periodsn males. Complex segregation results demonstrate that unconjugated hyperbilirubinemia exhibits a precise mode of inheritance in which a

ajor recessive gene with a frequency of 0.45 is responsible for higher serum bilirubin values.Conclusions. This major recessive gene accounts only for a part of the serum bilirubin concentration, thus implying additional, environ-ental factors for the clinical appearance of GS.2007 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.

eritance

eatt

eywords: Bilirubin; Complex segregation analysis; Gilbert syndrome; Inh

. Introduction

Unconjugated hyperbilirubinemia syndromes are selec-ive disorders of bilirubin metabolism, resulting in

ccumulation of pigment in the blood [1]. Isolated uncon-ugated hyperbilirubinemia can be defined as the repeatedbservations of total serum bilirubin concentration clearly in

∗ Corresponding author at: Department of Surgical and Gastroenterolog-cal Sciences, University of Padua, Ospedale Busonera, Via Gattamelata,5100 Padua, Italy. Tel.: +39 049 821 5617; fax: +39 049 821 5619.

E-mail address: lajos.okolicsanyi@unipd.it (L. Okolicsanyi).

dw

ofagf

590-8658/$30 © 2007 Editrice Gastroenterologica Italiana S.r.l. Published by Elseoi:10.1016/j.dld.2006.12.019

xcess of the values found in normal individuals of the samege and sex, with direct-reacting bilirubin comprising lesshan 15–20% of the total. Clinical forms vary widely fromhe mild, asymptomatic hyperbilirubinemia of Gilbert syn-rome (GS) to the severe Crigler–Najjar syndrome (type I)here kernicterus is frequent.In GS unconjugated hyperbilirubinemia occurs without

vert signs of haemolysis and other abnormalities of liver

unction. Increased bilirubin levels are caused by a reducedctivity of the liver microsomal enzyme uridine-diphospho-lucuronate glucuronosyl-transferase (UGT), which has dif-erent isoforms, one of which, the UGT1A1, catalyses the

vier Ltd. All rights reserved.

3 nd Live

tptUamtlipa

whTghthasobhtbidgbbbmdufG

cNio

2

2

[saasyD

nu

2

wpNt4siottTtdlohietfGcwpbmsfupaa5fel

2

pa[smn

52 M. Clementi et al. / Digestive a

ransformation of the lipid-soluble bilirubin IX-� to the moreolar, water soluble bilirubin, through its conjugation withwo molecules of glucuronic acid. The gene encoding forGT1A1 is located in region q37 of chromosome 2 (long

rm) and is regulated by the promoter TATAA box that directsRNA transcription by controlling RNA polymerase II and

ranscription factor 2D [2]. Population and family studieseft in the past a degree of uncertainty regarding the mode ofnheritance of GS, although it has been hypothesized in theast an autosomal dominant pattern of transmission (#143500s reported in the OMIM catalog) [3–5].

More recently, Bosma et al. observed [2] that patientsith GS are homozygous for the TATAA box polymorphism,aving two extra bases of thymidine–adenine, instead of aATAA element with six repeats. The same authors also sug-ested that this abnormality is not sufficient by itself to revealyperbilirubinemia and subicterus. The allele frequency ofhe longer A(TA)7TAA sequence in the normal populationas been estimated to be 0.40, allowing a frequency of thebnormal promoter element homozygosity of about 16% ofubjects. However, less than half of the expected proportionf homozygotes actually have increased serum unconjugatedilirubin levels [4,6], so that the estimated prevalence ofyperbilirubinemia is only around 5% in the adult popula-ion, without taking into account the sex differences in serumilirubin [4]. Furthermore, during the last years, three stud-es on association of serum bilirubin with coronary heartisease showed that serum bilirubin segregates as a majorene with the rarer genotype associated with high serumilirubin levels [7–9]. Extensive family studies of serumilirubin concentrations, including a sufficiently large num-er of healthy subjects following exclusion of factors thatay influence serum bilirubin concentration (alcohol and

rug consumption, manifest haemolysis or liver disease) andsing appropriate procedures of genetic analysis, are there-ore needed in order to clarify the genetic heterogeneity ofS [4,5].Taking this approach, we measured serum bilirubin

oncentrations in a large, homogeneous population fromorth-Eastern Italy and then selected 112 nuclear families

n which a complex segregation analysis was performed inrder to define the mode of inheritance of GS.

