Immunogenicity and Tolerability of Recombinant Serogroup B Meningococcal Vaccine Administered With or Without Routine Infant Vaccinations According to Different Immunization Schedules:  A Randomized Controlled Trial FREE

Meningococcal disease occurs with an incidence of 0.2 to 14 per 100 000 population in industrialized countries.¹ Effective vaccines based on capsular polysaccharides for meningococcal serogroups A, C, W-135, and Y are available,² but serogroup B capsular polysaccharide is antigenically similar to human neural cell glycopeptides containing polysialic acid and poorly immunogenic in humans.

Serogroup B Neisseria meningitidis (MenB) vaccines based on detergent-extracted outer membrane vesicles (OMVs) containing outer membrane proteins, including porin protein A, proved effective in clinical trials and controlled a clonal MenB outbreak in New Zealand.^3- 4 However, the high strain specificity of these vaccines limits their usefulness, especially in infants and young children.^5- 6

To develop a vaccine with broader protection, novel antigens identified by sequencing the MenB genome—factor-H binding protein variant 1 (fHbp1), Neisserial adhesin A (NadA), and Neisseria heparin binding antigen (NHBA)—were combined with OMVs from the New Zealand epidemic strain NZ98/254 in a multicomponent serogroup B meningococcal vaccine (4CMenB).⁷ To facilitate manufacture and enhance protein stability, fHbp1 and NHBA were combined with other proteins identified during genome sequencing to create 2 fusion proteins.^8- 9 Small studies in infants demonstrated immunogenicity of 4CMenB against reference strains.^8- 9

In this study we assessed the immunogenicity and reactogenicity of 4CMenB in a large cohort of infants when given in 2 different schedules, and concomitantly or separately from routine vaccines.

This phase 2b, open-label, parallel-group, randomized controlled trial was conducted between August 2008 and July 2010 in Belgium (6 centers), United Kingdom (4 centers), Czech Republic (4 centers), Germany (25 centers), Italy (5 centers), and Spain (16 centers). Appropriate regulatory authorities in each participating country granted approval for the trial. Participants were healthy, full-term, 2-month-old infants with a birth weight of 2.5 kg or greater, whose parent(s) or legal guardian(s) provided written informed consent. In the United Kingdom, participants were recruited by information letters sent via National Health Service child health computer databases; participants in other study centers were recruited through pediatric hospitals or private practices. Exclusion criteria were any history of meningococcal infection or close contact with someone experiencing meningococcal disease; prior receipt of meningococcal B or C vaccines or any of the nationally recommended routine infant vaccines; acute or chronic illness; known reactions to vaccine components; known or suspected immune disease or impairment, including the administration of steroids; receipt of antibiotics 6 days prior to enrollment; receipt of blood products; or planned receipt of nonstudy vaccines. Investigator-defined options for race/ethnicity were chosen by the participants' parents for demographic description of the study population.

The 4CMenB vaccine consisted of 50 μg each of fHbp1, NadA, and NHBA fusion proteins, 25 μg of detoxified OMV from N meningitidis strain NZ98/254, 1.5 of mg aluminum hydroxide, and histidine 10mM in 0.5 mL of water for injection. Participants also received a combined diphtheria, tetanus, acellular pertussis, inactivated polio, hepatitis B, and Haemophilus influenzae type b vaccine (DTaP-HBV-IPV/Hib) (Infanrix Hexa; GlaxoSmithKline) and 7-valent pneumococcal glycoconjugate vaccine (PCV7) (Prevnar; Wyeth Pharmaceuticals). These are referred to as routine vaccines, although they do not necessarily reflect the routine vaccines used in each study country.

The study design allowed assessment of 3 primary 4CMenB schedules: ages 2, 4, and 6 months, together with routine infant vaccines (concomitant); 2, 4, and 6 months, with routine vaccines given separately at 3, 5, and 7 months (intercalated); and 2, 3, and 4 months, concomitantly with routine infant vaccines (accelerated). A control group received DTaP-HBV-IPV/Hib and PCV7 only at 2, 3, and 4 months. These immunization time points were chosen to reflect schedules currently used in different countries. All vaccines were administered by intramuscular injection in the anterolateral thigh; 4CMenB and routine vaccines given concomitantly were administered in opposite limbs.

Participants were randomized to the 4 vaccination groups in a 2:2:1:1 ratio according to a centrally held randomization list (Biostatistics and Clinical Data Management, Novartis) using a fixed block size of 6. The unequal allocation ratio was chosen to efficiently assess primary as well as secondary objectives. This was an open-label study; ie, study personnel and parents both knew which vaccines were being administered.

