Sources of heterogeneity in hepatitis B virus (HBV) seroprevalence estimates from U.S. adult incarcerated populations: A systematic review and meta-regression analysis.

HBV seroprevalence estimates from U.S. incarcerated populations are relatively high. However, the usefulness of these estimates for guiding HBV-related correctional healthcare policy is limited by wide variation in estimates across studies and little understanding of the sources of this variation. The authors systematically reviewed studies indexed from 1975-2005, meeting pre-specified criteria and reporting HBV seroprevalence estimates from U.S. adult incarcerated populations. Using meta-regression techniques, the authors investigated report type, geographical region, serum collection year, facility type, serum source, sampling procedures, sample characteristics, and measurement procedures as potential study-level sources of heterogeneity in prevalence estimates for common HBV seromarkers. In bivariable meta-regression analyses, mean age ≥31 years was strongly associated with increased HBsAg prevalence (POR=2.6), and serum collection year before 1991 was strongly associated with increased prevalence of any positive marker (POR=2.0). Other moderate-to-strong associations were observed, but these were considered less certain because of small numbers of observations, influence of single studies, or potential confounding. Potential sources of heterogeneity should be considered when comparing HBV seroprevalence estimates in adult U.S. incarcerated populations and when developing HBV screening and vaccination protocols in correctional settings.