Potential conflict of interest: Nothing to report.
Supported by the Taiwan Centers for Disease Control (grants AIDS-97-1002 and EU099059).
Preliminary analyses of these data were presented as Abstract O-556 at the 21st European Congress of Clinical Microbiology and Infectious Diseases, London, United Kingdom, March 31-April 3, 2012.
The purpose of this prospective cohort study was to compare the serologic response between human immunodeficiency virus (HIV)-infected men who have sex with men (MSM) receiving two and three doses of hepatitis A virus (HAV) vaccine and HIV-uninfected MSM receiving two doses of HAV vaccine. Between June 2009 and December 2010, 582 MSM aged 18 to 40 years who were seronegative for HAV were enrolled in the study. HIV-infected MSM received either two doses of HAV vaccine (1,440 enzyme-linked immunosorbent assay units) (n = 140) with the second dose given at week 24 or three doses (n = 225) with the second and third dose given at weeks 4 and 24, respectively, while HIV-uninfected MSM (n = 217) received two doses. The primary endpoint was seroconversion at week 48. The geometric mean concentration (GMC) of anti-HAV antibody was determined at weeks 48 and 72. At week 48, the seroconversion rate was 75.7%, 77.8%, and 88.5% in intention-to-treat analysis for two-dose HIV-infected, three-dose HIV-infected, and two-dose HIV-uninfected MSM, respectively. The GMC of anti-HAV antibody at week 48 for three-dose HIV-infected MSM (2.29 ± 0.73 log10 mIU/mL) was significantly higher than that for two-dose HIV-infected MSM (1.94 ± 0.66; P < 0.01), but was lower than HIV-uninfected MSM (2.49 ± 0.42; P < 0.01). Multivariate analysis revealed higher CD4 counts (adjusted odds ratio [AOR] for per 50 cells/μL increase, 1.13; 95% confidence interval [CI], 1.05-1.21) and undetectable plasma HIV RNA load (AOR, 1.90; 95% CI, 1.10-3.28) before HAV vaccination were predictive of seroconversion in HIV-infected patients. Conclusion: Serologic response rate to three and two doses of HAV vaccine was similar in HIV-infected MSM, which was lower than that in HIV-uninfected MSM receiving two doses. HAV vaccination in HIV-infected patients with a higher CD4 count and suppression of HIV replication increased the seroconversion rate. (HEPATOLOGY 2013)
Hepatitis A virus (HAV) infection that is transmitted via a fecal-oral route occurs worldwide, especially in countries where sanitary and hygienic conditions are not maintained appropriately. In countries with improved sanitation and access to HAV vaccination, the incidence of HAV infection has declined significantly. The annual incidence of HAV infection has decreased from 12 per 100,000 population in 1995 to 0.6 per 100,000 population in 2009 in the United State after HAV vaccine was licensed in 1995.1 In Finland, the incidence of HAV infections ranged from 0.3 to 3.6 per 100,000 between 1990 and 2007, and most of the cases seemed to be travel-related.2
Most HAV infections are self-limited; however, severe symptoms and complications associated with acute hepatitis A increase with age.3, 4 In addition, persons with chronic hepatitis B virus (HBV) or hepatitis C virus (HCV) infection who develop superinfection with HAV are at increased risk for fulminant hepatitis and associated mortality.5, 6 In human immunodeficiency virus (HIV)-infected patients who acquire HAV, the duration of HAV viremia would be longer compared with that of healthy persons,7 which may increase the risk of transmission to others. Therefore, HAV vaccination is recommended for HIV-infected patients, injecting drug users, men who have sex with men (MSM), hemophiliacs, patients with chronic liver disease (e.g., chronic HBV or chronic HCV infection), and persons planning to travel to or reside in endemic areas for HAV transmission.8
The response rate of HAV vaccination is almost 100% among HIV-uninfected persons who receive two standard doses of HAV vaccine.9, 10 However, the immunogenicity to HAV vaccination is impaired in HIV-infected patients, and the seroconversion rate is estimated to be 68% to 88.2% after two doses of HAV vaccine (each dose containing 720 or 1,440 enzyme-linked immunosorbent assay [ELISA] units) and 76% to 90% after three doses of HAV vaccine (each containing 720 ELISA units).9-14 Furthermore, the antibody titers of HIV-infected patients following HAV vaccination (100-636 mIU/mL) are significantly lower compared with those of HIV-uninfected persons (>1,000 mIU/mL).9, 10 In a recently published randomized clinical trial in HIV-infected patients, the three-dose schedule of HAV vaccination (each dose containing 1,440 ELISA units) tended to achieve a higher seroconversion rate than the two-dose schedule at week 72 (78.3% versus 61.2%; P = 0.07) and produce higher and more durable antibody titers after follow-up for 18 months.15
In this study, we compared the immunogenicity of two different doses of HAV vaccination (two doses versus three doses) in HIV-infected patients and evaluated whether an additional dose of HAV vaccination in HIV-infected patients could achieve a comparable serologic response to HIV-uninfected subjects.
Subjects and Methods
Setting and Study Population.
This prospective cohort study was conducted at a university hospital that is a major hospital designated for HIV care and provision of voluntary counseling and testing for HIV and sexually transmitted infections in Taiwan, where the overall seroprevalence of HAV in HIV-infected MSM was 50.5%, which rapidly increased in persons aged 30 to 40 years, suggesting an opportunity for HAV vaccination in persons younger than 40 years.16, 17 Subjects with and without HIV infection who were seronegative for HAV and aged 18 to 40 years were enrolled from June 25, 2009, to December 31, 2010. HIV-infected MSM were recruited from the outpatient clinics, while HIV-uninfected MSM were recruited from voluntary counseling and testing services. Exclusion criteria included failure to provide informed consent, active infection, malignancy, splenectomy, and concurrent use of systemic corticosteroids, chemotherapy, immunoglobulin, or immunomodulation agents.
In Taiwan, HIV-infected patients are provided free access to HIV care that includes monitoring of CD4 count and plasma HIV RNA load, and combination antiretroviral therapy (cART) that was introduced in Taiwan in 1997. CART was defined as the use of at least three agents from at least two classes of antiretroviral agents according to the local treatment guidelines for adults with HIV infection. The study was approved by the Research Ethics Committee of the hospital and subjects gave written informed consent (NCT registration no. 01102296).
HAV Vaccination Schedule.
HIV-infected subjects were sequentially enrolled to receive two doses and three doses of HAV vaccine (1440 ELISA units) (HAVRIX 1440; GlaxoSmithKline, Biologicals, Rixensart, Belgium). Enrollment of HIV-infected subjects to receive three doses of HAV vaccine began after completion of enrollment of HIV-infected subjects to receive two doses of vaccine. In the two-dose vaccination schedule in HIV-infected and HIV-uninfected subjects, HAV vaccine was administered at week 0 and week 24, while in the three-dose schedule, HAV vaccine was administered at week 0, week 4, and week 24. The subjects were contacted by cell phone to inquire reactions following vaccination. Anti-HAV antibody was determined at week 24 (before the last dose was administered) and week 48. The primary endpoint was seroconversion at week 48. The secondary endpoints were seroconversion at week 24 and the geometric mean concentration (GMC) of anti-HAV antibody at weeks 48 and 72. Patients who had no serum available for determination of anti-HAV antibody at week 24 were considered as nonresponders. At week 48, those patients who had no serum samples available, but had achieved seroconversion at week 24, were considered as responders; those without seroconversion at week 48, or having no serologic data available before the last dose of vaccination at week 24 were considered as nonresponders (intention-to-treat [ITT] analysis using last-observation-carried-forward principle). Sensitivity analyses were performed in patients who had available results of anti-HAV antibody at different time points (per-protocol [PP] analysis).
Matched Case-Control Study.
Because the baseline characteristics were statistically significantly different in age and immunologic and virologic characteristics between the two-dose HIV-infected and three-dose HIV-infected group (Table 1), pairs from the two groups were selected that were matched for CD4 count (±20 cells/μL) and age (±2 years) to better assess the serologic responses to HAV vaccination between the two groups receiving different doses of HAV vaccine.
Table 1. Baseline Characteristics of Subjects Who Received Two or Three Doses of HAV Vaccine
Two-Dose HIV-infected Group (n = 140)
Three-Dose HIV-infected Group (n = 225)
Two-Dose HIV-Uninfected Group (n = 217)
Abbreviation: PVL, plasma HIV RNA load.
