Derivation and validation of a new score to predict long-term survival after sudden cardiac arrest

Background: There is insufficient information about the long-term prognosis of sudden cardiac arrest (SCA) survivors. We therefore derived a clinical score (Sudden Cardiac Arrest-mortality score, SCA-MS) that predicts long-term mortality in patients surviving to hospital discharge and validated it in an independent cohort of SCA survivors.

Methods: A total of 1433 SCA survivors data were collected, who were discharged from the hospitals of the University of Pittsburgh Medical Center between 2002 and 2012. The overall cohort was randomly divided into two near equal cohorts used for the derivation and validation of the SCA-MS, respectively.

Results: The derivation cohort included 768 patients and identified serum potassium level>4.2 mg/dL at admission, the presence of atrial fibrillation at any time during the index hospitalization, and the presence of asystole or pulseless electrical activity as the initial documented rhythm as independent predictors of long-term mortality. Based on the multivariable modeling result, one point was assigned to each one of these variables to create the SCA-MS that ranged from 0 to 3. In the validation cohort, the SCA-MS was predictive of long-term mortality (hazards ratio = 1.69, 95% confidence interval 1.50-1.91, P < 0.001) per 1-point increment in the SCA-MS.

Conclusions: We describe a new clinical score that predicts long-term survival after SCA based on serum potassium levels at the admission, presence of atrial fibrillation, and documented rhythm of SCA.