Epidemiology and Outcomes of Hepatitis C Infection in Elderly US Veterans

H. B. El-Serag; J. Kramer; Z. Duan; F. Kanwal


J Viral Hepat. 2016;23(9):687-696. 

In This Article


Data Sources

This study was approved by Baylor College of Medicine's Institutional Review Board, and all procedures conformed to the ethical guidelines of the 1975 Declaration of Helsinki. We used data from the VHA HCV Clinical Case Registry (CCR), which contains health information for all known HCV-infected patients from 128 VHA facilities nationwide. Data elements in the CCR include date of birth, laboratory test results, outpatient and inpatient VHA pharmacy data and inpatient and outpatient diagnoses and procedure codes.[7] We examined data from October 1 1996 through January 1 2010.

Study Population

The study cohort included patients with CHC, defined as a positive test for HCV RNA in plasma by qualitative or quantitative assays between 1 October 1999 and 31 December 2009. We included patients between 20 and 85 years of age. We excluded patients with less than 1 year of follow-up in the VHA to minimize bias related to incomplete ascertainment of patients' cirrhosis and HCC. We defined the date of earliest positive HCV RNA as the index date for this analysis.

Study Exposure and Outcomes

Age at HCV index date was examined as a continuous variable and subsequently categorized at 20–49, 50–64 and 65–85 years. The primary outcomes of the study were new (incident) cases of cirrhosis (ICD-9 codes 571.2, 571.5, 571.6) and/or HCC (ICD-9 code 155.1) that were first recorded after 1 year of the HCV index date. Patients with outcomes recorded in the first year of follow-up were excluded because they might have been existing (i.e. prevalent) cases. We also examined death as an outcome that was extracted from the VA Vital Status file. The ICD-9 code-based definitions for both cirrhosis and HCC were validated in our previous studies against detailed medical record reviews and shown to have a high positive predictive value.[8]

Potential Confounders

We ascertained several risk factors that may be associated with age at HCV index date and an accelerated or decreased progression to cirrhosis and HCC in patients with CHC: period of service (World War I/II, Vietnam era, post-Vietnam era), race (African American, non-Hispanic White, Hispanic, other or unknown), gender, diabetes, alcohol use, obesity, HIV infection, hepatitis B virus (HBV) infection, and receipt and success of antiviral treatment (i.e. HCV treatment and SVR). We identified HIV, diabetes and alcohol use by the presence of outpatient or inpatient ICD-9 diagnosis codes recorded during 1 year before or after the HCV index date. We defined patients with HBV coinfection as subjects with a positive HBV surface antigen test and classified a patient as obese if the BMI closest to the HCV index date was 30 kg/m2 or greater. We defined antiviral treatment as at least 1 filled prescription of interferon or pegylated interferon with or without ribavirin any time after HCV index date. We defined SVR as last RNA test being negative at least 12 weeks after treatment completion (SVR12), as previously described.[9] Patients with 'No SVR' were defined as those having their last RNA test positive any time after end of treatment. Any treatment occurring after a gap of > 65 days was ignored, and only, the first course with interferon was considered.

Data Analysis

We first examined demographic and clinical differences among the age groups (20–49, 50–64, 65–85 years), using chi-square test. We also conducted a logistic regression model to examine the effect of age on treatment receipt while adjusting for other age-related factors, such as comorbidities. To calculate the incidence rates of cirrhosis, we divided the number of incident cases by a denominator of patient years' (PY) follow-up after excluding patients who had prevalent diagnosis of cirrhosis. For the HCC incidence-rate calculation, we divided the number of incident cases by a denominator of PY follow-up that excluded patients who had a prevalent diagnosis of HCC. The study follow-up ended at the time of HCC, patient's death, last visit in the VHA or 1 January 2010. We calculated the incidence rates per 1000 PY of follow-up for newly diagnosed cirrhosis and HCC for each age group. We also conducted a secondary analysis on 65–74 and 75–85 age groups.

We used Kaplan–Meier method to estimate the survival probabilities of cirrhosis, HCC and death starting 12 months after the HCV index date till the end of the follow-up period. We used the log-rank test to evaluate the differences among these probabilities by age group. To illustrate change in the risk of events over time, cumulative hazard plots were generated using the negative logarithm of survival probabilities. Cumulative hazard is a measure of sustained risk of event over time, the larger the value, the higher the risk. We constructed Cox proportional hazards models to examine the association between age groups and time to cirrhosis, HCC or death, separately, while adjusting for potential confounders (period of service, race, gender, diabetes, alcohol use, obesity, HIV infection, HBV infection and antiviral treatment). In these models, we also examined the effect of treatment and SVR12 as time dependent variable on the study outcomes. To further understand the effect of treatment by age, we created an interaction term of treatment with and without an SVR12 with the 3 age groups and examined this term in adjusted Cox proportional hazards models. In a secondary analysis and to minimize selection bias by treatment, we examined the effect of the interaction of SVR12 compared with no SVR by age group on incident cirrhosis, HCC and mortality among only those who received treatment. The results of these regressions were expressed as hazard ratios (HRs) and corresponding 95% confidence intervals (CIs). The proportional hazard assumption was tested and fulfilled in all models.