Early Vaccine-type Pneumococcal Acute Otitis Media Does not Predispose to Subsequent Otitis When Compared With Early Acute Otitis Media Due to Other Bacterial Etiology

Abstract

Background: 

Prevention of acute otitis media (AOM), and especially recurrence and biofilm formation, by pneumococcal conjugate vaccines has been hypothesized to be due to prevention of early episodes triggering the vicious cycle. We tested the specific role of vaccine-type pneumococcal AOM in this hypothesis.

Methods: 

In the phase III randomized, double-blind Finnish otitis media Vaccine Trial conducted in 1995–1999, children received pneumococcal conjugate vaccine 7 or hepatitis B vaccine as control at 2, 4, 6, and 12 months of age and were followed for AOM. Myringotomy with middle ear fluid aspiration was performed in AOM, and samples were cultured. We compared control-vaccinated children with confirmed vaccine-type or 6A-AOM with those with AOM due to other confirmed etiology within 2–6 months of age (early AOM) and followed for subsequent AOM from 6–24 months of age.

Results: 

Eight hundred thirty-one children were enrolled in the Finnish otitis media control arm. Before 6 months of age, 34 children experienced vaccine-type–AOM, and 40 children experienced AOM of other bacterial etiology. The subsequent AOM incidences were 1.9 (95% CI, 1.5–2.4) and 2.1 (1.7–2.5) in these subgroups, respectively. However, the subsequent incidences were lower if no bacteria were detected at AOM (1.5, 1.2–1.8) or if there was no early AOM (1.1, 1.1–1.2).

Conclusions: 

Early vaccine-type AOM was not associated with a higher risk of subsequent AOM compared with AOM due to other confirmed bacterial etiology. These data do not support any specific role of vaccine-type pneumococcus in the hypothesis.

Otitis media (OM) is a common disease in infants and young children. While up to one-third of children do not experience any acute OM (AOM), most children have at least 1 disease episode. However, some 20% of children have recurrent and/or complex otitis media resulting in repeated disease episodes and physician consultations with high antimicrobial consumption, otitis media surgery, and potential hearing defects.¹

The prevention of otitis media, and especially the recurrent otitis media and formation of biofilm disease, by the pneumococcal conjugate vaccines (PCVs) has been hypothesized to be due to prevention of early episodes and thus, preventing the vicious cycle² and otitis episodes due to any etiology.

We evaluated this hypothesis further using our Finnish otitis media (FinOM) vaccine trial data to explore whether early vaccine-type pneumococcal AOM was critical for development of later otitis media encounters.

METHODS

The FinOM vaccine trial was a phase III randomized, double-blind controlled trial conducted in 1995–1999 in Finland.³ Children received PCV7 or hepatitis B vaccine as control at 2, 4, 6, and 12 months of age. In the current analysis, we evaluated control-vaccinated children only and, therefore, the study design is a prospective cohort study, not a clinical trial.

Active follow-up for AOM was conducted in specific study clinics from 2–24 months of age. Myringotomy with middle ear fluid aspiration was performed in each case of AOM. Middle ear fluid samples were cultured, bacteria including Streptococcus pneumoniae were identified and serotyped using serologic methods.¹

Intention-to-treat follow-up started at the first vaccine dose at 2 months of age and ended at 24 months of age. For the current analysis, we split the follow-up time into 2 periods:

1. Two to 5 months of age to define the children exposed to AOM due to the various otitis pathogens; specifically, those with culture-confirmed PCV7 vaccine-type or serotype 6A AOM before 6 months of age were considered exposed and those with AOM due to other confirmed etiology before 6 months of age but not PCV7/6A AOM as unexposed
2. Six to 24 months of age for outcome follow-up of subsequent AOM episodes.

The primary outcome was AOM due to any etiology in the FinOM Vaccine Trial (from 6–24 months of age). Any tympanostomy tube placement from 6 months up to 11–13 years of age was assessed as a secondary outcome.⁴ ClinicalTrials.gov Identifier: NCT00378417.

Statistical Analysis

Exact Poisson 95% confidence intervals (95% CI) were calculated for the incidences.

RESULTS

In total, 831 children were enrolled in the FinOM control arm; 825 with follow-up beyond 6 months of age were included in the primary outcome analysis.

Before 6 months of age, 34 children experienced 40 episodes of culture-confirmed vaccine-type or 6A AOM (including 22 episodes due to 23F, 7 due to 19F and 4 due to 14 and 6B each), and 40 children experienced AOM due to other confirmed bacterial etiology (including 8 episodes of nonvaccine serotype AOM: 2 due to 19A and 39 each, and 1 due to 9N, 10, 11, and 15). The subsequent AOM incidences were 1.9 (95% CI, 1.5–2.4) and 2.1 (1.7–2.5) in these subgroups, respectively. The overall AOM incidences between 6–24 months of age in the different exposure categories are shown in Table 1 and incidences of tympanostomy tube placements from 6 months up to 11–13 years of age in Figure 1.

