Four years of universal pneumococcal conjugate infant vaccination in Germany: impact on incidence of invasive pneumococcal disease and serotype distribution in children

Abstract

Introduction: Vaccination with pneumococcal conjugate vaccine (PCV) for all children <2 years was recommended in Germany in July 2006. Initially PCV7 was exclusively used; PCV10 became available from April 2009 and PCV7 was replaced by PCV13 in December 2009.

Objective: To compare the incidence and serotype distribution of invasive pneumococcal disease (IPD) for pneumococcal meningitis and non-meningitis IPD in children from 2007 to 2010 with reference to the pre-vaccination period from 1997 to 2001.

Methods: Nationwide surveillance of IPD for children <16 years in Germany was based on two independent reporting sources: active surveillance in paediatric hospitals and passive web-based surveillance through microbiological laboratories. Serotyping was performed using the Neufeld Quellung reaction.

Case definition: isolation of Streptococcus pneumoniae from a normally sterile body site. IPD incidence was estimated by capture-recapture analysis. Rate ratios comparing post- to pre-vaccination incidence were calculated as well as PCV7 and non-PCV7 serotype specific incidences.

Results: While PCV7 incidence decreased by 88% (95%CI: 83 to 91) in children <16 years both in pneumococcal meningitis and non-meningitis IPD, an increase in Non-PCV7 serotypes was observed which was more pronounced in non-meningitis cases (168%; 95%CI: 140-257) than in pneumoccocal meningitis (65%; 95%CI: 23-123). The changes in incidence after four years were: <16 years: -35% (95%CI: -49 to -19), <2 years: -46% (95%CI: -61 to -27) for pneumococcal meningitis and+11% (95%CI: -4 to +29) and -26% (95%CI: -41 to -7) for non-meningitis IPD respectively.

Conclusion: Infant PCV7 vaccination in Germany prompted a decrease in the incidence of pneumococcal meningitis similar to that observed in England/Wales. In non-meningitis IPD the decrease was smaller and confined to the age group <2 years with no change or an increase in incidence in other age groups pointing to potential ascertainment bias due to increased blood-culturing.