- Email: [email protected]
Persistence of seroprotection 10 years after primary hepatitis A vaccination in an unselected study population
Vaccine 25 (2007) 927–931
Persistence of seroprotection 10 years after primary hepatitis A vaccination in an unselected study population Pamela Rendi-Wagner a,b , Maria Korinek a , Birgit Winkler a , Michael Kundi c , Herwig Kollaritsch a,b , Ursula Wiedermann a,b,∗ a
c
Department of Specific Prophylaxis and Tropical Medicine, Center for Physiology and Pathophysiology, Medical University Vienna, Vienna, Austria b Center for Travel Medicine, Vienna, Austria Institute of Environmental Health, Center for Public Health, Medical University Vienna, Vienna, Austria Received 31 July 2006; received in revised form 25 August 2006; accepted 29 August 2006 Available online 18 September 2006
Abstract Hepatitis A vaccines have been demonstrated to be highly immunogenic. Mathematical models have predicted antibodies to persist for at least 20–25 years. Most of these studies have been conducted in young and healthy study populations. We aimed to evaluate long-term immunity 10 years following complete primary immunization according to a 3-dose schedule (HavrixTM 720 El.U at months 0, 1, 6–12) in an adult and unselected study population. In total, 999 (98.3%) of 1016 vaccinees (mean age 54.7 ± S.D. 13.0), tested 10 years after primary vaccination, still had protective antibody levels (≥10 mIU/ml) as measured by ELISA. An anti-HAV titer cut off level of 11,400 mIU/ml was calculated to differentiate between vaccine-induced and infection-induced titer levels. The vaccine-induced geometric mean titer (GMT) was 406.1 mIU/ml (95% CI: 369.2–446.7 mIU/ml), showing an age-related trend, the 10-years seroprotection rate (SPR) was 97.9%. Females exhibited significantly higher GMTs than male vaccinees (p < 0.001). The only parameter predicting a titer below 10 mIU/ml 10 years after vaccination was the body mass index (p = 0.001). This study confirms that protection following primary hepatitis A vaccination persists for more than 10 years. © 2006 Elsevier Ltd. All rights reserved. Keywords: Hepatitis A vaccines; Long-term protection; Adults; Antibody-persistence
1. Introduction Hepatitis A is a food- and waterborne infectious disease causing worldwide approximately 1.5 million hepatitis A cases per year [1,2]. Incidence rates are closely related to standards of hygiene and sanitation. For quite some time hepatitis A has been regarded the most common vaccine-preventable infection in travelers [3,4]. To protect non-immune individuals from infection with hepatitis A virus (HAV), several ∗ Corresponding author at: Department of Specific Prophylaxis and Tropical Medicine, Center for Physiology and Pathophysiology, Medical University Vienna, Kinderspitalgasse 15, A-1090 Vienna, Austria. Tel.: +43 1 40490 64890; fax: +43 1 40490 64899. E-mail address: [email protected] (U. Wiedermann).
0264-410X/$ – see front matter © 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2006.08.044
inactivated hepatitis A vaccines became available more than 10 years ago. These proved to be safe, effective, and highly immunogenic. Hepatitis A vaccination has primarily been recommended for individuals of well-recognized high-risk groups by national and international health authorities. Due to the extensive use of hepatitis A vaccines in travel medicine, long-term protection and antibody kinetics gained increasing attention. Serological follow-up studies of vaccinees confirmed anti-HAV antibody persistence up to 10 years [5,6]. Using these data, antibody decline models have predicted protective levels of antibodies to persist for more than 20 years [7–12]. However, most of these previous serological studies have been conducted under ideal, carefully controlled clinical trial conditions in healthy and young study population groups and may not necessarily be valid for the general pop-
928
P. Rendi-Wagner et al. / Vaccine 25 (2007) 927–931
ulation. Immunogenicity results, however, obtained under everyday field conditions may differ, since controlled trials do not reflect reality in daily vaccine use [13]. Therefore, the purpose of the current study was to assess long-term immunity and persistence of seroprotection 10 years following the primary vaccination course with an inactivated hepatitis A vaccine under field conditions in an everyday vaccination center setting, in a large unselected group of vaccine recipients with no upper age limits.
2. Material and methods 2.1. Study design and population This study is a descriptive evaluation of anti-HAV immunity (at about) 10 years following a complete primary threedose immunization schedule. Volunteers (≥15 years old at the time of primary hepatitis A vaccination) of both sexes who were seeking pre-travel immunization advice were invited to be included in this study. According to their immunization documentation, all subjects enrolled proved to be vaccinated by HavrixTM 720 El.U (GlaxoSmithKline Biologicals, Rixensart, Belgium), the only HAV vaccine licensed in Austria at time of vaccination, which has been replaced only few years ago by HavrixTM 1440 El.U applied in a 2-dose schedule. The only exclusion criteria had been contraindications as mentioned in the prescribing information, including hypersensitivity to any component of the vaccine, severe febrile illness as well as documented history of hepatitis A infection in the past. The vaccine dose had been administered intramuscularly into the deltoid region according to a 0-, 1-, and 6- to 12-month schedule. A total of 1016 individuals fulfilling all inclusion criteria were enrolled in this study. Subject recruitment (for blood sampling) commenced in January 2003, the last subject was enrolled in April 2006. The hepatitis A vaccine used was formulated to contain no less than 720 ELISA units (El.U) of hepatitis A antigen (strain HM175) per 1 ml dose, adsorbed onto 0.5 mg of aluminium hydroxide. Serum samples were tested over time (3 years) quantitatively for the presence of anti-HAV antibodies using a commercially available ELISA (Elecsys® 2010 anti-HAV, Roche Diagnostics) at the Institute of Virology, Medical University Vienna. Subjects with antibody levels below 10 mIU/ml were considered seronegative [14]. Only positive titers were used for geometric mean titer (GMT) calculation. 2.2. Statistical analysis In regard to smoking habits, data were divided into groups of smokers (>10 cigarettes per day at the time of blood sampling), non-smokers and persons with unknown smoking habits. Body mass index (BMI) was calculated on the basis of reported height and weight (weight in kg/(height in m)2 ) at the time of blood sampling.
