Cardiac arrest is a major healthcare burden; survival has historically been poor. Contemporary therapies have been associated with improvements in out-of-hospital cardiac arrest survival in other healthcare systems.
To determine temporal trends and predictors of 30-day and overall survival in patients with cardiac arrest who survive to hospital admission in Scotland.
We conducted a consecutive individual patient linkage study using data from the Scottish Morbidity Records held by the Information Services Division, National Health Service Scotland. We identified all patients who had an index cardiac arrest hospitalisation from 1st January 1990 to 31st December 2014. The primary outcome was 30-day mortality. Generalised additive models were used to estimate temporal trends. Year of admission was the primary explanatory variable, adjusted for age, sex, comorbidities (stroke, myocardial infarction, heart failure), Scottish Index of Multiple Deprivation (SIMD, a national deprivation score), and angiography within 30 days of admission. Cox regression models were constructed for overall mortality, adjusting for the above variables.
In total, 47,692 patients had an index hospitalisation with cardiac arrest in Scotland between 1990 and 2014. The mean age was 69±16 years; 45% (n=21,257) were female. Most patients (n=24,867, 52.4%) were in top two SIMD quintiles (greater deprivation). Incidence was lowest in 1990 (27 per 100,000 population), rising until 1998 (47 per 100,000) before declining (mean 30 per 100,000 for 2010-2014) (Figure 1A). Overall rates of angiography and PCI at 30 days were low (5.4% and 3.1% respectively), albeit higher in more recent years (14.1% and 9.6% respectively for 2010-2014).
Thirty-day mortality was high but decreased over time (73.7% from 1990-1994 to 63.1% from 2010-2014, p<0.001). This effect persisted after adjustment for covariates. The predicted 30-day mortality for a 69 year old patient was highest in 1990 (75%, 95% confidence interval [CI] 73-77%) and lowest in 2014 (66%, 95% CI 64-68%) (Figure 1B). Age, heart failure and greater deprivation were other independent predictors of 30-day mortality. In contrast, previous myocardial infarction and angiography within 30 days were predictors of reduced 30-day mortality. On Cox regression analysis, age (hazard ratio [HR] 1.15, 95% CI 1.14-1.16 per 10 years) and heart failure (HR 1.11, 95% CI 1.08-1.14) were independent predictors of overall mortality. Thirty-day angiography (HR 0.29, 95% CI 0.28-0.31) was the strongest independent predictor of survival, likely reflecting selection bias, in addition to year of admission, previous myocardial infarction and less deprivation.
Hospital episodes with cardiac arrest in Scotland peaked in the late 1990s and have since fallen. Thirty-day mortality remains high, with an improvement over time that is independent of baseline patient characteristics.