Prediction of mortality of APACHE-II and APACHE-III scores in patients admitted to the intensive care unit (ICU) for intoxication

Objective

To study the characteristics and mortality in intoxicated patients admitted to the ICU, and prediction of mortality for usual prognostic indexes.

Methods

A multicenter study conducted from 2008 to 2013 in three Spanish hospitals. We collected clinical and demographics data. The differences between observed-to-predicted mortality were analyzed whith the Hosmer-Lemeshow test. Discrimination was evaluated by área under the ROC curve. P < 0.005 was statistically significant.

Results

We studied 119 patients. Mean age: 44.42 ± 13.85 years. Drug intoxication in 77.3% of patients, alcohol in 16.8% and caustic 9.2%. 78.3% was a suicide attempt. The mean Glasgow Coma Scale in 72.5% were < 8 points. 69.7% required mechanical ventilation. The mortality of patients hospitalized for caustic ingestion was 54.5% and 1.9% in the rest (p < 0.001).

The multivariate analysis showed that with SAPS-3 equal gravity (OR: 1.19 (1.02-1.39)) the mortality of patients who ingested caustics was higher (OR: 560.34 (11.64-26.973.83)) than the rest.

The mortality predicted by SAPS-3 (general equation) was 26.98% and the observed 6.7%, Hosmer-Lemeshow (H = 36.47) (p < 0.001). With the APACHE-II and APACHE-III scores no statistically significant differences were observed, being the predicted mortality for APACHE-II of 7.57% (Hosmer-Lemeshow Test: H = 6.96, no difference is) and for APACHE-III of 8.15% (Hosmer-Lemeshow Test: C = 3.51, not discrepancies).

Conclusions

APACHE-II and APACHE-III scores may adequately predict the probability of death in intoxicated patients, in contradistinction to the SAPS-3 score that overestimates the mortality. The ICU admission by intoxication is rare. A high percentage of patients have decreased level of consciousness and required mechanical ventilation. The mortality o f patients hospitalized for caustic ingestion is higher to the rest.