Letting the arm hang unsupported raises systolic blood pressure readings by about 6.5 mmHg

There are exactly six ways to order three arm positions ^[1][1]. The six randomized groups therefore covered every possible sequence ^[6]. Each participant was also their own comparison, so differences in age, body size or baseline pressure between people drop out of the estimate ^[6]. Rotating the order deals with what is left. If one position were always measured first, any change in pressure over the session would be credited to the arm.

The trial held most of the other American Heart Association steps fixed. The association's guidance covers cuff size, back support, feet flat with legs uncrossed, and arm position, with the middle of the cuff at mid-heart level and the arm resting on a desk or table ^[9]. In the trial, each cuff was sized to the participant's upper arm, and backs and feet were supported ^[7][8]. Participants first emptied their bladders and walked for two minutes, to mimic arriving at a clinic, then sat for five minutes ^[7]. A digital device took three sets of three readings, 30 seconds apart ^[8]. That is nine readings per person ^[2].

Tammy Brady, the senior author and medical director of the pediatric hypertension program at Johns Hopkins Children's Center, said arm position makes a "huge difference" in getting an accurate measurement ^[11]. According to the team, the JAMA Internal Medicine report reinforces guidance that the arm be firmly supported on a desk or other surface ^[1][12].

Against the association's rough normal of 120/80, 6.5 mmHg is about 5% of a systolic reading ^[3]. The association also counts nearly half of U.S. adults as having elevated blood pressure ^[13]. The researchers' warning about wrong diagnoses rests on the average shift ^[4]. An average of 6.5 mmHg could come from a modest rise in everyone or a large rise in a few, and those two patterns would move different numbers of people across a cutoff. The release does not report how many participants crossed a diagnostic threshold when the arm left the desk.

I think the evidence supports making a supporting surface at heart height the default. The researchers describe two shortcuts in clinics, the lap and the unsupported arm, and both pushed readings in the same direction ^[2][3][10]. The remedy is already written into the guidance ^[9]. The condition on that view is that each person was measured on a single visit under research conditions ^[6][7].