Relationships between cold-temperature indices and all causes and cardiopulmonary morbidity and mortality in a subtropical island.

This study aimed to identify optimal cold-temperature indices that are associated with the elevated risks of mortality from, and outpatient visits for all causes and cardiopulmonary diseases during the cold seasons (November to April) from 2000 to 2008 in Northern, Central and Southern Taiwan. Eight cold-temperature indices, average, maximum, and minimum temperatures, and the temperature humidity index, wind chill index, apparent temperature, effective temperature (ET), and net effective temperature and their standardized Z scores were applied to distributed lag non-linear models. Index-specific cumulative 26-day (lag 0-25) mortality risk, cumulative 8-day (lag 0-7) outpatient visit risk, and their 95% confidence intervals were estimated at 1 and 2 standardized deviations below the median temperature, comparing with the Z score of the lowest risks for mortality and outpatient visits. The average temperature was adequate to evaluate the mortality risk from all causes and circulatory diseases. Excess all-cause mortality increased for 17-24% when average temperature was at Z=-1, and for 27-41% at Z=-2 among study areas. The cold-temperature indices were inconsistent in estimating risk of outpatient visits. Average temperature and THI were appropriate indices for measuring risk for all-cause outpatient visits. Relative risk of all-cause outpatient visits increased slightly by 2-7% when average temperature was at Z=-1, but no significant risk at Z=-2. Minimum temperature estimated the strongest risk associated with outpatient visits of respiratory diseases. In conclusion, the relationships between cold temperatures and health varied among study areas, types of health event, and the cold-temperature indices applied. Mortality from all causes and circulatory diseases and outpatient visits of respiratory diseases has a strong association with cold temperatures in the subtropical island, Taiwan.