. Materials and methods

.1. Subjects

This was an ancillary investigation of the M.I.COL study10]. A random control sample of 3.059 unrelated healthyubjects living in the district of Montegrotto Terme (Padua),rural area of the Veneto Region, North-Eastern Italy, were

sked to participate in this study, and gave their informed con-ent. There were 1.639 males (age 44.5 ± 13.9, range 18–89ears), and 1.420 females (age 45.1 ± 15.0, range 18–85).ata from this sample were used in this study to obtain the

Taam

r Disease 39 (2007) 351–355

ormal population values (distribution by sex and age) to besed for the segregation analysis.

.2. Clinical data recording

All the participants (controls and family samples) under-ent routine clinical and laboratory examination, and aersonal interview using the M.I. COL questionnaire [10].one was taking any drug, including over-the-counter drugs,

here was no excessive alcohol consumption (less than0 g/day in males and 20 g/day in females) and cigarettemoking (less than 10 cigarettes/day) potentially capable tonterfere with liver function and bilirubin metabolism at timef blood sampling. All were healthy at the physical examina-ion. After an overnight fasting, venous blood samples wereaken and analysed by an autoanalyzer (SMAC II Analyzer,echnicon Instrument Corp.). For each individual, serum

otal and direct-reacting bilirubin values were recorded in aatabase, and serum indirect bilirubin obtained by difference;iver ultrasound was also performed in each subject. Nonef these subjects evidenced overt haemolysis on standardaematologic tests or routine liver enzyme abnormalities;n addition, none had hepatobiliary alterations at ultrasoundxamination. Taking these data together, concurrent hepa-obiliary and haematologic disorders potentially responsibleor increased bilirubin levels could be reasonably ruled out.S was then defined according to previously well-established

riteria [1]: hyperbilirubinemia in both sexes was consideredhen total serum bilirubin was greater than 17.5 �mol/l. Inarticular, in all of these selected individuals direct reactingilirubin never exceeded 30% of total. From hyperbilirubine-ic probands, observed in the gastroenterology unit, we then

elected 112 nuclear families (parents and siblings) derivedrom the same geographical area, resulting in 427 individ-als overall, which agreed to participate in this study. Botharents were available in 17 families, only one parent wasvailable in 49 families and 46 families had no parents avail-ble. Fathers ranged in age from 46 to 77 years (mean age7.9), mothers from 44 to 82 (mean age 61.9), and siblingsrom 12 to 76 (mean age 44.7). Relatives evidencing liver dis-ase and/or anomaly in standard haematologic tests or routineiver enzyme were excluded from the study.

.3. Analytical methods of complex segregation analysis

Pedigrees of nuclear families were analysed by com-lex segregation analysis (CSA) using the mixed model [11]s formulated and implemented in the POINTER program12]. POINTER was used on a SUN Sparc Classic X work-tation running the UNIX operating system. In the mixedodel a major gene, a transmissible multifactorial compo-

ent and a residual random source of variation are assumed.

he major locus (ML) component is determined by twolleles producing three possible genotypes. Phenotypes aressumed normally distributed around the mean values of eachajor-locus genotype and have equal variances. Complex

M. Clementi et al. / Digestive and Liver Disease 39 (2007) 351–355 353

Table 1Population sample

Age groups 18–30 31–40 41–50 51–60 61–89 18–89

Males 14.5 ± 8.3* 12.8 ± 5.2* 12.9 ± 5.4* 13.4 ± 6.0* 12.1 ± 4.8* 13.2 ± 6.1*

N 294 347 433 359 206 1639F .5 ± 4.7N 9

D years).

