Effect modification of ozone-related mortality risks by temperature in 97 US cities

Many time-series studies have characterized the relationship between short-term ozone exposure and adverse health outcomes, controlling for temperature as a confounder. Temperature may also modify ozone effects, though this has been largely under-investigated. In this study, we explored whether temperature modifies the effect of short-term ozone exposure on mortality. We used the database developed for the National Morbidity and Mortality Air Pollution Study to estimate ozone mortality risks in 97 US cities in May through September, 1987-2000. We treated temperature as a confounder as well as an effect modifier by estimating risks at low, moderate, and high temperature categories. When temperature was treated as a confounder, a 10-ppb increase in daily 24-h ozone was associated with a 0.47% (95% CI: 0.19%-0.76%) increase in mortality. When we assessed effect modification by temperature, the interaction between ozone and temperature was not statistically significant. However, there was a U-shaped pattern in mortality risk, which was greater at the low (75th percentile) temperature levels than moderate temperature levels. At the high temperature category, a 10% increase in AC prevalence mitigated mortality risk associated with 10-ppb of ozone exposure by -0.18% (95% CI: -0.35%, -0.02%). Furthermore, ozone mortality risk in the high temperature category increased as we restricted our analyses to hotter days. On days where temperatures exceeded the 75th, 90th, and 95th percentile temperatures, a 10-ppb increase in ozone was associated with a 0.65% (95% CI: 0.20%-1.09%), 0.83% (95% CI: 0.17%-1.48%), and 1.35% (95% CI: 0.44%-2.27%) increase in mortality, respectively. These results suggested that high temperatures may exacerbate physiological responses to short-term ozone exposure.