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

The sponsors had no role in the design, data collection, analysis, interpretation, writing or decision to publish this report. Ethical approval for the study was from the Gambian Government and The Gambia Government/MRC Ethics Committee (L2008.05, SCC1084v2).

2670 people born in 1986–90 (and hence possibly in the GHIS) were recruited, but reliable linkage to the original GHIS records of vaccination in infancy was possible for only 753 (28%). Recruitment was approximately balanced between the four study areas: Basse 905 (34%), Mansakonko 633 (24%), Farafenni 572 (21%), and Gambisara 560 (21%). Of all participants, 1751 (66%) were female and 1681 (63%) were recruited from villages rather than towns. Regardless of HBV vaccination history, the prevalence of current hepatitis C virus infection among these young adults, as indicated by anti-HCV antibody, was 0.5% (13/2598).

The prevalence of current hepatitis B infection, as indicated by HBsAg positivity, was 0.8% (2/255) in those fully vaccinated with 3 or 4 doses in infancy, 12.4% (59/475) in those unvaccinated, and 17.9% (4/23) among the few who had been only partially vaccinated (heterogeneity p < 0.00001) (Table 1). The crude odds ratio for HBsAg positivity comparing those fully vaccinated versus those unvaccinated was 0.06, suggesting 94% vaccine efficacy (with 95% CI 77-99%). Logistic regression was used to assess the impact of potential confounders on the association between vaccination status and HBsAg positivity, but adjustment for recruitment area, age, sex, and urban or rural location did not alter the odds ratio. Comparing fully vaccinated and unvaccinated subjects, the prevalence of past or present infection (as indicated by HBsAg positivity or IgG antibody to the HBV core protein) was 27% (70/255) vs 56% (267/475), p < 0.00001. Thus, although the prevalence of past or present infection among young adults is approximately halved by infant vaccination, many of the fully vaccinated participants had at some time been infected.

Among those in either group (vaccinated or not) who were currently infected (HBsAg-positive), only 8% (5/65) had IgM antibody to the HBV core protein (indicating that the onset of infection was only recent). Of these 65 samples, 17 (26%) also tested positive for the early antigen (HBeAg-positive, indicating intense infection), 46 (71%) had antibody to the early antigen, and only 2 (3%) had neither the early antigen nor antibodies to it detected.

HBsAg positivity was strongly associated with ALT elevation, indicating some impairment of liver function (Table 2). Ignoring vaccination status and comparing all HBsAg-positive samples vs matched HBsAg-negative samples, the odds ratio for having ALT above the upper limit of normal (ULN) was 7.1 (26/218 vs 10/472, p < 0.0001), and the odds ratio for having chronic active hepatitis, as indicated by ALT > 2xULN, was 34.1 (9/218 vs 1/472). Although the numbers were small, chronic active hepatitis was, as expected, even more highly correlated with intense HBV infection, as indicated by HBeAg positivity.

Among those born since 1990 (since the beginning of the nationwide vaccination program, which reaches most but not all of the population) the overall prevalence of current HBV infection, as indicated by HBsAg positivity, was only 0.5% (24/4613), about as low as among the fully vaccinated controls born in 1986–90. When stratified by 5-year age group there was an increase from 0.2% in those <5 years of age to 1.8% in those >15 years of age. (Table 3) Those born since 1997 were 1–9 (median 5) years of age when we surveyed them, and among them 9/3506 (0.3%) were HBsAg-positive.

There were only five HBsAg-positive children born since 1990 for whom we could find records showing they had definitely been vaccinated in infancy against HBV, and these were accounted for by only two mothers (three siblings in one family with an HBsAg positive mother and two in another family with an untested mother). There were no other associations with HBsAg positivity apart from the decrease with year of birth.

The consistency of the estimates of the efficacy of full infant vaccination in the present survey, in previous surveys of GHIS participants and (with historical controls) in surveys of the Gambian village of Keneba is shown in Table 4. A review of sero-prevalence rates in unvaccinated populations shows that the pre-vaccination HBV prevalence rates in Keneba and in unvaccinated GHIS controls are reasonably representative of those in many other parts of West Africa [29, 30]: Figure 2. (In one other Gambian village, Manduar, the pre-vaccination HBV prevalence rate was somewhat higher.)

The nationwide vaccination program that began in 1990 has been more effective than might have been predicted from extrapolation of the GHIS results. The program offers a birth dose of HBV vaccine plus a three-dose regimen of HBV-DPT vaccine, and has resulted in a massive decrease in the prevalence of chronic HBV infection between those born in the early 1980s and those born in the present century. Transmission from mothers may well account for the majority of the few remaining chronic infections in infants. As most chronic infection is acquired in early childhood, the significant downward trend in prevalence by year of birth since 1990 implies increasing effectiveness in the years since adoption in 1990 of universal HBV vaccination in The Gambia. This is partly because Gambian EPI coverage rates are better now than in the early 1990s, and partly because of herd immunity among children. High levels of infant vaccination will, however, continue to be needed, as for many decades there will still be significant numbers of people chronically infected with HBV in The Gambia and its neighbours.

The major decrease in chronic hepatitis B virus infection because of infant vaccination has been accompanied by a substantial decrease in chronic hepatitis C virus infection, probably due to improved sterile precautions in medical use of needles. Among the controls in a case–control study of adults in The Gambia there was a substantial prevalence of HCV (around 5%, 10/190) among those born before the 1950s but a much lower prevalence (0.5%, 1/192) among those born later [28]. In the present survey of young adults born during 1986–1990 the prevalence of HCV was again only 0.5% (13/2598), confirming the low HCV prevalence in recent generations. The present study tested much larger numbers of Gambians than any previous study and involved a representative community sample. Comparison of the results among older adults in other studies with those among young adults in the present study suggests a birth cohort effect involving almost a ten-fold decrease in the prevalence of HCV in The Gambia. This suggests that, in The Gambia as in many other countries, the mid-century epidemic of iatrogenic HCV infection from unsterilized needles has been controlled by the sterile procedures of the last quarter of the century. So, although HBV and HCV are causes of almost all liver cancer and cirrhosis deaths today in The Gambia, both should largely vanish over the coming decades.