Temporal rise in air temperature may increase heat-related MI risk
Temporal variations in the triggering of myocardial infarction by air temperature in Augsburg, Germany, 1987–2014
Introduction and methods
Short-term exposure to environmental factors, such as air temperature can be a trigger for acute myocardial infarction (MI) [1,2]. Although epidemiological data have shown adverse effects on CVD by increased mortality and morbidity with both high and low temperatures (i.e. heat and cold) [3,4], it remains controversial whether air temperature and MI occurrence are associated.
To foresee the impact of climate change on future health, temporal variations in the link between air temperature and health have become one of the most critical issues [5,6]. This study therefore examined the temporal variations in the association between short-term temperature exposure and occurrence of MI over a 28-years period, using a time-varying distributed lag non-linear model (DLNM).
This time-stratified case-crossover study was based on data from the MONICA/KORA MI registry, a population-based MI registry in Augsburg, Germany from Jan 1987 to Dec 2014. MI events (fatal [n=13.177] and non-fatal [n=14.133]) were further stratified based on admission type (incident and recurrent events) and infarction type (STEMI and NSTEMI for survivors of 24-h hospital stay). Data on daily 24-h average meteorological variables were obtained from monitoring stations in Augsburg. For each participant, the exposure on the day that the person experienced an MI (‘case’ day) was compared with exposures on days at the same day of the week during the same month (‘control’ day). Temperature-MI associations were estimated for two sub-periods (1987-2000 and 2001-2014) separately. The reference temperature (minimum MI temperature [MMIT]) was 18.4˚C, defined as the temperature that gave the minimum risk on total MI events over the whole study period. To identify vulnerable subgroups, information was obtained on sex, age groups (25-64 and 65-74 years), place of residence (city and counties), living alone, history of diabetes mellitus, hyperlipidemia, and pre-existing CVD (angina, coronary heart disease, hypertension, and stroke) for each event. For non-fatal events, additional information was available on education level, smoking status, and obesity.
Time-varying association of air temperature and MI occurrence
- From 1987-2014, significant increasing MI risks were observed at low temperatures (<18.4˚C) for total (RR: 1.26, 95%CI: 1.08-1.47), fatal (RR: 1.33, 95%CI: 1.06-1.65), incident (RR: 1.24, 95%CI: 1.04-1.49), and STEMI (RR: 1.52, 95%CI: 1.11-2.07) events, compared to MMIT whereas MI risks were not significantly different at high temperatures (>18.4˚C).
- Cold effects on MI risk were seen within 5 days in most MI groups, whereas heat effects were immediately observed within 2 days.
- Temporal variations in the association between air temperature and MI were similar for all MI groups when comparing the exposure-response curves of 2 subperiods (1987-2000 and 2001-2014), with the exception for NSTEMI events (P=0.03).
- From 2001-2014, significant heat effects were observed for total, non-fatal, recurrent, and NSTEMI events.
- When comparing risks between 1989-2000 and 2001-2014, significantly higher heat-related MI risks were seen in 2001-2014 for recurrent (RR: 0.82, 95%CI: 0.56-1.19 vs. RR: 1.37 95%CI: 1.04-1.80, P=0.03) and NSTEMI (RR: 0.83 95%CI: 0.56-1.21 vs. RR: 1.31 95%CI: 1.06-1.61, P=0.04) events.
- When looking at separate time periods (1987-2000 and 2001-2014), no significant changes in cold effects were observed.
Time-varying temperature-MI association in specific subgroups
- Significant heat-related MI risks were found in individuals with pre-existing CVD (RR: 1.16, 95%CI: 1.00-1.34) in 2001-2014, as well as in those with diabetes mellitus (RR: 1.33, 95%CI: 1.06-1.67) and hyperlipidemia (RR: 1.23, 95%CI: 1.03-1.46).
- For current smokers, both cold-related and heat-related MI risks increased from 1987-2000 to 2001-2014.
This time-stratified case-crossover study over 28 years showed higher risk estimates of heat-related MI occurrence in 2001–2014 compared to 1987–2000 for recurrent and NSTEMI events, especially in subjects with diabetes mellitus and hyperlipidemia.