Estimates of incidence of early-onset group B streptococcal (EOGBS) infection are based on blood or cerebrospinal fluid culture-proven cases, which can be falsely negative and hence underestimate the true burden of disease. Probable EOGBS infection can be defined as colonisation by group B streptococci accompanied by features of clinical sepsis. Data collected prospectively in the UK over 1 year for neonates who required a septic screen in the first 72 h of life indicated a combined rate of definite and probable EOGBS infection of 3.6 per 1000 livebirths. This estimate indicates a much greater disease burden in the UK than that suggested by figures of culture-proven sepsis, and lends support to the need for prevention strategies.