P < 0.05 between the three-dose HIV-infected group versus the HIV-uninfected group;
P < 0.05 between the two-dose HIV-infected group versus the HIV-uninfected group (two-tailed chi-square and exact measures of association between two-dose and three-dose HIV-infected groups).
Serum samples were collected before vaccination (week 0) and at weeks 24, 48, and 72. The determinations of anti-HAV antibody of serum samples were performed at the central laboratory of the hospital via chemiluminescence immunoassay according to the manufacturer's protocol (ARCHITECT HAVAb-IgG, Abbott Diagnostics, Wiesbaden, Germany). The laboratory staff was unaware of the HIV serostatus and the doses of HAV vaccine the subjects had received. The anti-HAV antibody titers were determined using a commercially available enzyme-linked immunosorbent assay (ELISA) method (ETI-AB-HAVK PLUS; DiaSorin, Saluggia, Italy). Seropositivity was defined as an anti-HAV antibody titer >20 mIU/mL. Plasma HIV RNA load was quantified using the Cobas Amplicor HIV-1 Monitor test (Cobas Amplicor version 1.5, Roche Diagnostics, Indianapolis, IN) with a lower detection limit of 40 copies/mL. CD4 lymphocyte count was determined using the FACFlow system (BD FACSCalibur, Becton Dickinson, San Jose, CA).
All statistical analyses were performed using SPSS version 17.0 (SPSS, Chicago, IL). Categorical variables were compared using a Fisher's exact test or chi-square test. Noncategorical variables were compared using a Mann-Whitney U test. Factors with P value ≤0.2, or with biological significance were included for multivariate analysis. Logistic regression analysis was used to determine the factors associated with HAV seroconversion. All comparisons were two-tailed and a P value <0.05 was considered significant. In HIV-infected subjects, the noninferiority in terms of seroconversion rate following two-dose HAV vaccination to three-dose vaccination would be concluded if the lower boundary of the two-sided 95% confidence interval (CI) (one-sided α = 0.025) for the difference in the seroconversion rate between the two groups was at least −0.1 (i.e., the noninferiority margin was set to 10%).
AOR, adjusted odds ratio; cART, combination antiretroviral therapy; CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; GMC, geometric mean concentration; HAV, hepatitis A virus; HBsAg, HBV surface antigen; HBV, hepatitis B virus; HCV, hepatitis C virus; HIV, human immunodeficiency virus; ITT, intention-to-treat; MSM, men who have sex with men; PP, per-protocol.
During the 18-month study period, 582 subjects were enrolled: 140 HIV-infected MSM received two doses of HAV vaccine; 225 HIV-infected MSM received three doses; and 217 HIV-uninfected MSM received two doses (Fig. 1). In total, 43 (7.4%) subjects did not receive the last dose of HAV vaccine: eight (5.7%) in the two-dose HIV-infected group; 12 (5.3 %) in the three-dose HIV-infected group; and 23 (10.6%) in the two-dose HIV-uninfected group. The baseline characteristics of the subjects are shown in Table 1. HIV-uninfected subjects who were enrolled from voluntary counseling and testing services were significantly younger than HIV-infected subjects (Table 1). In HIV-infected subjects, the three-dose group was younger than the two-dose group. The seroprevalences of HBV (HBV surface antigen [HBsAg]-positive) and HCV (anti-HCV antibody–positive) were similar between the two-dose and three-dose HIV-infected groups (HBV, 13.7% versus 14.1%; HCV, 5.7% versus 5.4%; P > 0.99); both the HBV and HCV seroprevalences were significantly higher than those of the HIV-uninfected group (HBV seroprevalence, 6.5%; HCV seroprevalence, 0.9%). The mean baseline CD4 count was lower (452 versus 538 cells/μL), while the baseline plasma HIV RNA load was higher (3.0 versus 2.5 log10 copies/mL) in the three-dose HIV-infected group compared with that of the two-dose HIV-infected group (both comparisons, P < 0.01). The proportion of subjects receiving cART before HAV vaccination in three-dose group and two-dose group was 58.2% and 67.1%, respectively (P = 0.12); that of the HIV-infected patients starting cART after vaccination in three-dose group and two-dose group was 15.6% and 11.4%, respectively (P = 0.27).
