Immune response to a new hepatitis B vaccine in healthcare workers who had not responded to standard vaccine: randomised double blind dose-response studyBMJ 1997; 314 doi: http://dx.doi.org/10.1136/bmj.314.7077.329 (Published 01 February 1997) Cite this as: BMJ 1997;314:329
- Jane N Zuckerman, heada,
- Caroline Sabin, lecturer in medical statistics and epidemiologyb,
- M Craig Fiona, project managerc,
- A Williams, director of clinical developmentc,
- Arie J Zuckerman, professor of medical microbiologyd
- a Academic Unit of Travel Medicine and Vaccines, Royal Free Hospital School of Medicine, London NW3 2PF
- b Department of Primary Care and Population Sciences, Royal Free Hospital School of Medicine, London NW3 2PF
- c Medeva Scientific and Regulatory Affairs, Evans House, Regent Park, Leatherhead, Surrey KT22 7PQ
- d WHO Collaborating Centre for Reference and Research on Viral Diseases, Royal Free Hospital School of Medicine, London NW3 2PF
- Correspondence to: Dr J N Zuckerman
- Accepted 15 November 1996
Objective: To evaluate the immunogenicity and reactogenicity of a new triple S recombinant hepatitis B vaccine in a cohort of healthy people in whom currently licensed hepatitis B vaccines had persistently not induced an immune response.
Design: Single centre, randomised, double blind, dose-response study.
Setting: Research vaccine evaluation centre at a teaching hospital.
Subjects: 100 healthcare workers aged 18-70 years with a history of failure to seroconvert after at least four doses of a licensed hepatitis B vaccine containing the S component.
Intervention: Each subject was randomly allocated two doses of 5, 10, 20, or 40 µg of a new hepatitis B vaccine two months apart.
Main outcome measures: Immunogenicity of the four doses. Seroconversion and seroprotection were defined as an antibody titre >10 IU/l and >100 IU/l respectively against an international antibody standard.
Results: 69 subjects seroconverted after a single dose of the vaccine. After the booster vaccination one other subject seroconverted, bringing the overall seroconversion rate to 70%. Fifteen subjects given 5 µg of vaccine, 19 given 10 µg, 16 given 20 µg, and 20 given 40 µg seroconverted. Seroconversion rates in the four antigen dose groups were 60% (15/25), 76% (19/25), 64% (16/25), and 80% (20/25). After the booster dose there was no significant dose-response effect on the overall seroconversion rate, although the small sample size meant that a clinically important dose-response could not be ruled out.
Conclusion: A single dose of 20 µg of the vaccine was as effective as two doses of either 40 µg or 20 µg of this vaccine formulation in terms of seroconversion, seroprotection, and geometric mean titres.
Up to 15% of healthy people do not respond to currently licensed hepatitis B vaccines
Incorporation of the pre-S1 and pre-S2 components with the S antigen overcame this non-response in 69% of healthcare workers with a history of persistent non-response to conventional hepatitis B vaccines
Significantly higher geometric mean titre levels were obtained with increased dosage of vaccine
A single dose of 20 µg of the new vaccine seems to be effective in terms of seroconversion, seroprotection, and geometric mean titres
Hepatitis B and its sequelae, which include chronic liver disease, cirrhosis, and hepatocellular carcinoma, is a major public health problem throughout the world. Many millions of people are estimated to become infected every year worldwide, and about 350 million chronic carriers constitute the primary reservoir of infection. Hepatitis B is transmitted primarily by blood to blood contact as well as sexually.
Systematic vaccination of individuals at risk of exposure to the virus has been the main method of controlling the morbidity and mortality associated with hepatitis B. The first hepatitis B vaccine was manufactured by the purification and inactivation of hepatitis B surface antigen obtained from the plasma of chronic hepatitis B virus carriers.1 2 3 This was soon followed by the production of hepatitis B surface antigen using recombinant DNA techniques and expression of the S component in yeast cells.4
All studies of the antibody response to currently licensed plasma derived hepatitis B vaccines and hepatitis B vaccines prepared by recombinant DNA technology have shown that between 5% and 10% or more of healthy immunocompetent subjects do not mount an antibody response to the surface antigen component present in these preparations (non-responders) or that they respond poorly (hyporesponders).5 6 7 8
The exact proportion depends partly on the definition of non-responsiveness or hyporesponsiveness, generally less than 10 IU/l or 100 IU/l respectively, against an international antibody standard.