GMTs were calculated together with 95% CI within age subgroups. Confidence intervals of GMTs were calculated based on the assumption of a log-normal distribution of titers. Exact confidence intervals of seropositivity rates were based on the binominal distribution. Regression analysis of log neutralizing titers was performed with age, gender, smoking habits and BMI as independent variables. Cut-off for natural infection induced immunity was calculated according to the following procedure: distribution of log-titers obtained in this study was assumed to be the result of a mixture of two normal distributions, one of vaccineinduced antibody levels and one from infection induced levels; parameters of these distributions were estimated by non-linear regression under the only constraint that the mean of the former is below 3 (i.e. below 1000 mIU/ml) and that of the latter is above 3. Censoring at a titer of 30,000 mIU/ml was accounted for in the analysis (i.e. the observed value of the distribution function at values above 30,000 mIU/ml was set as the integral of values above this level for natural infection and vaccine-induced immunity). Fit was excellent (pseudo R2 = 0.98); cut-off was then computed as the titer that results in an equal fraction of false negatives and false positives (with respect to natural immunity). Let Φ(c; μv ; σ v ) be the value of the distribution function for vaccine induced immunity and Φ(c; μi ; σ i ) for infection induced immunity at a log-titer of c. For any cut-off c sensitivity to detect natural infection can be estimated as 1 − (c; μi ; σ i ) while specificity is given by (c; μv ; σ v ). Hence solution of the equation Φ(c; μi ; σ i ) + Φ(c; μv ; σ v ) = 1 yields the value c for which the fraction of false negatives equals that of false positives. This value was estimated as 4.057, equivalent to a HAV antibody titer of 11,400 mIU/ml. A value of 11,428 mIU/ml (rounded to 11,400 mIU/ml) was previously also estimated in [15], due to the practically equal distribution function in both (independent) studies. Based on this calculation, the study population was categorized in subgroups according to HAV antibody levels (“seronegative”: <10 mIU/ml; “vaccine-induced antibody titer”: ≥10 and <11,400 mIU/ml; “infection-induced immunity”: ≥11,400 mIU/ml).
3. Results A total of 1016 individuals (52.9% females; mean age at time of blood sampling 54.7 ± S.D. 12.9) who were primo vaccinated against hepatitis A at mean 9.9 years (± S.D. 0.8) ago according to a 3-dose immunization schedule (0, 1, 6–12 months) were included in this study. Subject recruitment commenced in January 2003, the last subject was enrolled in April 2006. Demographic features of all subjects analyzed are reported in Table 1. All included vaccinees completed their hepatitis A primary vaccination course between September 1992 and November 1996.
P. Rendi-Wagner et al. / Vaccine 25 (2007) 927–931
929
Table 1 Characteristics of study population and subgroups according anti-HAV titer concentration Total study population
n (%) Age (±S.D.)a Vaccination interval (±S.D.) Sex (female, %) Smoker/non-smoker (%)a,b BMI (±S.D.)a (n = 863) a b
1016 (100) 54.67 (±12.95) 9.9 (±0.81) 52.9 9.94/50.60 25.31 (±4.15)
Anti-HAV <10 mIU/ml
<11400 mIU/ml
≥11400 mIU/ml
17 (1.67) 53.83 (±10.44) 9.88 (±0.74) 64.7 0/35.29 31.629 (±7.97)
796 (78.35) 52.51 (±13.17) 9.88 (±0.80) 51.8 10.30/48.49 25.17 (±4.13)
203 (19.98) 63.23 (±7.72) 10.09 (±0.81) 56.2 9.35/60.10 25.41 (±3.53)
At the time of blood sampling. Remaining %: smoking habit unknown.
3.1. Wild-virus induced anti-HAV immunity Overall, 98.3% of the vaccinees attained protective levels of anti-HAV (≥10 mIU/ml). As illustrated by Fig. 1, the distribution of anti-HAV log-titer levels reveals two peaks. Distinction between vaccine induced and wild-virus induced antibody titers was done according to the calculated cut-off level, revealing that 20.0% of the vaccinees exhibited titers ≥11,400 mIU/ml and thus having acquired immunity by natural infection (Fig. 1). Detailed demographic characteristics regarding this subgroup are summarized in Table 1. When comparing vaccinees with titers below and subjects with antibody concentrations above 11,400 mIU/ml, the parameter “age” indicates a significant difference (52.5 ± 13.1 years versus 63.3 ± 7.7 years, p < 0.001). The proportion of subjects per age group with titers ≥11,400 mIU/ml is rising with age and is highest in subjects >70 years of age (40.0%) (Fig. 2). 3.2. Vaccine-induced anti-HAV immunity 3.2.1. Seroprotection rate (SPR) For further analysis of vaccine-induced immunity, titer levels of the subgroup with titer levels ≥11,400 mIU/ml were excluded. The mean age of the subgroup with
Fig. 1. Distribution of anti-HAV log-titer levels with titer cut-off level at 11,400 mIU/ml to differentiate between vaccine-induced and infectioninduced antibody levels.
assumed vaccine-induced antibody levels (anti-HAV <11,400 mIU/ml) was 52.5 years (S.D. ± 13.2 years) (Table 1). In total, the vaccine-induced seroprotection rate (SPR) was 97.9%. Among the 17 subjects who did not exhibit protective anti-HAV titers, 11 persons were older than 50 years. No significant difference in SPR was observed comparing subjects above and below 50 years of age (p = 0.71). When dividing subjects into two groups according to seroprotection rate (above and below 10 mIU/ml) and investigating the risk of having a titer below 10 mIU/ml, the only significant result was obtained for BMI (p = 0.001). No significant gender specific difference in SPR was observed (females: 97.4% versus males: 98.5%; p = 0.3 χ-test). 3.2.2. Geometric mean titer concentrations (GMT) Overall, the vaccine-induced GMT (analyzing only antibody concentrations <11,400 mIU/ml) was 406.1 mIU/ml (95% CI: 369.2–446.7 mIU/ml) with an age-related trend as illustrated by Fig. 3. Comparing gender specific GMTs, females exhibited significantly higher concentrations of antibodies than male vaccinees (501.6 mIU/ml versus 323.8 mIU/ml, respectively; p < 0.001) (Table 2, Fig. 3). Regression analysis of log anti-
Fig. 2. Proportion of subjects with anti-HAV titer concentration ≥11,400 mIU/ml (“infection-induced” antibodies) per age group.
