Outdoor temperature as a confounder of blood pressure in randomized trials

Temperature is a time varying factor that is associated with blood pressure and has the potential to be a confounder in randomized crossover trials.

We assessed the effect of temperature on blood pressure within the Home Air Filtration for Traffic-Related Air Pollution study, a randomized crossover trial of air purifiers in homes. We assessed the association of blood pressure with both outdoor and indoor temperatures using 190 paired blood pressure measurements that were not affected by the presence of air purifiers in the preceding month.

Mean (standard error) peripheral systolic blood pressure was observed to change by +0.24 (0.09) mmHg per 1 °F colder outdoor temperature and by -0.28 (0.07) mmHg per 1 °F warmer (P < 0.001). Results were similar, although somewhat smaller for peripheral diastolic blood pressure. Associations were nearly identical for central blood pressures. We found much smaller effects for indoor temperatures, none of which approached statistical significance, as indoor temperatures were more stable and did not change as much over the study period.

Despite the cooling and warming of outdoor temperatures being associated with increases and decreases in blood pressure, we believe that outdoor temperature may also be a surrogate marker of particulate matter air pollution. Regardless of the underlying mechanism, randomized trials using blood pressure as outcomes would be well advised to measure temperature and consider it a potential time-varying confounder in the analyses.