Differences in long-term mortality after out-of-hospital, in-hospital and intensive care unit cardiac arrests in finland

Methods We performed a retrospective registry based study on adult patients treated in five Finnish university hospitals’ ICUs between 2003 and 2013. We included patients with admission diagnosis of cardiac arrest according to the Acute Physiology and Chronic Health Evaluation III (APACHE III) and patients with positive value for Therapeutic Intervention Scoring System (TISS-76) item “cardiac arrest and/or countershock within past 48 hours”. Patients were classified by cardiac arrest location as out-of-hospital cardiac arrests (OHCA), in-hospital cardiac arrests (IHCA) and intensive care unit cardiac arrests (ICU-CA). Data on long-term outcome were retrieved from Finnish Population Register Centre. We excluded re-admissions and cases with missing baseline data. We calculated crude long-term mortality using life tables for the whole population and for patients who survived initial 30 days. The median time to death with interquartile range (IQR) was calculated for the nonsurvivors and compared between cardiac arrest location groups using Kruskal-Wallis test with pairwise comparison. In order to determine independent association of cardiac arrest location with mortality, we used Cox regression with case mix adjustment using pre-admission functional status, Simplified Acute Physiology Score II (SAPS II), chronic comorbidities, age, admission type (emergency and operative) and admission year.