Sterol transporter adenosine triphosphate–binding cassette transporter G8, gallstones, and biliary cancer in 62,000 individuals from the general population

Authors

  • Stefan Stender,

    1. Department of Clinical Biochemistry, Rigshospitalet, Copenhagen, Denmark
    2. Copenhagen University Hospital, University of Copenhagen, Copenhagen, Denmark
    Search for more papers by this author
  • Ruth Frikke-Schmidt,

    1. Department of Clinical Biochemistry, Rigshospitalet, Copenhagen, Denmark
    2. The Copenhagen General Population Study, Herlev Hospital, Copenhagen, Denmark
    3. Copenhagen University Hospital, University of Copenhagen, Copenhagen, Denmark
    Search for more papers by this author
  • Børge G. Nordestgaard,

    1. The Copenhagen General Population Study, Herlev Hospital, Copenhagen, Denmark
    2. Department of Clinical Biochemistry, Herlev Hospital, Copenhagen, Denmark
    3. The Copenhagen City Heart Study, Bispebjerg Hospital, Copenhagen, Denmark
    4. Copenhagen University Hospital, University of Copenhagen, Copenhagen, Denmark
    5. Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark
    Search for more papers by this author
  • Anne Tybjærg-Hansen

    Corresponding author
    1. Department of Clinical Biochemistry, Rigshospitalet, Copenhagen, Denmark
    2. The Copenhagen General Population Study, Herlev Hospital, Copenhagen, Denmark
    3. The Copenhagen City Heart Study, Bispebjerg Hospital, Copenhagen, Denmark
    4. Copenhagen University Hospital, University of Copenhagen, Copenhagen, Denmark
    5. Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark
    • Professor, Chief Physician, Department of Clinical Biochemistry KB3011, Section for Molecular Genetics, Rigshospitalet, Copenhagen University Hospital, Blegdamsvej 9, DK-2100 Copenhagen, Denmark
    Search for more papers by this author
    • fax: +45 3545 4160


  • Potential conflict of interest: Nothing to report.

Abstract

Gallstone disease, a risk factor for biliary cancer, has a strong heritable component, but the underlying genes are largely unknown. To test the hypothesis that ABCG8 (adenosine triphosphate–binding cassette transporter G8) Asp19His (D19H) genotype predicted risk of gallstones and biliary cancer in the general population, we studied 62,279 white individuals from The Copenhagen City Heart Study and The Copenhagen General Population Study, randomly selected to reflect the adult Danish population aged 20 to 80+ years. Endpoints were recorded from January 1976 through May 2009. During a mean follow-up of, respectively, 31 and 4.4 years, 3124 participants developed symptomatic gallstone disease and 30 developed biliary cancer. The multifactorially adjusted hazard ratio for symptomatic gallstone disease was 1.9 (95% confidence interval, 1.7-2.1) in DH heterozygotes (prevalence, 12%), and 3.3 (2.3-4.6) in HH homozygotes (0.4%) versus noncarriers (P for trend <0.001). Mean age at onset of symptomatic gallstone disease was 56 years for noncarriers, 54 for DH heterozygotes, and 52 for HH homozygotes (P for trend <0.001). The fraction of all gallstones attributed to D19H was 11%. The multifactorially adjusted hazard ratio for biliary cancer was 4.0 (1.9-8.4) in DH heterozygotes and HH homozygotes combined versus noncarriers (P< 0.001). The fraction of all biliary cancers attributed to the D19H genotype was 27%. Finally, D19H genotype associated with stepwise increases in plasma levels of alanine aminotransferase and gamma glutamyltransferase of up to 14% and 25% in HH homozygotes, and with corresponding stepwise reductions in plasma levels of total and low-density lipoprotein cholesterol of up to 5% versus noncarriers (all comparisons, P for trend <0.001). Conclusion: In this general population cohort, ABCG8 D19H genotype was an important predictor of both symptomatic gallstone disease and biliary cancer. (HEPATOLOGY 2011:53:640-648)

