Abstract
Seven-valent pneumococcal conjugate vaccine (PCV7) was introduced to Sweden in 2009 and replaced by pneumococcal non-typeable Haemophilus influenzae protein D conjugate vaccine (PHiD-CV) or 13-valent PCV (PCV13) from late 2009. A retrospective cohort study assessed the impact of PCVs on otitis media/acute otitis media (OM) in children aged ≤5 years (NCT02742753) living in Skåne (PCV7 then PHiD-CV) or Västra Götalandsregionen (PCV7 then PCV13) between 2005 and 2013 using linked regional and national databases. Time-series analyses described differences between pre-PCV and post-PCV eras. Adjusted age-period-cohort (APC) predictive models estimated vaccine effectiveness and OM incidence ratios between PCV cohorts. Time-to-first OM diagnosis was estimated in ≤2 year-olds by survival analysis using a Cox proportional hazards model. Descriptive interrupted time-series analyses showed OM incidence in ≤2 year-olds declined by 42% (Skåne) and 25% (Västra Götalandsregionen) after PHiD-CV/PCV13, respectively, versus pre-PCV, but baseline OM incidence and duration of PCV7 use differed between regions. In adjusted APC models, OM incidence decreased after PHiD-CV by 9.9% (95% confidence interval [CI]: 4.4; 15.1, p < .001) and PCV13 by 2.3% (95%CI: -3.2; 7.6, p = .401) compared with pre-PCV. Both PHiD-CV and PCV13 decreased the risk of first OM diagnosis: hazard ratio (95%CI) for PHiD-CV relative to pre-PCV 0.67 (0.65; 0.69); 0.87 (0.85; 0.89) for PCV13 relative to pre-PCV; p < .001 for both comparisons. Within the limitations of this study conducted in two large Swedish regions, descriptive time-series analyses showed that OM incidence rates declined following the introduction of PHiD-CV and PCV13; however, this reduction only reached statistical significance for PHiD-CV in the adjusted APC models.