Although the study by Hawkins et al1 provides valuable new information regarding the epidemiologic features of persistent Staphylococcus aureus bacteremia (pSAB), it leaves certain important questions unanswered and is potentially misleading regarding others. First, the authors state that vancomycin drug levels were “not determined ” in their study, precluding comment on the relationship between vancomycin levels and outcomes, which, as noted by the authors,1 is an important and currently controversial topic. It is hard to imagine that vancomycin levels truly were not obtained for their subjects as a part of routine clinical care. Presumably, in their study, “not determined ” means “not extracted during the medical record review.” This would seem to be a significant omission but one that could be readily remedied by a focused search of laboratory records. Such an analysis might provide insights into whether more aggressive vancomycin dosing can prevent pSAB or the associated adverse clinical outcomes. Second, regarding clinical outcomes, the abstract emphasizes that these were significantly worse among patients with pSAB, yet the multivariable analysis found no independent association of pSAB with crude mortality (the only outcome for which multivariable analysis was reported). It would be important to know whether the other outcomes (hospital duration, sepsis, and attributable mortality) retained their significant associations with pSAB in multivariable analysis. If so, these results should be presented. If not, the emphasis on adverse outcomes would seem unwarranted, at least from the perspective of the currently implied causal relationship.
Johnson JR. Persistent Staphylococcus aureus Bacteremia. Arch Intern Med. 2008;168(7):772–773. doi:10.1001/archinte.168.7.772