Year: 2017
Type: article
Abstract: To the Editor—In the 1 March 2017 issue of Clinical Infectious Diseases, Britt et al [1] reported on a cohort study of hospitalized Veterans Affairs (VA) patients treated with standard-, medium-, or high-dose daptomycin (6, 8, or ≥10 mg/kg total body weight, respectively) for vancomycin-resistant Enterococcus bloodstream infection. Although the authors conclude that high-dose daptomycin was associated with improved survival and microbiologic clearance in patients with such infection, we do not believe their data can be interpreted without additional information beyond what was presented in the article. There are several issues concerning the definition and analyses of the primary outcome. It seems that follow-up was truncated at 30 days, and it is not clear how the primary outcome of “overall survival” (P = .002 for standard- and P = .003 for standard-medium-dose daptomycin) differs from the secondary outcome of “30-day all-cause mortality” (P = .042 for daptomycin dose), because the curve presented [1, fig 1] tracks survival only to 30 days. Of the 911 patients analyzed in the study, 843 were no longer being followed up at 30 days. However, because 272 patients died, this suggests that 571 patients were censored at only 30 days. It is unclear why so many patients were lost to follow-up within such a short time frame. We find it surprising that 63% of subjects would be lost to follow-up in a large, nationalized healthcare system such as the VA. Of those subjects who did not have the event within 30 days, if any subsequent encounter in the VA system occurred during the study period of 2004–2014, we would suggest their outcome be updated with the information at last follow-up. If no additional information about vital status was available in the VA system beyond 30 days, subjects who died after 30 days but before 2014 could still be tracked in the National Death Index [2] to augment the VA data in lieu of censoring at 30 days. The data presented for the important secondary outcome of “time to microbiologic cure” are incomplete. The study compares microbiologic clearance among the different dosing strategies [1, fig 2]. Because the duration of bacteremia can strongly influence which dosing strategy a clinician chooses, one should evaluate for such an association, but this was not presented. The authors also do not provide the absolute number of subjects who reached the end point of microbiologic cure, which would allow us to determine how many subjects were censored in this analysis. Finally, it is unclear why the authors used the stepwise approach to build the final Cox model. This strategy may be appropriate if the goal is predictive modeling, but when testing a specific hypothesis we believe this is a starting point at best and should not replace carefully examining other covariates or potential confounders central to the hypothesis. For example, time to admission, polymicrobial bacteremia, and Acute Physiology and Chronic Health Evaluation (APACHE) score were strongly associated with overall survival in the unadjusted analysis. However, these variables were not in the final multivariable model, and we are not told how they are associated with daptomycin dose, how adjusting for them individually affects the daptomycin association with the primary outcome, or how forcing them into the final model affects the results. We applaud the authors for attempting to answer an important and highly relevant clinical question. However, given the data presented, we do not believe their findings and conclusions are justified. Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
Source: Clinical Infectious Diseases
DOI: 10.1093/cid/cix551
Institutions: University of Michigan, Michigan Medicine
Language: English
Subfield: Pharmacology
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2625366002