Background: Based on studies using mercury sphygmomanometers, arm support is recommended when measuring blood pressure (BP) to keep the cuff at heart level and prevent flexion of arm muscles. Still, BP is commonly measured without arm support, e.g. with the arm resting on the lap or hanging on the side, and the impact when using automated BP devices is unknown. Methods: We conducted a randomized crossover trial of among community dwelling adults in Baltimore, MD. Participants were randomly assigned to sets of triplicate BP measurements with the arm positioned in 3 ways: supported on a desk (desk1 [reference]); hand supported on lap (lap), and arm hanging at side unsupported (side). To account for intrinsic BP variability, all participants underwent a 4th set of BP measurements with the arm supported on a desk (desk2). Between each set of triplicate measurements participants walked for 2 minutes, then rested for 5 minutes prior to BP measurement with a WelchAllyn ProB 2000, using an appropriately sized cuff. We conducted difference-in-difference analysis overall and stratified by baseline systolic BP (SBP≥ vs. < 130mmHg) and age (≥ vs. <60yrs) Results: N=134, mean age 56 (SD 17) years, 78% Black race, and 52% female. 48 participants had SBP ≥130mmHg, 53% were ≥60 years. Overall, mean desk1 and desk2 BP was 126/74 each (Δ= -0.12; Table). Lap and side were greater than desk1 (Table) with the difference in differences significantly greater for both conditions [lap: SBP Δ 3.8 mmHg (95% CI:2.5, 5.2)/ DBP Δ 4.1 (95% CI:3.1, 5.0); side: SBP Δ 6.4 (95% CI:5.0, 7.9)/ DBP Δ 4.6 (95% CI:3.6, 5.5)]. The patterns were generally consistent across subgroups by SBP and age. Conclusion: Commonly used arm positions, i.e. the arm resting on the lap or hanging on the side, resulted in substantial overestimation of BP reading when using an automated device which may lead to patient misdiagnosis and overestimation of the population prevalence of hypertension. Every effort should be made to optimize arm support, even in resource-limited settings.
Read full abstract