Gallstone disease is one of the most common and costly diseases in the Western World.1,2 Apart from risk factors such as age, obesity, and female sex, genetic factors account for at least 25% of the risk of gallstone disease.3,4 Nevertheless, despite the clinical and economic impact of gallstones, lithogenic (i.e., gallstone-forming) genetic variation remains largely unknown.5

A recent genomewide association scan identified the sterol transporter adenosine triphosphate–binding cassette transporter G8 (ABCG8) as a susceptibility factor for gallstone disease in humans.6 The association was attributed to a common variant that leads to the substitution of aspartate with histidine at amino acid residue 19 (D19H; rs11887534). It has been hypothesized that D19H constitutes a gain-of-function variant that increases hepatobiliary cholesterol efflux, resulting in precipitation of cholesterol gallstones.6-8 Gallstone disease is a risk factor for biliary cancer, a rare but highly lethal disease.9

Whether D19H affects the risk of gallstone disease and biliary cancer in the general population is currently unknown. This question is clinically important because (1) genetic variants identified in case-control studies often overestimate risk compared to studies of the general population10, 11; (2) D19H has the potential to become the first lithogenic genetic variant included in presymptomatic genetic screening for gallstone disease and/or biliary cancer in the general population; and (3) the identification of specific genetic variants predisposing to gallstones and biliary cancer might also lead to new nonsurgical interventions that extend our current limited strategies for the prevention of these diseases.

We tested the clinical impact of the D19H genotype in 62,279 individuals from the Danish general population followed from January 1976 through May 2009, of whom 3124 developed symptomatic gallstone disease and 30 developed biliary cancer. Extensive clinical biochemistry was available on all participants as well as information on well-known risk factors for gallstone disease. First, we determined whether D19H predicted risk of symptomatic gallstone disease and/or biliary cancer in the general population; on the basis of these results, we calculated the population-attributable risk. Subsequently, we determined the absolute 10-year risk of symptomatic gallstone disease as a function of D19H genotype stratified by sex, age, and body mass index (BMI). Finally, to explore the effect of D19H genotype on liver and pancreas damage, and on the putative effects on cholesterol excretion, we determined the association of the D19H genotype with plasma liver biochemistry and with plasma lipid and lipoprotein levels.

Abbreviations

BMI, body mass index; CCHS, Copenhagen City Heart Study; CGPS, Copenhagen General Population Study; CI, confidence interval; HDL, high density lipoprotein; HR, hazard ratio; LDL, low density lipoprotein.

Patients and Methods

Participants.

We included participants in two similar studies of the Danish general population: The Copenhagen City Heart Study (CCHS) and The Copenhagen General Population Study (CGPS). Combining these two studies yielded a total of 62,279 participants, of whom 3124 developed gallstone disease, 2084 underwent cholecystectomy, and 30 developed biliary cancer.

The studies were approved by institutional review boards and Danish ethical committees (KF100.2039/91, H-KF01-144/01), and were conducted according to the Declaration of Helsinki. Written informed consent was obtained from all participants. There was no overlap of individuals between the two studies. All participants were white and of Danish descent, and were randomly selected to reflect the adult Danish population aged 20 to ≥80 years.

The CCHS12-14 is a study of the Danish general population initiated in 1976-1978 with follow-up examinations in 1981-1984, 1991-1994, and 2001-2003. In the present analysis, we included 10,388 participants who gave blood for DNA analysis at the 1991-1994 and/or 2001-2003 examinations. Of these, 633 developed symptomatic gallstone disease, 380 underwent cholecystectomy, and 24 developed biliary cancer during follow-up.

The CGPS12, 14 is a study of the Danish general population initiated in 2003 and still recruiting. We included the first 51,891 participants from this study in the present analysis. Of these, 2491 developed symptomatic gallstone disease, 1704 underwent cholecystectomy, and six developed biliary cancer during follow-up.

Follow-up time for symptomatic gallstone disease for each participant in either study began at the establishment of the national Danish Patient Registry (January 1, 1976) or on the participant's birthday, whichever came later. Follow-up time for biliary cancer began at the time of blood sampling for DNA (1991-1994 or 2001-2003 for CCHS; 2003 and onward for CGPS). For all endpoints, follow-up ended at the date of death, occurrence of event, emigration, or on May 17, 2009 (last update of the registry), whichever came first. Mean follow-up was 31 years for symptomatic gallstone disease and 4.4 years for biliary cancer, and was 100% complete, i.e., none were lost to follow-up.

