Abstract
High ambient heat can directly influence blood pressure (BP) through the vasodilation of the skin vasculature and indirectly by affecting urinary volume and electrolyte levels. We evaluated the direct and urine electrolyte-mediated effects of ambient temperature on BP. We pooled 5,624 person-visit data from a community-based stepped-wedge randomized control trial in southwest coastal Bangladesh from December 2016 to May 2017. Same-day ambient temperature data from local weather stations were linked to participant BP and urine electrolytes using geo-locations of their residential addresses. We implemented causal mediation analyses using the product methods of coefficients with linear mixed models under the sequential ignorability assumption. Separate models were run for each urinary electrolyte mediator (sodium, potassium, calcium, and magnesium), followed by combined models to evaluate the natural direct and electrolyte-mediated indirect effects of temperature on BP. Models had participant-level random intercepts and were adjusted for age, sex, body mass index (BMI), religion, exercise, smoking status, sleep hours, alcohol consumption, urine creatinine, time trend, household assets, drinking water salinity, and seasonality. For the combined mediators (sodium, potassium, calcium, and magnesium), for every 5°C increase in average daily temperature: the direct effect on systolic BP was -1.42 (95 % CI: -1.94, -0.92) mmHg and urine sodium mediated effect was -0.12 (95 % CI: -0.20, -0.05) mmHg; while urine potassium mediated effect was 0.15 (95 % CI: 0.08, 0.25) mmHg; urine calcium-mediated effect 0.06 (95 % CI: 0.01, 0.12) mmHg; and urine magnesium mediated effect -0.00 (95 % CI: -0.03, 0.02) mmHg. We detected similar associations for diastolic BP, pulse pressure, and mean arterial pressure. We found a significant inverse direct effect of ambient temperature on BP compared to minimally mediated urine electrolyte effects. Further studies are needed to uncover the underlying mechanisms of ambient heat and BP associations and to describe the clinical consequences of these associations.