930
P. Rendi-Wagner et al. / Vaccine 25 (2007) 927–931
Table 2 Number of subjects and vaccine-induced geometric mean titer levels (±95% CI; min, max) per categories (gender, age: <50 and >50 years; smoking habit, BMI <25 and >25) Male
GMT 95% LCI 95% UCI Minimum Maximum u-Test
323.71 282.46 370.99 20 11000
Total number
384
Female
501.38 440.29 570.95 10 10800 p < 0.001 412
Age
Smoker
<50 years
>50 years
513.01 451.92 582.36 10 9900
347.74 304.60 396.98 25 11000
335.06 246.64 455.18 10 5800
479
82
p < 0.001 317
Fig. 3. Vaccine-induced geometric mean antibody levels (±95% CI) (<11,400 mIU/ml) of male (white bars) and female (grey bars) vaccinees per age category.
HAV titers with age, BMI, gender, and smoking habits as predictor variables revealed a significant positive correlation for younger age (p = 0.003) and female gender (p = 0.0006).
4. Discussion Hepatitis A vaccines are known to be highly immunogenic inducing seroconversion rates of approximately 100% [6]. Indeed, a substantial amount of data have been generated on immunogenicity and long-term persistence of hepatitis A vaccine-induced antibodies, however, one has to keep in mind, that such studies were obtained in carefully monitored clinical trial situations with either healthy and young study populations, limited number of subjects or selected cluster of vaccine recipients. In most of these studies a selected cohort of subjects was followed at fixed times without taking into account common variables encountered in clinical practice, or long-term protection was evaluated by mathematic analyses [8–10]. The gain of the present study implies: (1) the considerable sample size of more than 1000 vaccine recipients, (2) the broad spectrum of unselected subjects with no upper age limits (mean age 55 years), reflecting the mean age of reallife population of vaccinees, increasingly including elderly travelers, and (3) the actual recruitment time interval of 10 years after primary hepatitis A vaccination for monitoring long-term immunity.
Non-Smoker
424.60 370.35 486.80 25 10800 p = 0.289 386
BMI <25
>25
475.54 416.83 542.53 20 10000
354.35 301.57 416.36 10 10800 p < 0.001
370
298
Testing more than 1000 post-vaccination anti-HAV titers, we observed an overall high seropositivity rate in an unselected population, showing particular high antibody levels in the older age groups. Based on our data, a cut-off level (anti-HAV ≥11,400 mIU/ml) was calculated to differentiate with high probability between vaccine-induced and infectioninduced antibody levels, which is also confirmed by Holzmann et al. who have followed up antibody levels up to 60 months post-vaccination [15]. Applying this cut-off level, the proportion of subjects with assumed naturally acquired anti-HAV antibodies constitute 20% of all vaccine recipients, all older than 40 years of age, thus, confirming previous population-based surveys on hepatitis A seroprevalence rates in Austria [16,17]. However, it can be assumed that the true seroprevalence rate among the Austrian population most probably exceeds that of our study population. Nevertheless, this proportion refers to the group of unscreened hepatitis A vaccinees under field conditions. In this context, it appears noteworthy, that more than 40% of all vaccinees older than 60 years of age would not have needed their hepatitis A shot due to a preexisting naturally acquired immunity. Hence, for this age group, pre-vaccination screening for anti-HAV immunity may be worth considering. Excluding anti-HAV antibody concentrations above 11,400 mIU/ml, the vaccine-induced seroprotection rate was still at a very high level (98%) with a geometric mean titer concentration of 406 mIU/ml. This is comparable to follow-up data of various hepatitis A vaccination studies [5,6]. With respect to the serum protection rates no other risk factors than the BMI were observed. Surprisingly, no effects of higher age were noticed. The reason for the lack of a significant age-related correlation of SPRs might be due to the overall low number of seronegative subjects in our study population, which may mask such an effect. This result confirms that a very small group of vaccinees exhibits an insufficient immune response to hepatitis A vaccination—a phenomenon that has extensively been described for other vaccines, such as hepatitis B [13,18,19]. However, since antibody concentrations were only measured 10 years post-vaccination, at this time point it is unclear, whether
P. Rendi-Wagner et al. / Vaccine 25 (2007) 927–931
these subjects lack the ability to build up protective antibody levels a priori or, if titers have decreased at a higher rate compared to the other vaccine recipients. When performing regression analysis for log titer values it became, however, evident that higher age and male gender were significantly associated with lower antibody concentration. This observation confirms the significant effect of age and gender on the immune response, at least in terms of GMTs, as also described for other vaccines [13]. However, since most of the observed antibody levels still proved to be far above detection limit, thus protective, this observation appears to be rather insubstantial regarding long-term protection. According to the consensus statement of “The International Consensus Group on Hepatitis A Virus Immunity”, HAV booster vaccination after a complete primary vaccination course in healthy individuals seems unnecessary, since evidence is accumulating that HAV vaccination elicits immunological memory persisting even after the loss of detectable antibodies [5,20–22]. This is supported by experimental studies in chimpanzees showing that persistence of detectable anti-HAV antibodies is not absolutely required for protective immunity and thus indicating that also cellular mechanisms play an important role [23–25]. In this respect current investigations in a subgroup of our study population will evaluate functional as well as phenotypical characteristics of cellular immune responses prior and after HAV booster vaccination. Taken together, our study is one of the few long-term studies showing that after a full course of primary hepatitis A vaccination protective immunity exceeds by far the 10-year booster interval, currently advised by the manufactures’ prescribing information and most health authorities. Our data clearly support the modification of vaccination recommendation as previously defined by the International Consensus Group on Hepatitis A Virus Immunity.
