Abstract
In the era of highly active antiretroviral therapy (HAART), co-morbidities such as HIV/hepatitis B (HBV) coinfection have appeared. As the uncertainty in relation to the effect of coinfection on the natural history of HIV disease remains, a retrospective cohort study was conducted to investigate the potential impact of HBV infection on AIDS progression, all-cause mortality, on the rate of antiretroviral treatment changes after HAART initiation and on the response to HAART in an HIV infected population in Greece. Furthermore, we supplemented our results by pooling them with previous studies in a meta-analysis. HIV seropositive patients have been retrospectively tested for hepatitis B surface antigen (HBsAg) using stored sera. Data on AIDS diagnosis and death were obtained from the national HIV/AIDS registry. Multivariable analyses were performed using Poisson and logistic regression models. The influence of HBV coinfection on the number of treatment changes after starting HAART was ass ...
In the era of highly active antiretroviral therapy (HAART), co-morbidities such as HIV/hepatitis B (HBV) coinfection have appeared. As the uncertainty in relation to the effect of coinfection on the natural history of HIV disease remains, a retrospective cohort study was conducted to investigate the potential impact of HBV infection on AIDS progression, all-cause mortality, on the rate of antiretroviral treatment changes after HAART initiation and on the response to HAART in an HIV infected population in Greece. Furthermore, we supplemented our results by pooling them with previous studies in a meta-analysis. HIV seropositive patients have been retrospectively tested for hepatitis B surface antigen (HBsAg) using stored sera. Data on AIDS diagnosis and death were obtained from the national HIV/AIDS registry. Multivariable analyses were performed using Poisson and logistic regression models. The influence of HBV coinfection on the number of treatment changes after starting HAART was assessed using Poisson regression with random effects. The performance of Poisson modeling was compared with variations of Cox models proposed for the analysis of recurrent data. For meta-analysis purposes, a comprehensive search for articles was conducted and data from relevant reports were abstracted. The studies were evaluated for publication bias and heterogeneity. Combined estimates along with their 95% confidence intervals (CI) were calculated using fixed and random effects models. Data on 1,729 individuals were used. The prevalence of HBsAg was 6.19%. The multivariable analysis showed that progression to AIDS or death was not associated with the HIV/HBV coinfection. Furthermore, virological response to HAART was comparable in all groups irrespective of the HBsAg status. Multivariable analyses did not reveal a statistically significant effect of coinfection status on the frequency of treatment changes. However, the evidence against an association was slight and a mild effect could not be ruled out. Antiretroviral drug family, calendar period, prior exposure to antiretrovirals and AIDS were independently associated with the number of changes. Totally nine studies investigated the impact of concurrent HBV infection on AIDS incidence and they were included in a meta-analysis. There was no evidence of an association assuming either a random or a fixed effects model. Eleven studies evaluated the relationship between HBV infection and overall mortality in HIV seropositive patients. The analysis yielded a significant result with a pooled effect estimate of 1.36 (95% CI: 1.12-1.64). Publication bias or heterogeneity were absent in both meta-analyses. HBV infection had no detectable effect on AIDS incidence or death in a cohort of HIV positive individuals in Greece. However, adding our numbers to previously published similar studies in a meta-analysis for a total of 12,382 cases, we discovered a significant effect of HIV/HBV coinfection on overall mortality. The statistical power of meta-analysis revealed a modest hepatitis B effect on total death rate, probably through an increase in liver related mortality. HIV/HBV coinfection was not found to affect significantly the frequency of treatment changes after the beginning of HAART. However, a small effect leading to a slight increase in the rate of treatment changes could not be definitely excluded. It seems that due to the limited number of studies on this issue, the topic warrants further investigation.
show more