Vaccine effectiveness of the pneumococcal Haemophilus influenzae protein D conjugate vaccine (PHiD-CV10) against clinically suspected invasive pneumococcal disease: a cluster-randomised trial

Abstract

Vaccine effectiveness of pneumococcal conjugate vaccines against culture-confirmed invasive pneumococcal disease has been well documented. In the Finnish Invasive Pneumococcal disease (FinIP) trial, we reported vaccine effectiveness and absolute rate reduction against laboratory-confirmed invasive pneumococcal disease (confirmation by culture or antigen or DNA detection irrespective of serotype). Here, we assessed vaccine effectiveness of PHiD-CV10 against clinically suspected invasive pneumococcal disease in children by use of diagnoses coded in hospital discharge registers. For this phase 3/4 cluster-randomised, double-blind trial, undertaken between Feb 18, 2009, and Dec 31, 2011, in municipal health-care centres and the Tampere University Vaccine Research Centre (Finland), we randomly assigned (2:2:1:1) 78 clusters into PHiD-CV10 three plus one, PHiD-CV10 two plus one, control three plus one, control two plus one groups (26:26:13:13 clusters) to give PHiD-CV10 in either three plus one or two plus one schedule (if enrolled before 7 months of age; infant schedules), two plus one (if enrolled between 7 and 11 months; catch-up schedules), and two doses at least 6 months apart (if enrolled between 12 and 18 months; catch-up schedules). Children were eligible if they had not received and were not anticipated to receive any of the study vaccines and had no general contraindications to vaccinations. We collected all inpatient and outpatient discharge notifications from the national hospital discharge register with International Classification of Diseases (ICD) 10 diagnoses compatible with invasive pneumococcal disease or unspecified sepsis, and verified data with patient files. We excluded invasive pneumococcal disease cases confirmed by positive culture or DNA/RNA detection from normally sterile body fluid. The primary objective was to estimate vaccine effectiveness against all register-based non-laboratory-confirmed invasive pneumococcal disease or unspecified sepsis and patient-file verified non-laboratory-confirmed invasive pneumococcal disease in infants younger than 7 months at enrolment. Masked follow-up lasted from the date of the first vaccination to Dec 31, 2011. Vaccine effectiveness was calculated against all episodes. This trial is registered with ClinicalTrials.gov, numbers NCT00861380 and NCT00839254. We enrolled 47,366 children. On the basis of ICD-10 diagnoses, we recorded 264 episodes of register-based non-laboratory-confirmed invasive pneumococcal disease or unspecified sepsis, of which 102 were patient-file verified non-laboratory-confirmed invasive pneumococcal disease. The vaccine effectiveness was 50% (95% CI 32-63) in the 30,527 infants with three plus one and two plus one schedules combined and the absolute incidence rate reduction was 207 episodes per 100,000 person-years (95% CI 127-286). The vaccine effectiveness against the patient-file verified non-laboratory-confirmed invasive pneumococcal disease was 71% (95% CI 52-83) in infant three plus one and two plus one schedules combined. The absolute rate reduction was 142 episodes per 100,000 person-years (95% CI 91-191) in infant cohorts. This vaccine-probe analysis is the first report showing the effect of pneumococcal conjugate vaccines on clinically suspected invasive pneumococcal disease. The absolute rate reduction was markedly higher compared with laboratory-confirmed invasive pneumococcal disease, which implies low sensitivity of the laboratory-based case definitions and subsequently higher public health effect of pneumococcal conjugate vaccines against invasive pneumococcal disease than previously estimated. GlaxoSmithKline Biologicals SA and National Institute for Health and Welfare (THL), Finland.