Serum Bilirubin levels in the U.S. Population: Gender Effect and Inverse Correlation With Colorectal Cancer Stephen D. Zucker,' Paul S. H0rn,~93and Kenneth E. Sherman' Bilirubin, the primary end product of heme catabolism, is a key marker of liver and hematological disorders, and important cytoprotective properties have been ascribed to this bile pigment. The Third National Health and Nutrition Examination Survey, a comprehensive assessment of health and nutrition in the United States, was analyzed to determine the demographics and correlates of serum bilirubin levels in the general population. Men and women aged 17 and older were included in the weighted analysis, representing a total of 176,748,462 subjects. The mean serum total bilirubin in the adult population is 0.62 f 0.003 mg/dL (SEM),with a 97.5% cut-off of 1.4 mg/dL. Serum bilirubin levels are significantly higher in men (0.72 2 0.004) than in women (0.52 f 0.003 mg/dL) and are lower in non-Hispanic blacks (0.55 f 0.005 mg/dL) compared with nowHispanic whites (0.63 f 0.004 mg/dL) and Mexican Americans (0.61 f 0.005 mg/dL). Bilirubin concentrations are unrelated to body weight but are reduced in active smokers. Individuals with a history of nondermatological malignancy exhibit significantly lower serum bilirubin concentrations compared with those who do not have a history of nondermatological cancer. In particular, each 1-mg/dL increase in serum bilirubin is associated with a markedly decreased prevalence of colorectal cancer (OR = 0.257; 95% CI 0.254-0.260). In conclusion, serum bilirubin levels vary significantly with gender, race, and smoking status. The observed inverse correlation between serum bilirubin concentrations and a history of nondermatological malignancy, particularly colorectal cancer, warrants further investigation of a potentially 2004;40:827- 835.) important chemopreventive function of bilirubin. (HEPATOLOGY
B
ilirubin, the principal degradation product of heme catabolism, is produced by heme oxygenase-mediated oxidation of heme to form biliverdin, followed by biliverdin reductase catalyzed reduction to bilirubin (Fig. 1). Bilirubin is cleared from the circulation by the liver, where it is conjugated by the bilirubin. specific 1Al isoform of the microsomal enzyme uridinediphosphoglucuronate (UDP)-glucuronosyltransferase to
form water-soluble monoglucuronide and diglucuronide metabolites that are actively secreted into bile. Unconjugated hyperbilirubinemia develops when heme breakdown is accelerated (e.g., hemolysis) and/or hepatic UDP-glucuronosyltransferase activity is diminished (e.g., Gilbert and CriglerNajjar syndromes), while an increase in serum levels of conjugated bilirubin occurs when secretion of bilirubin glucuronides by the liver is impaired (e.g., Dubin-Johnson syndrome) or obstruction to bile flow occurs (k, cholestasis). is considered Although conjugated hyperbilirubinemia Abbreviations: UDP, uridine-diphosphoglucuronate; NHANES HI, Third National Health and Nutrition Examination Survq; GI,gastrointestinal; BMI, body innocuous, markedly elevated serum concentrations of mass index; OR, oddc ratio. unconjugated bilirubin can cause neurological injury in From the 'Division of Digestive Diseases and the 'Department ofMathematical newborns' and pigment gallstones in adults.2 O n the Sciences, University of Cincinnati, Cincinnati, OH; and the 3 Veterans Affairs other hand, more modest levels of unconjugated hyperbiMedical Center, Cincinnati, OH. Received April 6, 2004; acceptedJune 2, 2004. lirubinemia appear to be harmless. An example is Gilbert Supported by National Institutes ofHealth researchgrants DK063954 (S. D. Z ) syndrome, an inherited deficiency in hepatic bilirubin and A149508 (K E. S.); an Ohio Division Supported Pilot Research Grantjbm and a CancerPrevention Research Grant conjugating activity caused by a polymorphism in the the American Cancer Society (S. D. 2.); from the Cancer Research and Prevention Foundation (S. D. 2.). No financial promoter TATA element of the gene encoding UDPdisciosures are necessaryfor any of the authors. glucuronosyltransferase.1 While the Gilbert polymorAddress reprint requests to: Stephen D. Zucker, M.D., Division of Digestive Diseases, Universityof Cincinnati,231 Bethesda Avenue, M.L. 0595, Cincinnati, phism is present in up to 43% of the general population,2 OH 45267-0595. E-mail:
[email protected];f m : 513-558-1744. phenotypic expression as overt hyperbilirubinemia is esCopyright Q 2004 by the American Association for the Study of Liver Diseases. timated to be ~ Y o - ~ O Ywith O , ~serum bilirubin concentraPublished online in WiIq InterScience (www.interscience.wilq.coinj. DO1 10.1002fhep.20407 tions ranging as high as 5 times that of normal. 827
828
ZUCKER, HORN, AND SHERMAN
NADPH
NADP’
NADPH
HEPATOLOGY, October 2004
NADP’
Bilirubin IXu I
HEPATOCYTE
BILE Fig. 1. Metabolic pathway for bilirubin production and elimination. Bilirubin IXa, the main physiologic isomer of unconjugated bilirubin, is produced by the sequential oxidation of heme through a biliverdin lXa intermediate. Bilirubin is taken up by the hepatocyte, where it is convetted to the water-soluble bilirubin diglucuronide and secreted into bile. NADPH, the reduced form of nicotinamide-adenine dinucleotide; NADP, nicotinamide-adenine dinucleotide phosphate; HO, heme oxygenase; BVR, biliverdin reductase; UGTlA1, bilirubin-specific uridine-diphosphoglucuronate-glucuronosyltransferase isoform; MRP2, multidrug resistance protein-2.
