This is the authors' abstract. We don't add key points for this paper.
Purpose Recurrent bloodstream infections (BSI) represent a significant clinical challenge, yet recurrence patterns and predictive factors remain poorly defined. We aimed to identify risk factors for BSI recurrence and characterize its microbiological and temporal patterns. Methods We conducted a population-based retrospective cohort study of all patients with confirmed BSI in 2017 in Skåne, southern Sweden. Recurrent BSI was defined as a positive blood culture within one year of the index BSI, either > 48 h apart for different species or > 14 days for identical species. Patients with recurrent BSI were included along with controls matched on blood culture date who experienced a single BSI and were discharged alive. We performed multivariable logistic regression, analysed species concordance between index and recurrent BSI, and examined time to recurrence by index bacterial group. Results Of 3,918 BSI episodes, 416 (10.6%) recurred within one year, with a median time to recurrence of 63 days (IQR 31–151); 60% occurred within 90 days. Independent risk factors included higher Charlson Comorbidity Index (CCI) (OR 13.9, 95% CI 4.11–50.4 for ≥ 6), immunosuppression (OR 1.96, 1.24–3.12), invasive procedures within the previous 6 months (OR 3.83, 2.12–7.15), undrained infection site at discharge (OR 2.22, 1.08–4.71) and septic shock (OR 4.05, 1.52–11.9). Conclusions Bloodstream infection (BSI) recurrence affects over 10% of patients, most often within three months. Independent risk factors included CCI ≥ 6, immunosuppression, recent invasive procedures, undrained infection sites and septic shock. These findings support individualized follow-up in high-risk patients, and further studies should target modifiable risk factors.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Mattisson et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: