Prevalence of Type 2 Diabetes Mellitus among Persons with Hepatitis C Virus Infection in the United States

Approximately 2.7 million persons in the United States have chronic hepatitis C virus (HCV) infection (1). Chronic HCV infection may lead to cirrhosis and hepatocellular carcinoma and is a leading cause of liver transplantation in the United States (2). Hepatitis C virus infection may also cause extrahepatic illnesses, including essential mixed cryoglobulinemia, sporadic porphyria cutanea tarda, and thyroid disease, all of which may reflect interactions between HCV and the host immune system (2-7). An increased prevalence of diabetes mellitus among persons with HCV infection has also been observed (8-15). The data linking HCV infection and diabetes mellitus are derived from several recent clinic-based, casecontrol studies that leave several important questions unanswered (8-15). Most of these reports did not consider such factors as body mass index, illicit drug use, and socioeconomic status, which have been associated with both conditions and thus could confound the relationship (1, 16). In addition, because the studies were based principally in referral centers, the relationship may be restricted to persons with severe forms of the diseases. For example, since the liver is crucial to carbohydrate metabolism and glucose homeostasis, diabetes may occur more often in anti-HCVpositive persons simply because of hepatocyte dysfunction (17). Discovery of an increased prevalence of diabetes in the general population among persons with HCV infection and less severe liver disease might suggest an alternate mechanism, such as an HCV-related autoimmune process. Similarly, it has not been determined whether HCV infection results in an increased occurrence of type 1 or type 2 diabetes. Answers to these questions could shed light on the biological mechanisms involved. Both HCV infection and diabetes have been carefully evaluated in a representative sample of the general population of the United States through the Third National Health and Nutrition Examination Survey (NHANES III) (18, 19). We sought to test the hypothesis that persons with HCV infection have an increased prevalence of type 2 diabetes after adjustment for important confounding variables, including age, body mass index, poverty level, and history of drug and alcohol use. Methods Survey Design and Study Sample The NHANES III was conducted from 1988 to 1994 by the National Center for Health Statistics of the Centers for Disease Control and Prevention and is described in detail elsewhere (18, 19). In brief, the survey used a stratified, multistage probability cluster sampling design to obtain a representative sample of the U.S. civilian, noninstitutionalized population. It was designed to oversample Mexican-Americans and African-Americans; in our analysis, we used sampling weights to account for this fact. Approximately 34 000 persons who were at least 2 months of age at the time of the evaluation were sampled at 89 randomly selected locations throughout the United States. Persons selected for evaluation were interviewed at their residence by using a questionnaire that collected information on demographic characteristics, medical history, current and past medication use, and other risk behaviors. Ninety-one percent (30 818) of participants also underwent physical examination and laboratory assessment at a mobile examination center. Plasma glucose levels were measured and HCV antibody testing was performed in examined persons who were at least 20 years of age or 6 years of age, respectively. The institutional review board at the Centers for Disease Control and Prevention approved the study, and all participants provided written informed consent (18, 19). Because plasma glucose testing was performed only in persons older than 20 years of age, we restricted our analysis to persons 20 years of age or older at the time of examination. Of the 18 825 persons older than 20 years of age who were interviewed, 16 573 (88%) also had a complete physical examination and laboratory analysis and were thus deemed eligible for our analysis. Persons were included in our investigation if they had complete evaluations for diabetes and HCV infection. In particular, each household was randomly assigned to a morning, afternoon, or evening evaluation, and participants were instructed to abstain from intake other than water for a specific period of time. Of the 8158 persons assigned to a morning session, 7439 (91%) completed an 8- to 24-hour fast, whereas 2467 of 8415 (29%) persons assigned to a later appointment fasted for 8 to 24 hours. In addition, 562 persons who did not fast but reported use of antidiabetic medication were included in the study sample. Of the 10 468 eligible persons, 627 were excluded from analysis because of indeterminate or missing plasma glucose levels (n =212), indeterminate or missing anti-HCV information (n =290), or a history of diabetes that was unsubstantiated by hyperglycemia or use of antidiabetic