Sera obtained before the first vaccination and 30 (−4/+7) days after the third dose of 4CMenB (or routine vaccines in the control group) were shipped to the Novartis Vaccines Serology Laboratory, Marburg, Germany, for blinded immunology assessments. Human complement serum bactericidal activity (hSBA) was assessed against 3 target strains chosen to determine the immunogenicity of individual vaccine components—strain 44/76-SL for fHBP1, strain 5/99 for NadA, and strain NZ98/254 for OMV. The hSBA was expressed as interpolated titers according to reciprocal serum dilutions yielding 50% or greater killing of the target strain after 60 minutes of incubation compared with growth at time 0. An interpolated titer of 1:5 or greater represented 95% confidence that participants achieving this titer had a protective hSBA titer (≥1:4).¹⁰

Because no hSBA strain specific for NHBA was available at the time of analysis, antibodies to NHBA were measured by enzyme-linked immunosorbent assay and expressed as geometric mean concentrations (GMCs).

Responses to routine vaccines were determined for participants in the accelerated and control groups only, according to availability of serum, using standard correlates of protection to interpret responses. Geometric mean titers (GMTs) of hSBA, routine vaccine antigen-specific IgG GMCs, and the between-group ratios of GMT and GMC were calculated (with 2-sided 95% CIs) using 2-way analysis of variance with a factor for vaccine group and immunization center.

Parents recorded local injection site reactions (pain, erythema, swelling, and induration) and systemic reactions (ie, fever [axillary temperature ≥38°C], change in eating habits, sleepiness, unusual crying, vomiting, diarrhea, irritability, rash) for 7 days after each vaccination. Injection site pain was classified by the parents as mild (minimal discomfort when the child's leg was touched), moderate (obvious discomfort when the leg was touched), or severe (pain on movement of the leg). Erythema, induration, and swelling were summarized by maximal severity (1-25 mm, 25-50 mm, and >50 mm, respectively).

Adverse event recording was enhanced by telephone contact in the week after study vaccination. Safety follow-ups were completed 6 months after the last dose of 4CMenB (and at age 10 months in the control group). All serious adverse events reported during the study were recorded. Determination of the relationship between adverse events and the study vaccine was made by a study investigator's judgment based on temporal relationship and biological plausibility criteria.

The primary outcome was assessment of percentages of participants with hSBA titers of 1:5 or greater 1 month after immunization with 4CMenB in the concomitant and accelerated groups, vaccine response being sufficient if the lower limit of the 2-sided 95% Clopper-Pearson¹¹ CI of this percentage was 70% or greater for all 3 reference strains. In addition, hSBA titers were log transformed and their GMTs and 2-sided 95% CIs calculated. The prespecified population for this analysis was the modified intention-to-treat population, including only those who received a study vaccine and provided evaluable serum samples before and after immunization.

The main secondary outcomes were noninferiority of immune responses to 4CMenB and routine vaccines administered together compared with separately. The percentage of 4CMenB recipients achieving hSBA titers of 1:5 or greater in the concomitant group, obtained from a categorical linear model, was considered noninferior to that in the intercalated group if the lower limits of the 95% CI for the differences in these percentages (concomitant minus intercalated) was greater than −10%.¹² The same criteria were used to determine noninferiority between the control and accelerated groups for the response to routine vaccines.

Additionally, a post hoc noninferiority criterion for the geometric mean ratios of 4CMenB response (intercalated divided by concomitant) obtained from analysis of variance adjusting for center was defined as the lower limit of the 95% CI being greater than 0.5 for all 3 strains. For pertussis antigens, noninferiority of the accelerated compared with the control group was achieved if the lower limit of the 2-sided 95% CI for the ratio of the antigen-specific IgG GMCs after vaccination was greater than 0.67. All antigen thresholds were chosen based on traditional use and acceptance by regulatory authorities. Responses to antigens administered at the same time were analyzed in a per-protocol population, ie, those who received all designated vaccines on time and provided sera within the scheduled window (23-55 days post vaccination) and who received no prohibited medications (systemic antibiotics or systemic/high-dose inhaled corticosteroids).

Modified intention-to-treat analysis was performed as per-randomized group; the per-protocol population was analyzed as per-immunization course received.

