Abstract
Background
The association between hepatitis C virus (HCV) infection and the occurrence of type II diabetes remains controversial. Prospective studies are needed to assess its causal temporality.
Methods
A cohort of 21 559 adults enrolled from seven townships in Taiwan during 1991–1992 and followed till the end of 2010. Incident diabetes over a study time period from 2000 to 2010 was ascertained through computerized linkage with the National Health Insurance database and the National Death Certification profiles. Cox's proportional hazards models were used to estimate hazard ratios (HRs) with 95% confidence intervals (CIs). Antibodies against HCV (anti-HCV) were tested for all participants, and serum HCV RNA levels were measured for anti-HCV seropositives.
Results
During 180 244 person-years of follow-up, there were 1917 incident diabetes cases recorded. The cumulative risk for diabetes was 10.9% for anti-HCV seronegatives and 16.7% for anti-HCV seropositives respectively. The HR for diabetes of anti-HCV seropositivity was 1.53 (95% CI: 1.29–1.81) compared with anti-HCV seronegatives after adjustment for risk predictors. The adjusted HRs were 1.63 (1.31–2.02) for anti-HCV seropositives with positive HCV RNA compared to anti-HCV seronegatives (P<.001).
Conclusion
Chronic HCV infection was associated with an increased risk for diabetes after adjustment for other risk predictors.
Discussion Only
Medscape -
Full Text Article
Liver International
In this study, our data suggested that HCV infection brought significant increasing risk for the occurrence of diabetes after controlling several risk factors. Whereas the findings from NHANES were inconsistent, and their most recent report did not demonstrate an association of HCV infection with diabetes or with insulin resistance. NHANES is a large survey that consists of interviews, health examinations and laboratory data collected from a stratified, multistage probability cluster sample to obtain a nationally representative sample of the noninstitutionalized civilian US population. The first investigation using the NHANES data that evaluated the relationship between HCV infection and diabetes used the survey collected in the cycles of 1988–1994.[12] The investigators found that among individuals who were 40 years of age or older, those with HCV infection had an increased likelihood of having diabetes with the odds ratio of 3.8 (95% CI: 1.8–7.8) after controlling risk factors.[12] However, during the later study cycles (1999–2004 and 2005–2008), a statistically significant association was not found.[10] In the NHANES survey, the prevalence of obesity, insulin resistance and diabetes increased steadily over the same time period.[25] Thus, it is possible that estimates of the relationship between HCV infection and diabetes became attenuated in the presence of this epidemic of obesity and increased numbers of metabolic factors in recent decades. More recently, investigators using NHANES data have found that chronic hepatitis C patients, defined as anti-HCV seropositives with positive HCV RNA levels, had a two-fold risk for insulin resistance or diabetes,[9] suggesting that serum HCV RNA testing which enables the differentiation of past HCV infection with spontaneous clearance of HCV RNA (undetectable HCV RNA) and chronic HCV infection (detectable HCV RNA) might be important.
Finally, the most recent report from NHANES, using data from the 1999 through 2010 surveys failed to find an association between HCV infection and diabetes.[11] However, they did find a positive relationship between insulin resistance and elevated liver enzymes.[11] The elevated serum liver enzyme levels may be as result of either HCV infection or other conditions such as metabolic syndrome or steatohepatitis. It is difficult to disentangle the association of these factors and diabetes by cross-sectional study design in which participants were tested for their serum levels of liver enzymes and diabetes at the same time. Compared with the participants in NHANES, the participants in our study had lower educational levels, lower BMI and a lower percentage of cigarette smoking and alcohol drinking habits than the participants in NHANES. In other words, our study population seemed to have relatively fewer metabolic risk factors for diabetes. The major risk factor for HCV infection was illicit drug use in NHANES, whereas iatrogenic factors were the main transmission route for HCV infection in our population.[17,26] Thus, differences in the findings from our study from those in the NHANES study might be caused by differences in the characteristics of the study populations as well as in the study design.