Information on diagnoses of symptomatic gallstone disease (the World Health Organization International Classification of Diseases, 8th edition, codes 574 and 575, and International Classification of Diseases, 10th edition, codes K80 and K81) was collected from the national Danish Patient Registry and the national Danish Causes of Death Registry. The national Danish Patient Registry has information on all patient contacts with all clinical hospital departments and outpatient clinics in Denmark. The national Danish Causes of Death Registry contains data on the causes of all deaths in Denmark, as reported by hospitals and general practitioners. Information on diagnoses of biliary cancer (International Classification of Diseases, 10th edition, codes C22.1, C23, and C24) was collected from the Danish Cancer Registry,15 and from the national Danish Patient Registry.

Symptomatic gallstone disease was cholelithiasis or cholecystitis. Biliary cancer was cancer of the intrahepatic bile ducts, gallbladder, or extrahepatic bile ducts. Information on cholecystectomies was also collected from the national Danish Patient Registry on all participants. The validity of cholecystectomy data in the national Danish Patient Registry is 99.9%.16

Genotyping and Biochemical Analysis.

An ABI-Prism 7900HT Sequence Detection System (Applied Biosystems) and a TaqMan-based assay was used to genotype for ABCG8 D19H (rs11887534). Genotyping was verified by sequencing of 50 individuals with each genotype. Plasma biochemical tests indicating liver or pancreas damage (alanine aminotransferase; gamma glutamyltransferase; bilirubin; coagulation factors II, VII, and X; and pancreas amylase), and plasma levels of total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides were measured using standard hospital assays (Konelab and Boehringer Mannheim). LDL cholesterol was calculated with the use of the Friedewald equation if the triglyceride level was less than 4 mmol/L (354 mg/dL), and was measured directly for higher triglyceride levels.

Other Covariates.

BMI was expressed as weight in kilograms divided by height in meters squared. Physical activity in leisure time, use of hormone replacement therapy in women only, use of lipid-lowering drugs, and alcohol intake were dichotomized and defined as physical activity (4 hours or more per week of light physical activity in leisure time versus less than 4 hours), hormone replacement therapy (yes/no), alcohol consumption (consumption at least twice weekly versus less than twice weekly), and use of lipid-lowering drugs (yes/no).

Statistical Analysis.

Stata software, version 10.1 (Stata Corp., College Station, TX), was used for all analyses. Two-sided probability values less than 0.05 were considered significant. The Mann-Whitney U test or Kruskal Wallis analysis of variance was used for continuous variables, and Pearson's chi-square test was used for categorical variables. For trend tests, subjects were classified according to D19H genotype and coded as 0 for noncarriers, 1 for heterozygotes and 2 for homozygotes. Power calculations were performed using NCSS-PASS software.

Kaplan-Meier curves and log-rank tests evaluated the cumulative incidence of symptomatic gallstone disease and biliary cancer as a function of age and ABCG8 D19H genotype. We used Cox proportional hazards regression models adjusted for sex or multifactorially to analyze age-at-event, using left truncation (delayed entry) and age as a time scale. Thus, age was automatically adjusted for. Proportionality of hazards over time was assessed by plotting −ln(−ln[survival]) versus ln(analysis time). There was no suspicion of nonparallel lines. Risk of symptomatic gallstone disease was analyzed in the period 1976-2009. Because (1) both studies combined include the same registration period, (2) because genotype status does not change over time, and (3) because gallstone disease rarely leads to premature death, the use of a Cox regression model instead of a logistic regression model is justified.17 However, if we used logistic regression, the odds ratios for symptomatic gallstone disease were similar to the hazard ratios (HRs) reported. In addition, if we analyzed the two studies separately, the results were similar to those reported for the combined study. Because biliary cancer leads to rapid death, risk of biliary cancer was analyzed prospectively using Cox regression from the time of blood sampling for DNA in either study.