At week 24, before the last dose of HAV vaccine was administered, the seroconversion rate was 37.4% for the two-dose HIV-infected group, 62.2% for the three-dose HIV-infected group, and 52.6% for the HIV-uninfected group (P < 0.05) (Fig. 2). At week 48, the seroconversion rates were statistically significantly lower in the HIV-infected groups compared with the HIV-uninfected group in ITT analysis (two-dose HIV-infected group, 75.7%; three-dose HIV-infected group, 77.8%; HIV-uninfected group, 88.5%). In PP analysis, the seroconversion rates for the three groups were 81.7% (two-dose HIV-infected group), 81.8% (three-dose HIV-infected group), and 97.9% (HIV-uninfected group). The seroconversion rate for the two-dose HIV-infected group was not inferior to that of the three-dose HIV-infected group, with a difference of −0.02 (95% CI, −0.069 to 0.110) between the two groups in the ITT population and −0.0003 (95% CI, −0.086 to 0.086) in the PP population.
In the multivariate analysis of all three groups after adjustment for doses of HAV vaccination, age, positive HBsAg, and positive anti-HCV antibody, HIV-infected subjects had a statistically significantly lower seroconversion rate than HIV-uninfected subjects, with an adjusted odds ratio (AOR) of 0.46 (95% CI, 0.28-0.75) (Table 2). In multivariate analysis among HIV-infected subjects after adjustment for age, baseline CD4 count, HIV RNA load, doses of HAV vaccine, positive HBsAg, and positive anti-HCV antibody, factors that were independently associated with seroconversion were higher CD4 counts (AOR for per 50 cells/μL increase, 1.13; 95% CI, 1.05-1.21) and undetectable plasma HIV RNA load (<40 copies/mL) (AOR, 1.90; 95% CI, 1.10-3.28) before vaccination (Table 2).
Table 2. Multivariate Analysis of Factors Associated With Seroconversion After HAV Vaccination in HIV-Infected and -Uninfected MSM
AOR (95% CI)
Abbreviation: PVL, plasma HIV RNA load.
Multivariate analysis after adjustment for doses of HAV vaccination, age, positive HBsAg, and positive anti-HCV antibody.
Multivariate analysis after adjustment for age, baseline CD4 count, HIV RNA load, dose of HAV vaccine, HBsAg positivity, and anti-HCV antibody positivity.
The seroconversion rate did not differ significantly between HIV-infected subjects with and those without HBV or HCV infection. In the three-dose HIV-infected group, the seroconversion rates were 71.0% (22/31) and 79.4% (150/189) for subjects with positive HBsAg and those with negative HBsAg, respectively (P = 0.35), and were 66.7% (8/12) and 78.7% (166/211) for subjects with positive anti-HCV antibody and those with negative anti-HCV antibody, respectively (P = 0.47). In the two-dose group, the seroconversion rate was 78.9% (15/19) and 75.8% (91/120) for subjects with positive HBsAg and those with negative HBsAg, respectively (P = 0.79); 62.5% (5/8) and 76.5% (101/132) for subjects with positive anti-HCV antibody and those with negative anti-HCV antibody, respectively (P = 0.40).
In the matched study, 114 pairs of HIV-infected subjects who received either two doses or three doses of HAV vaccine were identified; their clinical characteristics are shown in Table 3. The seroconversion rates at week 48 were 78.1% and 84.2% for the two-dose HIV-infected group and three-dose HIV-infected group, respectively, in ITT analysis (P = 0.23), with a difference of −0.06 (95% CI, −0.040 to 0.163). In PP analysis, the seroconversion rates were 81.6% and 91.7% for the two-dose HIV-infected group and three-dose HIV-infected group, respectively (P = 0.04). Therefore, one additional dose of hepatitis A vaccination in HIV-infected patients was associated with a statistically significantly higher seroconversion rate in PP analysis (AOR, 2.50; 95% CI, 1.03-6.07), but not in ITT analysis (AOR, 1.44; 95% CI, 0.73-2.85) (Table 4).
Table 3. Clinical Characteristics of 114 Matched Pairs of HIV-Infected Subjects Who Received Two Doses and Three Doses of HAV Vaccine
Two-Dose Group (n = 114)
Three-Dose Group (n = 114)
Abbreviation: PVL, plasma HIV RNA load.