stbdemdoiogasstddwdocsc

2

afbcti

agemiaR[vc

3

3

vBwRanarfdd

TS

H

N

T

T

Uhv

emales 10.6 ± 4.8 10.4 ± 5.1 10289 284 29

istribution of total serum bilirubin levels (�mol/l) by sex and age groups (

egregation analysis is a general method for evaluating theransmission of a trait/disease within pedigrees. It proceedsy testing models of varying degrees of generality, both toetermine whether a Mendelian locus is likely to exert a largeffect on the phenotype (trait or disease) and to estimate theagnitude of genetic sources of variation in the trait. These

ata are useful to start linkage analysis, and to provide a modeln which to base parametric linkage methods. The major lim-tations of CSA are that a large amount of a very specific typef data is generally needed and that CSA is unable to distin-uish between the effect of a single locus that underlies a traitnd the effects of two or more independently acting loci withimilar transmission patterns. For example, if many reces-ive loci are involved in determining a complex trait/disease,hey would be detected as if they were a single locus with aisease–allele frequency equalling that of the sum of severalisease alleles. The resulting overestimate of allele frequencyould result both in an overestimate of the probability ofetecting a locus, as well as in misspecification of the modelf inheritance. This error will reduce the power to detect eachontributing locus. Notwithstanding these limitations, manytudies have used CSA methodology to identify the geneticomponent and the inheritance model in complex diseases.

.4. Normalization of data

The analysed data showed differences between sexes andmong age groups (Table 1), thus they have been adjustedor these effects using the control sample. As requested

y POINTER for a quantitative trait, a transformation wasarried out to eliminate skewness and to obtain a normal dis-ribution with mean = 0 and variance = 1. This transformations considered appropriate to guard against false inference of

3

s

able 2egregation analysis of log-transformed standardized residuals (family sample)

ypothesis U V Q

o familiar effect 0 1 (0)

ests on ML dominance:High values: ML recessive .03 1.02 .4488High values: ML dominant .03 1.08 .0085High values: ML additive .02 1.01 .0544High values: ML general .03 1.02 .4488

ests on the transmission probability τ2:High values: ML general −.01 .98 .4400

: The overall phenotypic mean; V: variance; Q: the gene frequency at major loomozygous means; D: the degree of dominance which corresponds to the relative dariance attributable to the multifactorial component; τ: the transmission probabilit

10.1 ± 5.1 10.3 ± 4.2 10.4 ± 4.8310 238 1420

*p < 0.01.

major gene [13] although restricts power to detect majorene segregation. Data were then corrected performing lin-ar regression of serum bilirubin values on age, separately forales and females. Residual values were computed for each

ndividual as the difference between the observed phenotypend the predicted value on the basis of the regression model.esiduals were then standardized using the program Skumix

14] that estimates an appropriate value of p for standardizedalues that removes skewness in the presence of up to threeommingled distributions.

. Results

.1. Statistics and data adjustment

Table 1 shows the distribution of total serum bilirubinalues in males and females in the population sample.ilirubin values in all age classes considered in this studyere significantly lower (p < 0.01) in females than in males.egression analysis showed a slight inverse correlation withge in males (BILmales = 0.655 + 4.111/age), while age hado impact on bilirubin levels in females. Effects of agend gender on observed phenotype (serum bilirubin) wereemoved by performing linear regression on age separatelyor males and females. Skumix [14] was used to transformata and obtain a distribution which did not show significanteviation from normality.

.2. Segregation analysis

Segregation analysis of serum bilirubin log transformedtandardized residuals with the program POINTER is shown

T D τ2 −2lnL+C

1209.55

1.74 (0) 1190.213.10 (1) 1204.042.47 (.5) 1206.521.74 0 1190.21

1.70 0 .44 1189.16

cus which favours high values of the trait; T: the displacement betweenisplacement of the heterozygous mean; H: the proportion of the phenotypicies.