Interaction of D19H genotype with other risk factors for gallstones (age, sex, BMI, physical activity, parity, hormone replacement therapy, and alcohol intake) were evaluated by including two-factor interaction terms between genotype and other risk factors, one at a time, in the Cox regression model.

Population-attributable risk was calculated as previously described.18 Absolute 10-year risks for symptomatic gallstone disease by ABCG8 D19H genotype, age (<40 years, 40-60 years, >60 years), and BMI (<25 kg/m2 [normal], 25-30 kg/m2 [overweight], >30 kg/m2 [obese]) were estimated with the use of the regression coefficients from a Poisson regression model for women and men separately, and presented as estimated incidence rates (number of events per 10 years) in percent.19, 20

Results

Clinical Characteristics.

Persons with symptomatic gallstones were older and more likely to be female than those without gallstones (Table 1). Those with gallstones also had a higher BMI, were less physically active, had borne more children, more often used hormone replacement therapy, drank less alcohol, and were more often on lipid-lowering therapy than persons without a gallstone diagnosis. The crude incidence rates of symptomatic gallstones and biliary cancer were, respectively, 16 (20 in the CCHS and 15 in the CGPS) and 1 (1.9 in the CCHS and 0.4 in the CGPS) per 10,000 person-years.

Table 1. Characteristics of Individuals in the General Population* by Disease Status
CharacteristicNo eventGallstone DiseaseBiliary Cancer
  • Values are median (interquartile range) or percentage. Physical activity was 4 hours or more per week of light physical activity in leisure time. Alcohol consumption was at least twice weekly. P values as calculated by Mann-Whitney U test or Pearson's chi-square test.

  • *

    The Copenhagen City Heart Study and The Copenhagen General Population Study combined.

  • In women only.

  • P < 0.001 versus participants with no event.

  • §

    P < 0.05 versus participants with no event.

Number of individuals (%)59,155 (95)3124 (5)30 (0.05)
Age (years)57 (46-66)60 (50-70)66 (61-72)
Women (%)547257
BMI (kg/m2)25 (23-28)27 (24-30)27 (25-29)§
Physical activity (%)463627§
Mean parity (number of births)1.82.01.4
Hormone replacement therapy (%)132018
Alcohol consumption (%)494050
Lipid-lowering therapy (%)8110

Genotyping identified 54,526 (87.5%) noncarriers, 7506 (12.1%) heterozygotes, and 247 (0.4%) homozygotes (Table 2). Genotype frequencies did not deviate from those predicted by the Hardy-Weinberg equilibrium (P = 0.50). Risk factors for symptomatic gallstone disease did not differ by D19H genotype, and were thus unlikely to confound any association of D19H with risk of symptomatic gallstone disease (compare Table 1 with Table 2).

Table 2. Characteristics of Individuals in the General Population* as a Function of ABCG8 D19H Genotype
CharacteristicDD NoncarriersDH HeterozygotesHH HomozygotesP Value
  • Values are median (interquartile range) or percentage. Physical activity was 4 hours or more per week of light physical activity in leisure time. Alcohol consumption was at least twice weekly. P values as calculated by Kruskal-Wallis analysis of variance or Pearson's chi-square test.

  • *

    The Copenhagen City Heart Study and The Copenhagen General Population Study combined.

  • In women only.

Number of individuals (%)54,526 (87.5)7,506 (12.1)247 (0.4)NA
Age (years)57 (46-67)57 (47-67)57 (45-68)0.57
Women (%)5555530.36
BMI (kg/m2)26 (23-29)26 (23-28)26 (23-28)0.83
Physical activity (%)4645430.37
Mean parity (number of births)1.81.81.70.54
Hormone replacement therapy (%)1413110.71
Alcohol consumption (%)4849480.72

Risk of Gallstone Disease.

