Temporal trends in acute coronary syndrome hospitalisation rates: a comparison of administrative diagnosis coding and cardiac biomarker testing

Abstract Funding Acknowledgements Type of funding sources: Foundation. Main funding source(s): Victor Chang Cardiac Research Institute. Background/aims Accurate monitoring of trends in acute coronary syndromes (ACS) provides important information on the impacts of primary and secondary prevention. Use of cardiac troponin (cTn) assays with increasing sensitivity and specificity during the past 20 years have likely increased the rates of diagnosis of myocardial infarction (MI), particularly NSTEMI. We aimed to determine the correlation of ICD-coded ACS hospitalisation rates with biomarker-classified ACS rates to inform population-level monitoring of ACS trends in Australia. Methods We used linked hospitalisation, emergency department, mortality, and pathology data from all tertiary hospitals in Western Australia (WA). We identified all patients aged >20years with a principal diagnosis of STEMI or NSTEMI (ICD-10-AM I21.0-I21.4) or unstable angina (UA) (ICD-10-AM I20.0) from 2001-2017. Creatine kinase (CK) and cTn test results including high-sensitivity (hs-Tn) assays were classified as elevated based on the ESC/AHA Universal Definition using the upper reference limit value for each assay. We calculated age-standardised rates (ASRs) for ICD-10 coded STEMI, NSTEMI and UA hospitalisations, and those which were cTn positive (cTn+), or any biomarker positive including CK (Bm+). We assessed annual ASR ratios over time (reference: ICD-coded rates) and estimated annual % changes in admissions using Poisson regression models adjusting for 5-year age-group and sex. Results The cohort comprised 143,231 ICD-coded ACS hospitalisations for 91,701 patients. 60.5% of the patients were male and 95.2% of the ACS hospitalisations had linked cardiac biomarker records. Despite slightly lower ASRs, Bm+ based rates were consistent with patterns in ICD-coded admissions in both MI sub-types (Figure) with no significant differences stratified by sex. Average annual decline in STEMI admissions was -3.7% (95% CI -4.1%,-3.4%) and -3.2% (95% CI -3.6%, -2.9%) in ICD-coded and Bm+ hospitalisations respectively. The annual rise in ICD-coded NSTEMI ASRs was 4.1% (95% CI 3.9%, 4.4%) which correlated with the 4.4% (95% CI 4.1%, 4.7%) annual rise in Bm+ ACS hospitalisations. The age-standardised rate ratios of ICD-coded to Bm+ were 1.27 and 1.28 for STEMI and NSTEMI respectively in 2001 and remained consistent at 1.29 and 1.30 for STEMI and NSTEMI respectively in 2017, although this was not the case for cTn+:ICD-coded rate ratios. There was an inverse trend between ASRs of ICD-coded NSTEMI and UA over the study period which correlated with use of increasingly sensitive cTn assays. The proportion of UA admissions with Bm+ was consistently low (21%) over the study period. Conclusions Despite shifts from CK to cTn and hs-Tn assays during our study period, the present study highlights the validity of using linked hospitalisation data to monitor long term trends in ACS based on cardiac biomarker positivity. Figure legend. Age-standardised hospitalisation rates by classification status in ACS subtypes.