Efficacy and effectiveness of infant vaccination against chronic hepatitis B in the Gambia Hepatitis Intervention Study (1986-90) and in the nationwide immunisation program

Abstract

Background: Gambian infants were not routinely vaccinated against hepatitis B virus (HBV) before 1986. During 1986-90 the Gambia Hepatitis Intervention Study (GHIS) allocated 125,000 infants, by area, to vaccination or not and thereafter all infants were offered the vaccine through the nationwide immunisation programme. We report HBV serology from samples of GHIS vaccinees and unvaccinated controls, and from children born later.

Methods: During 2007-08, 2670 young adults born during the GHIS (1986-90) were recruited from 80 randomly selected villages and four townships. Only 28% (753/2670) could be definitively linked to their infant HBV vaccination records (255 fully vaccinated, 23 partially vaccinated [1-2 doses], 475 not vaccinated). All were tested for current HBV infection (HBV surface antigen [HBsAg]) and, if HBsAg-negative, evidence of past infection (HBV core-protein antibody [anti-HBc]). HBsAg-positive samples (each with two age- and sex-matched HBsAg-negative samples) underwent liver function tests. In addition, 4613 children born since nationwide vaccination (in 1990-2007) were tested for HBsAg. Statistical analyses ignore clustering.

Results: Comparing fully vaccinated vs unvaccinated GHIS participants, current HBV infection was 0.8% (2/255) vs 12.4% (59/475), p < 0.0001, suggesting 94% (95% CI 77-99%) vaccine efficacy. Among unvaccinated individuals, the prevalence was higher in males (p = 0.015) and in rural areas (p = 0.009), but adjustment for this did not affect estimated vaccine efficacy. Comparing fully vaccinated vs unvaccinated participants, anti-HBc was 27.4% (70/255) vs 56.0% (267/475), p < 0.00001. Chronic active hepatitis was not common: the proportion of HBsAg-positive subjects with abnormal liver function tests (ALT > 2 ULN) was 4.1%, compared with 0.2% in those HBsAg-negative. The prevalence of antibodies to hepatitis C virus was low (0.5%, 13/2592). In children born after the end of GHIS, HBsAg prevalence has remained low; 1.4% (15/1103) in those born between 1990-97, and 0.3% (9/35150) in those born between 1998-2007.

Conclusions: Infant HBV vaccination achieves substantial protection against chronic carriage in early adulthood, even though approximately a quarter of vaccinated young adults have been infected. This protection persists past the potential onset of sexual activity, reinforcing previous GHIS findings of protection during childhood and suggesting no need for a booster dose. Nationwide infant HBV vaccination is controlling chronic infection remarkably effectively.

Figure 1

Flow diagram of the study.

Figure 2

Prevalence of chronic HBV infection in The Gambia, by year of birth: born before routine infant HBV vaccination; vaccinees and controls born during the Gambia Hepatitis Intervention Study (GHIS) in 1986–90; born since nationwide infant HBV vaccination. Key: Number positive for HBV surface antigen (HBsAg)/number surveyed, % positive, and range of ages in years (y) when surveyed. Open circles = unvaccinated, closed circles = vaccinated in infancy (or supposed to be, in the Gambian EPI). Keneba: Unvaccinated historical controls (one village, children surveyed in 1984 at ages 1–19 years). West Africa median: Median of survey-specific HBsAg prevalences in 83 population surveys of unvaccinated children or adults conducted between 1990–2009 in West Africa (Merrill & Hunter, 2011). GHIS: Vaccinees and concurrent controls in the Gambia Hepatitis Intervention Study of children born in 1986–90, and surveyed for the present study in 2007–08, at ages 17–22 years, 0 · 8% vs 12 · 4% p < 0 · 00001. EPI: Population sample of children born since 1990 and surveyed for the present study in 2007–08 to assess the effectiveness of nationwide infant HBV vaccination (in the Gambian Expanded Programme on Immunisation), 1990–97 births 1 · 4% vs 1998–2007 births 0 · 3%, p < 0 · 00001.