3 nd Live

ihphwwmqoF4

4

ooioiymcm

tmrHgati[crctpmlooitclacaihgca

iwtblwsblttsdHtaeta

wycmmeicabrCofpipiwrlo

ettabcoi

54 M. Clementi et al. / Digestive a

n Table 2. The hypothesis of no familial effect and theypothesis of no major locus (q = 0) were rejected when com-ared with the major locus model (χ2 = 19.34, p < .0001). Theypotheses on the dominance parameter for the major geneere tested: dominant (d = 1) and additive major gene (d = .5)ere also rejected when compared to the general major-locusodel (χ2 = 13.83, p = .0002; χ2 = 16.31, p < .0001). A fre-

uent major recessive gene (q = .45) that controls high valuesf SB gave the best fit and considered the model for this study.or this model the major locus accounted in our sample for5% of the total phenotypic variance.

. Discussion

GS is considered to be a genetically transmitted defectf bilirubin conjugation, since its strikingly frequent familialccurrence has been assumed for many years [3–5]. However,ts precise mode of inheritance is still unclear. In this epidemi-logical study carried out in a resident healthy population,ncluding 112 nuclear families, by complex segregation anal-sis we confirmed that hyperbilirubinemia exhibits a preciseode of inheritance, where a major autosomal recessive gene

ontrolling the high values of serum bilirubin is the best fittingodel.During the last years, molecular studies have suggested

hat the presence of two extra bases in the A(TA)6TAA pro-oter region of the gene UGT1A1 is responsible for the

educed UGT1A1 activity leading to hyperbilirubinemia [2].owever, as the prevalence of the high risk A(TA)7TAAenotype was estimated to be 0.40 and the prevalence of GSbout 5%, in normal population the reduced expression ofhe UGT1A1 gene caused by a defect of the promoter regions not sufficient by itself to determine hyperbilirubinemia6,15]. Therefore, this should be considered a multifactorialondition requiring the influence of other (genetic or envi-onmental) factors acting in modifying the serum bilirubinoncentration in otherwise healthy persons, homozygous forhe promoter defect. In accordance with this concept, it hasreviously been shown that one third of the subjects with GSay have a concurrent generalized defect of the hepatocel-

ular uptake or alternatively, of the intracellular translocationf organic anions, such as bilirubin, bromosulfon-phtaleinr indocyanine green [5]. As shown by Persico et al. [15],n GS subjects, a second abnormality in the function of theransport system for organic anion uptake coexists with defi-ient glucuronidation, in accordance with the observation thatess than half of the subjects homozygous for the (TA)7TAAbnormality actually have the elevated plasma bilirubin levelsharacteristics of GS. In addition, nearly 50% of individu-ls with GS have also a shortened erythrocyte life span [3],ndicating that bilirubin overproduction derived from a mild

aemolytic state may become evident once the conjugatinglucuronyl-transferase activity is reduced [16], as seen duringoncurrent stress conditions that affect the haeme oxygenasectivity [17].

o(pg

r Disease 39 (2007) 351–355

Sex and age are additional factors that may be involvedn the regulation of bilirubin metabolism. In our study, dataere obtained in healthy subjects randomly collected from

he same rural area, to define the normal population valuesy age and sex. According to recent data [18], we found thatevels of serum bilirubin were always higher in men than inomen; furthermore, whereas in women they remain con-

tant throughout age, in men the highest values are reachedetween 18 and 30 years. The discrepancy in serum bilirubinevels between men and women does not appear to be relatedo sex differences in the distribution of the TATA-box geno-ype [19–21], but rather to the opposite effects, inhibitory andtimulatory, exerted by sex hormones, testosterone, and estra-iol and progesterone, respectively, on UGT activity [22].owever, the role of testosterone seems to be prevalent given

he fact that in men a peak of bilirubin levels was seen mostlyfter puberty, in contrast with women showing constant lev-ls with age groups. More recently, it has been demonstratedhat the hepatic uptake of organic anions is sex-, but notge-dependent [15].