TABLE 1.: Incidence of AOM in the Different Exposure Categories Based on AOM due to Different Microbes Before 6 Months of Age, Intention-to-treat Follow-up

FIGURE 1.: Incidence of tympanostomy tube placements between 6 months to 13 years of age in the different exposure categories based on AOM due to different microbes before 6 months of age. MEF indicates middle ear fluid.

The subsequent AOM incidences were lower if no bacteria were detected in early AOM, and especially if child had no confirmed AOM within the first 6 months of age (Table 1).

We also performed various sensitivity analyses resulting in similar results (Table 2). In these, we evaluated PCV 7 vaccine types only as the exposure (ie, not including 6A), sensoring the follow-up at first tympanostomy tube placement (for AOM analysis), inclusion of the first AOM episode only to define the exposure (not all episodes within 6 months as in the primary analysis), and excluding subjects with otitis media diagnosed outside the study clinics before age of 6 months of age and those with abnormal tympanic membrane findings suggesting otitis media at enrolment to avoid bias in the exposure classification, and finally, the use of the age of 12 months as the cut-point for the definition of exposure and outcome follow-up.

TABLE 2.: Sensitivity Analyses for the Incidence of AOM in the Different Exposure Categories Based on AOM due to Different Microbes, Intention-to-treat Follow-up

DISCUSSION

Early vaccine-type pneumococcal AOM was not associated with a higher incidence of subsequent AOM compared with AOM due to other confirmed bacterial etiology. These data do not support any specific role for vaccine-type pneumococcus in this hypothesis, but rather subjects with early AOM due to any bacterial etiology are prone to experience also later disease.

Currently, the PCVs are the only vaccines available to prevent early AOM. Influenza vaccines have been shown to reduce influenza-related AOM,⁵ but the vaccine is licensed for children 6 months of age and older; therefore, direct protection against early AOM is not possible. Development of vaccines against other otitis pathogens is ongoing.

Although the vaccine efficacy of PncCRM7 against any AOM in clinical trials was low, only 6%–7%,^3,6,7 the vaccine efficacy estimates against recurrent AOM (defined as 3 within 6 months or 4 within 12 months) and tympanostomy tube placements have been higher, especially in long-term follow-up.^4,8

In our earlier reanalysis on recurrence of AOM, we found high vaccine efficacy for the PncCRM7 vaccine against recurrent vaccine-type AOM, but failed to show efficacy against overall AOM of increasing rank.⁹ This demonstrates the importance of disease replacement; reduction of vaccine-type AOM leads to an increased disease burden due to other pathogens in AOM resulting in a low net impact in the total population.⁹ On the contrary, observational before-after PCV introduction data show reduction, not replacement, in AOM due to other pathogens in a selected subgroup of complex disease.¹⁰ In this kind of study design, multiple sources of bias should be borne in mind including care-seeking behavior, diagnostic changes (both clinical and laboratory), administrative changes, and so on.

Nevertheless, with large-scale PCV vaccination programs, it is possible to further prevent early vaccine-type AOM, even before the age of first dose of the vaccination, due to the development of the indirect impact and eventual eradication of the vaccine-type circulation in carriage.¹¹ However, the replacement by other pathogens, including nonvaccine pneumococci,¹² is also likely to be enhanced during the vaccination programs.

The serotypes for the PCVs were originally selected based on their importance in causing invasive pneumococcal disease. The relative importance was due the common carriage of these serotypes and their relative propensity to cause disease (case-to-carrier ratio) compared with other pneumococcal serotypes. While clear differences in the case-to-carrier ratios have been observed in invasive pneumococcal disease,¹³ the differences in AOM are more subtle.¹⁴ Thus, also pneumococcal serotypes not included in the vaccine are capable of causing otitis media.

The evaluation whether the early AOM is a true risk factor for subsequent disease per se or a mere indicator of a frail subject is difficult when all the risk factors are not known and the incidence shows major variation due to, for example, viral respiratory infection epidemics, seasonality, and age.

PCVs have been shown to elicit long-lasting antibodies, which probably contribute to the long-term protection.¹⁵ Thus, the impact of the PCVs does not only rely on prevention of early episodes.

The active prospective follow-up with detection of AOM confirmed by myringotomy underlines the specificity of the pathogen-specific results in our study. The high incidence of AOM observed in the trial, on the other hand, speaks for high sensitivity of the case detection. Our various sensitivity analyses showed that the study findings were robust. The limitations of this study include the low number of pathogen-specific AOM episodes resulting in quite wide CIs and the potentially poor generalizability to other settings with different AOM etiology and risk factors. However, it should be noted that the FinOM trial is the single trial globally that has assessed etiology in all cases of AOM longitudinally in an unselected cohort.

We postulate that the long-term impact of PCVs on otitis media cannot be fully explained by the prevention of early vaccine-type episodes, but also sustained vaccine-type antibody persistence and the development of the indirect impact with eradication of the vaccine-type carriage contribute to the protection against AOM, including complex disease. To fully prevent later otitis media and its sequelae, prevention of early AOM irrespective of etiology is important.