Short‐Term Effects of Personal‐Level Environmental Temperature on Ambulatory Blood Pressure in Patients With Hypertension: A Multicity Panel Study

Figures

• 

  Figure 1. Association of PET with ambulatory BP and BP variability among patients with hypertension. Changes and 95% CIs of BP and BP variability against per 1 °C decrease of PET in the exposure windows of lag 0 day, lag 1 day, lag 2 day and 2‐day moving average were estimated using a generalized linear mixed model adjusted age, sex, current regular drinking, physical activity level, body mass index, antihypertensive medication, diabetes, dyslipidemia, fine particulate matter, relative humidity, and random intercepts composed of subject level nested within city level. Models were further adjusted for mean BP when BP variability was treated as the outcome, and BP variability was calculated as the SD of BP. BP indicates blood pressure; DBP, diastolic blood pressure; lag0d, lag 0 day; lag1d lag 1 day; lag2d, lag 2 day; MA2d, 2‐day moving average; PET, personal‐level environmental temperature; and SBP, systolic blood pressure.

• 

  Figure 2. Exposure–response curve of PET and 24‐hour, daytime, and nighttime ambulatory BP in the exposure window of lag 0 day. Generalized additive mixed model was used, with adjustment for age, sex, current regular drinking, physical activity level, body mass index, antihypertensive medication, diabetes, dyslipidemia, fine particulate matter, relative humidity, and random intercepts composed of subject level nested within city level. PET was processed by cubic spline functions, with degrees of freedom selected according to the Akaike information criterion. BP indicates blood pressure; DBP, diastolic blood pressure; df, degrees of freedom; lag0d, lag 0 day; PET, personal‐level environmental temperature; and SBP, systolic blood pressure.

• 

  Figure 3. Association of PET with ambulatory BP and BP variability among patients with hypertension during the daytime and nighttime. Changes and 95% CIs of BP and BP variability against per 1 °C decrease of PET in the exposure windows of lag nighttime, concurrent daytime, and concurrent nighttime were estimated using generalized linear mixed model adjusted for age, sex, current regular drinking, physical activity level, body mass index, antihypertensive medication, diabetes, dyslipidemia, fine particulate matter, relative humidity, and random intercepts comprised of subject level nested within city level. Models were further adjusted for mean BP when BP variability was treated as the outcome, and BP variability was calculated as the SD of BP. BP indicates blood pressure; DBP, diastolic blood pressure; PET, personal‐level environmental temperature; and SBP, systolic blood pressure.

• 

  Figure 4. Association of PET with ambulatory BP and BP variability among patients with hypertension stratified by BP control status. Changes and 95% CIs of BP and BP variability against per 1 °C decrease of PET in the exposure windows of lag 0 day, lag 1 day, lag 2 day, and 2‐day moving average were estimated using a generalized linear mixed model adjusted for age, sex, current regular drinking, physical activity level, body mass index, antihypertensive medication, diabetes, dyslipidemia, fine particulate matter, relative humidity, and random intercepts composed of subject level nested within city level. Models were further adjusted for mean BP when BP variability was treated as the outcome, and BP variability were calculated as the SD of BP. BP indicates blood pressure; DBP, diastolic blood pressure; lag0d, lag 0 day; lag1d lag 1 day; lag2d, lag 2 day; MA2d, 2‐day moving average; PET, personal‐level environmental temperature; and SBP, systolic blood pressure. *Denotes statistical significance between 2 groups.