Age, years, mean (SD)
≤20, n (%)
21-30, n (%)
31-40, n (%)
HBsAg-positive, n (%)
Anti-HCV–positive, n (%)
Baseline CD4 count, cells/μL, mean (SD)
<200, n (%)
200-350, n (%)
>350, n (%)
Baseline PVL, mean (SD) log10copies/mL
PVL ≤40 copies/mL, n (%)
PVL >5 log10copies/mL, n (%)
Cart at vaccination, n (%)
CD4 at week 48, cells/μL, mean (SD)
PVL at week 48, log10 copies/mL, mean (SD)
PVL at week 48, ≤40 copies/mL, n (%)
Seroconversion at week 24, n (%)
Seroconversion at week 48, n (%)
Table 4. Factors Associated With Seroconversion in ITT Analysis Using the Last Observation Carried Forward Principle, and PP Analysis in the 114 Matched Pairs of HIV-Infected Patients
AOR (95% CI)
Three-dose versus two-dose vaccination
CD4 count (per 50 cells/μL increase)
Three-dose versus two-dose vaccination
CD4 count (per 50 cells/μL increase)
Compared with the two-dose HIV-infected group, the GMC of anti-HAV antibody was statistically significantly higher for the three-dose HIV-infected group (week 48, 2.29 ± 0.73 versus 1.94 ± 0.66 log10 mIU/mL, P < 0.01; week 72, 2.08 ± 0.68 versus 1.78 ± 0.56 log10 mIU/mL, P<0.01) (Fig. 3). The proportion of HAV antibody titer that was >20 mIU/mL at weeks 48 and 72 was 88.6% (109/123) and 86.6% (110/127), respectively, for the two-dose HIV-infected group and 89.2% (182/204) and 86.9% (173/199), respectively, for the three-dose HIV-infected group (data not shown).
The GMC in the three-dose HIV-infected group was significantly lower than that of the two-dose HIV-uninfected group (week 48, 2.29 ± 0.73 versus 2.49 ± 0.42 log10 mIU/mL, P < 0.01; week 72, 2.08 ± 0.68 versus 2.23 ± 0.45 log10 mIU/mL, P = 0.02) (Fig. 3). The proportion of HAV antibody titer that was >20 mIU/mL at weeks 48 and 72 for HIV-uninfected group was 100% (172/172) and 100% (147/147), respectively.
HAV vaccination did not cause intolerable adverse effects in either group of subjects, with the most adverse effect being mild tenderness at the local injection site in 24 hours of vaccination that was reported in 51.6% of all subjects (HIV-infected versus HIV-uninfected, 51.7% versus 51.6%, P = 0.98) (data not shown).
In this prospective cohort study of HAV vaccination in HIV-infected and HIV-uninfected MSM, we found that an additional dose of HAV vaccination in HIV-infected patients failed to achieve a comparable serologic response rate to HIV-uninfected persons. While the three-dose HAV vaccination schedule achieved a higher serologic response rate than the two-dose HAV vaccination schedule in PP analysis in HIV-infected matched pairs, the difference was not statistically significant in ITT analysis.
The strength of our study is that we enrolled a large number of subjects consisting of HIV-infected as well as HIV-uninfected subjects to evaluate the serologic responses to two different doses of HAV vaccination. Furthermore, the completion rates of follow-up for seroconversion and antibody titers were high at weeks 48 and 72. Several limitations need to be acknowledged, however. This study was not a randomized clinical trial, and the HIV-infected subjects who were enrolled to receive three doses of HAV vaccine and those who received two doses of HAV vaccine differed in several important clinical characteristics. Although we selected subjects in the two groups that were comparable in terms of age, CD4, and plasma HIV RNA load for comparisons of serologic responses, we were only able to demonstrate the superiority of a three-dose HAV vaccination schedule to a two-dose schedule in PP analysis instead of ITT analysis.
