To assess the efficacy of a 7-valent pneumococcal polysaccharide-meningococcal outer membrane protein complex conjugate vaccine (PncOMPC) against acute otitis media (AOM), 1666 infants were randomly assigned to receive either PncOMPC or control vaccine (hepatitis B vaccine) at 2, 4, 6, and 12 months of age. Of the 835 children assigned to receive PncOMPC, 187 received a 23-valent pneumococcal polysaccharide vaccine (PncPS) at 12 months of age instead. Whenever AOM was diagnosed, middle ear fluid was aspirated for bacterial culture. In the PncOMPC and control groups, there were 110 and 250 AOM episodes, respectively, in children between 6.5 and 24 months of age that could be attributed to vaccine serotypes, which indicates a vaccine efficacy of 56% (95% confidence interval, 44%-66%). The serotype-specific efficacy ranged from 37% for 19F to 82% for 9V. The 2 boosters seemed to provide equal protection against AOM, but PncPS induced markedly higher antibody concentrations. The efficacy of PncOMPC was comparable to that of the recently licensed pneumococcal conjugate vaccine.
Prevention of pneumococcal infections remains a public health challenge. Although the highest rates of pneumococcal disease occur in children <2 years of age [1–3], an effective vaccine for prevention of disease in this population has long been lacking. The situation is now rapidly changing. The recently licensed pneumococcal conjugate vaccine containing 7 pneumococcal polysaccharides conjugated to protein CRM197 (PncCRM) has shown excellent efficacy against invasive disease  and moderate efficacy against acute otitis media (AOM)  caused by the serotypes included in the vaccine. Other pneumococcal conjugate vaccines with different carrier proteins have also been tested in clinical trials [6–8], but no data on their efficacy have been published yet.
The Finnish Otitis Media (FinOM) Vaccine Trial offered us a unique opportunity for parallel evaluation of the efficacy of 2 pneumococcal conjugate vaccines in prevention of serotype-specific pneumococcal AOM in young children. We have previously published efficacy results for the PncCRM vaccine  that has recently been granted marketing authorization in several countries. This report describes the efficacy, immunogenicity, and safety profiles of the other study vaccine, a pneumococcal conjugate with the meningococcal outer membrane protein complex as the carrier protein (PncOMPC).
The FinOM Vaccine Trial was a prospective, randomized, double-blind cohort study designed to evaluate the efficacy of two 7-valent conjugate vaccines in prevention of AOM caused by vaccine serotypes. Both vaccines were used in parallel and compared with the same control vaccine (hepatitis B vaccine).
Participants. All infants reaching the age of 2 months in the communities of Tampere, Kangasala, and Nokia, Finland, were eligible to participate in the study. The enrollment period was from December 1995 through April 1997. Those enrolled were followed from 2 to 24 months of age at 8 study clinics established specifically for the purpose (figure 1).
The study was conducted in accordance with the Declaration of Helsinki (as amended in Hong Kong, 1989). The initial protocol and its amendments were approved by the ethics committee of the National Public Health Institute of Finland (KTL), by the National Agency for Medicines, and by the relevant local health authorities. Written informed consent was obtained from a parent/guardian before enrollment in the study.
Vaccines and vaccinations. The pneumococcal vaccine prepared by Merck (PncOMPC) consisted of 1 µg of capsular polysaccharides of pneumococcal serotypes 4 and 14, 1.5 µg of type 9V, 2 µg of types 18C and 19F, 3 µg of type 23F, and 5 µg of type 6B, each individually conjugated to the outer membrane protein complex of Neisseria meningitidis serogroup B. The hepatitis B vaccine (Recombivax HB; Merck) contained 5 µg of recombinant hepatitis B surface protein.
