HIV, hepatitis B and hepatitis C coinfection in Kenya

There are few data regarding hepatitis and HIV coinfection in Africa. In 378 HIV seropositive individuals in Nairobi, 23 (6%) were hepatitis B virus (HBV) and HIV coinfected, four (1%) were hepatitis C virus (HCV) and HIV coinfected and one patient was infected with all three viruses. Coinfected individuals were more likely to be men and older; a lack of HBV vaccination was a risk factor for HIV/HBV coinfection (P = 0.001) and tenofovir containing regimens appeared most effective at reducing HBV viral load.

Although rates of coinfection with at least two of hepatitis B virus (HBV), hepatitis C virus (HCV) and HIV are well reported in numerous studies from Europe and America, there are few data on the prevalence of coinfection in African populations ^[1-7]. The 2003 Kenyan Demographic and Health Survey reported that 6.7% of Kenyan adults are infected with HIV ^[8] and other small studies have suggested that the seroprevalence of HBV or HCV is approximately 5% ^[9,10]. As both HBV and HCV are transmitted through routes similar to HIV, coinfection would be expected to be common.

To investigate this further, we performed a prospective study of patients attending the Aga Khan University Hospital (AKUH), Nairobi, Kenya. All HIV-positive patients over the age of 13 years (both from the in-patient and out-patient settings) were included, after written informed consent was obtained. Statistical analysis was undertaken using SPSS, version 8.0. Ethical approval was granted by ethical bodies at both the AKUH and collaborators in London.

Two millilitres of blood was obtained from each subject by venepuncture for serological HCV and HBV tests. The samples were collected and processed using standard methods, stored at -20°C and analysed in batches. Routine blood tests such as CD4 (Beckman Coulter Flow cytometry), viral load (Nucleisense technique), and liver function tests (aspartate and alanine aminotransferase) were also performed simultaneously. HepBsAg was assayed using enzyme-linked immunosorbent assay (ELISA) kits (Enzygnost) and individuals found to be HBsAg positive also had an HBeAg test. HCV antibodies were detected using fourth-generation ELISA kits (DRG International, Inc. Mountainside, New Jersey, USA/Biokit Labsystems, Helsinki, Finland).

As prospective data were not routinely collected, a study questionnaire was completed with the assistance of one of the investigators, with simple questions regarding age, history of intravenous drug use, HBV vaccination and sexual orientation.

A total of 378 consecutive HIV-positive individuals were recruited and, of these, there were 209 men (55%) and 169 women (45%) with a median age of 39.5 years (range 13-65). A total of 351 (92.8%) had HIV infection alone, 23 (6.1%) were infected with both HIV and HBV and a further four (1.1%) had HIV and HCV coinfection. One individual (0.3%) was infected with all three viruses. Of the HIV and HBV coinfected patients, four (17%) were HBeAg positive.

We compared the characteristics of the HIV positive individuals with those who were HIV and HBV coinfected (Table 1). There were no differences with regards to CD4 count, viral load and liver function tests (AST and ALT levels) but coinfected patients were older (P = 0.05) and more likely to be male (P = 0.02). Of interest, however 57% of individuals had CD4 counts below 200 cells/μl, a fact that may limit the validity of antibody responses measured and increase false negatives.

Table 1

In the questionnaire study, all individuals were heterosexual apart from a single homosexual male (who had HIV infection alone). No individual admitted to intravenous drug use. As may be expected, previous hepatitis B vaccination appeared in this cohort to protect against HBV infection. No patient with previous hepatitis B vaccination developed subsequent hepatitis B infection whereas 23 of 309 (7.4%) who had not been vaccinated were coinfected with HIV and HBV (P = 0.001).

This study also highlighted that HCV infection was uncommon in this high-risk population, which was generally representative of the general population (with a slightly increased number of men which is observed in urban settings such as this). HIV viral load had no effect on the presence of both HBV/HIV and HIV/HCV co infections.

Of those who were infected with both HIV and HBV, nine received zidovudine, lamivudine and efavirenz, six received tenofovir and emtricitabine and one received zidovudine, lamivudine and nevirapine. The group that received tenofovir had a greater reduction in HBV DNA viral load than the group that received lamivudine (P = 0.0031). The effects of treatment on a more widespread scale, and the effects of hepatitis on HIV and HIV on hepatitis in this setting remain to be established.

Acknowledgements

All authors participated in the design of the study, data collection, analysis and interpretation and all authors helped write and approved the final manuscript.