Comparing to anti-HCV seronegatives, the adjusted HR was 1.39 for anti-HCV seropositives with seronegative HCV RNA, and 1.63 for anti-HCV seropositives with seropositive HCV RNA respectively. It showed a trend effect on the risk of diabetes for patients who had replication of HCV. Whereas, among anti-HCV seropositives, the risk of diabetes was only increased slightly when comparing to HCV RNA seronegatives and RNA seropositives, suggesting that the lipid profiles may be altered once individuals infected by HCV. However, the mechanisms still need to be further evaluated. Our study suggested that HCV infection has a major public health impact, not only for hepatic diseases but also for extrahepatic diseases.
Our cohort is recognized as a natural history cohort because most of the participants did not have the experience of antiviral treatment as a result of its high cost and adverse effects. Until October 2003, only patients with abnormal ALT levels (>82 U/L) and moderate fibrosis proven by liver biopsy could be reimbursed for treatment by the National Health Insurance. However, there are still huge gaps of self-awareness, referral and linking to care.[27] Although it is an important clue to compare the patients with or without antiviral treatment for further clarification of the associations between HCV infection and diabetes in the future.
In addition, we identified diabetes cases through a computerized data linkage with the National Health Insurance database, which provided coverage for 96% of the total population for Taiwan (23 million) in 2000, 98% in 2005 and 99.6% in 2009.[22] A limitation of our study is that we enrolled the participants and collected the questionnaires and blood samples during 1991–1992, whereas the index date for the linkage with the National Health Insurance database for the study described in this study was January 1, 2000 when the claims data were released. Life style and biochemical markers in the REVEAL-HCV enrolees might have changed over the period 1991–1992 to 2000. In particular, there might have been new infection with HCV during the 10 years of follow-up. However, establishing the incidence of HCV is very difficult because most infections are initially asymptomatic. In Taiwan the incidence of transfusion-associated hepatitis decreased as a result of the effectiveness of a series of donor screening intervention.[28] Also the presence of such cases would result in an underestimate of the effects of HCV on the incidence of diabetes. In this study, we defined the patients with diabetes with the stringent criteria: patients with at least one hospital admission code with diabetes diagnosis or with three or more outpatient visits code for diabetes,[21,22] which made the findings conservatively.
During the long-term follow-up, BMI may change when individuals change their life style, it is important to consider the follow-up changes on anthropometrical parameters. In our study, the anti-HCV seropositives were asked and invited for regular health examinations every 6–12 months. We obtained BMI data during their follow-up, and it showed similar results in the four models of multivariate analyses. The stratification analysis showed that chronic HCV infection may increase the risk for diabetes by comparing anti-HCV seronegatives. The changes in BMI during follow-up of each person were highly correlated (r=.84, P<.0001), and the correlation was even higher after grouping by <23 and ≥23 (r=.98, P<.0001). Another limitation of our study is that we did not measure blood glucose in our study. Thus, data of blood glucose are not available. However, serum level of triglycerides is highly correlated with blood glucose. We included triglyceride levels as an alternative marker for glucose levels in the analyses.
In a previous study, anti-HCV seropositives had increased mortality from diabetes (106.1 per 100 000 person-years) than anti-HCV seronegatives (60.8 per 100 000 person-years).[5] These results are consistent with the results from a database study of deaths in patients with diagnosed chronic hepatitis C infection which showed a 1.77 times increased risk of death from diabetes compared to the general population in the USA.[29] Although the evidence from these two studies of mortality provided an indication of a relationship between HCV and diabetes, it could not determine whether the infection increased the risk for the incidence of diabetes or worsened the prognosis for those with diabetes. The current longitudinal study has estimated the incidence of diabetes in those with and without anti-HCV seropositivity. Although data on steatosis in our study were lacking, we stratified by surrogate measures for steatosis including BMI, ALT and triglycerides.
In conclusion, our data suggested that HCV infection has increased risk for the occurrence of diabetes after controlling other risk factors. Chronic hepatitis C patients may benefit from antiviral treatment to decrease their risks for diabetes. More affordable prices of the effective drugs are required to increase the accessibility for the patients in need.