The study vaccine was administered intramuscularly at 2, 4, 6, and 12 months of age. From 3 November 1997 onward, for the children randomized to receive PncOMPC, the fourth dose of the conjugate vaccine was replaced by a 23-valent pneumococcal polysaccharide vaccine (PncPS) that consisted of serotypes 1, 2, 3, 4, 5, 6B, 7F, 8, 9N, 9V, 10A, 11A, 12F, 14, 15B, 17F, 18C, 19F, 19A, 20, 22F, 23F, and 33F (Pneumovax23; Merck). A diphtheria-tetanus toxoids-pertussis vaccine with a whole-cell pertussis component, combined with a Haemophilus influenzae type b conjugate vaccine (DTP-Hib), was administered concomitantly with the first 3 doses of the study vaccine, and an inactivated poliovirus vaccine (Imovax; Aventis Pasteur) was administered with the fourth dose. In 4 study clinics, the carrier protein in the DTP-Hib conjugate combination was CRM197 (Tetramune; Wyeth Lederle Vaccines), and in the other 4, it was tetanus toxoid (TetrAct-HIB; Aventis Pasteur).
Objectives and outcomes. The primary objective of our study was to determine the protective efficacy of the pneumococcal conjugate vaccine against all episodes of culture-confirmed pneumococcal AOM caused by vaccine serotypes, compared with control vaccine. The secondary objective was to determine the efficacy against the first episode of vaccine serotype-specific AOM. AOM was defined as a visually abnormal tympanic membrane (with regard to color, position, and/or mobility) suggesting middle ear effusion, with at least 1 of the following symptoms or signs of acute infection: fever, earache, irritability, diarrhea, vomiting, acute otorrhea not caused by otitis externa, and other symptoms of respiratory infection [5, 9].
Episodes of AOM were evaluated in different categories: all episodes; culture-confirmed pathogen-specific episodes; episodes caused by the vaccine serotypes, by serotypes that cross-react with those serotypes, and by other pneumococcal serotypes or groups; and episodes caused by individual serotypes. For the overall and pathogen-specific categories, a new episode was considered to have started if ⩾30 days had elapsed since the beginning of the previous episode. For the serotype-specific category, a new episode was considered to have started if 30 days had elapsed since the beginning of the previous episode caused by the same serotype or, regardless of the time interval, if the new episode was caused by a different serotype. If >1 serotype was recovered from the middle ear fluid at the same time, this was considered to be only 1 episode within the categories of AOM caused by vaccine-specific, cross-reactive, and other serotypes (altogether, 2 and 5 cases in the PncOMPC and control groups, respectively). The follow-up period for per-protocol analyses started 14 days after the third study vaccine dose was administered and ended on the day of discontinuation or on the day of the final visit at the age of 24 months.
Follow-up. The study subjects made 8 scheduled visits to the study clinic. In 3 study clinics, blood samples for determination of pneumococcal antibodies were obtained at 7 months of age; in 1 clinic, samples were obtained at 7 and 13 months; and in the remaining 4 clinics, samples were obtained at 13 months. Immunizations and follow-up for AOM were performed in the study clinics. Adverse events were recorded throughout the follow-up period in questionnaires and in study diaries filled in by the parents during the 3 days after vaccination. An adverse event was defined as “serious” if it was fatal, life-threatening, or permanently disabling or if it required inpatient hospitalization.
Parents were encouraged to bring their child to the study clinic if the child had a respiratory infection or symptoms suggesting AOM. Myringotomy with aspiration of middle ear fluid was performed if AOM was diagnosed at the visit.
Laboratory methods. Middle ear fluid samples were plated immediately on selective sheep blood agar containing 5 µg/mL gentamicin and on enriched chocolate agar plates. Streptococcus pneumoniae, H. influenzae, and Moraxella catarrhalis were identified by standard procedures. Pneumococcal serotyping was done by counterimmunoelectrophoresis and latex agglutination  and confirmed by the quellung reaction when needed, with antisera obtained from Statens Serum Institut (Copenhagen). IgG antibodies to the capsular polysaccharides of the 7 pneumococcal serotypes in the study vaccine were determined by EIA [6, 10, 11].
