The aim of the IVANHOE study was to determine the real-world impact of the rotavirus vaccine, controlling for epidemic-to-epidemic variation in disease burden. A population-based prospective cohort study was conducted in Brest City and 7 suburban districts (CUB area), North-western Brittany, France (210,000 inhabitants; 5500 births per year). The vaccination program started in May 2007 for a 2-year period for all infants born in the Brest birth zone through pediatricians, public outpatient clinics and general practitioners. To determine vaccine impact we monitored trends in hospitalizations for rotavirus-specific diarrhea using an active hospital-based surveillance system initiated 5 years before vaccine introduction. The number of hospitalizations for rotavirus-specific diarrhea during the 2008/2009 epidemic in infants less than 2 years of age whose parents lived within the CUB area was modelled as a function of (1) the number of hospitalizations in infants 2-5 years of age to control for epidemic-to-epidemic variation and (2) vaccine introduction. A total of 4684 infants received at least one dose. Of these, 2635 lived within the CUB area. Vaccine coverage for a complete schedule in the CUB area was 47.1%. Poisson modelling revealed a reduction by a factor of 2.04 (1.56-2.66) in the number of hospitalizations during the last epidemic season (2008/2009), the number of observed cases being equal to 30, against an expected number of 61. Relative risk reduction for hospitalizations for rotavirus diarrhea was 98% (95% CI: 83-100%). We observed a noticeable impact of vaccination on rotavirus diarrhea hospitalizations within 2 years of vaccine introduction integrating for the first time rotavirus epidemics variation. The trial is registered with ClinicalTrials.gov, number, NCT00740935.