It is notable that, in some nonmammalian vertebrates (cg., birds, reptiles, and amphibians), heme degradation terminates at the heme oxygenase step, and the resultant biliverdin is efficiently cleared and excreted by the liver and kidneys. The evolutionary development of the energetically costly process of bilirubin production and elimination suggests that this bile pigment may serve an important physiological function. Analogous to theories regarding hemochromatosis gene mutations,’ the high prevalence of the Gilbert polymorphism raises the possibility that modestly increased levels of circulating bilirubin are adaptive. Support for this hypothesis is derived from numerous epidemiological studies demonstrating an inverse correlation between serum bilirubin concentrations and the risk of cardiovascular disease,* and it has been suggested that the beneficial effects of this bile pigment are a consequence of its potent antioxidant properties.5.6 Indeed, recent data indicate that intracellular redox cycling of bilirubin to biliverdin may be important for cellular cytoprotection.7 To elucidate the potential physiological functions of bilirubin, a detailed understanding of the demographics and determinants of serum bilirubin levels is required. T o this end, we conducted an analysis of the Third National Health and Nutrition Examination Survey (NHANES III), a large clinical database weighted to reflect the general United States population.
Methods Survey Design. The NHANES 111survey is a periodic assessment of health and nutrition conducted by the Cen-
ters for Disease Control and Prevention between 1988 and 1994.8 This analysis of noninstitutionalized United States civilians consists of household interviews, physical examinations, and blood sampling from 40,000 persons at 89 locations. It employed a complex, stratified, multistage probability cluster design in which older individuals, children, Mexican Americans, and African Americans were specifically overrepresented to insure a prespecified minimum sample size. Sampling weights are used to adjust the cohort distribution to one similar to that of the United States population as a whole. The data used for the present analysis focused on the 20,216 subjects aged 17 years and older who were assessed at home and in a mobile examination center. Excluding those individuals without serum bilirubin data, a total of 16,865 subjects were available for analysis. Information regarding a prior history of malignancy was derived from patient-based questionnaire responses. Gastrointestinal (GI) malignancies were classified as including cancers of the esophagus, stomach, small and large intestine, liver, gallbladder, and pancreas. Non-GI malignancies comprised all other malignancies except for skin cancer. Individuals who reported a history of skin cancer only were included in the noncancer cohort. Laboratory Methods. Serum samples were collected in a standardized manner from fasting individuals as described in the White Sands Clinical Laboratory’s Collection Procedures and Specimen Requirements ManuaL8 Specimens were transported on dry ice and stored at -70°C. Serum total bilirubin concentrations were determined using 2,5-dichlorophenyldiazonium tetrafluoroborate in a strongly acidic medium. The intensity of the azobilirubin color was measured photometrically using a Hitachi 737 automated autoanalyzer (Boehringer Mannheim Diagnostics, Indianapolis, IN). This method has a minimum detection limit of 0.1 mg/dL and is linear to a maximum bilirubin concentration of 30 mgldL (Manual for Medical Technicians, U.S. Department of Health and Human Services, 1996). StatisticaZ Analyses. Algorithms developed for the NHANES I11 database provide weighted estimates of the entire United States population at the time the study was performed. For our analyses, sampling weights were determined from the WTPFHXG parameter, which is used for appropriate weighting of study participants who underwent mobile and home examination and answered medical history questions from household interviews.8 Weighted sample estimates incorporated selection probabilities and included adjustments for noncoverage and nonresponse. Sample means, standard errors, and frequency histograms, as well as all statistical models and tests, are based on these weighted observations. Statistical
HEPATOLOGY, Vol. 40, No. 4, 2004
Enlm Population
A 20
B
Men vs. Women
"1 'aili
2o
P
I0
Fig. 2. Frequency distribution histograms of serum total bilirubin in the adult United States population. (A) Serum total bilirubin concentrations for the entire adult United States population plotted over 0.1mg/dL intervals. (6) Serum bilirubin levels stratified by sex (grey bars, women; dark bars, men).
calculations were performed using the SAS software package (SAS Institute, Inc., Cary, NC) and account for sampling weights when determining variance estimates. Weighted Student's t test was used to establish mean differences. Weighted linear models were employed when bilirubin was the dependent variable, because values are continuous. Weighted logistic regression models were used when the independent variable was dichotomous.