Acknowledgments The skilful technical assistance and administrative support of Dr. Eva Jeschko, Erika Garner-Spitzer, and Brigitte Laaber is gratefully acknowledged. This study was financially supported by the Centre for Travel Medicine, Vienna.
References [1] Koff RS. Hepatitis A. Lancet 1998;351:1643–9. [2] World Health Organization. Hepatitis A vaccines. Weekly Epidemiol Rec 2000;75:38–44. [3] Mutsch M, Spicher VM, Gut C, Steffen R. Hepatitis A virus infections in travelers, 1988–2004. Clin Infect Dis 2006;42:490–7. [4] Jong EC. Risks of hepatitis A and B in the traveling public. J Travel Med 2001;8(Suppl 1):S3–8.
931
[5] Van Herck K, Van Damme P, Lievens M, Stoffel M. Hepatitis A vaccine: indirect evidence of immune memory 12 years after the primary course. J Med Virol 2004;72:194–6. [6] Bell BP. Hepatitis A vaccine. Pediatr Infect Dis J 2000;19:1187–8. [7] Andr´e F, Van Damme P, Safary A, Banatvala JE. Inactivated hepatitis A vaccine: immunogenicity, efficacy, safety, and review of official recommendations of use. Expert Rev Vaccines 2002;1:9–23. [8] Wiedermann G, Kundi M, Ambrosch F, Safary A, D’Hondt E, Delem A. Inactivated hepatitis A vaccine: long-term antibody persistence. Vaccine 1997;15:612–5. [9] Wiedermann G, Kundi M, Ambrosch F. Estimated persistence of antiHAV antibodies after single dose and booster hepatitis A vaccination (0–6 schedule). Acta Trop 1998;69:121–5. [10] Van Herck K, Beutels P, Van Damme P, Beutels M, Van den Dries J, Briantais P, et al. Mathematical models for assessment of long-term persistence of antibodies after vaccination with two inactivated hepatitis A vaccines. J Med Virol 2000;60:1–7. [11] Bovier PA, Bock J, Loutan L, Farinelli T, Glueck R, Herzog C. Longterm immunogenicity of an inactivated virosome hepatitis A vaccine. J Med Virol 2002;68:489–93. [12] Van Herck K, Renard D, Molensberghs G, Van Damme P. Model-based estimates of long-term persistence of vaccine-induced hepatitis A antibodies. In: Margolis HS, Alter MJ, Liang TJ, Dienstag JL, editors. Viral hepatitis and liver disease. London: International Medical Press; 2002. p. 56–9. [13] Rendi-Wagner P, Kundi M, Stemberger H, Wiedermann G, Holzmann H, Hofer M, et al. Antibody-response to three recombinant hepatitis B vaccines: comparative evaluation of multicenter travel-clinic based experience. Vaccine 2001;19:2055–60. [14] Wiedermann G, Ambrosch F, Andr´e FE, Hondt ED, Delem A, Safary A. Persistence of vaccine-induced antibody to hepatitis A virus. Vaccine 1992;10(Suppl. 1):S129–31. [15] Holzmann H, Croy C, Kundi M, Popow-Kraupp T. Quantitative determination of hepatitis A virus antibodies in vaccinees: is it necessary? Proceedings of the third annual meeting of the European society for clinical virology (ESCV), September 1–5. 1999 [Abstract 59]. [16] Frisch-Niggemeyer W, Kunz C. The incidence of antibodies to hepatitis A in people from Vienna and certain Austrian provinces. Wien Klin Wochenschr 1979;91:230–3. [17] Prodinger WM, Larcher C, Solder BM, Geissler D, Dierich MP. Hepatitis A in Western Austria—the epidemiological situation before the introduction of active immunisation. Infection 1994;22:53–5. [18] Alper CA, Kruskall MS, Marcus-Bagley D, Craven DE, Katz AJ, Brink SJ. Genetic prediction of non-response to hepatitis B vaccine. N Engl J Med 1989;321:708–12. [19] Hollinger FB. Factors influencing the immune response to hepatitis B vaccine, booster dose guidelines, and vaccine protocol recommendations. Am J Med 1989;87(Suppl 3A):36–40. [20] Van Damme P, Banatvala J, Fay O, Iwarson S, McMahon B, Van Herck K, et al. Hepatitis A booster vaccination: is there a need? Lancet 2003;362:1065–71. [21] Van Herck K, Van Damme P. Inactivated hepatitis A vaccine-induced antibodies: follow-up and estimates of long-term persistence. J Med Virol 2001;63(1):1–7. [22] Werzberger A, Kuter B, Nalin D. Six years follow-up after hepatitis A vaccination. N Engl J Med 1998;338(16):1160. [23] Lemon SM. Immunologic approaches to assessing the response to inactivated hepatitis A vaccine. J Hepatol 1993;18(Suppl. 2):15–9. [24] Cederna JB, Klinzman D, Stapleton JT. Hepatitis A virus-specific humoral and cellular immune responses following immunization with a formalin-inactivated hepatitis A vaccine. Vaccine 2000;18:892–8. [25] Purcell RH, D’Hondt E, Bradbury R, Emerson SU, Govindarajan S, Binn L. Inactivated hepatitis A vaccine: active and passive immunoprophylaxis in chimpanzees. Vaccine 1992;10(Suppl 1):148–51.
