Changes in causes of death over time after treatment for invasive aspergillosis.

Abstract

Assessment of response to invasive aspergillosis (IA) therapy has been challenging in treatment trials. The causes of death over 12 weeks were categorized prospectively by a blinded data review committee using a priori defined criteria in participants in a randomized comparative trial of voriconazole versus amphotericin B as first-line therapy of proven or probable IA. Death occurred in 98 of 277 patients during the 12-week course of study. Seventy-three of the 98 deaths (74%) occurred in the first 6 weeks; 25 deaths occurred during the second 6 weeks. Of the 73 deaths during the first 6 weeks, 50 (68%) were judged to be attributable to IA. Of the 25 deaths during the second 6 weeks, only 6 (24%) were judged to be attributable to IA. Fifty of the 56 deaths (89%) attributable to IA occurred during the first 6 weeks. These data suggest that most deaths due to IA occur during the first 6 weeks after the start of therapy and 6 weeks may be a better interval to judge the effectiveness of antifungal therapy because most deaths after 6 weeks are due to causes related to the underlying disease and its treatment rather than due to IA. Attributable mortality when assessed using a priori definitions and conducted in a blinded manner by a central data review committee can be useful in the assessment of IA therapy.