A post hoc analysis was performed that used all available data and included imputed values to replace missing data.¹³ Each missing value was replaced with a set of plausible values that represent the uncertainty about the correct values to impute. Variables included in the imputation model were actual vaccine administered, center, age, birth weight, weight, height, sex, race, hSBA titers for strains 44/76-SL, 5/99, and NZ98/254 (at baseline and following the third dose), and IgG GMC against NHBA using the adjusted Wald test proposed by Agresti and Coull.¹⁴

The sample size calculation for the primary outcome was determined using a simulation approach to estimate that 300 participants in the accelerated group would yield 95% power to demonstrate that the lower limit of the 2-sided 95% CI for the true percentage of participants with hSBA titer of 1:5 or greater was 70% or greater. A sample size of 600 in the concomitant and intercalated groups was calculated to provide 87% power to demonstrate noninferiority in response rates for all 3 MenB strains (assuming only 80% provided evaluable sera, response rates to 4CMenB given together with routine immunizations were similar to those in previously published studies,^8- 9 and that response rates to 4CMenB given without routine immunizations were 2% higher for each strain).

Individuals who received at least 1 dose of vaccine and provided postbaseline safety data were included in the safety analyses, for which results were reported descriptively with no formal statistical analyses.

Statistical analysis was performed using SAS version 9.1.

A total of 1885 participants were enrolled (United Kingdom, 561; Italy, 371; Germany, 317; Czech Republic, 283; Belgium, 248; Spain, 105), of whom 1810 (96%) completed the study. Of these, 1636 were included in the modified intention-to-treat immunogenicity analysis (Figure 1) and 1599 in the per-protocol immunogenicity analysis. Median age at enrollment was 68.7 days for all groups, 48.3% to 53.4% of each group were boys, and 93.1% of all participants were white.

Primary Outcome. After immunization with 4CMenB and routine vaccines together at either 2, 4, and 6 or 2, 3, and 4-months, 99% or more of participants had hSBA titers of 1:5 or greater for strains 44/76-SL and 5/99 in the modified intention-to-treat analysis (Figure 2 and Table 1). For strain NZ98/254, percentages were 79.0% (417/528 [95% CI, 75.2% to 82.4%]) and 81.7% (219/268 [95% CI, 76.6% to 86.2%]) following concomitant or accelerated schedules. Lower limits of the 95% CIs were greater than 70% for all 3 tested reference strains, thereby meeting the predefined criteria of a sufficient immune response. Results from multiple imputation analysis were similar (Table 1).

Secondary Outcomes. The difference in the percentage of participants with hSBA titers of 1:5 or greater in the concomitant group minus the intercalated group met the prespecified noninferiority criterion for 44/76-SL and 5/99 but not NZ98/254 (−7.1% [95% CI, −11.7% to −2.6%]) (Table 1). Although the lower limit of the 95% CI for the geometric mean concomitant to intercalated hSBA ratios were all above 0.5, thus meeting the prespecified noninferiority criterion, it is notable that the upper limit of the 95% CI for these ratios were all less than 1 (Table 2).

The geometric mean ratios (concomitant divided by accelerated) of hSBA titers were similar for strains 44/76-SL and NZ98/254 (confidence intervals crossing 1) but higher for strain 5/99 (1.61 [95% CI, 1.41 to 1.84]). IgG GMCs against NHBA 1 month after the third 4CMenB vaccination were 4342 (95% CI, 4067-4635) when 4CMenB was administered separately from routine vaccines, compared with 3211 (95% CI, 2949-3495) to 3332 (95% CI, 3120-3558) when administered together (Table 3).

The percentages of participants achieving the threshold of response for DTaP-HBV-IPV/Hib components were noninferior in the accelerated group compared with the control group (lower limits of 95% CI of the differences in these values being greater than −10%), with the exception of pertactin (difference, −4% [95% CI, −11% to 3%]). Noninferiority was demonstrated for 6 of the 7 serotypes in the pneumococcal vaccine (serotypes 4, 9V, 14, 18C, 19F, 23F) but not for serotype 6B (difference, −5% [95% CI, −14% to 3%]), because the lower limit of the difference in the percentage of participants with antibody response greater or equal to the prespecified cutoff level was −14% (Table 4).

The numbers of participants experiencing adverse reactions after immunization are shown in the eTable. Throughout the study, fewer than 1% of all participants experienced severe erythema, swelling, or induration at either of the vaccination sites. However, 12% to 16% in the concomitant and accelerated groups, respectively, experienced severe local pain after a dose of 4CMenB, compared with 1% to 3% after doses of DTaP-HBV-IPV/Hib or PCV7 in the control group.

