3333Contemporary mortality rates in myocardial infarction patients in Sweden: a tale of two registries

Background Sweden has contributed to the understanding of the long-term prognosis after myocardial infarction (MI) utilising the quality registry SWEDEHEART, including patients admitted to heart intensive care, and the National Patient Registry (PAR), based on administrative records for Swedish hospitals. As registration procedures differ between the registries, and not all MI patients are admitted to heart intensive care, MI patients identified in SWEDEHEART and PAR, respectively, will yield different cohorts of patients. This may result in different epidemiological research findings regarding prognosis after MI. Purpose To study MI populations identified in SWEDEHEART and PAR, respectively, and investigate potential differences in mortality outcome. Methods Patients hospitalised with an MI primary diagnosis (ICD-10 I21) between 2002 and 2015 were identified using SWEDEHEART and PAR. The analysis time started at the date of hospital admission and survivors were followed for 365 days. Kaplan-Meier analysis was used to estimate survival by cohort category controlling for age and gender. Results Excluding cases with invalid data (n=1,905), 225,612 and 282,118 SWEDEHEART and PAR patients, respectively, were identified. We found 213,367 patients in both SWEDEHEART and PAR, whereas 12,245 and 68,751 patients were unique to SWEDEHEART and PAR, respectively. The one-year survival probability after MI in the SWEDEHEART population was 0.841, compared to 0.788 in PAR (Figure). This discrepancy can be explained by high mortality among patients not covered by SWEDEHEART and persists after controlling for age and gender (Table). To what extent differences in registration procedures and other patient characteristics can explain the mortality difference is an area for further research. One-year survival by age and gender Age ± 1 year Male Female PAR S.H. Diff. PAR S.H. Diff. 65 0.922 0.936 0.015 0.919 0.936 0.017 70 0.893 0.909 0.016 0.889 0.908 0.019 75 0.829 0.858 0.029 0.834 0.860 0.026 80 0.743 0.783 0.040 0.768 0.800 0.033 85 0.625 0.677 0.052 0.662 0.705 0.042 One-year survival after MI Conclusion Estimated one-year survival for MI patients differs by up to 5 percentage points depending on the registry used. Although further research is needed to fully understand these differences, epidemiological findings regarding MI prognosis should be interpreted in light of registry type used and population represented. Acknowledgement/Funding Region Östergötland