Predictors of Short- and Long-term Survival in HIV-Infected Patients Admitted to the ICU: Short-term Outcome
Overall in-ICU and in-hospital mortality rates were 20.6% and 39.0%; other mortality rates were as follows: respiratory failure group, 16.7% and 33.9%; neurologic failure group, 23.2% and 41.1%; heart failure group, 25.0% and 68.8%; severe sepsis group, 38.9% and 58.3%; and miscellaneous admission causes group, 12.1% and 24.2%. Significant differences were found across admission groups for in-ICU (p=0.026) and in-hospital mortality rates (p=0.002). Mortality rates were lowest in the respiratory and miscellaneous groups, and highest in the neurologic, cardiac, and sepsis groups. Flexible Bronchoscopy
The mean time between ICU discharge and hospital discharge or death was 18.3±15.8 days (median, 20 days).
Results of the univariate and multivariate analysis for short-term outcomes in all 354 patients are presented in Table 2.
In the respiratoiy admission group, the in-ICU mortality rate was 14% for PCP patients vs 19% for patients with other causes of respiratory failure (p, not significant [NS]).
In the PCP subgroup, nonmechanically ventilated patients had an in-ICU mortality rate of 5.1% vs 40.0% for mechanically ventilated patients. Mortality rates in patients with PCP did not differ by CPAP use (12.5% vs 14.5% in CPAP nonmechanically ventilated patients) but increased significantly between the two CPAP duration categories (6.25% vs 27.7% for the <5 days and >5 days groups, respectively; p=0.03). The in-ICU mortality rate in patients who required mechanical ventilation after failure of CPAP was 46.2%. In-hospital mortality rates were 25.3% for PCP patients and 41.1% for patients with other respiratory causes of admission (p=0.03).
No significant differences were found across the three neurologic admission subgroups (toxoplasmosis, intracerebral space-occupying lesions, and meningitis) for either in-ICU mortality (27.1%, 25%, and 6.25%, respectively) or in-hospital mortality (35.6%, 50%, and 50%, respectively). In-ICU mortality rates varied as a function of the need for mechanical ventilation (16.7% and 26.15%; p=0.03) and the duration of mechanical ventilation (20.7% and 30.7%; p=0.05), but not between the two admission Glasgow score categories (<8, >8) (23.8% and 28.6; p=0.6). However, the admission Glasgow score was significantly associated with the in-hospital outcome (p=0.03). Univariate analysis showed that the SAPS I (p=0.005) and the duration of mechanical ventilation (p=0.005) were significantly associated with the in-ICU outcome.
Table 2—Short-term Outcome
|Univariate p Value||Univariate p Value||Multivariate Analysis|
|p Value||OR (95% Cl) ‘|
|HIV disease stage||0.08||0.000001||0.016|
|AIDS in ICU||1.62(1.10-2.40)|
|AIDS prior to ICU||2.63 (1.77-3.88)|
|For AIDS patients, time since AIDS diagnosis, d||0.17||0.00008||0.04|
|CD4 count, cells X109/L||0.56||0.37|
|No. of previous opportunistic infections||0.03||0.008|
|Admission cause group||0.08||0.00001|
|For patients with MV, duration of MV, d||0.000001||0.08||0.0001|