Pediatric HIV–HBV Coinfection in Lusaka, Zambia: Prevalence and Short-Term Treatment Outcomes

Abstract

Hepatitis B virus (HBV) is endemic in Africa, where it may occur as an HIV coinfection. Data remain limited on HIV–HBV epidemiology in Africa, particularly in children. Using programmatic data from pediatric HIV clinics in Lusaka, Zambia during 2011–2014, we analyzed the prevalence of chronic HBV coinfection (defined as a single positive hepatitis B surface antigen [HBsAg] test) and its impact on immune recovery and liver enzyme elevation (LEE) during the first year of antiretroviral therapy. Among 411 children and adolescents, 10.4% (95% confidence interval, 7.6–14.1) had HIV–HBV. Coinfected patients were more likely to have World Health Organization stage 3/4, LEE and CD4 <14% at care entry (all p < 0.05). During treatment, CD4 increases and LEE incidence were similar by HBsAg status. HBsAg positivity decreased (11.8% vs. 6.6%; p = 0.24) following HBV vaccine introduction. These findings support screening pediatric HIV patients in Africa for HBV coinfection. Dedicated cohorts are needed to assess long-term outcomes of coinfection.

BACKGROUND

Worldwide, 240 million individuals have chronic hepatitis B virus infection (CHB), with most occurring in Asia and sub-Saharan Africa (SSA) [1]. Individuals who develop CHB in childhood have a 25% lifetime risk of premature death from complications of chronic liver disease, including hepatocellular carcinoma [2]. In SSA, an estimated 3 million CHB patients are dually HIV infected [3]. HIV coinfection increases the risk of CHB and accelerates hepatitis B virus (HBV)-related liver disease progression[4]. However, there are limited data on HIV–HBV epidemiology in SSA, particularly among children. Most SSA countries incorporated HBV vaccine into childhood immunization programs, but some reports suggest that CHB remains relatively common among HIV-infected children. We described the prevalence of HIV–HBV among children born in the pre- and post-vaccine eras and investigated the short-term treatment outcomes of pediatric HIV–HBV in Zambia.

METHODS

We analyzed data from Zambia’s capital Lusaka from 1 January 2011 to 28 February 2014, following dissemination of the Ministry of Health’s 2010 HIV guidelines recommending hepatitis B surface antigen (HBsAg) testing at enrollment [5]. We categorized pediatric patients as children (0–9 years of age) or adolescents (10–19 years of age) [6] and excluded adults. To reduce selection bias, we excluded patients from clinics where <10% of children were tested. Per 2010 national guidelines, patients were antiretroviral therapy (ART) eligible if they were <2 years old, 2–5 years old with a CD4 <25% or 5–19 years old with a CD4 ≤350 cells/mm [5, 7]. In addition, all patients with World Health Organization (WHO) stage 3/4 were ART eligible regardless of age or CD4%/count.

Using multivariable logistic regression, we identified patient factors associated with HBsAg screening at enrollment. Among HBsAg-tested patients, we defined CHB as a single positive HBsAg test and determined patient characteristics associated with coinfection in multivariable models. Among the group that was HBsAg tested and initiated ART, we assessed the association of CHB with the odds of CD4 >25% at 1 year of treatment. We also described the incidence of liver enzyme elevation (LEE; defined as alanine aminotransferase [ALT] >45 U/L for patients <12 months old and >33 U/L for all others) among ART recipients with normal initial ALT. We compared the incidence of events per 1000 person-years of treatment between those with and without CHB.

We explored CHB prevalence over calendar time in relation to the Ministry’s introduction of HBV vaccine in 2005 [8]. We compared HBsAg positivity between patients born before (1991–2005) and after (2006–2013) vaccine availability using a chi-square test. Statistical analyses were performed with SAS version 9.3 (SAS Institute, Cary, NC). Ethics committees of the University of Zambia and University of North Carolina at Chapel Hill approved the use of programmatic data for this analysis.

RESULTS

Within 10 facilities, 2,010 HIV-infected children/adolescents enrolled in HIV care. Of those, 411 (20.9%) were HBsAg tested and comprised the analysis cohort. The odds of HBsAg testing was increased for adolescents (adjusted odds ratio [AOR]: 2.50; 95% confidence interval [CI]: 1.92–3.27); however, those with WHO stage 3/4 were less likely to be tested (AOR: 0.64; 95% CI: 0.48–0.84). Testing was unrelated to sex, body mass index, CD4 count, LEE or tuberculosis (all p > 0.05).

In the analysis cohort, 265 (64.5%) were female, median CD4 was 15.6% (interquartile range: 10.2–23.8) and 43 were HBsAg positive for a prevalence of 10.5% (95% CI, 7.6–14.1). In adjusted analysis, adolescence, LEE, CD4 <14% and WHO stage 3/4 were associated with CHB coinfection (all p < 0.05; Table 1).