The cumulative incidence of symptomatic gallstone disease as a function of age increased stepwise by D19H genotype, with HH homozygotes having the highest risk and DH heterozygotes an intermediate risk compared with noncarriers (Fig. 1; P for trend <0.001). At the age of 60 years, 15% of HH homozygotes and 9% of DH heterozygotes had symptomatic gallstone disease compared with 4% of noncarriers. Mean age at onset of symptomatic gallstone disease was 56 years for noncarriers, 54 years for DH heterozygotes, and 52 years for HH homozygotes (Fig. 1; P for trend <0.001). HRs for symptomatic gallstone disease as a function of D19H genotype in the general population adjusted for age and sex were 1.9 (95% confidence interval [CI], 1.7-2.1) for DH heterozygotes and 3.3 (95% CI, 2.3-4.6) for HH homozygotes compared with noncarriers (Fig. 1). After adjustment for all other risk factors for gallstone disease in Table 1 (age, sex, BMI, physical activity, mean parity, hormone replacement therapy in women, and alcohol consumption), as well as for use of lipid-lowering therapy (mainly statins), results were similar. There were no significant interactions between D19H genotype and any of the abovementioned risk factors on risk of symptomatic gallstone disease.

Figure 1.

Cumulative incidence of symptomatic gallstone disease (graph at top) and mean age at onset (table at bottom) as a function of ABCG8 D19H genotype in the general population. DD, noncarriers; DH, heterozygotes; HH, homozygotes.

Among individuals with symptomatic gallstone disease, 2084 of 3124 (68%) underwent cholecystectomy. The HR for cholecystectomy as a function of D19H genotype in the general population adjusted for age and sex was 2.0 (95% CI, 1.8-2.3) for DH heterozygotes and 3.3 (95% CI, 2.2-4.9) for HH homozygotes compared with noncarriers, similar to HRs for gallstone disease. D19H genotype did not associate with overall mortality.

Based on the frequency of D19H and the observed HRs for symptomatic gallstone disease as a function of genotype, the fraction of all gallstones that could be attributed to D19H was 11%.

Absolute 10-year risk of symptomatic gallstone disease increased from men to women, with increasing age, increasing BMI, and by D19H genotype (Fig. 2). The highest absolute 10-year risk of gallstone disease was 13% in HH homozygous women above 60 years of age who were obese (BMI > 30 kg/m2) versus 4% in noncarriers in the same risk stratum. The equivalent risks in men were 9% and 2%.

Figure 2.

Absolute 10-year risk of developing gallstone disease as a function of ABCG8 D19H genotype, by sex, age (<40 years, 40-60 years, >60 years), and body mass index (<25 kg/m2, normal weight; 25-30 kg/m2, overweight; >30 kg/m2, obese).

Risk of Biliary Cancer.

The cumulative incidence of biliary cancer as a function of age increased by D19H genotype (Fig. 3; P< 0.001). The HR for biliary cancer as a function of D19H genotype adjusted for age and sex was 4.0 (95% CI, 1.9-8.4) for DH heterozygotes and HH homozygotes combined compared with noncarriers (Fig. 3) (3.8 [95% CI, 1.8-8.1] in DH heterozygotes and 10.3 [95% CI, 1.4-77.5] in HH homozygotes). Results were similar after multifactorial adjustment. Given the observed HR of 4.0, and assuming a power of 90%, the chance of a type I error (false positive result) was 4%. The fraction of all biliary cancers attributed to D19H was 27% (for clinical characteristics of patients, see Table 3).

Figure 3.

Cumulative incidence of biliary cancer in the general population as a function of ABCG8 D19H genotype. DD, noncarriers; DH+HH, heterozygotes and homozygotes combined.

Table 3. Characteristics of the 30 Participants With Biliary Cancer
Subject NumberD19H GenotypeSex (F/M)BMI (kg/m2)Age at Baseline ExaminationAge at Diagnosis of Gallstone DiseaseAge at CholecystectomyAge at Diagnosis of Biliary CancerICD-10 CodeAge at Death
  1. BMI is in kg/m2. All ages are presented in years. Empty cells for age at gallstone diagnosis, age at cholecystectomy, or age at death indicate that no event took place. F, female; DD, noncarriers; DH, heterozygotes; HH, homozygotes; ICD-10, International Classification of Diseases, 10th edition; M, male.