Persistence of seroprotection 10 years after primary hepatitis A vaccination in an unselected study population Pamela Rendi-Wagner a,b , Maria Korinek a , Birgit Winkler a , Michael Kundi c , Herwig Kollaritsch a,b , Ursula Wiedermann a,b,∗ a
c
Department of Specific Prophylaxis and Tropical Medicine, Center for Physiology and Pathophysiology, Medical University Vienna, Vienna, Austria b Center for Travel Medicine, Vienna, Austria Institute of Environmental Health, Center for Public Health, Medical University Vienna, Vienna, Austria Received 31 July 2006; received in revised form 25 August 2006; accepted 29 August 2006 Available online 18 September 2006
Abstract Hepatitis A vaccines have been demonstrated to be highly immunogenic. Mathematical models have predicted antibodies to persist for at least 20–25 years. Most of these studies have been conducted in young and healthy study populations. We aimed to evaluate long-term immunity 10 years following complete primary immunization according to a 3-dose schedule (HavrixTM 720 El.U at months 0, 1, 6–12) in an adult and unselected study population. In total, 999 (98.3%) of 1016 vaccinees (mean age 54.7 ± S.D. 13.0), tested 10 years after primary vaccination, still had protective antibody levels (≥10 mIU/ml) as measured by ELISA. An anti-HAV titer cut off level of 11,400 mIU/ml was calculated to differentiate between vaccine-induced and infection-induced titer levels. The vaccine-induced geometric mean titer (GMT) was 406.1 mIU/ml (95% CI: 369.2–446.7 mIU/ml), showing an age-related trend, the 10-years seroprotection rate (SPR) was 97.9%. Females exhibited significantly higher GMTs than male vaccinees (p < 0.001). The only parameter predicting a titer below 10 mIU/ml 10 years after vaccination was the body mass index (p = 0.001). This study confirms that protection following primary hepatitis A vaccination persists for more than 10 years. © 2006 Elsevier Ltd. All rights reserved. Keywords: Hepatitis A vaccines; Long-term protection; Adults; Antibody-persistence
1. Introduction Hepatitis A is a food- and waterborne infectious disease causing worldwide approximately 1.5 million hepatitis A cases per year [1,2]. Incidence rates are closely related to standards of hygiene and sanitation. For quite some time hepatitis A has been regarded the most common vaccine-preventable infection in travelers [3,4]. To protect non-immune individuals from infection with hepatitis A virus (HAV), several ∗ Corresponding author at: Department of Specific Prophylaxis and Tropical Medicine, Center for Physiology and Pathophysiology, Medical University Vienna, Kinderspitalgasse 15, A-1090 Vienna, Austria. Tel.: +43 1 40490 64890; fax: +43 1 40490 64899. E-mail address: [email protected] (U. Wiedermann).
0264-410X/$ – see front matter © 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2006.08.044
inactivated hepatitis A vaccines became available more than 10 years ago. These proved to be safe, effective, and highly immunogenic. Hepatitis A vaccination has primarily been recommended for individuals of well-recognized high-risk groups by national and international health authorities. Due to the extensive use of hepatitis A vaccines in travel medicine, long-term protection and antibody kinetics gained increasing attention. Serological follow-up studies of vaccinees confirmed anti-HAV antibody persistence up to 10 years [5,6]. Using these data, antibody decline models have predicted protective levels of antibodies to persist for more than 20 years [7–12]. However, most of these previous serological studies have been conducted under ideal, carefully controlled clinical trial conditions in healthy and young study population groups and may not necessarily be valid for the general pop-
928
P. Rendi-Wagner et al. / Vaccine 25 (2007) 927–931
ulation. Immunogenicity results, however, obtained under everyday field conditions may differ, since controlled trials do not reflect reality in daily vaccine use [13]. Therefore, the purpose of the current study was to assess long-term immunity and persistence of seroprotection 10 years following the primary vaccination course with an inactivated hepatitis A vaccine under field conditions in an everyday vaccination center setting, in a large unselected group of vaccine recipients with no upper age limits.
2. Material and methods 2.1. Study design and population This study is a descriptive evaluation of anti-HAV immunity (at about) 10 years following a complete primary threedose immunization schedule. Volunteers (≥15 years old at the time of primary hepatitis A vaccination) of both sexes who were seeking pre-travel immunization advice were invited to be included in this study. According to their immunization documentation, all subjects enrolled proved to be vaccinated by HavrixTM 720 El.U (GlaxoSmithKline Biologicals, Rixensart, Belgium), the only HAV vaccine licensed in Austria at time of vaccination, which has been replaced only few years ago by HavrixTM 1440 El.U applied in a 2-dose schedule. The only exclusion criteria had been contraindications as mentioned in the prescribing information, including hypersensitivity to any component of the vaccine, severe febrile illness as well as documented history of hepatitis A infection in the past. The vaccine dose had been administered intramuscularly into the deltoid region according to a 0-, 1-, and 6- to 12-month schedule. A total of 1016 individuals fulfilling all inclusion criteria were enrolled in this study. Subject recruitment (for blood sampling) commenced in January 2003, the last subject was enrolled in April 2006. The hepatitis A vaccine used was formulated to contain no less than 720 ELISA units (El.U) of hepatitis A antigen (strain HM175) per 1 ml dose, adsorbed onto 0.5 mg of aluminium hydroxide. Serum samples were tested over time (3 years) quantitatively for the presence of anti-HAV antibodies using a commercially available ELISA (Elecsys® 2010 anti-HAV, Roche Diagnostics) at the Institute of Virology, Medical University Vienna. Subjects with antibody levels below 10 mIU/ml were considered seronegative [14]. Only positive titers were used for geometric mean titer (GMT) calculation. 2.2. Statistical analysis In regard to smoking habits, data were divided into groups of smokers (>10 cigarettes per day at the time of blood sampling), non-smokers and persons with unknown smoking habits. Body mass index (BMI) was calculated on the basis of reported height and weight (weight in kg/(height in m)2 ) at the time of blood sampling.