Results The distribution of fasting serum bilirubin levels in the United States population was determined using weighted estimates (based on the WTPFHX6 parameter) representative of 176,748,462 individuals (Fig. 2A). It is notable that the use of alternative weighting parameters (e.g.,
ZUCKER, HORN, AND SHERMAN
829
WTPFEX6) had no significant effect on the results. As shown in Table 1, the mean serum total bilirubin concentration in United States adults is 0.62 ? 0.003 mg/dL (SEM; 95% CI 0.61-0.62 mg/dL), ranging from 0-7.9 mg/dL. The 2.5 and 97.5 percentile cut-offs for the weighted frequency estimate of total bilirubin (normal range) are 0.2 mg/dL and 1.4 mg/dL, respectively. When stratified by gender, the mean bilirubin concentration among women (0.52 ? 0.003 mg/dL; 95% CI 0.520.53) is significantly lower (P < .0001) than in men (0.72 I+_ 0.004 mg/dL; 95% CI 0.71-0.73), with normal values ranging from 0.2-1.2 mg/dL and 0.3-1.7 mg/dL, respectively (Fig. 2B). These weighted estimates are representative of a total of 91,935,790 women and 84,812,672 men. Although age is not an independent factor with regard to bilirubin levels across the entire United States population, the mean serum bilirubin concentration decreases by 0.0017 mg/dL per year of life in men (P< .OOOl). No association between age and serum bilirubin levels was identified in women. The effect of race-defined in the NHANES I11 survey by the DMARETHN variable as non-Hispanic white, non-Hispanic black, and Mexican American-on serum total bilirubin levels also was analyzed.*Among non-Hispanic whites, the mean serum bilirubin concentration is 0.63 2 0.004 mg/dL, with a normal range of 0.3 to 1.4 mg/dL (see Table 1). Non-Hispanic blacks exhibit a mean serum bilirubin of 0.55 I+_ 0.005 mg/dL (normal range 0.2-1.3 mg/dL), while in Mexican Americans the mean is 0.61 ? 0.005 mg/dL (normal range 0.2-1.6 mg/dL). As determined via painvise t test, mean serum bilirubin levels were the same in non-Hispanic whites and Mexican Americans, and both of these groups had higher levels than those found in non-Hispanic blacks. When substratified by gender, racial differences in the serum bilirubin concentration in men mirror the overall findings. In contrast, bilirubin levels in Mexican American women are significantly lower than in non-Hispanic whites but are significantly higher than in non-Hispanic blacks (Table 2). Prior studies have indicated that serum bilirubin levels inversely correlate with cigarette smoking' and obesity. l o Consistent with these data, our assessment of the NHANES I11 database reveals that bilirubin levels are significantly lower (P< .OOOl) in active smokers (0.60 ? 0.005 mg/dL; 95% CI = 0.60-0.61) versus nonsmokers (0.63 2 0.004 mg/dL; 95% C.I. = 0.63 - 0.64), a finding which holds for both men (0.68 f 0.005 vs. 0.78 2 0.007 mg/dL) and women (0.50 2 0.004 vs. 0.54 ? 0.004 mg/dL). O n the other hand, Pearson correlation failed to demonstrate a meaningful association between bilirubin and body mass index (BMI), either overall ( r =
830
ZUCKER, HORN, AND SHERMAN
HEPATOLOGY, October 2004
Table 1. Demographics of Serum Total Bilirubin Concentration in the United States Adult Population Welghted Mean (m/dL)
SE (mg/dL)
95% CI
Number of Subjects
Adult population
0.62
0.003
0.61-0.62
16,865
Sex Men Women
0.72 0.52
0.004 0.003
0.71-0.73 0.52-0.53
7,892 8,973
Ethnicity Non-Hispanic whites Non-Hispanic blacks Mexican Americans
0.63 0.55 0.61
0.004 0.005 0.005
0.62-0.64 0.54-0.56 0.60-0.62
6,973 4,568 4,646
Smoking status Nonsmokers Active smokers
0.63 0.60
0.004 0.005
0.63-0.64 0.60-0.61
8,652 8,213
Malignancy No prior history All non-dermatologic GI Colorectal Non-GI
0.62 0.58 0.52 0.52 0.58
0.003 0.014 0.017 0.018 0.016
0.61-0.62 0.55-0.60 0.49-0.56 0.48-0.55 0.55-0.62
16,233 630 106 83 524
P Value
