Trends in Incidence and Outcomes of Cardiac Arrest Occurring in Swedish ICUs

Abstract

OBJECTIVE: To determine temporal trends in the incidence of cardiac arrest occurring in the ICU (ICU-CA) and its associated long-term mortality. DESIGN: Retrospective observational study. SETTING: Swedish ICUs, between 2011 and 2017. PATIENTS: Adult patients (≥18 yr old) recorded in the Swedish Intensive Care Registry (SIR). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: ICU-CA was defined as a first episode of cardiopulmonary resuscitation and/or defibrillation following an ICU admission, as recorded in SIR or the Swedish Cardiopulmonary Resuscitation Registry. Annual adjusted ICU-CA incidence trend (all admissions) was estimated using propensity score-weighted analysis. Six-month mortality trends (first admissions) were assessed using multivariable mixed-effects logistic regression. Analyses were adjusted for pre-admission characteristics (sex, age, socioeconomic status, comorbidities, medications, and healthcare utilization), illness severity on ICU admission, and admitting unit. We included 231,427 adult ICU admissions. Crude ICU-CA incidence was 16.1 per 1,000 admissions, with no significant annual trend in the propensity score-weighted analysis. Among 186,530 first admissions, crude 6-month mortality in ICU-CA patients was 74.7% (95% CI, 70.1–78.9) in 2011 and 68.8% (95% CI, 64.4–73.0) in 2017. When controlling for multiple potential confounders, the adjusted 6-month mortality odds of ICU-CA patients decreased by 6% per year (95% CI, 2–10). Patients admitted after out-of-hospital or in-hospital cardiac arrest had the highest ICU-CA incidence (136.1/1,000) and subsequent 6-month mortality (76.0% [95% CI, 73.6–78.4]). CONCLUSIONS: In our nationwide Swedish cohort, the adjusted incidence of ICU-CA remained unchanged between 2011 and 2017. More than two-thirds of patients with ICU-CA did not survive to 6 months following admission, but a slight improvement appears to have occurred over time.