Table 1.

Characteristics associated with HBsAg positivity among 411 HIV-infected children and adolescents in Zambia screened for HBV at enrollment

Characteristic	Number (%) HBsAg positive	Crude odds ratio (95% CI)	Adjusted Odds ratio (95% CI)
Age, in years
0–9	9 (7.4)	Reference	Reference
10–19	34 (11.7)	1.59 (0.78–3.56)	2.80 (1.13–7.93)
Sex
Male	18 (12.3)	Reference	Reference
Female	25 (9.4)	0.74 (0.39–1.41)	0.62 (0.28–1.37)
Body mass index
≥−2 z score	24 (8.2)	Reference	Reference
<−2 z score	12 (16.7)	2.26 (1.04–4.64)	1.57 (0.68–3.48)
WHO clinical stage
1 or 2	24 (9.0)	Reference	Reference
3 or 4	19 (15.0)	1.78 (0.93–3.36)	2.30 (1.07–4.91)
Tuberculosis
No	42 (10.7)	Reference	Reference
Yes	1 (5.9)	0.75 (0.08–3.13)	0.43 (0.04–2.20)
CD4 percentage
≥14%	14 (5.9)	Reference	Reference
<14%	29 (16.9)	3.15 (1.65–6.28)	3.34 (1.58–7.56)
ALT, in U/L
Normal	27 (8.0)	Reference	Reference
Elevated	15 (22.4)	3.32 (1.64–6.55)	2.96 (1.29–6.58)

After enrollment, 257 (62.5%) patients initiated ART and contributed 198.1 person-years of follow-up. During the first year of ART, patients experienced an average increase of 196 CD4 cells/mm³ or 8.3%. Baseline CD4 <14% was associated with reduced odds of achieving CD4 ≥25% at 1 year (AOR: 0.06; 95% CI: 0.01–0.44); however, CHB did not predict reduced immune recovery (AOR: 1.39, 95% CI: 0.38–4.93). At enrollment, 17 (16.7%) had LEE, and during treatment, there were 13 incident LEEs (1 among HBsAg positives and 12 among HBsAg negatives), including 7 that were grade 2 or greater, [9] for an overall incidence rate of 65.6 per 1000 person-years. Incidence of LEE was similar by HBsAg status (65.0 vs. 65.7, p = 1.000). Children born following implementation of routine HBV vaccination had a trend toward lower HBsAg positivity (11.8% vs. 6.6%; p = 0.24).

DISCUSSION

In urban Zambia, among those screened for HBV, approximately 10% of pediatric HIV-infected patients had CHB, with slightly lower prevalence among those born since HBV immunization began. Compared with HBsAg-negative patients, CHB coinfected patients had more advanced HIV disease and evidence of LEE at enrollment; however, short-term ART outcomes appeared to be similar.

This is the first report from Zambia on pediatric HIV–HBV, and the relatively large size of the cohort allowed us to analyze ART outcomes, something lacking in recent publications from SSA. Our analysis has limitations, primarily arising from incomplete data, a common programmatic issue in settings like ours. Dissemination of HBsAg-testing guidelines was suboptimal [10] and only one-fifth of pediatric patients were HBsAg tested; therefore, patients included in analysis may represent a select population. However, those with LEE at enrollment were not more likely to be screened, suggesting that testing did not occur because of suspected liver disease. Adolescents were more likely to be screened as some receive care in the adult HIV clinic where there is greater awareness of HBV. We also acknowledge that small sample sizes in longitudinal analyses reduced our statistical power to detect small differences in CD4 change and incident LEEs by HBsAg status.

These results build on an emerging literature on HIV–HBV in SSA. Our observation that CHB coinfected patients had LEE, lower CD4 and more advanced clinical status is similar to some SSA reports, [11–14] but differs from others [11, 13, 15–17]. Significant liver disease is uncommon in pediatric CHB; however, the natural history of HIV–HBV is less well described, particularly in SSA. CHB did not appear to impact early treatment outcomes, but these results require confirmation in other settings. Our observed HIV–HBV prevalence of 10.5% is consistent with reported prevalence of 1.2–12.1% in other SSA settings [11–15, 17–22]. We observed lower HBsAg positivity among children born in the era of HBV immunization, similar to other low- and middle-income countries [12, 23]. HBsAg testing was inconsistent in our program;[10] therefore, these data are not strong evidence of HBV vaccine effectiveness, and further investigation is needed in this area.

In summary, CHB coinfection was common among HIV-infected children and adolescents in Zambia, supporting the need for dedicated screening and treatment guidelines for developing countries, and cohort studies to characterize the long-term ART outcomes of these at-risk patients.

FUNDING

This work was supported by the Fogarty International Center of the U.S. National Institutes of Health [K01TW009998].