RESUMEN
STUDY OBJECTIVE: To indicate predictors of witnessed hypothermic cardiac arrest. METHODS: We conducted a retrospective analysis of 182 patients with severe accidental hypothermia (i.e., with core body temperature of ≤28 °C) who presented with preserved spontaneous circulation at first contact with medical services. We divided the study population into two groups: patients who suffered hypothermic cardiac arrest (HCA) at any time between encounter with medical service and restoration of normothermia, and those who did not sustain HCA. The analyzed outcome was the occurrence of cardiac arrest prior to achieving normothermia. Hemodynamic and biochemical parameters were analyzed with regard to their association with the outcome. RESULTS: Fifty-two (29%) patients suffered HCA. In a univariable analysis, four variables were significantly associated with the outcome, namely heart rate (p < 0.001), systolic blood pressure (p = 0.03), ventricular arrhythmia (p = 0.001), and arterial oxygen partial pressure (p = 0.002). In the multivariable logistic regression the best model predicting HCA included heart rate, PaO2, and Base Excess (AUROC = 0.78). In prehospital settings, when blood gas analysis is not available, other multivariable model including heart rate and occurrence of ventricular arrhythmia (AUROC = 0.74) can be used. In this study population, threshold values of heart rate of 43/min, temperature-corrected PaO2 of 72 mmHg, and uncorrected PaO2 of 109 mmHg, presented satisfactory sensitivity and specificity for HCA prediction. CONCLUSIONS: In patients with severe accidental hypothermia, the occurrence of HCA is associated with a lower heart rate, hypoxemia, ventricular arrhythmia, lower BE, and lower blood pressure. These parameters can be helpful in the early selection of high-risk patients and their allocation to extracorporeal rewarming facilities.