Effect of pneumococcal Haemophilus influenzae protein D conjugate vaccine (PHiD-CV10) on outpatient antimicrobial purchases: a double-blind, cluster randomised phase 3-4 trial

Background Antimicrobial drugs are frequently prescribed to children for respiratory tract infections such as otitis, tonsillitis, sinusitis, and pneumonia. We assessed the effect of the ten-valent pneumococcal Haemophilus influenzae protein D conjugate vaccine (PHiD-CV10; GlaxoSmithKline) on antimicrobial purchases. Methods In this nationwide phase 3-4 cluster-randomised, double-blind trial, children younger than 19 months were randomly assigned to receive PHiD-CV10 in 52 of 78 dusters or hepatitis B or A vaccine as control in 26 dusters according to three plus one or two plus one schedules (infants younger than 7 months) or catch-up schedules (children aged 7-18 months). The main objective for the antimicrobial treatment outcome was to assess vaccine effectiveness against outpatient prescriptions of antimicrobial drugs recommended by national treatment guidelines for acute otitis media in Finland in children who received at least one dose of study vaccine before 7 months of age. Masked follow-up lasted from the date of first vaccination (from Feb 18, 2009, through Oct 5, 2010) to Dec 31, 2011. We obtained data on all purchased antimicrobial prescriptions through the benefits register of the Social Insurance Institution of Finland. This and the nested acute otitis media trial are registered at ClinicalTrials.gov, numbers NCT00861380 and NCT00839254. Findings More than 47 000 children were enrolled. In 30 527 infants younger than 7 months at enrolment, 98436 outpatient antimicrobial purchases were reported with incidence of 1.69 per person-year in the control dusters. Analysis of the main objective included 91% of all antimicrobial purchases: 31 982 in the control and 57 964 in the PHiD-CV10 dusters. Vaccine effectiveness was 8% (95% CI 1-14) and the incidence rate difference 0.12 per person-year corresponding to the number needed to vaccinate of five (95% CI 3-67) to prevent one purchase during the 2 year follow-up for combined PHiD-CV10 three plus one and two plus one infant schedules. The vaccine effectiveness was identical for the two infant schedules. In the catch-up schedules, the vaccine effectiveness was 3% (95% CI 4 to 10). Interpretation Despite low relative rate reductions the absolute rate reductions were substantial because of the high incidence of the outcome. This reduction would lead to over 12 000 fewer antimicrobial purchases per year in children younger than 24 months in Finland (birth cohort of 60 000 children).