Measuring mortality in Staphylococcus aureus bloodstream infections: are 3 months of follow-up enough?

Abstract

Hanses et al. [1] recently published data from a retrospective study of patients with Staphylococcus aureus bacteremia (SAB). The authors observed a considerable additional mortality due to SAB beyond 90 days and concluded that long-term survival data should be taken into account in outcome studies involving patients with SAB. Our experience with the INSTINCT (Invasive S. aureus Infection Cohort) study [2] suggests that it is the severity of the underlying disease which is responsible for the vast majority of additional deaths beyond 90 days. In Fig. 1, the cumulative survival of 200 patients with SAB is shown for patients that were recruited from 1 January 2006 to 21 December 2009 at the University Hospital Cologne as part of the INSTINCT study. One-year follow-up data were available for all patients, and patients were stratified by the severity of underlying disease, as expressed by the Charlson weighted index of comorbidity (WIC) [3]. Figure 1 clearly shows that beyond 90 days, there is little change in the survival of patients with mild underlying disease (WIC 2), whereas an additional 13% of patients with severe underlying disease (WIC [2) die. Another often-mentioned rationale for extending the follow-up beyond 90 days is the recurrence of SAB, which is either due to relapse of the original infection or to reinfection. Chang et al. [4] used pulsed-field gel electrophoresis to differentiate relapse and reinfection and found that relapse occurs mostly within 2 months after completion of antibiotic therapy. However, even with molecular typing techniques, relapse and reinfection cannot be differentiated reliably, since patients can be reinfected by colonizing strains, resulting in a possible overestimation of relapse. Therefore, we propose that a follow-up period of 90 days is sufficient to evaluate SAB-related complications in most patients. Longer follow-up periods are warranted when long treatment courses are monitored as, for example, in suppression therapy.