Osteopenia in patients with HIV infection is not associated with elevated sTNFR2 levels.

in the absence of deep organ involvement [6]—perhaps because drug penetration into infected organs is poorer, and drug concentration levels fall below those that are required to kill Candida species. To address the issue of homogeneity of the study population, analyses were also performed on subpopulations with can-didemia or with positive cultures of specimens from deep, normally sterile sites— populations from which patients with symptomatic candiduria were excluded [2]. These analyses showed highly statistically significant differences by day 10 in overall response (80% vs. 40%;), P ! .001 clinical response (82% vs. 45%;), P ! .001 mycological response (86% vs. 47%; P !), Candida-attributable mortality rates .001 by day 33 (5% vs. 21%;), and P p .033 the rate of culture-confirmed clearance of the infection, which remained more than twice as fast among those receiving combination therapy, compared with those receiving monotherapy (hazard ratio, 2.4;). P p .011 The 2 deaths on day 1 referred to by Herbrecht et al. [1] resulted, respectively, from a tension pneumothorax (a complication of mechanical ventilation) and a discontinuation of life-support systems at the request of the family because of the patient's irremediable underlying pathol-ogies. Therefore, these deaths contributed to the overall mortality rate but not the Candida-attributable mortality rate. It is usual practice in antifungal trials to base efficacy assessments on a modified intent-to-treat population, as we described [2]. By applying the primary efficacy variable to an intent-to-treat population composed of all patients who received at least 1 dose of study drug and by taking a conservative approach to the analysis of patients with no post-treatment disease as-sessment—such patients (including the 3 referred to in [1]) being classified as experiencing treatment failure—we showed an overall response by day 10 in 51% of the placebo group (35 of 69 patients) and 79% of the Mycograb group (54 of 68 patients)—a highly statistically significant difference (). P ! .001 All patients had culture-positive can-didiasis before starting therapy. This was a prerequisite for study entry. Because it takes several days for Candida to grow on culture, there was inevitably a lag between the time the culture was performed, the positive culture result, obtaining patient consent, and the day of study entry. The protocol allowed 3 days for this process, and samples were not recultured on day 0 for all patients. The protocol required all patients to have clinical sepsis at the time of study entry. If …