Long-term Vaccine Impact on Invasive Pneumococcal Disease Among Children With Significant Comorbidities in a Large Australian Birth Cohort

Background: Little is known about long-term invasive pneumococcal disease (IPD) incidence in children with risk factors (RFs) in populations with high coverage pneumococcal conjugate vaccine (PCV) programs. We measured IPD burden and changes with PCV use in children by RF status. Methods: A retrospective cohort of all live births in 2001–2012 in New South Wales, Australia was linked to IPD, hospitalization and death data. RFs were identified from International Classification of Diseases codes in linked hospitalizations. For each RF adjusted hazard ratios (aHRs, using Cox models), population attributable fractions (PAFs) and changes post-PCV relative to baseline for IPD were calculated. Results: One-thousand two-hundred fifty-one IPD cases occurred in ~1.1 million children in 12-year study cohort. The 75,404 children (6.8% of cohort) with RFs accounted for 255 (20.4%) IPD cases [rate (per 100,000 person-years) of 61 compared with 14 in no RFs]. Asthma was most common RF (n = 41,074; 3.6%) but highest IPD risk was in 2452 children (0.2%) with immunosuppression, splenic dysfunction or breach in cerebrospinal fluid barrier (aHR~20; PAF 0.7–1.8%) versus asthma (aHR 5.3; PAF 14.8%). Compared with 2001–2004 birth cohort (baseline), IPD incidence in PCV-eligible 2009–2012 birth cohort was 78% (95% confidence interval: –72% to –82%) less in children without RFs. IPD declined nonsignificantly (13%; 95% confidence interval: –70% to +138%) in highest IPD risk group, but by 67% (–43% to –82%) in children with other RFs. Conclusions: By 8 years of universal PCV, IPD incidence reduced significantly in all children except in the 0.2% at highest risk, for whom antibiotic prophylaxis and additional vaccine doses are recommended but compliance and effectiveness remain uncertain.