Abstract
No studies have quantified the impact of pre-culture antibiotic use on the recovery of individual blood-borne pathogens or on population-level incidence estimates for Streptococcus pneumoniae. We conducted bloodstream infection surveillance in Thailand during November 2005-June 2008. Pre-culture antibiotic use was assessed by reported use and by serum antimicrobial activity. Of 35,639 patient blood cultures, 27% had reported pre-culture antibiotic use and 24% (of 24,538 tested) had serum antimicrobial activity. Pathogen isolation was half as common in patients with versus without antibiotic use; S. pneumoniae isolation was 4- to 9-fold less common (0.09% versus 0.37% by reported antibiotic use; 0.05% versus 0.45% by serum antimicrobial activity, P < 0.01). Pre-culture antibiotic use by serum antimicrobial activity reduced pneumococcal bacteremia incidence by 32% overall and 39% in children < 5 years of age. Our findings highlight the limitations of culture-based detection methods to estimate invasive pneumococcal disease incidence in settings where pre-culture antibiotic use is common.