The epidemiology of invasive Haemophilus influenzae infections was evaluated in Ontario between 1989 and 2007 to assess the impact of the introduction of the conjugate H. influenzae serotype b (Hib) vaccine in the early 1990 s on Hib and non-Hib serotypes in both vaccinated and unvaccinated cohorts as well as the possibility of "strain replacement" with non-vaccine H. influenzae strains. Data were collected by the provincial Public Health Laboratories-Toronto, Ontario Agency for Health Protection and Promotion, which performed almost all serotyping on invasive (blood, CSF, other sterile sites) H. influenzae strains isolated in the province during the study period. Temporal trends for Hib, other typeable strains, and non-typeable H. influenzae were evaluated by Poisson regression, controlling for the specimen submissions. Prior to infant Hib vaccination, the most commonly observed serotype was serotype b (64.9%). Subsequently, 70.3%, 13.6%, and 9.4% of isolates were non-typeable, serotype f, and serotype b, respectively. Infant Hib vaccination resulted in a decrease in Hib incidence in all age groups (pooled IRR 0.432) and marked increases of non-typeable and serotype f H. influenzae in children aged