A population-based nested case-control study was conducted in order to characterize patient factors and microbial species associated with recurrent bacteraemia.

All patients with bacteraemia in a Danish region during 1996–2006 were investigated. Recurrence was defined based on pathogen identity, site of infection and time frame, and not restricted to homologous pathogens.

We identified 8672 patients with first-time bacteraemia, of whom 1003 (12%) had a recurrence within 1year. The proportion of mono-microbial bacteraemia was similar for first (86%) and recurrent episodes (84%). An unknown focus was common in both episodes (22.7 and 29.1%, respectively). Independent predictors of a recurrence (incidence rate ratio, 95% confidence interval) included health care-associated (2.4; 1.9–3.0) and nosocomial bacteraemia (2.1; 1.8–2.6), poly-microbial Gram-positive bacteraemia (2.7; 1.6–4.6), and fungaemia (2.2; 1.4–3.5), a Charlson co-morbidity index score of 1–2 (1.7; 1.4–2.1), inappropriate empirical antimicrobial chemotherapy (1.3; 1.1–1.5), a gastro-intestinal tract focus (2.3; 1.7–3.0), a liver/biliary tract focus (2.7; 2.0–3.6), an iv-catheter focus (2.0; 1.4–2.8), endocarditis (2.7; 1.6–4.3), and an unknown focus (1.9; 1.5–2.3).

This study showed recurrent bacteraemia to be common and the following risk factors were identified: a health care-associated or nosocomial origin, poly-microbial or fungal aetiology, a focus within the abdomen, endocardium, iv-catheter-related or unknown, a Charlson co-morbidity index score of >1 and inappropriate empirical antimicrobial chemotherapy.