Clustering of Environmental Parameters and the Risk of Acute Myocardial Infarction

Abstract

The association between days with similar environmental parameters and cardiovascular events is unknown. We investigate the association between clusters of environmental parameters and acute myocardial infarction (AMI) risk in Singapore. Using k-means clustering and conditional Poisson models, we grouped calendar days from 2010 to 2015 based on rainfall, temperature, wind speed and the Pollutant Standards Index (PSI) and compared the incidence rate ratios (IRR) of AMI across the clusters using a time-stratified case-crossover design. Three distinct clusters were formed with Cluster 1 having high wind speed, Cluster 2 high rainfall, and Cluster 3 high temperature and PSI. Compared to Cluster 1, Cluster 3 had a higher AMI incidence with IRR 1.04 (95% confidence interval 1.01–1.07), but no significant difference was found between Cluster 1 and Cluster 2. Subgroup analyses showed that increased AMI incidence was significant only among those with age ≥65, male, non-smokers, non-ST elevation AMI (NSTEMI), history of hyperlipidemia and no history of ischemic heart disease, diabetes or hypertension. In conclusion, we found that AMI incidence, especially NSTEMI, is likely to be higher on days with high temperature and PSI. These findings have public health implications for AMI prevention and emergency health services delivery during the seasonal Southeast Asian transboundary haze.