GMTs were calculated together with 95% CI within age subgroups. Confidence intervals of GMTs were calculated based on the assumption of a log-normal distribution of titers. Exact confidence intervals of seropositivity rates were based on the binominal distribution. Regression analysis of log neutralizing titers was performed with age, gender, smoking habits and BMI as independent variables. Cut-off for natural infection induced immunity was calculated according to the following procedure: distribution of log-titers obtained in this study was assumed to be the result of a mixture of two normal distributions, one of vaccineinduced antibody levels and one from infection induced levels; parameters of these distributions were estimated by non-linear regression under the only constraint that the mean of the former is below 3 (i.e. below 1000 mIU/ml) and that of the latter is above 3. Censoring at a titer of 30,000 mIU/ml was accounted for in the analysis (i.e. the observed value of the distribution function at values above 30,000 mIU/ml was set as the integral of values above this level for natural infection and vaccine-induced immunity). Fit was excellent (pseudo R2 = 0.98); cut-off was then computed as the titer that results in an equal fraction of false negatives and false positives (with respect to natural immunity). Let Φ(c; μv ; σ v ) be the value of the distribution function for vaccine induced immunity and Φ(c; μi ; σ i ) for infection induced immunity at a log-titer of c. For any cut-off c sensitivity to detect natural infection can be estimated as 1 − (c; μi ; σ i ) while specificity is given by (c; μv ; σ v ). Hence solution of the equation Φ(c; μi ; σ i ) + Φ(c; μv ; σ v ) = 1 yields the value c for which the fraction of false negatives equals that of false positives. This value was estimated as 4.057, equivalent to a HAV antibody titer of 11,400 mIU/ml. A value of 11,428 mIU/ml (rounded to 11,400 mIU/ml) was previously also estimated in [15], due to the practically equal distribution function in both (independent) studies. Based on this calculation, the study population was categorized in subgroups according to HAV antibody levels (“seronegative”: <10 mIU/ml; “vaccine-induced antibody titer”: ≥10 and <11,400 mIU/ml; “infection-induced immunity”: ≥11,400 mIU/ml).
3. Results A total of 1016 individuals (52.9% females; mean age at time of blood sampling 54.7 ± S.D. 12.9) who were primo vaccinated against hepatitis A at mean 9.9 years (± S.D. 0.8) ago according to a 3-dose immunization schedule (0, 1, 6–12 months) were included in this study. Subject recruitment commenced in January 2003, the last subject was enrolled in April 2006. Demographic features of all subjects analyzed are reported in Table 1. All included vaccinees completed their hepatitis A primary vaccination course between September 1992 and November 1996.
P. Rendi-Wagner et al. / Vaccine 25 (2007) 927–931
929
Table 1 Characteristics of study population and subgroups according anti-HAV titer concentration Total study population
n (%) Age (±S.D.)a Vaccination interval (±S.D.) Sex (female, %) Smoker/non-smoker (%)a,b BMI (±S.D.)a (n = 863) a b
1016 (100) 54.67 (±12.95) 9.9 (±0.81) 52.9 9.94/50.60 25.31 (±4.15)
Anti-HAV <10 mIU/ml
<11400 mIU/ml
≥11400 mIU/ml
17 (1.67) 53.83 (±10.44) 9.88 (±0.74) 64.7 0/35.29 31.629 (±7.97)
796 (78.35) 52.51 (±13.17) 9.88 (±0.80) 51.8 10.30/48.49 25.17 (±4.13)
203 (19.98) 63.23 (±7.72) 10.09 (±0.81) 56.2 9.35/60.10 25.41 (±3.53)
At the time of blood sampling. Remaining %: smoking habit unknown.
3.1. Wild-virus induced anti-HAV immunity Overall, 98.3% of the vaccinees attained protective levels of anti-HAV (≥10 mIU/ml). As illustrated by Fig. 1, the distribution of anti-HAV log-titer levels reveals two peaks. Distinction between vaccine induced and wild-virus induced antibody titers was done according to the calculated cut-off level, revealing that 20.0% of the vaccinees exhibited titers ≥11,400 mIU/ml and thus having acquired immunity by natural infection (Fig. 1). Detailed demographic characteristics regarding this subgroup are summarized in Table 1. When comparing vaccinees with titers below and subjects with antibody concentrations above 11,400 mIU/ml, the parameter “age” indicates a significant difference (52.5 ± 13.1 years versus 63.3 ± 7.7 years, p < 0.001). The proportion of subjects per age group with titers ≥11,400 mIU/ml is rising with age and is highest in subjects >70 years of age (40.0%) (Fig. 2). 3.2. Vaccine-induced anti-HAV immunity 3.2.1. Seroprotection rate (SPR) For further analysis of vaccine-induced immunity, titer levels of the subgroup with titer levels ≥11,400 mIU/ml were excluded. The mean age of the subgroup with
Fig. 1. Distribution of anti-HAV log-titer levels with titer cut-off level at 11,400 mIU/ml to differentiate between vaccine-induced and infectioninduced antibody levels.
assumed vaccine-induced antibody levels (anti-HAV <11,400 mIU/ml) was 52.5 years (S.D. ± 13.2 years) (Table 1). In total, the vaccine-induced seroprotection rate (SPR) was 97.9%. Among the 17 subjects who did not exhibit protective anti-HAV titers, 11 persons were older than 50 years. No significant difference in SPR was observed comparing subjects above and below 50 years of age (p = 0.71). When dividing subjects into two groups according to seroprotection rate (above and below 10 mIU/ml) and investigating the risk of having a titer below 10 mIU/ml, the only significant result was obtained for BMI (p = 0.001). No significant gender specific difference in SPR was observed (females: 97.4% versus males: 98.5%; p = 0.3 χ-test). 3.2.2. Geometric mean titer concentrations (GMT) Overall, the vaccine-induced GMT (analyzing only antibody concentrations <11,400 mIU/ml) was 406.1 mIU/ml (95% CI: 369.2–446.7 mIU/ml) with an age-related trend as illustrated by Fig. 3. Comparing gender specific GMTs, females exhibited significantly higher concentrations of antibodies than male vaccinees (501.6 mIU/ml versus 323.8 mIU/ml, respectively; p < 0.001) (Table 2, Fig. 3). Regression analysis of log anti-
Fig. 2. Proportion of subjects with anti-HAV titer concentration ≥11,400 mIU/ml (“infection-induced” antibodies) per age group.