Non-responders remain susceptible to infection with hepatitis B virus.9 Several factors adversely affect the antibody response to hepatitis B surface antigen, including the site and route of injection, sex, advancing age, being overweight, immunosuppression, and immunodeficiency, but the mechanisms underlying non-responsiveness to the S component of hepatitis B surface antigen in humans remain largely unexplained. However, evidence is accumulating that different HLA-DR alleles are associated with specific low responsiveness in different ethnic populations. Considerable experimental evidence is available that the ability to produce antibody in response to specific protein antigens is controlled by dominant autosomal class II genes of the major histocompatibility complex in mice.10 11 12 Much effort has been devoted to overcoming class II linked non-responsiveness to current hepatitis B vaccine.13 14 15
The pre-S1 and pre-S2 domains have an important immunogenic role in augmenting hepatitis B surface antibody responses, preventing the attachment of the virus to hepatocytes and eliciting antibodies that are effective in clearing viruses, stimulating cellular immune responses, and circumventing genetic non-responsiveness to the S antigen.14 15 16 17 18 Thus several studies indicate that pre-S components should be included in new recombinant or synthetic vaccines. For example, the pre-S2 region is more immunogenic in T and B cells than are the S regions in mice,13 19 as is the case with pre-S1 in mice14 and humans,20 and circumvents S region non-responsiveness to antibody production.
Indeed, Milich et al showed that in mice the independent genetic regulation of immune responses to pre-S1, pre-S2, and S regions of hepatitis B surface antigen that are linked to the major histocompatibility complex would assure fewer genetic non-responders to a vaccine containing all three antigenic regions. Studies in humans of experimental recombinant hepatitis B vaccines containing all three S components of the viral envelope polypeptides showed enhanced immunogenicity of such preparations compared with conventional yeast derived vaccines.21 22 23 We investigated the immunogenicity of a new hepatitis B vaccine containing the three antigenic components of both surface antigen subtypes adw and ayw in a group of healthcare staff who persistently showed no response to the conventional vaccine.
Subjects and methods
The vaccine used was manufactured by Evans Medical, a subsidiary of Medeva (Leatherhead, Surrey). The vaccine is a third generation vaccine containing pre-S1, pre-S2, and S antigenic components of both viral surface antigen subtypes adw and ayw. All three antigenic components are produced in a continuous mammalian cell line, the mouse c127 clonal cell line, after transfection of the cells with recombinant hepatitis B surface antigen DNA.24 25 The vaccine was presented as an aluminium hydroxide adjuvant preparation of purified antigenic proteins in 1.0 ml of isotonic saline.
The study was a single centre, randomised, double blind dose-response study using four doses (5, 10, 20, and 40µg) of the vaccine. Subjects were allocated randomly to receive in a double blind manner one of the four antigen doses. The same antigen dose of vaccine was given on two occasions intramuscularly into the deltoid two months apart. The primary objectives were to assess the immunogenicity and reactogenicity of the four doses of the vaccine and the kinetics of the immune response. A secondary objective was to evaluate the hepatitis B surface antibody response to vaccination. Non-seroconversion was defined as the presence of <10 IU/l hepatitis B surface antibody.
The sample size calculation assumed a 5% placebo response, although such a response was considered to be unlikely. The calculated group sizes necessary to have 90% power to detect a 45% response in the active groups (P<0.05; two tailed) gave group sizes of 23, which were made up to 28 to allow for up to five subjects dropping out in each treatment group.
Before inclusion each subject was screened for the presence of hepatitis B core and surface antibodies, and haematological and biochemical profiles were measured; women of childbearing age had a pregnancy test. Subjects were excluded if they were seropositive for hepatitis B core antibodies or had abnormal results in liver function tests. Fourteen days after the first visit each subject was then randomly allocated by a double blind method to receive one of the four antigen doses of the vaccine; 25 subjects were included in each dose group. Owing to minor discrepancies between the hepatitis B antibody titres obtained specifically at screening for this study and those previously documented in subjects' study files at the Royal Free Hospital School of Medicine both an intention to treat and per protocol population arose.
The intention to treat population included all randomised subjects who received the first vaccination at their second visit. The per protocol population consisted of the patients in the intention to treat population who did not violate any of the specified exclusion criteria. The analysis reported in this paper is based on the intention to treat population. However, when the analysis was repeated on the per protocol population the conclusions were essentially unchanged.
A standard physical examination was carried out on screening and at two and six months after the initial dose. Serum was collected for antibody evaluation at these times. The proportions of people who seroconverted and who were seroprotected at six months were considered to be primary end points. Geometric mean titres of hepatitis B surface antibody at both time points and seroconversion and seroprotection rates at two months were considered to be secondary end points.
Every subject was observed for 15 minutes after each vaccination, and any local or systemic reactions were recorded. Side effects and body temperatures were recorded on diary cards on the day of vaccination and for three days thereafter. Symptoms were divided into local, general, or other and were graded according to severity.
The seroconversion rate and geometric mean titres were measured to evaluate the immunogenicity in each group for all time points at which blood samples were taken.
Seroconversion was defined as the presence of hepatitis B surface antibody titres >10 IU/l while hepatitis B surface antibody titres of >100 IU/l were considered to be seroprotective. The serological hepatitis B surface antibody titrations using an international standard were carried out at the Royal Free Hospital School of Medicine using a commercial kit (Bioelisa, Biokit, Barcelona, Spain). The cut off point was in the range of 0-10 IU/l. Only subjects below 10 IU/l on the basis of this test were considered to be non-responders and were enrolled in the study.
The hepatitis B surface antibody titrations were repeated at the end of the study on all the serum samples which were maintained frozen at -20°C using another commercially available test kit (Ausab, Abbott Laboratories, North Chicago, USA).
The proportions of subjects seroconverting and protected at two and six months in each dose group were studied by logistic regression. Odds ratios and 95% confidence intervals are presented for the effect of a doubling in the dose of vaccine given–that is, an increase from 5 to 10 µg and from 10 to 20µg, etc. Serological responses among the four dose groups, allowing for the presence of a trend in the response according to the dose of vaccine given, were further compared by linear regression with logarithmically transformed data. Geometric mean titres with confidence intervals for each dose group are presented. Because of the wide variation in the within individual differences in titres between months 2 and 6, the sign rank test was used to assess whether there was any overall change in titres after the booster dose.
One hundred subjects were entered into the study, 25 in each of the four antigen dose groups. At screening, 14 subjects had hepatitis B surface antibody titres of >10 IU/l and were not included in the per protocol population. Of the 86 per protocol subjects, 22, 23, 20, and 21 subjects were entered into the four antigen dose groups respectively. All 100 subjects completed the study and received two doses of the vaccine.
The median age at entry was 38 years, and the median number of immunisations each subject had received of commercially available hepatitis B vaccines containing the S component was 5. There were no significant differences in age or sex distribution within the four antigen dose groups. Table 1 summarises the demographic details of the volunteers.
Fifteen per cent of subjects in each dose group experienced a reaction after their first vaccination. Overall, after each vaccination the incidence of local symptoms was higher than that of general symptoms. Most local symptoms were classed as either mild or moderate in severity, with only 11 severe events–predominantly pain at the site of injection–being reported for the 200 doses of vaccine. The incidence of these across the four antigen dose groups was 2, 4, 3, and 2 respectively.
General symptoms, including chills, dizziness, headache, nausea, and diarrhoea, occurred within 24 hours of immunisation and were transient. No serious adverse events occurred during the study and no dose related incidence was seen for either local or general symptoms.
Two months after a single dose of vaccine 69 subjects seroconverted with a hepatitis B surface antibody titre of >10 IU/l. The seroconversion rates for each of the four antigen dose groups ranged from 52% to 84% in the four dose groups (table 2). The overall rate of seroconversion four months after the booster dose was 70%. Seroconversion rates in the four dose groups were 60%,76%, 64%, and 80% respectively.
Although a significant dose-response effect on seroconversion rates was seen at two months (P=0.03) with a 56% increase in the seroconversion rate for each doubling in the dose of vaccine given, by six months this effect had become non-significant (P=0.24). However, the confidence interval for the odds ratio was wide and the upper limit of the confidence interval suggested that an odds ratio as high as 1.86 could be consistent with the trial results.
The overall rate of arbitrary seroprotection, defined as the presence of hepatitis B surface antibody titres of >100 IU/l, was 27% (27/100) after the first immunisation and 26% (26/100) after the booster. At two months seroprotection rates ranged from 12% to 44% in the four dose groups. At six months they were 4%, 24%, 40%, and 36% respectively (table 3). There was a significant trend with dose at both two and six months (P=0.02 and 0.006 respectively), although it seems that the booster dose may be less effective in achieving a sustained serological response, as several subjects who were protected at two months had lost their protection by six months.
There was a significant dose-response effect on geometric mean titre levels at two months (P=0.002) (table 4). Administration of a booster dose of the vaccine at two months did not produce a corresponding rise in geometric mean titres within any of the antigen dose groups. Within subjects, variation was wide in the change in titres between the two visits, with changes ranging from a drop of 16 800 IU/l to an increase of 6600 IU/l in two subjects with initially high titres. However, the median change was zero and there was no systematic pattern to the changes (P=0.19, sign rank test). Despite this, the strong dose-response effect remained at six months (P=0.009) (table 4).
There were no significant differences in the hepatitis B surface antibody titres obtained by the two different assay systems used.
