Meta‐analysis of a Possible Signal of Increased Mortality Associated with Cefepime Use

validity of guidelines requires an analysis of their recommendations—including early removal—and a comprehensive assessment of the reportedly beneficial outcomes that form the basis of these guidelines (Table 1). The authors also suggest that our multivariate analysis does not answer the question of whether early removal affects outcome. Because our question was “whether early CVC removal ... was associated with the beneficial outcomes that form the basis for current recommendations for early CVC removal” [3, p 296], we performed logistic regression analysis, a robust method for examining outcome predictors. Our consistent findings that early removal did not predict poor outcome answer our question convincingly. Surprisingly, the authors of both letters support current guidelines [4, 5], notwithstanding the fact that the concerns they raise about our analysis apply to many of the studies from which these guidelines were derived and despite their serious limitations, including small sample size, retrospective data collection, inclusion of patients without candidemia and/or without CVC, and no time points for early removal [3]. In contrast, our report examined the largest series of candidemia cases, predefined time points for early removal, and assessed 12 end points using prospective protocol-based evaluation. Our consistent findings across outcomes are a testimony to the robustness of our analysis. How should our findings be interpreted at the bedside? The compelling evidence that we present that patients with candidemia are unlikely to suffer adverse outcomes if their CVC is not promptly removed provides clinicians with the opportunity to individualize their decision when managing diverse populations, instead of blindly following guidelines. Indeed, one size does not fit all.