The importance of bacterial sepsis in intensive care unit patients with acquired immunodeficiency syndrome: implications for future care in the age of increasing antiretroviral resistance

Crit Care Med. 2001 Mar;29(3):548-56. doi: 10.1097/00003246-200103000-00013.


Objective: To describe the clinical characteristics and outcomes of patients with acquired immunodeficiency syndrome (AIDS) admitted to the intensive care unit (ICU).

Design: An observational cohort study with retrospective chart review.

Setting: ICU of an urban university medical center.

Patients: Consecutive ICU admissions of patients with AIDS at an urban university medical center between December 1993 and June 1996.

Interventions: None.

Measurements and main results: For each patient, we recorded ICU admission diagnosis, clinical characteristics, and outcome. Among 129 ICU admissions of patients with AIDS, 102 (79%) were admitted for infections, of which (45%) had infections caused by bacteria. Pseudomonas aeruginosa, Staphylococcus aureus, and other enteric pathogens were the most frequent isolates. Pneumonia accounted for 65% of 102 admissions for infections. Overall hospital mortality was 54%, but mortality was higher (68%) for patients with bacterial sepsis. Neutropenia was associated with differences in unadjusted survival rates, whereas CD4 counts were not. Independent predictors of hospital mortality included increasing acute physiology scores and severity of sepsis.

Conclusions: In our ICU, among patients with AIDS, sepsis resulting from bacterial infection is now a more frequent cause of admission than Pneumocystis carinii pneumonia. Severity of illness and the presence of severe sepsis were the clinical predictors most associated with increased mortality. Patients who are not receiving or responding to highly active antiretroviral therapy may become as likely to be admitted to an ICU with a treatable bacterial infection as with classic opportunistic infections. Therefore, broad-spectrum empirical antibacterial therapy is particularly important when the etiology of infection is uncertain.

Publication types

  • Research Support, Non-U.S. Gov't
  • Research Support, U.S. Gov't, Non-P.H.S.

MeSH terms

  • AIDS-Related Opportunistic Infections / drug therapy*
  • AIDS-Related Opportunistic Infections / microbiology
  • AIDS-Related Opportunistic Infections / mortality*
  • Academic Medical Centers
  • Adult
  • Anti-Bacterial Agents / therapeutic use
  • Anti-HIV Agents / therapeutic use*
  • Antiretroviral Therapy, Highly Active
  • Bacterial Infections / drug therapy*
  • Bacterial Infections / microbiology
  • Bacterial Infections / mortality*
  • CD4 Lymphocyte Count
  • Critical Care* / methods
  • Critical Care* / trends
  • District of Columbia / epidemiology
  • Drug Resistance, Microbial
  • Female
  • Hospital Mortality
  • Humans
  • Infection Control
  • Length of Stay / statistics & numerical data
  • Logistic Models
  • Male
  • Middle Aged
  • Neutropenia / complications
  • Predictive Value of Tests
  • Prognosis
  • Retrospective Studies
  • Risk Factors
  • Sepsis / drug therapy*
  • Sepsis / microbiology
  • Sepsis / mortality*
  • Survival Analysis
  • Treatment Outcome


  • Anti-Bacterial Agents
  • Anti-HIV Agents