Removal of central venous catheters from patients with candidemia.

Sir—We read with interest the recent article " Should Vascular Catheters Be Removed from All Patients with Candide-mia? An Evidence-Based Review " by Nucci and Anaissie [1], who attempt to clarify a controversial issue related to the clinical management of this infection. As the authors of a study [2] cited in their article, we were surprised by the interpretation of some of our data, and we would like to add our comments on this topic. First, in our retrospective study [2], removal of central venous catheters (CVCs) was associated with a significant decrease in the mortality rate (OR, 0.62; 95% CI, 0.38–0.99;). The mortality rate P p .047 among patients whose catheters were not removed was 65%, whereas, among patients whose catheters were removed, it was 26%. This measure's impact on mortality was statistically modest on multi-variate analysis. However, the majority of our study population, as well as other patient populations with candidemia, were critically ill patients who had a high risk of dying of causes other than Candida infection. In a number of these patients, even the most effective medical treatment is unable to change the poor prognosis, as confirmed by our data on the efficacy of adequate antifungal treatment. We agree with Nucci and Anaissie [1] that, in the absence of adequate prognosis stratifica-tion, the high mortality rate may have resulted from selection bias (i.e., if catheters had been removed from patients with better prognoses, the mortality rate in this group would have been lower). They propose resolving this issue by means of a prospective, randomized trial. We would like to raise doubt about whether performing this kind of study is ethical. In fact, according to their analysis, у3 studies have shown that CVC removal has some benefit in patients with candidemia. Because candidemia is associated with a high mortality rate, it would be more acceptable , with respect to ethical issues, to perform a retrospective, case-control study. Second, we found that the mortality rate in the intensive care unit (ICU) differed from the rates in the surgical and medical wards, which Nucci and Anaissie [1] considered " puzzling. " Actually, the authors inverted the factors of our comparison between surgical/medical wards and the ICU; therefore, they reported an unlikely result. In fact, we found, as expected , that the mortality rate in the surgical and medical wards was significantly lower than that insized the lack of correlation …