NON-LINEAR RELATIONSHIPS BETWEEN WITHIN-VISIT BLOOD PRESSURE VARIABILITY, MEAN BLOOD PRESSURE, SEX, AND AGE IN UNTREATED MAY MEASURMENT MONTH 2019 PARTICIPANTS

Abstract

Objective: This study aimed to explore change in four commonly used measures of blood pressure variability (BPV) by mean blood pressure (BP), age, and sex in May Measurement Month (MMM) 2019 global survey participants. Design and method: We performed an observational analysis of BP readings from MMM 2019 participants. Adults (18 years and older) were surveyed via convenience sampling from 92 countries in May 2019. Each participant had three seated BP readings. Participants who reported taking antihypertensive medication or who lacked all three readings were excluded. Four indices of BPV were calculated: the standard deviation (SD), coefficient of variation (CV), absolute difference between the second and first BP readings (AR), and variation independent of the mean (VIM). A regression model was created for each index. Each model was adjusted for mean BP, age, sex, and an interaction between age and sex. Age and mean BP were each included as natural cubic splines with five knots. Results: Data from 922,259 participants were included. The mean age was 42.6 (SD 15.9) and 50.8% were female. Systolic SD and AR increased with higher mean BP up to 150 mmHg, while systolic CV and VIM decreased with higher BP across the BP range (Figure 1). Up to 90 mmHg, diastolic CV and VIM decreased with increased mean BP. All diastolic indices increased with higher mean BP above 90 mmHg. Systolic BPV was higher in females between 35 and 70 years (Figure 2), with little difference between sexes within diastolic BPV. Diastolic BPV was lowest between 40 and 50 years of age and highest in the youngest and oldest age groups. Systolic BPV was lowest between 25 and 30 years for females, and at around 45 years for males. Conclusions: All within-visit BPV indices examined showed non-linear variation with mean BP levels and age, particularly for diastolic BPV. Studies should consider non-linear effects when investigating and modelling BPV.