To determine the genetic component of serum bilirubin,e studied SB inheritance, using complex segregation anal-sis [11]. Data have been adjusted to reduce skewness thatould infer the presence of a major gene, even if transfor-ation is known to reduce the power to detect an existingajor gene segregation. In our study, SB has been consid-

red as a continuous trait, while in the clinical approach SBs usually dichotomized, with values higher than 17.5 �mol/lonsidered as feature of GS. Taking this approach we wereble to overcome a possible epidemiological bias determinedy the different distribution of serum bilirubin concentrationelated to age and sex, as shown in Table 1. The results ofSA suggest that a major recessive gene with a frequencyf 0.45 is responsible of higher SB values. This observedrequency is strikingly similar to the A(TA)7TAA frequencyreviously estimated by Bosma (0.44), and obtained by divid-ng the reported median prevalence of GS (0.06) by the meanrevalence (0.137) of the homozygous A(TA)7TAA mutationn six series [2]. Our estimated frequency is also in agreementith that of 0.50 estimated by Persico [15], that has been

elated to an impaired organic anion uptake, possibly regu-ated by genetic abnormalities in the basolateral transporterf unconjugated bilirubin.

The major recessive gene responsible of higher SB lev-ls determined by our analysis accounts only for a part ofhe SB determination, so that an additional, environmen-al component is also necessary for determining the clinicalppearance of GS, as also suggested by others [15]. It muste emphasised that by CSA we could not distinguish the spe-ific genes involved, but only their effect. The high frequencyf the major recessive gene may also explain the uncertaintyn determining the inheritance model. In fact the proportion

f homozygotes for the rarer allele (high SB value) is 0.20q2), while the heterozygotes frequency is 0.49 (2pq), and therobability of matching homozygote/homozygote, homozy-ote/heterozygote, heterozygote/heterozygote is extremely

nd Live

fctt

dito

CTpm

A

i

R

[

[

[

[

[

[

[

[

[

[

[

[

M. Clementi et al. / Digestive a

requent in the population allowing the possibility of affectedhildren from both affected (simulating autosomal dominantransmission) and unaffected parents (autosomal recessiveransmission).

In summary, our results, based on both epidemiologicalata and genetic analysis, support the concept that GS isndeed a distinct genetically inherited clinical entity, evenhough evidenced by the concurrence of age-, sex-related andther environmental factors.

Practice points

• The precise mode of inheritance of Gilbertsyndrome, an unconjugated familial hyper-bilirubinemia, is unknown.

• Using complex segregation analysis in alarge, homogeneous resident populationfrom North-Eastern Italy we showed thatunconjugated hyperbilirubinemia exhibits aprecise mode of inheritance where a majorrecessive gene with a frequency of 0.45 isresponsible for higher serum bilirubin val-ues.

• However, this major recessive gene accountsonly for a part of the serum bilirubin con-centration, thus implying additional, environ-mental factors for the clinical evidence of thiscondition.

Research agenda

• Epidemiological studies investigating addi-tional, environmental components necessaryfor determining the clinical appearance of GSare needed.

onflict of interest statementhe authors disclose any commercial association that mightose a conflict of interest in connection with the submittedanuscript.

cknowledgment

The financial support of Telethon, Italy (LF, grant E.1253)s gratefully acknowledged.

eferences

[1] Okolicsanyi L, Fevery J, Billing B, Berthelot P, ThompsonRPH, Schmid R, et al. How should mild isolated unconjugated

[

r Disease 39 (2007) 351–355 355

hyperbilirubinemia be investigated? Semin Liver Dis 1983;3:36–41.

[2] Bosma PJ, Chowdhury JR, Bakker C, Gantla S, de Boer A, OostraBA, et al. The genetic basis of the reduced expression of bilirubinUDP-glucuronosyl-transferase-1 in Gilbert’s syndrome. N Engl J Med1995;333:1171–5.

[3] Powell LW, Hemingway E, Billing BH, Sherlock S. Idiopathic unconju-gated hyperbilirubinemia (Gilbert’s syndrome): a study of 42 families.N Engl J Med 1967;277:1108–12.

[4] Okolicsanyi L, Nassuato G, Muraca M, Orlando R, Iemmolo RM,Lirussi F, et al. Epidemiology of unconjugated hyperbilirubinemia:revisited. Semin Liver Dis 1988;8:179–82.