1HHF286277 77C24.977
2DHF2452 5757C24.058
3DHF246259 74C24.075
4DHF3367  71C24.971
5DHM3168  81C23.981
6DHF2771828283C24.883
7DHM2972  79C24.080
8DHF2973767677C24.177
9DHF3078797979C24.981
10DHF2285  87C23.987
11DHM2887  91C23.991
12DDM2745  48C24.051
13DDF3845  61C24.0 
14DDM284648 58C24.058
15DDF3150 5151C24.051
16DDM2550  66C24.9 
17DDM255469 70C23.971
18DDM2761  68C24.068
19DDF3061 6767C23.968
20DDF2062  75C24.977
21DDM2663  79C24.9 
22DDF2563  66C23.966
23DDF2664 7272C24.072
24DDF2667  77C23.978
25DDF2669  80C24.082
26DDM2872  74C22.174
27DDF3872  78C24.979
28DDM2978 8080C24.180
29DDM2481  82C24.186
30DDM2882  89C23.990

Plasma Liver Biochemistry.

D19H associated with stepwise increases in plasma levels of gamma glutamyltransferase and alanine aminotransferase of, respectively, 3% and 2% for heterozygotes, and 25% and 14% for homozygotes compared with noncarriers (Fig. 4; P for trend <0.001). Alcohol intake was equally distributed among the different D19H genotypes and therefore did not confound these associations (Table 2). Associations were similar after exclusion of all patients with symptomatic gallstone disease.

Figure 4.

Baseline liver parameters, plasma lipid and lipoprotein levels, and use of lipid-lowering therapy as a function of ABCG8 D19H genotype. ALAT, alanine aminotransferase; GGT, gamma glutamyltransferase; LDL-C, LDL cholesterol; HDL-C, HDL cholesterol. P values were determined by nonparametric trend test. For factor II-VII-X, data were available on 56,284 participants. For pancreas amylase, data were available on 45,203 participants. To convert cholesterol in millimoles per liter to milligrams per deciliter, multiply by 38.6; to convert triglycerides in millimoles per liter to milligram per deciliter, multiply by 88. DD, noncarriers; DH, heterozygotes; HH, homozygotes.

Plasma Lipids and Lipoproteins.

D19H genotype associated with modest stepwise reductions in plasma levels of total cholesterol and LDL cholesterol of 1.6% and 2.4%, respectively, in DH heterozygotes, and of 3.5% and 4.5%, respectively, in homozygotes compared with noncarriers (Fig. 4; P for trend <0.001). Associations were similar after exclusion of all patients with symptomatic gallstone disease. A total of 5117 persons were on lipid-lowering therapy, mainly statins. Genotype associated with a stepwise reduction in use of lipid-lowering therapy, from 8.5% in noncarriers to 7.1% in homozygotes (Fig. 4; P for trend = 0.002). There was no association of D19H with HDL cholesterol or triglycerides.

Discussion

The principal findings of this study of 62,279 individuals from the general population, of which 3124 developed symptomatic gallstone disease and 30 developed biliary cancer, are: (1) ABCG8 D19H genotype associated with an approximate 2-fold to 3.5-fold increase in risk of symptomatic gallstone disease in more than 12% of the general population. (2) D19H genotype associated with stepwise reductions in mean age at onset of symptomatic gallstone disease of up to 4 years in homozygotes. (3) A total of 11% of symptomatic gallstones in the general population could be attributed to D19H genotype. (4) Heterozygous and homozygous carriers for D19H had a four-fold increased risk of biliary cancer. (5) A total of 27% of biliary cancers could be attributed to D19H. (6) D19H genotype associated with stepwise increases in plasma alanine aminotransferase and gamma glutamyltransferase, and with a modest stepwise decrease in total and LDL cholesterol levels. These observations in the general population are novel and clinically applicable.

