-
Views
-
Cite
Cite
David Andes, Nasia Safdar, John W. Baddley, Bart-Jan Kullberg, Peter G. Pappas, Reply to Anaissie and Nucci and to Cisneros et al, Clinical Infectious Diseases, Volume 55, Issue 6, 15 September 2012, Pages 894–895, https://doi.org/10.1093/cid/cis523
Close - Share Icon Share
Extract
Tothe Editor—We thank Drs Anaissie and Nucci for their interest in our presentation of the patient-level analysis of data from major candidemia and invasive candidiasis treatment trials [1]. They object to the interpretation of the study observation regarding the impact of central venous catheter (CVC) management. We applaud their attempts over the years to carefully assess the impact of this treatment strategy. In fact, analysis of the current data was based largely upon their prior editorials suggesting the importance of patient severity of illness and other potentially confounding disease factors in interpreting the impact of this strategy [2]. The data from this analysis found that removal of the CVC was associated with a >10% reduction in mortality and clinical success across a wide range of APACHE II scores. Furthermore, sensitivity analyses did not identify an interaction of CVC and other cofactors of importance for determining outcome.
Our colleagues provide a list of additional critiques of the analysis and on this basis conclude the data are insufficient to consider a recommendation for CVC removal in many patients. First, they note a different result from their recent analysis of CVC removal among 2 of the studies used in our investigation [3]. They suggest this is due to the lack of timing information linked to CVC removal. They hypothesize that CVC removal beyond 24–48 hours will not favorably impact patient outcome. We have some concern that our colleagues did not identify the impact of CVC removal because of the smaller sample size and the grouping of patients with CVC removal beyond 24–48 hours in the “CVC not removed” portion of the analysis. In fact, more than a quarter of the patients counted in the “catheter not removed” group indeed underwent catheter removal, but the device was extirpated after 48 hours of diagnosis. This misclassification bias potentially diluted an effect on outcome. These points are thoughtfully discussed in an accompanying editorial [4]. Additionally, similar analysis of the identical database found catheter removal linked to favorable outcome [5].