930
P. Rendi-Wagner et al. / Vaccine 25 (2007) 927–931
Table 2 Number of subjects and vaccine-induced geometric mean titer levels (±95% CI; min, max) per categories (gender, age: <50 and >50 years; smoking habit, BMI <25 and >25) Male
GMT 95% LCI 95% UCI Minimum Maximum u-Test
323.71 282.46 370.99 20 11000
Total number
384
Female
501.38 440.29 570.95 10 10800 p < 0.001 412
Age
Smoker
<50 years
>50 years
513.01 451.92 582.36 10 9900
347.74 304.60 396.98 25 11000
335.06 246.64 455.18 10 5800
479
82
p < 0.001 317
Fig. 3. Vaccine-induced geometric mean antibody levels (±95% CI) (<11,400 mIU/ml) of male (white bars) and female (grey bars) vaccinees per age category.
HAV titers with age, BMI, gender, and smoking habits as predictor variables revealed a significant positive correlation for younger age (p = 0.003) and female gender (p = 0.0006).
4. Discussion Hepatitis A vaccines are known to be highly immunogenic inducing seroconversion rates of approximately 100% [6]. Indeed, a substantial amount of data have been generated on immunogenicity and long-term persistence of hepatitis A vaccine-induced antibodies, however, one has to keep in mind, that such studies were obtained in carefully monitored clinical trial situations with either healthy and young study populations, limited number of subjects or selected cluster of vaccine recipients. In most of these studies a selected cohort of subjects was followed at fixed times without taking into account common variables encountered in clinical practice, or long-term protection was evaluated by mathematic analyses [8–10]. The gain of the present study implies: (1) the considerable sample size of more than 1000 vaccine recipients, (2) the broad spectrum of unselected subjects with no upper age limits (mean age 55 years), reflecting the mean age of reallife population of vaccinees, increasingly including elderly travelers, and (3) the actual recruitment time interval of 10 years after primary hepatitis A vaccination for monitoring long-term immunity.
Non-Smoker
424.60 370.35 486.80 25 10800 p = 0.289 386
BMI <25
>25
475.54 416.83 542.53 20 10000
354.35 301.57 416.36 10 10800 p < 0.001
370
298
Testing more than 1000 post-vaccination anti-HAV titers, we observed an overall high seropositivity rate in an unselected population, showing particular high antibody levels in the older age groups. Based on our data, a cut-off level (anti-HAV ≥11,400 mIU/ml) was calculated to differentiate with high probability between vaccine-induced and infectioninduced antibody levels, which is also confirmed by Holzmann et al. who have followed up antibody levels up to 60 months post-vaccination [15]. Applying this cut-off level, the proportion of subjects with assumed naturally acquired anti-HAV antibodies constitute 20% of all vaccine recipients, all older than 40 years of age, thus, confirming previous population-based surveys on hepatitis A seroprevalence rates in Austria [16,17]. However, it can be assumed that the true seroprevalence rate among the Austrian population most probably exceeds that of our study population. Nevertheless, this proportion refers to the group of unscreened hepatitis A vaccinees under field conditions. In this context, it appears noteworthy, that more than 40% of all vaccinees older than 60 years of age would not have needed their hepatitis A shot due to a preexisting naturally acquired immunity. Hence, for this age group, pre-vaccination screening for anti-HAV immunity may be worth considering. Excluding anti-HAV antibody concentrations above 11,400 mIU/ml, the vaccine-induced seroprotection rate was still at a very high level (98%) with a geometric mean titer concentration of 406 mIU/ml. This is comparable to follow-up data of various hepatitis A vaccination studies [5,6]. With respect to the serum protection rates no other risk factors than the BMI were observed. Surprisingly, no effects of higher age were noticed. The reason for the lack of a significant age-related correlation of SPRs might be due to the overall low number of seronegative subjects in our study population, which may mask such an effect. This result confirms that a very small group of vaccinees exhibits an insufficient immune response to hepatitis A vaccination—a phenomenon that has extensively been described for other vaccines, such as hepatitis B [13,18,19]. However, since antibody concentrations were only measured 10 years post-vaccination, at this time point it is unclear, whether
P. Rendi-Wagner et al. / Vaccine 25 (2007) 927–931
these subjects lack the ability to build up protective antibody levels a priori or, if titers have decreased at a higher rate compared to the other vaccine recipients. When performing regression analysis for log titer values it became, however, evident that higher age and male gender were significantly associated with lower antibody concentration. This observation confirms the significant effect of age and gender on the immune response, at least in terms of GMTs, as also described for other vaccines [13]. However, since most of the observed antibody levels still proved to be far above detection limit, thus protective, this observation appears to be rather insubstantial regarding long-term protection. According to the consensus statement of “The International Consensus Group on Hepatitis A Virus Immunity”, HAV booster vaccination after a complete primary vaccination course in healthy individuals seems unnecessary, since evidence is accumulating that HAV vaccination elicits immunological memory persisting even after the loss of detectable antibodies [5,20–22]. This is supported by experimental studies in chimpanzees showing that persistence of detectable anti-HAV antibodies is not absolutely required for protective immunity and thus indicating that also cellular mechanisms play an important role [23–25]. In this respect current investigations in a subgroup of our study population will evaluate functional as well as phenotypical characteristics of cellular immune responses prior and after HAV booster vaccination. Taken together, our study is one of the few long-term studies showing that after a full course of primary hepatitis A vaccination protective immunity exceeds by far the 10-year booster interval, currently advised by the manufactures’ prescribing information and most health authorities. Our data clearly support the modification of vaccination recommendation as previously defined by the International Consensus Group on Hepatitis A Virus Immunity.