The incidence of symptomatic gallstone disease and biliary cancer, and the frequency of D19H, in our study are comparable to previous studies in Western countries.6, 21, 22 The incidence rates of mainly biliary cancer were lower in the CGPS compared to the CCHS. This might reflect the fact that participants in the CGPS were enrolled relatively recently (from 2003 and onward). Morbidity associated with occult biliary cancer might thus have precluded those participants from showing up at the baseline examination in the CGPS. Risk estimates for symptomatic gallstone disease in the general population for heterozygotes are consistent with those reported in a genomewide association scan and in subsequent case-control studies, whereas the risk estimate is somewhat more conservative in homozygotes than previously reported.6, 23-25 In contrast, our risk estimate for incident biliary cancer is markedly higher than the odds ratio of 1.8 (95% CI 1.1-2.8) reported by Srivastava et al.26 in a recent case-control study. Possible explanations for this include a higher background risk of biliary cancer in India than in Western countries,9, 21 a different etiology (bacterial infection of bile), and a frequency of D19H among controls of 23%, which is twice that reported in Western populations.

We found that D19H associates with reductions in plasma total and LDL cholesterol levels. In hepatocytes, ABCG8 forms half of a transmembrane sterol transporter that effluxes sterols into bile.27 Because biliary supersaturation with cholesterol is a pivotal event in cholesterol gallstone formation and D19H is associated with gallstones, it has been hypothesized that D19H constitutes a gain-of-function variant that increases biliary cholesterol efflux.6-8 Overexpression of ABCG8 in mice leads to reduced levels of plasma sterols.28 By analogy, a stepwise reduction in levels of plasma cholesterol as a function of D19H genotype in humans would support the gain-of-function hypothesis. Indeed, many,7, 8, 29, 30 but not all,31-35 previous studies did report reduced plasma sterols. We detected a modest but statistically highly significant stepwise reduction in plasma LDL cholesterol levels as a function of D19H genotype. These results therefore support the gain-of-function hypothesis.

Use of lipid-lowering therapy has recently been associated with a reduction in risk of gallstone disease.36, 37 In the present study, use of lipid-lowering agents (mainly statins) was more frequent among subjects with gallstones than among subjects without gallstones and therefore did not suggest a protective effect of these agents that could have lowered the incidence of gallstone disease. We found a modest stepwise reduction in use of lipid-lowering therapy as a function of D19H genotype. Regardless of whether we restricted the follow-up period to the period 1976-1990 (i.e., prior to the advent of statins) or excluded all participants on lipid-lowering therapy during the study, the hazard ratios for symptomatic gallstone disease as a function of D19H genotype were largely unchanged. Together, these results indicate that the increased risk of symptomatic gallstone disease observed in carriers of D19H could not be explained by their lower use of lipid-lowering therapy.

Although functional studies aimed at testing whether D19H constitutes a gain-of-function variant that increases biliary cholesterol efflux are clearly warranted, a clinically interesting question is whether the putative increase in biliary cholesterol efflux can be counteracted medically. An affirmative answer to this question could resurrect the widely abandoned concept of nonsurgical treatment for gallstones, at least in the subset of D19H-positive patients. Candidate drugs could include statins and ezetimibe, both of which have been associated with reduced biliary concentrations of cholesterol and with reduced risk of gallstone formation.36, 38

A limitation of the present study is that we studied only white individuals and thus results may not apply to other ethnic groups. We hypothesized that the risk of misclassification of symptomatic gallstone disease would be very limited. Hazard ratios for symptomatic gallstone disease and for cholecystectomy were similar, suggesting that misclassification was not a major problem.

In conclusion, ABCG8 D19H is the first genetic variant to predict risk of symptomatic gallstone disease and biliary cancer in the general population. The clinical impact appears to be substantial in that 11% of all symptomatic gallstones and 27% of all biliary cancers in our study can be attributed to D19H genotype. The identification of genetic variants predisposing to gallstones and biliary cancer might lead to design of new nonsurgical interventions that extend our currently limited strategies for the prevention of these diseases.

Acknowledgements

We thank senior technician Mette Refstrup, Department of Clinical Biochemistry, Rigshospitalet, Copenhagen University Hospital, for her persistent attention to the details of the large-scale genotyping. Mette Refstrup did not receive any compensation for her contribution. We are indebted to the staff and participants of the Copenhagen City Heart Study and the Copenhagen General Population Study for their important contributions to our study.

Ancillary