2005 HSR&D National Meeting Abstract
1054 — Evaluation of Predictive Tools to Define Need for ICU in Patients with Community-Acquired Pneumonia
Restrepo M (South Texas Health Care System/Audie L. Murphy Division)
Mortensen E (South Texas Health Care System/Audie L. Murphy Division)
Anzueto A (South Texas Health Care System/Audie L. Murphy Division)
Pugh J (South Texas Health Care System/Audie L. Murphy Division)
Community-acquired pneumonia (CAP) is the leading infectious cause of death in the United States. There are several predictive rules to define low risk group status and the need for intensive care unit (ICU) admission. Our objective was to examine the most commonly used prediction rules to assess their prediction of mortality and the need for ICU admission due to CAP.
A retrospective cohort study was conducted at two tertiary teaching hospitals. Eligible subjects were admitted with a diagnosis of CAP between 1/1/1999 and 12/31/2001, had a chest x-ray consistent with, and an ICD-9 diagnosis of, CAP. Subjects were excluded if they were “comfort measures only” or transferred from another acute care hospital. Patients were stratified between severe and non-severe CAP according to Pneumonia Severity Index (PSI), revisited American Thoracic Society (rATS) criteria, and the British Thoracic Society (CURB-65 rule).
Data was abstracted on 787 subjects at the two hospitals, and 154 (19.6%) were admitted to the ICU. Mortality was 9.2% at 30-days and 13.6% at 90-days. 11% of the patients were intubated and 6% received vasopressors. When comparing CURB-65 and PSI class for 30-day mortality, 20/461 (4%) vs. 16/409 (3.9%) were low risk; 29/187 (15%) vs. 26/266 (10%) were moderate risk; and 20/116 (17%) vs. 30/112 (27%) were high risk. When comparing CURB-65 and PSI class for the need of ICU admission, 63/461 (14%) vs. 44/409 (11%) were low risk; 47/187 (25%) vs. 62/266 (23%) were moderate risk; and 36/116 (31%) vs. 48/112 (43%) were high risk. ATS severity criteria predicted 30-day mortality in 21/74 (28%) and ICU admission in 70/74 (95%).
This study confirms the power of the modified ATS rule to predict the need for ICU admission. The CURB-65 rule may be used as an alternative tool to the PSI for the detection of low risk patients, but is not a good rule to define ICU admission.
Described rules are imperfect and have significant limitations due to the difficulty of application to individual patients. Further studies are needed to develop clinical prediction tools for high-risk patients requiring ICU admission.