Combination antimicrobial therapy in patients with Staphylococcus aureus bacteraemia-a post hoc analysis in 964...

The evidence for using combination antimicrobial therapy (CoRx) in Staphylococcus aureus bacteraemia (SAB) is limited. We aimed to investigate whether CoRx is associated with higher survival or lower SAB-related late complications.

We performed a post hoc analysis of a prospective SAB cohort study. CoRx was defined as a cell wall-active antistaphylococcal agent plus either rifampicin, a fluoroquinolone, fosfomycin or an aminoglycoside. To adjust for survivor bias multivariable Cox models that included CoRx as a time-dependent covariable were calculated.

Of 964 evaluable patients, 512 (53%) received CoRx, most of them (301/512, 59%) rifampicin-containing CoRx. All-cause mortality after 30 and 90 days was similar for the two groups, although the patients in the CoRx group had more often endocarditis, deep-seated or disseminated infections and severe sepsis/septic shock. For the entire cohort, only age, comorbidity and severe sepsis/septic shock were associated with a higher mortality and infectious disease consultation, but not CoRx with a lower mortality. However, in the subgroup of patients with implanted foreign bodies or devices CoRx was independently associated with a lower mortality at 30 days (hazard ratio 0.6, 95% confidence interval 0.3-1.1) and at 90 days (hazard ratio 0.6, 95% confidence interval 0.4-0.9). SAB-related late complications in this subgroup occurred in 15 (10.6%) of 142 patients in the monotherapy group vs. nine (4.5%) of 202 patients in the CoRx group (p 0.03).

In a setting of optimized management of adult patients with SAB secured by infectious disease consultations, this observational study could not prove CoRx to be independently associated with improved survival or reduced late complications in the entire cohort. However, administration of CoRx may be associated with lower mortality and fewer SAB-related late complications in the subgroup of patients with implanted foreign bodies or devices. Prospective randomized trials should be performed to prove this benefit.