Factors affecting the immune response
Inactivated hepatitis B vaccines have been available since 1981; they are derived from plasma or from yeast recombinant DNA. These vaccines have been shown to be both immunogenic and safe, but 5-10% of healthy people do not respond to them 6 7 8 26 and variants of hepatitis B virus that are not neutralised by vaccine induced hepatitis B surface antibody have recently emerged. Several factors play a part in the failure to mount an antibody response to hepatitis B surface antigen. These include the site of injection, the deltoid area being preferred to the buttocks as fat lacks antigen presenting cells, resulting in a delay in presentation of antigen to T and B cells27; increasing age; sex; smoking; being overweight; immunosuppression; and immunogenetic makeup. However, after the identification of an immunodominant domain in the pre-S2 region of hepatitis B surface antigen28 and the observation that immunologically non-responsive mice developed antibodies corresponding to the pre-S epitope, vaccines have been prepared containing all three antigenic components–pre-S1, pre-S2, and S.
Guidelines for immunisation
The recommendations for immunisation against hepatitis B are well known, and healthcare workers are at occupational risk of exposure to hepatitis B virus. A proportion are non-responders and remain susceptible to infection. Recent guidelines from the United Kingdom Department of Health on the immunisation of healthcare workers has led to improvements in immunisation programmes and consequently uptake of hepatitis B vaccine.29 In this study we evaluated the immunogenicity and reactogenicity of a new vaccine in healthcare workers who had previously failed to develop an immune response after multiple doses of currently available hepatitis B vaccines. The new vaccine preparation produced a response in a group of persistent non-responders, with an overall seroconversion rate of 69% after a single dose, and a significant dose-response effect was seen. After a further four months, however, there was no significant dose effect on the proportion of subjects who seroconverted–that is, hepatitis B surface antibody titre >10 IU/l in the four antigen dose groups. This suggests that it may be as effective to use a single dose of either 20 µg or 40 µg than two doses of, say, 10 µg or 20 µg. Although the dose-response effect six months later was not significant, seroconversion rates were highest in the group receiving the highest dose of vaccine (80%), and the confidence interval for the odds ratio suggests that we cannot rule out the possibility of an important dose-response effect. This was a phase II clinical trial and as such, several questions remain to be answered concerning the precise schedule of vaccination and the groups for whom it would be advocated. Phase III comparative clinical trials are currently being undertaken.
Seroprotection and seroconvertion rates
Although the aim of this study was to determine the immunogenicity of this vaccine in non-responders, 14 subjects included in the intention to treat population had a hepatitis B surface antibody titre on screening of >10 IU/l but <100 IU/l. Thus, the seroconversion rates quoted may be higher than expected. However, when the analysis was repeated including only the per-protocol population, seroconversion rates were only slightly lower (64% and 66% at two months and six months respectively), and the conclusions remained unchanged.
A few subjects (16% (5/31)) whose antibody titre remained below 10 IU/l after one dose of the vaccine did seroconvert after the booster dose. We could not determine the cause of this delayed immune response. An unexpected observation was that the administration of a booster dose did not significantly enhance the immune response in the patients overall. However, there was wide individual variability both in the titres at two and six months and in the change in titres after the booster dose.
No empirical data are available for the hepatitis B surface antibody titre required for protection against particular routes of infection or the size of the infectious inoculum. The minimal protective titre has been assumed almost universally to be 10 IU/l, and immunological memory is thought to ensure protection even after circulating antibody becomes undetectable.3 5 29 30 31 32 In the United Kingdom 100 IU/l is regarded as the desired antibody titre for complete seroprotection after immunisation of normal healthy subjects. Nevertheless, the need for booster inoculation after decay in antibody titres is still the subject of debate internationally.
The results of this trial have been based on the measurements of hepatitis B surface antibody titres but the pre-S1 and pre-S2 antibody response was not measured. Further analysis of the pre-S1 and pre-S2 antibody response will be undertaken when reproducible assays become available and have been validated.
A proportion of subjects (34% (30/86)) did not mount a hepatitis B surface antibody response after two doses of the vaccine. These subjects will receive a further dose of the vaccine and be followed up. Studies were also undertaken to investigate the relation between the humoral response to hepatitis B vaccine and immunogenetic profiles in this population group, the results of which have been submitted for publication.
We thank the doctors and others who referred non-responders to us, and the volunteers, without whom this study could not have been undertaken. The laboratory tests were carried out by the staff of the department of virology and the WHO Collaborating Centre for Reference and Research on Viral Diseases at the Royal Free Hospital and School of Medicine.
Funding: The study was supported by research grants available to the Academic Unit of Travel Medicine and Vaccines, the WHO Collaborating Centre for Reference and Research on Viral Diseases, the Violet Richards Charitable Trust; study costs were paid for by Medeva.
Conflict of interest: None.