medications (n =125). The remaining 9841 persons constitute the study sample (Figure 1). Figure 1. Determination of the study sample. Ascertainment of Diabetes Venous whole blood was drawn into a vacuum tube containing the glycolytic inhibitors potassium oxalate and sodium fluoride and was immediately centrifuged at 1500 g for 10 minutes, as described elsewhere (19). Plasma was frozen at 70 C and shipped to the University of Missouri Diabetes Diagnostic Laboratory, Columbia, Missouri, where plasma glucose testing was performed by using a modified hexokinase enzymatic method. During the 6 years of the survey, the within-assay and between-assay coefficients of variation were 1.6% to 3.7% (20). Type 1 and type 2 diabetes were classified according to previously defined criteria, a combination of the 1997 American Diabetes Association criteria and that used by the Early Treatment Diabetic Retinopathy Study group (21, 22). Persons were considered to have diabetes if they used insulin or oral hypoglycemic agents at the time of the survey or had a fasting plasma glucose level of 7.0 mmol/L or more ( 126 mg/dL). Persons in whom diabetes was diagnosed before 30 years of age, started receiving insulin therapy within 1 year of diagnosis, and reported insulin use at the time of the survey were categorized as having type 1 diabetes. All others who met the above criteria for diabetes were classified as having type 2 diabetes. Exposure Assessment Presence of antibody to HCV (anti-HCV) was assessed by using a second-generation enzyme immunoassay test (Abbott Laboratories, Chicago, Illinois). Positive specimens were tested in duplicate, and repeatedly positive samples were tested again by using the MATRIX assay (Abbott Laboratories). Specimens that were positive according to all three tests were considered to be anti-HCV positive. A sandwich radioimmunoassay (Abbot Laboratories, North Chicago, Illinois) was used for semiquantitative determination of hepatitis B surface antigen in human serum. Serum blood chemistries, including hematologic variables, were obtained by using a Hitachi Model 737 multichannel analyzer (Boehringer Mannheim Diagnostics, Indianapolis, Indiana) (19). Serum liver enzyme levels, including alanine aminotransferase levels, could not be determined from frozen plasma. Information on other covariates was collected during the interview and subsequent examination. Age, ethnicity, and socioeconomic status were categorized according to the survey design as suggested by the National Center for Health Statistics for analysis of NHANES III data (19). Age was analyzed in 10-year groups, and ethnicity was divided into four categories: non-Hispanic white, non-Hispanic black, Mexican-American, and other, which included other Hispanics, Asians, and Native Americans. Too few persons and potential heterogeneity in the other category prohibited its inclusion in analysis. Educational attainment and poverty level were used as proxy measures of socioeconomic status. Educational attainment was classified according to whether a participant had achieved greater than a high school diploma. Poverty level was calculated as a poverty income ratio of self-reported family income to a denominator based on poverty threshold, family size, and the calendar year of the interview. Poverty threshold values, which were standardized for inflation, were based on tables published annually by the U.S. Census Bureau (19). Participants with a poverty income ratio less than 1.0 were considered to be below the poverty level. Body mass index, measured in kg/m2, was assessed during the examination. Participants with a body mass index less than 25 kg/m2, 25 to 29.9 kg/m2, 30 to 34.9 kg/m2, and 35 kg/m2 or more were classified according to the National Heart, Lung, and Blood Institute as lean or normal, overweight, obese, or morbidly obese, respectively (23). Participants who indicated that any of their first-degree relatives had diabetes were considered to have a positive family history of diabetes. Cigarette smoking was categorized according to whether the person was a never, former, or current smoker at the time of the interview. Illicit drug use was assessed by questions about lifetime use of marijuana or cocaine (including crack cocaine), but no specific questions were asked about injection drug use. Excessive alcohol intake was defined as alcohol consumption of more than 50 g/d (approximately five drinks) during the past year. Statistical Analysis General descriptive analysis was performed to compare participants with and those without diabetes. For categorical variables, two-way tabulations calculating a Pearson chi-square statistic, corrected for complex survey design or clustered data, were used. For continuous variables, survey designcorrected t-tests were performed. Univariate and multivariate survey logistic regression techniques were used to determine the crude and adjusted odds ratios of type 2 diabetes with respect to HCV infection. Variables considered to be potential confounders in multiva