Sample size. The study was designed to have at least 80% power to show that the lower limit of the 2-sided 95% CI exceeds 30%, assuming that the true vaccine efficacy is at least 60% for first episodes of AOM caused by vaccine serotypes. Based on the epidemiological follow-up data from the preceding FinOM cohort study [1–3], the expected proportion of infants having at least 1 episode of pneumococcal AOM due to vaccine serotypes during the follow-up period from 6.5 through 24 months of age was conservatively set at 12.7%, and the discontinuation rate was set at 15%. With these assumptions, the sample size needed was estimated to be 727 subjects per group. In fact, in this study, >20% of the children in the control group experienced ⩾1 AOM caused by vaccine serotypes, the overall discontinuation rate was only 5%, and the number of enrolled subjects per group was 831–835, all of which increased the power of the study to greater than the prespecified 80%.
Randomization and blinding. Six letters corresponding to the 3 treatment options were randomly allocated to consecutive subject identification numbers, using an allocation ratio of 1:1:1 and a block size of 12. Individual treatment assignments were kept in sealed envelopes until each vaccination. Blinding was ensured by destruction of the letter code immediately after vaccination and by use of vaccinators who were not otherwise involved in the trial follow-up.
Statistical analysis. The relative risk for all episodes of AOM in a given category between the vaccine and parallel control groups was estimated using a generalized Cox regression model , with robust SEs for the parameter estimates . Vaccine efficacy was estimated by subtracting the relative risk from 1. Efficacy for the first AOM episodes was calculated using the standard Cox model.
The χ2 test was used to compare reactogenicity rates between the vaccine and control groups, and McNemar's test for matched pairs was used for comparison between the pneumococcal and concomitant vaccine injection sites after each dose. Overall rates for serious adverse events were compared between the treatment groups using person-time-based relative risk estimates with associated 95% CIs . Antibody concentrations are given as geometric mean concentrations (GMCs) with 95% CIs. Further details of the study design, setting, and procedures have been presented elsewhere .
Study cohort and demographic characteristics. Of the 2497 infants enrolled in the FinOM Vaccine Trial, 835 received PncOMPC, and 831 received the control vaccine (hepatitis B vaccine). The remaining 831 received PncCRM, and these results have been reported elsewhere . Of those enrolled, 805 (96.4%) in the PncOMPC group and 794 (95.5%) in the control group completed the follow-up as specified in the protocol (figure 1). Because compliance was excellent, the difference between the per-protocol and intent-to-treat data sets is negligible, and we therefore only report the efficacy and immunogenicity results of the per-protocol analysis. Demographic characteristics and distribution of established risk factors for AOM were similar in both groups (table 1).
Vaccine efficacy. During the follow-up period from 6.5 through 24 months of age, there were 110 AOM episodes attributed to vaccine serotypes in the PncOMPC group and 250 in the control group, yielding a vaccine efficacy of 56% (95% CI, 44%–66%; table 2). The efficacy calculated for the period between the first and second doses was 11% (95% CI, −9% to 53%); between the second and third doses, 57% (95% CI, 8%–80%); between the third and fourth doses, 47% (95% CI, 22%–64%); and between the fourth dose and the end of follow-up, 62% (95% CI, 48%–72%). The efficacy of vaccination for the prevention of a first episode of AOM caused by any of the serotypes included in the vaccine was 50% (95% CI, 36%–61%).
The estimates of vaccine efficacy against AOM episodes attributed to a specific vaccine serotype ranged from 37% for 19F to 79% for 6B and 82% for 9V (table 2). AOM episodes caused by serotypes that cross-react with the vaccine serotypes (6A, 9N, 18B, 19A, and 23A) were equally common in the 2 groups. The rate of episodes caused by pneumococcal serotypes other than the vaccine or cross-reactive serotypes was 27% higher (95% CI, −6% to 70%) in the PncOMPC group than in the control group.