Acknowledgments The skilful technical assistance and administrative support of Dr. Eva Jeschko, Erika Garner-Spitzer, and Brigitte Laaber is gratefully acknowledged. This study was financially supported by the Centre for Travel Medicine, Vienna.
References [1] Koff RS. Hepatitis A. Lancet 1998;351:1643–9. [2] World Health Organization. Hepatitis A vaccines. Weekly Epidemiol Rec 2000;75:38–44. [3] Mutsch M, Spicher VM, Gut C, Steffen R. Hepatitis A virus infections in travelers, 1988–2004. Clin Infect Dis 2006;42:490–7. [4] Jong EC. Risks of hepatitis A and B in the traveling public. J Travel Med 2001;8(Suppl 1):S3–8.
931
[5] Van Herck K, Van Damme P, Lievens M, Stoffel M. Hepatitis A vaccine: indirect evidence of immune memory 12 years after the primary course. J Med Virol 2004;72:194–6. [6] Bell BP. Hepatitis A vaccine. Pediatr Infect Dis J 2000;19:1187–8. [7] Andr´e F, Van Damme P, Safary A, Banatvala JE. Inactivated hepatitis A vaccine: immunogenicity, efficacy, safety, and review of official recommendations of use. Expert Rev Vaccines 2002;1:9–23. [8] Wiedermann G, Kundi M, Ambrosch F, Safary A, D’Hondt E, Delem A. Inactivated hepatitis A vaccine: long-term antibody persistence. Vaccine 1997;15:612–5. [9] Wiedermann G, Kundi M, Ambrosch F. Estimated persistence of antiHAV antibodies after single dose and booster hepatitis A vaccination (0–6 schedule). Acta Trop 1998;69:121–5. [10] Van Herck K, Beutels P, Van Damme P, Beutels M, Van den Dries J, Briantais P, et al. Mathematical models for assessment of long-term persistence of antibodies after vaccination with two inactivated hepatitis A vaccines. J Med Virol 2000;60:1–7. [11] Bovier PA, Bock J, Loutan L, Farinelli T, Glueck R, Herzog C. Longterm immunogenicity of an inactivated virosome hepatitis A vaccine. J Med Virol 2002;68:489–93. [12] Van Herck K, Renard D, Molensberghs G, Van Damme P. Model-based estimates of long-term persistence of vaccine-induced hepatitis A antibodies. In: Margolis HS, Alter MJ, Liang TJ, Dienstag JL, editors. Viral hepatitis and liver disease. London: International Medical Press; 2002. p. 56–9. [13] Rendi-Wagner P, Kundi M, Stemberger H, Wiedermann G, Holzmann H, Hofer M, et al. Antibody-response to three recombinant hepatitis B vaccines: comparative evaluation of multicenter travel-clinic based experience. Vaccine 2001;19:2055–60. [14] Wiedermann G, Ambrosch F, Andr´e FE, Hondt ED, Delem A, Safary A. Persistence of vaccine-induced antibody to hepatitis A virus. Vaccine 1992;10(Suppl. 1):S129–31. [15] Holzmann H, Croy C, Kundi M, Popow-Kraupp T. Quantitative determination of hepatitis A virus antibodies in vaccinees: is it necessary? Proceedings of the third annual meeting of the European society for clinical virology (ESCV), September 1–5. 1999 [Abstract 59]. [16] Frisch-Niggemeyer W, Kunz C. The incidence of antibodies to hepatitis A in people from Vienna and certain Austrian provinces. Wien Klin Wochenschr 1979;91:230–3. [17] Prodinger WM, Larcher C, Solder BM, Geissler D, Dierich MP. Hepatitis A in Western Austria—the epidemiological situation before the introduction of active immunisation. Infection 1994;22:53–5. [18] Alper CA, Kruskall MS, Marcus-Bagley D, Craven DE, Katz AJ, Brink SJ. Genetic prediction of non-response to hepatitis B vaccine. N Engl J Med 1989;321:708–12. [19] Hollinger FB. Factors influencing the immune response to hepatitis B vaccine, booster dose guidelines, and vaccine protocol recommendations. Am J Med 1989;87(Suppl 3A):36–40. [20] Van Damme P, Banatvala J, Fay O, Iwarson S, McMahon B, Van Herck K, et al. Hepatitis A booster vaccination: is there a need? Lancet 2003;362:1065–71. [21] Van Herck K, Van Damme P. Inactivated hepatitis A vaccine-induced antibodies: follow-up and estimates of long-term persistence. J Med Virol 2001;63(1):1–7. [22] Werzberger A, Kuter B, Nalin D. Six years follow-up after hepatitis A vaccination. N Engl J Med 1998;338(16):1160. [23] Lemon SM. Immunologic approaches to assessing the response to inactivated hepatitis A vaccine. J Hepatol 1993;18(Suppl. 2):15–9. [24] Cederna JB, Klinzman D, Stapleton JT. Hepatitis A virus-specific humoral and cellular immune responses following immunization with a formalin-inactivated hepatitis A vaccine. Vaccine 2000;18:892–8. [25] Purcell RH, D’Hondt E, Bradbury R, Emerson SU, Govindarajan S, Binn L. Inactivated hepatitis A vaccine: active and passive immunoprophylaxis in chimpanzees. Vaccine 1992;10(Suppl 1):148–51.