The seroconversion rate of HAV vaccination in HIV-infected patients has been shown to be lower than that in HIV-uninfected persons before the cART era.9, 10 A prospective study conducted in Australia during 1993-1995, which recruited 90 HIV-infected and 46 HIV-uninfected MSM to receive two doses of HAV vaccine (720 ELISA units), showed that the seroconversion rate in HIV-infected persons (88.2%) was lower than that in HIV-uninfected persons (100%) (P = 0.03).9 In the era of cART, adding a third dose to the standard two-dose HAV vaccination schedule has been tried to enhance immunogenicity in HIV-infected patients. In a clinical trial by Launay et al.15 in which 95 HIV-infected patients were randomized to receive three doses (n = 46) or two doses (n = 49) of HAV vaccine (1,440 ELISA units), the seroconversion rates were 82.6% and 69.4% for the three-dose group and two-dose group, respectively, at week 28 (primary endpoint; P = 0.13, ITT analysis) and 78.3% and 61.2%, respectively, at week 72 (P = 0.07). In our study, the subjects were all younger MSM with significantly higher CD4 counts than those in the study by Launay et al., and more than 60% of our HIV-infected subjects were receiving cART; regardless, three doses of HAV vaccination in our HIV-infected MSM failed to achieve a seroconversion rate or GMC comparable to that of HIV-uninfected MSM who received two doses of HAV vaccination at weeks 48 and 72, both in ITT and PP analyses.
In this study, we identified that a higher CD4 count and undetectable plasma HIV RNA load were predictive of serologic responses in HIV-infected adult patients (Table 2), which was similar to the findings of the retrospective study by Overton et al.19 that enrolled 268 HIV-infected patients who had received at least one dose of HAV vaccine (Havrix, 1,440 ELISA units) and to those of the prospective study by Weinberg et al.14 that enrolled 152 HIV-infected patients aged 2 to 21 years. In a subgroup analysis of our study, we found that for patients with a plasma HIV viral load of ≤40 copies/mL and a CD4 count of ≥200 cells/μL at baseline, the seroconversion rates were 81.2% (56/69) and 89.0% (73/82) in the two-dose group and three-dose group, respectively, compared with 88.5% (192/217) in the HIV-uninfected group in ITT analysis (P > 0.05 in two comparisons). The response rate could be further increased to 92.1% (58/63) in the three-dose group who had a CD4 count of ≥350 cells/μL and an HIV viral load of ≤40 copies/mL (data not shown). The findings of our subgroup analysis suggest that, to improve immunogenicity, HAV vaccination should be recommended for HIV-infected patients who have achieved good virologic and immunologic responses to cART. Although more studies are warranted to confirm our findings, we suggest that three doses of HAV vaccine could maximize the response rate in HIV-infected patients who have a CD4 count of ≥350 cells/μL.
HIV-infected adult patients who received two doses of HAV vaccine have been demonstrated to generate a lower anti-HAV antibody titer that was less durable compared with HIV-uninfected persons and higher GMCs over time among HIV-infected adults were associated with lower plasma HIV RNA loads.20 In our study, we also found that, despite an additional dose, the GMC for three-dose HIV-infected MSM remained lower than that for two-dose HIV-uninfected MSM at week 48 (2.29 ± 0.73 versus 2.49 ± 0.42 log10 mIU/mL, P < 0.01) and at week 72 (2.08 ± 0.68 versus 2.23 ± 0.45 log10 mIU/mL, P = 0.02). Whether further increase of doses and dosing frequency of HAV vaccine among HIV-infected patients, similar to those demonstrated in a recent HBV vaccination study by Launay et al.,21 can improve the efficacy warrants more studies. While our study and the aforementioned study by Launay et al.15 failed to demonstrate a statistically significantly higher seroconversion rate in HIV-infected patients who received three doses of HAV vaccination than those who received two doses, an additional dose of HAV vaccine did significantly increase the anti-HAV antibody titers (Fig. 3).15 Of those initial responders, Launay et al. found that 85% could maintain a protective antibody titer for a follow-up duration of 3.7 years.22
In conclusion, the serologic response rate to three and two doses of HAV vaccine was similar in HIV-infected MSM, which was lower than that in HIV-uninfected MSM who received two doses. Administration of HAV vaccine in HIV-infected patients with higher CD4 counts (preferably >200 cells/μL) and suppression of HIV replication increased the seroconversion rate.
We thank the subjects who participated in the study and Chin-Fu Hsiao (Division of Biostatistics and Bioinformatics, National Health Research Institutes, Taiwan) for statistical analyses.