Factors predictive of 30-day postoperative mortality in HIV/AIDS patients in the era of highly active antiretroviral therapy.

Abstract

Factors that predict HIV (human immunodeficiency virus)/AIDS patient postoperative mortality have remained poorly defined. The primary objective of this study was to identify factors predictive of short-term, postoperative mortality in HIV/AIDS patients. The secondary objective of this study was to develop a scoring system that would predict short-term postoperative mortality in HIV/AIDS patients. We retrospectively reviewed all HIV/AIDS patients who underwent surgical procedures in British Columbia, Canada, between April 1995 and March 2002. The primary outcome evaluated was 30-day postoperative mortality. Demographic, clinical, and hospitalization-related data were obtained and utilized to predict outcomes using a logistic regression model. A total of 2305 procedures were carried out on 1322 patients during the study period. Admissions were classified as urgent/emergent for 1311 procedures (57%) and the overall 30-day postoperative mortality was 9.5% (126 deaths). Urgent/emergent admission, older age, prior surgery, a CD4 cell count of ≤ 50 cells/mm, a hemoglobin level ≤ 120 g/L, and a white blood cell count >11 g/L within 90 days before the surgical procedure was predictive of an increased 30-day postoperative mortality in a multivariate model. Using these variables, we formulated the HIV Surgical Mortality Score (HSMS) to obtain the median-estimated probability of postoperative death. For accurate preoperative mortality risk stratification for HIV/AIDS patients, we have found that several clinical and laboratory variables must be evaluated. If appropriately validated, our proposed HSMS could be utilized to estimate the probability of short-term postoperative death among HIV/AIDS patients.