icon-    folder.gif   Conference Reports for NATAP  
  International AIDS Conference
Durban, South Africa
July 18-22 2016
Back grey_arrow_rt.gif
Mortality Prediction in HIV-Infected Persons with Critical Illness
  Reported by Jules Levin
IAS Durban 2016 July 18-22
Jason D. Goldman1,6, Kristina Crothers2, Christine T. Fong4 Maria Corcorran3, E. Lawrence Fuhrman5, Rui Lin5 Robert A. Black3, Kathleen M. Akgun7, Catherine L. Hough2, Mark M. Wurfel2, Shireesha Dhanireddy1
1. Division of Allergy and Infectious Disease; 2. Division of Pulmonary and Critical Care Medicine; 3. Department of Medicine;
4. Institute of Translational Health Sciences; 5. UW Medicine; and 6. School of Public Health; all at the University of Washington, Seattle, WA;
7. Department of Internal Medicine, Yale University School of Medicine, New Haven, CT


Despite declining mortality overall in HIV-infected (HIV+) persons, critical illness and intensive care unit (ICU) admission remain frequent. We compared hospital, 30-day and 1-year mortality prediction between the Acute Physiology and Chronic Health Evaluation II score (APACHE) and the Veterans Aging Cohort Study Risk Index (VACS) at ICU admission among critically ill HIV+ patients.
Methods: Electronic medical records, administrative data and state death data were reviewed for HIV+ persons admitted to an ICU at University of Washington or Harborview Medical Center. Logistic regression was used to model APACHE and VACS on the outcomes of hospital, 30-day and 1-year mortality. Model discrimination for each outcome was evaluated using the area under the curve (AUC), and comparisons of AUC made with DeLong method in Stata v13.
Results: We identified 4,914 HIV+ patients from 10/2007-12/2013; 8.5% developed critical illness resulting in an analytic cohort of 419 with ICU admission to medical (62%) or surgical (34%) services. Mean (standard deviation, SD) age was 47 (11) years, 17% were female, 60% were Caucasian, 149 of 356 (42%) had CD4 < 200 cells/μL, 161 of 363 (44%) had detectable HIV RNA and 42% were hepatitis C virus co-infected. On admission, the mean (SD) APACHE score and VACS Index were 20 (9) and 59 (31). Mortality in-hospital was 7%, at 30-days was 8%, and at 1-year was 20%. AUC for APACHE vs. VACS were similar for hospital, 30-day, and 1-year mortality, with a trend for improved discrimination of VACS Index over APACHE II score for 1-year mortality prediction (Table).
Conclusions: APACHE and VACS were moderate to good at predicting short and long-term mortality in HIV+ patients admitted to the ICU. These metrics can be used by clinicians, researchers and health administrators to predict mortality or standardize admission disease severity in HIV+ patients with critical illness.