The aggregate efficacy against any culture-confirmed pneumococcal AOM was 25% (95% CI, 11%–37%). The overall incidence of AOM was almost the same in the 2 groups (table 2).
After 3 doses of PncOMPC, 631 children received PncOMPC and 187 children received PncPS as the booster dose. The 2 types of boosters seemed to be equal in terms of protection against AOM; the efficacy against AOM caused by the 7 serotypes included in the vaccine was 60% (95% CI, 43%–72%) after the PncOMPC booster and 65% (95% CI, 34%–81%) after the PncPS booster (table 3).
Immunogenicity. After the primary series of PncOMPC, the GMCs of antibodies to the 7 vaccine serotypes ranged from 0.35 to 3.45 µg/mL (table 4). After the booster, the GMCs of antibodies depended heavily on the type of pneumococcal vaccine administered: substantially higher antibody concentrations were seen after administration of PncPS than after administration of PncOMPC. The difference was particularly pronounced for type 19F (5-fold difference), for which the antibody concentrations among the PncPS booster recipients were exceptionally high (GMC, 50.76 µg/mL). In the control group, GMCs were low at 7 and at 13 months of age, ranging from 0.07 for type 4 at 7 months to 0.36 µg/mL for type 19F at 13 months (data not shown).
Safety. PncOMPC caused local reactions within 3 days of each dose more often than the hepatitis B vaccine (table 5) but less often than the concomitantly administered DTP-Hib combination vaccine (P = .3 to P < .01; data not shown). Crying was more common in the PncOMPC group than in the control group after all 4 doses. The PncOMPC booster tended to cause local reactions more often than the PncPS booster.
Seven cases of serious adverse events or unexpected adverse reactions occurring within 7 days of vaccination were initially judged to be possibly related to the study vaccines. Of these, 1 case of febrile seizure, 1 case of contractions of the trunk without fever, and 1 case of pneumococcal meningitis caused by serogroup 7 (not included in the vaccine) occurred in the PncOMPC group. All 3 of these infants recovered uneventfully and did not present with any neurological abnormalities during follow-up. In addition, 1 child had an unexpectedly large local reaction (10.0 × 4.5 cm) after the first dose of PncOMPC. One child died during the study period, at the age of 10 months, 147 days after administration of the third dose of PncOMPC. The death was due to volvulus with resulting bowel necrosis and was assessed as being unrelated to the study vaccine.
There were no statistically significant differences in the occurrence of any diagnosis among individuals who experienced serious adverse events between the 2 vaccine groups or between the recipients of the different booster vaccines. The number of suspected infections requiring hospitalization was lower in the PncOMPC group (6 cases) than in the control group (13 cases) (relative risk, 0.46; 95% CI, 0.17–1.21). The only case of invasive pneumococcal infection (meningitis) in the PncOMPC group was caused by the nonvaccine serogroup 7, whereas 2 of the 3 cases of invasive pneumococcal infections in the control group were caused by vaccine serotypes (meningitis due to serogroup 23F and bacteremia due to serogroup 19F), and the third was caused by a nonvaccine serogroup (meningitis due to serogroup 15).
In the current study, PncOMPC showed an efficacy of 56% (95% CI, 44%–66%) against AOM episodes caused by serotypes included in the vaccine. The vaccine was indisputably immunogenic in infants, although the antibody concentrations reached were lower than those elicited by PncCRM . Remarkably, the PncPS booster induced higher antibody responses than did the PncOMPC booster. The safety of PncOMPC did not raise concerns. The local and systemic reaction rates were similar to those generally seen with conjugate vaccines [8, 15]. No serious adverse events were associated with the vaccine.