Fever (≥38.0°C) after any vaccination was described in 80% (501/624) in the concomitant group and 76% (243/318) in the accelerated group, compared with 51% (160/311) in control group and 71% (447/627) in the intercalated group. For the intercalated group there were twice as many immunization days, and therefore more opportunity for fever to be experienced. Rates of fever per 4CMenB dose are shown in the eTable.

Over the 8- to 10-month study, 166 serious adverse events were reported in 158 participants: 63 (10%) in the concomitant group, 57 (9%) in the intercalated group, 19 (6%) in the accelerated group, and 19 (6%) in the control group. Of these, 20 (9 concomitant, 7 intercalated, 3 accelerated, and 1 control) were thought to be possibly related to 4CMenB or routine vaccinations.

Four possibly related serious adverse events were seizures, 1 each following routine vaccination in the intercalated and accelerated groups, and 2 following vaccination with 4CMenB in the intercalated group, 1 of the latter being a febrile seizure 2 days after the second 4CMenB dose. Two hypotonic hyporesponsive episodes were reported, 1 within 12 hours of concomitant 4CMenB and routine vaccines and 1 within 6 hours after routine vaccines in the intercalated group. Two cases of Kawasaki disease were reported, 1 of which was considered possibly related to the study vaccine by an independent expert panel. Six children were observed in the hospital because they had experienced fever within 2 days after vaccination with 4CMenB. The remaining 5 possibly related serious adverse events were aseptic meningitis, retinal dystrophy (believed to be congenital), transient synovitis of the right hip, a transient hearing loss noted by a parent, and transient apnea after concomitant vaccination.

This study of more than 1800 infants shows that when administered together with routine vaccines to healthy infants, a primary immunization course of 4CMenB is immunogenic against 3 reference strains expressing 1 of 3 vaccine antigens. Furthermore, 4CMenB was immunogenic when administered in a 2-, 3-, and 4-month schedule, an important finding given the high rates of MenB disease in the first 6 months of life.^15- 16 These results suggest that there can be some flexibility in the incorporation of 4CMenB into the various immunization schedules used in different countries.

When administered alone in the intercalated group rather than with routine vaccines, 4CMenB elicited higher hSBA GMTs for all strains. Similarly, administration of 4CMenB in a 2-, 4-, and 6-month schedule, rather than an accelerated 2-, 3-, and 4-month schedule, resulted in higher hSBA GMTs for the 5/99 strain. However, immunization schedules need to balance the advantages of optimal immunogenicity and the practical need to minimize immunization visits and ensure early immunization. The clinical significance of any reduction in hSBA GMTs is uncertain, given that the proportion of participants with an hSBA titer of 1:5 or greater was at least 79% for all strains in all 4CMenB recipients. Lower hSBA GMTs may lead to shorter persistence of bactericidal antibodies following primary immunization, as observed with the serogroup C meningococcal vaccines^17- 19 and with the New Zealand MenB vaccine (MeNZB).^8,20 A follow-up study of the same participants assessing antibody persistence and the response to a booster dose of 4CMenB has been undertaken²¹ and will further inform the relative benefits of the different immunization schedules.

4CMenB immunogenicity in this study is consistent with previous small trials that found 4CMenB to be more broadly immunogenic when given with OMVs,^8- 9 possibly because of a synergistic or enhancing effect of the OMV. Furthermore, the OMV component of 4CMenB was used as a vaccine to control a clonal outbreak of MenB disease in New Zealand, with an apparent effectiveness of 85% in children aged 6 months to 5 years²² after inducing very similar hSBA titers against strain NZ98/254.²³ This provides confidence that the levels of functional antibody induced by this vaccine could indeed induce protection against strains expressing the vaccine antigens.

Concomitant 4CMenB impaired the immunogenicity of only 2 routine vaccine antigens: pertactin and pneumococcal serotype 6B. This is unlikely to be of clinical significance, because acellular pertussis vaccines that lack pertactin are known to be effective, and evidence suggests that pneumococcal disease resulting from the 6B serotype is lower in countries that have implemented vaccination with the PCV7 vaccine, although the putative response threshold of IgG level of 0.35 μg/mL or greater was achieved only by a relatively low percentage of the immunized population, possibly resulting from both direct protection and herd immunity.²⁴

Three strains expressing key vaccine antigens were used for hSBA analysis, demonstrating “proof of principle” immunogenicity of these antigens but allowing only limited conclusions regarding cross-protection against other naturally occurring strains. Also, data from assays using strains genetically engineered to express a range of fHbp1 subvariants²⁵ and from early phase 2 studies^8- 9 suggest that genotypic analysis of panels of invasive strains is not sufficient to predict the likely breadth of coverage of 4CMenB immunization against invasive meningococcal disease in infants.

