Blood pressure changes after stair climbing in busy and quiet outpatient clinic environments among Tanzanian adults: a randomised crossover study.

Introduction: Blood pressure (BP) measured soon after activity may differ from later seated readings. We examined serial BP changes after stair climbing in busy and quiet outpatient settings, without assuming that a particular waiting period represents complete physiological recovery. Methods: In this randomized crossover study at Mbeya Zonal Referral Hospital, Tanzania, adults with and without a previous hypertension diagnosis drew folded numbered slips to determine environmental order. Staff recording BP were not blinded. Participants ascended and descended 19 steps before measurements immediately and every 3 min for 30 min in each environment. Mixed-effects models included period, sequence, age, sex and body mass index, with session-specific random slopes and autoregressive residual correlation. Classification was compared with each participant's environment-specific average at 24, 27 and 30 min. Exploratory laboratory analyses accounted for reporting limits and used the available observations for each marker. Clinical sensitivity analyses examined alcohol, diabetes history and medication classes. Results: The analysis included 443 participants-223 with hypertension and 220 without a previous diagnosis-contributing 9,746 measurement occasions. At 15 min, adjusted busy-minus-quiet systolic BP differences were 2.87 mmHg (95% CI 1.39-4.35) and 4.07 mmHg (2.58-5.56), respectively. Classification disagreement at that time was 10.8% (7.3%-15.5%) in quiet and 7.6% (4.8%-11.9%) in busy settings among participants with hypertension; corresponding values without a previous diagnosis were 2.3% (1.0%-5.2%) and 5.9% (3.5%-9.8%). First-period-only systolic contrasts remained positive but were larger and less precise. Evidence for differences in trajectory shape between diagnostic groups depended on the covariate-adjustment specification. No routine laboratory marker showed a clear adjusted association with the signed SBP difference from the late-session average at 12 or 15 min. Environmental systolic estimates changed little after clinical adjustment within matching samples. Conclusions: BP changed during seated observation after stair climbing, and early readings disagreed more often with the late-session average. Persistent disagreement and within-session variability do not support 12-15 min as a universally sufficient resting interval. The absence of a pre-exercise resting reference and a no-exercise comparator limits attribution of the decline to recovery from activity.