The efficacy profile of PncOMPC was surprisingly similar to that of PncCRM . The 2 vaccines were comparable not only in their overall ability to reduce AOM caused by any of the vaccine serotypes but also in the variation of efficacy with serotype. Both vaccines prevented ∼80% of AOM episodes that could be attributed to type 6B but were clearly less efficacious against type 19F. However, although PncCRM was reasonably efficacious against AOM caused by serotypes that cross-react with those in the vaccine, PncOMPC provided no protection against AOM caused by these serotypes. Apparently because of this lack of effect on the cross-reactive serotypes, the reduction all types of pneumococcal AOM was only 25%, compared with 34% when PncCRM is used [4, 5].
The rate of AOM episodes caused by all other serotypes was higher among the recipients of both pneumococcal vaccines than in the control group (27% for PncOMPC and 33% for PncCRM). This finding is consistent with data from previous studies that show increased rates of carriage of nonvaccine serotypes among infants vaccinated with pneumococcal conjugate vaccines [16, 17]. Together, these findings strongly suggest that, although pneumococcal conjugate vaccines reduce carriage of vaccine serotypes, in doing so they open an ecological niche for serotypes not included in the vaccine . Although no statistically significant increase in nonvaccine serotypes has so far been observed in invasive disease , the full effects of pneumococcal conjugate vaccines on pneumococcal populations, both in normal flora and in disease, need to be closely followed.
In contrast to the similarity in efficacy, PncOMPC and PncCRM elicited markedly different antibody responses. The concentrations of antibodies to serotypes 6B, 9V, 14, 18C, and 23F after 3 doses of PncOMPC were lower than those seen in the parallel group of infants after 3 doses of PncCRM, and this difference persisted for serotypes 6B, 14, 18C, and 23F after administration of the conjugate booster . In spite of the lower antibody concentrations evoked by PncOMPC, protection against AOM caused by the vaccine serotypes was equal to that afforded by the higher concentrations achieved with PncCRM. Moreover, the full protective effect of PncOMPC was already achieved after the second dose, whereas 3 doses of PncCRM seemed to be needed for the same effect. This is consistent with the findings of other studies that show that the Hib-OMPC conjugate vaccine , in contrast to other Hib conjugate vaccines [21, 22], is protective in infants even after a single dose.
The lower antibody concentrations could explain why PncOMPC, unlike PncCRM, lacked protection against serotypes that cross-react with the vaccine serotypes. This interpretation is supported by the study of Väkeväinen et al. , which demonstrated that 2–6 times more anti-6B antibodies are needed for 50% opsonophagocytic killing of type 6A than for type 6B .
The different immunogenicity coinciding with equivalent efficacy results of the 2 vaccines used in the FinOM Vaccine Trial shows that antibody concentrations much lower than those elicited by PncCRM are sufficient to protect against pneumococcal disease. Furthermore, the present findings with regard to the levels of protection provided, for example, against serotype 19F, compared with protection provided against serotype 6B, indicate that prevention of AOM caused by different serotypes may require quite different concentrations of type-specific antibody in the serum.
Good but variable antibody responses to PncPS booster  or OMPC [7, 24] after priming with pneumococcal conjugate vaccines with diphtheria toxoid as carrier proteins have been reported. In the current study, the PncPS booster elicited considerably higher antibody responses than did the PncOMPC booster, and this difference was especially pronounced for type 19F. This finding is of special interest, because the use of PncPS as a booster would be less costly than use of the conjugate vaccines. However, before such a vaccination program is recommended, several aspects of a PncPS booster should be clarified through experimental research, including its immunogenicity after priming with other pneumococcal conjugate vaccines and its effects on immunological memory. Theoretically, boosting with a conjugate vaccine may be important for persistence of immunity, because a conjugate booster is considered to stimulate generation and expansion of high-affinity clones of B memory cells , whereas use of a polysaccharide booster may, after a short-lived antibody response, even result in depletion of the memory pool .