[5] Schmid R. Gilbert’s syndrome—a legitimate genetic anomaly? N EnglJ Med 1995;333:1217–8.

[6] Sampietro M, Lupica L, Perrero L, Romano R, Molteni V, FiorelliG. TATA-box mutant in the promoter of the uridine diphosphate glu-curonosyltransferase gene in Italian patients with Gilbert’s syndrome.Italian J Gastroenterol Hepatol 1998;30:194–8.

[7] Hunt SC, Wu LL, Hopkins PN, Williams RR. Evidence for a major geneelevating serum bilirubin concentration in Utah pedigrees. ArteriosclerThromb Vasc Biol 1996;16:912–7.

[8] Hunt SC, Kronenberg F, Eckfeldt JH, Hopkins PN, Myers RH, HeissG. Association of plasma bilirubin with coronary heart disease andsegregation of bilirubin as a major gene trait: the NHLBI family heartstudy. Atherosclerosis 2001;154:747–54.

[9] Lin JP, Cupples LA, Wilson PWF, Heard-Costa N, O’Donnel CJ.Evidence for a gene influencing serum bilirubin on chromosome 2qtelomere: a genomewide scan in the Framingham study. Am J HumGenet 2003;72:1029–34.

10] Attili AF, Carulli N, Roda E, Barbara B, Capocaccia L, Menotti A, et al.Epidemiology of gallstone disease in Italy: prevalence data of the Mul-ticenter Italian Study on Cholelithiasis (M.I. COL.). Am J Epidemiol1995;141:158–65.

11] Morton NE, MacLean CJ. Analysis of family resemblance III. Com-plex segregation analysis of quantitative trait. Am J Hum Genet1974;26:318–30.

12] Lalouel JM, Morton NE. Complex segregation analysis with pointers.Hum Hered 1981;31:312–21.

13] Lalouel JM, Rao DC, Morton NE, Elston RC. A unified model forcomplex segregation analysis. Am J Hum Genet 1983;35:816–26.

14] MacLean CJ, Morton NE, Elston LC, Yee S. Skewness in commingleddistributions. Biometrics 1976;32:695–9.

15] Persico M, Persico E, Bakker CTM, Rigato I, Amoroso A, Torella R, etal. Hepatic uptake of organic anions affects the plasma bilirubin levelin subjects with Gilbert’s syndrome mutations in UGT1A1. Hepatology2001;33:627–32.

16] Okolicsanyi L, Ghidini O, Orlando R, Cortellazzo S, Benedetti G, Nac-carato R, et al. An evaluation of bilirubin kinetics with respect to thediagnosis of Gilbert’s syndrome. Clin Sci Mol Med 1978;54:539–47.

17] Okolicsanyi L, Orlando R, Venuti M, Dal Brun G, Cobelli C, RuggeriA, et al. A modeling study of the effect of fasting on bilirubin kineticsin Gilbert’s syndrome. Am J Physiol 1981;240:R266–71.

18] Zucker SD, Horn PS, Sherman KE. Serum bilirubin levels in the U.S.population: gender effect and inverse correlation with colorectal cancer.Hepatology 2004;40:827–35.

19] Monaghan G, Ryan M, Seddon R, Hume R, Burchell B. Geneticvariation in bilirubin UPD-glucuronosyltransferase gene promoter andGilbert’s syndrome. Lancet 1996;347:78–81.

20] Biondi ML, Turri O, Dilillo D, Stival G, Guagnellini E. Contribu-tion of the TATA-box genotype (Gilbert syndrome) to serum bilirubinconcentrations in the Italian population. Clin Chem 1999;45:897–8.

21] Borlak J, Thum T, Landr O, Erb K, Hermann R. Molecular diagnosis

of a familial nonhemolytic hyperbilirubinemia (Gilbert’s syndrome) inhealthy subjects. Hepatology 2000;32:792–5.

22] Muraca M, Fevery J. Influence of sex and sex steroids on bilirubinuridine diphosphate-glucuronosyltransferase activity of rat liver. Gas-troenterology 1984;87:308–13.