Further information regarding the breadth of protection could be achieved by testing against a more extended panel of strains, as in previous 4CMenB infant studies,^8- 9 but the number of strains that can be tested by hSBA is limited by the volume of sera available from infants, the limited supply of suitable human complement, and the unsuitability of many strains for hSBA.²⁶ This, combined with the evolutionary and geographical variation in the target MenB antigens, makes selection of strain panels difficult.²⁷ The meningococcal antigen testing system²⁸ using an antigen-specific enzyme-linked immunosorbent assay (evaluating both immunological cross-reactivity and antigen expression) has an accuracy of 86% in predicting the hSBA response and might be suitable to assess the proportion of MenB strains likely to be covered by the vaccine on a regional basis. Recently presented data suggest that pooled sera from children immunized with the 4CMenB vaccine would be bactericidal against 78% (95% CI, 66% to 91%) of invasive strains in Europe²⁹ and against 76% (95% CI, 59% to 87%) of such strains in Australia.³⁰ However, the ability of these laboratory assays to accurately predict the effects of a MenB vaccine—possibly also on non-MenB pathogens expressing vaccine antigens—is unknown and can only be evaluated properly following implementation in a program with close surveillance.

Outer membrane vesicle vaccines (eg, MeNZB) have been administered to more than 3 million children across all age groups and are considered well tolerated and safe.²³ In trials of MeNZB to control clonal outbreaks in New Zealand, fever was the most pronounced adverse reaction,³¹ with no additional reactogenicity burden with concomitant administration of MenB and routine vaccines. In contrast, in this study 4CMenB appeared to be less reactogenic when administered separately rather than together with routine vaccines (although the study made no formal statistical comparisons of reactogenicity rates between groups). The majority of children in this study became febrile after receiving their first and second dose of concomitant 4CMenB and routine vaccines, although only 1 child in our study had a febrile convulsion (2 days following 4CMenB administration). There is therefore a need to compare the risk profile of a relatively common, but transient, postimmunization fever with the benefit of reducing the risk of an uncommon, but potentially fatal, meningococcal infection. This must be carefully communicated to parents in the event of routine implementation if the possibility of increased systemic reactogenicity is to be an accepted component in the control of this serious disease. Rates and magnitude of fever in this study are comparable with those seen with other pediatric vaccines, notably after the second dose of ASO3_B-adjuvanted split virion H1N1 influenza vaccine used in children aged 6 months to 5 years³² and whole-cell pertussis vaccine in 1 study,³³ and were actually lower than those observed in another study of the pertussis vaccine.³⁴

The open-label design could potentially bias reporting of adverse events if parents were more alert to subjective reactions such as irritability if their child received 4CMenB rather than routine vaccines. This would be less likely to influence objective measures such as temperature and extent of swelling or erythema. The 3 strains selected for hSBA analysis do not specifically evaluate the immunogenicity of NHBA, because the only strain expressing NHBA also contained porin protein A homologous to the OMV component. Concentrations of IgG specific to NHBA increased following administration of 4CMenB; however, this was also observed following administration of a vaccine containing the recombinant proteins without OMVs in early phase 2 studies. This increase in IgG GMCs was not associated with an increase in hSBA titers against NZ98/254,^8- 9 raising doubts about the bactericidal qualities of these antibodies. A naturally occurring strain allowing the assessment of NHBA-induced bactericidal activity has recently been identified³⁵ and will provide important additional information regarding the immunogenicity of this antigen.

In conclusion, 4CMenB was immunogenic, generally well tolerated, and showed minimal interference with routine vaccines in the first year of life. The flexibility in schedule allows it to be incorporated into a range of country-specific immunization schedules and for primary immunization to be completed in early infancy. If licensed, the decisions regarding vaccine introduction will require detailed assessment of potential vaccine coverage at a regional level and monitoring after implementation to determine the accuracy of such predictions. Nevertheless, this vaccine could potentially provide improved protection for infants against meningococcal disease beyond the protection provided by currently licensed vaccines.