Pneumococcal AOM seems to be a more difficult target for a vaccine than is invasive disease [4, 5], probably at least partly because of its mucosal localization. Therefore, it is likely that moderate protection against pneumococcal AOM caused by the vaccine serotypes will translate into an excellent efficacy profile against invasive disease caused by the same pathogens. This was demonstrated by the AOM efficacy results for the PncCRM arm in the FinOM Trial, combined with the findings for invasive disease in the trial in California of this vaccine .
In conclusion, the present results support findings published elsewhere  that significant protection against pneumococcal AOM can be obtained with a pneumococcal conjugate vaccine. Furthermore, 2 pneumococcal conjugate vaccines with different immunogenicity profiles can afford comparable protection against AOM caused by the vaccine serotypes. This does not speak against a decisive role of antibody concentrations in protection but shows that this role can be different for different serotypes.
Finnish Otitis Media Study Group
Principal investigator: Juhani Eskola.
Steering committee: Mervi Eerola, Tapani Hovi, Pekka Karma, Terhi Kilpi, Helena Käyhty, Maija Leinonen, P. Helena Mäkelä, Arto Palmu, Esa Ruokokoski, and Aino K. Takala.
Study coordination: Terhi Kilpi, Kari S. Lankinen, Aino K. Takala, Petri Mattila, and Arto Palmu.
Coordinating study nurses: Pirjo-Riitta Saranpää, Anna-Stina Leinonen, and Terhi Hulkko.
Study physicians: Wilhelm Bredenberg, Kaisu Hattela, Tuija-Leena Huupponen, Marja-Leena Hyypiä, Elina Hyödynmaa, Päivi Leinonen, Päivi Limnell, Merja Mölsä, Hanna Rautio, Auli Räsänen, Päivi Savikurki-Heikkilä, Heljä Savolainen, Anneli Siro, Ritva Syrjänen, Sirpa Vesa, and Sari Vikström.
Study nurses: Hannele Holli, Marja-Leena Hotti, Helena Jokinen, Marjo-Riitta Kauppinen, Eija Lahtinen, Johanna Laitinen, Ella Lehto, Taina Nissinen, Sirkka Oikarinen, Sirkka-Liisa Piirto, Minna Ranta, Päivi Sirén, Terttu Suikkanen, and Päivi Tervonen.
Vaccinators: Eija Kujanne, Hannamari Salonen, and Marjo Virkki.
Clinical laboratory samples: Arja Katila and Maire Selin.
Bacteriology: Elja Herva, Aili Hökkä, Tarja Kaijalainen, Eeva-Liisa Korhonen, Maija Leinonen, and Hilkka Ohukainen.
Immunology: Maijastiina Karpala, Minna Koivuniemi, Helena Käyhty, Hannele Lehtonen, Piia Pihlajamaa, Satu Rapola, Leena Saarinen, Merja Väkeväinen, and Heidi Åhman.
Virology: Soile Blomqvist, Tapani Hovi, and Marjaana Kleemola.
Otorhinolaryngology: Pekka Karma.
Data management: Marko Grönholm, Jaason Haapakoski, Eeva Koskenniemi, Satu Nahkuri, Esa Ruokokoski, and Matti Sarjakoski.
Biostatistics: Mervi Eerola, Jukka Jokinen, Mika Lahdenkari, and Jouko Verho.
Secretariat: Ulla Johansson and Paula Solukko.
Data and Safety Monitoring Board: Olli Ruuskanen (chair), Paul Fine, Jussi Mertsola, Richard Moxon, Patrick Olin, and Karin Prellner.
We are indebted to Elizabeth Horigan (National Institute of Allergy and Infectious Diseases, Bethesda, MD) and to Heikki Puhakka and Tapani Rahko (Department of Otorhinolaryngology-Head and Neck Surgery, Tampere University Hospital, Tampere, Finland), for their valuable contributions to the trial implementation; to Jeffrey L. Silber (Merck, West Point, PA), for useful comments on the manuscript; and to the personnel of Tampere, Nokia, and Kangasala health centers (Finland) and several employees of Tampere University Hospital, for collaborating with us and helping us with numerous practical issues.