Physiological influences on continuous finger and simultaneous intra-arterial blood pressure

Because of the clinical and experimental utility of continuous finger blood pressure measurements and the need for accuracy, we tested the performance of a new hydraulic device in 22 consecutive hypertensive subjects during physiological and pharmacological interventions. Ipsilateral brachial intra-arterial pressure was monitored during rest, Valsalva's maneuver, static handgrip, and mental arithmetic and after sublingual glyceryl trinitrate. In excess of 40,000 blood pressure values were analyzed. Average bias (intra-arterial minus finger blood pressure) was 8.2 +/- 17.0 mm Hg (mean +/- SD, P = NS) for systolic and 2.8 +/- 10.4 mm Hg (P = NS) for diastolic pressure. Two-way ANOVA of biases with subject and task factors showed a subject effect (P < .001). Intersubject and intrasubject standard deviations of bias were 13.8 and 9.8 mm Hg systolic and 8.7 and 5.7 diastolic, respectively. Linear drift (millimeters of mercury per minute) of finger pressure was greater (P < .001) for systolic than diastolic pressure during static exercise and math and after glyceryl trinitrate. Coefficients of determination for blood pressure ranged from 0.4 +/- 0.3 to 0.8 +/- 0.3 during the tasks. We conclude that (1) noninvasive finger blood pressure faithfully follows intra-arterial changes but with clinically relevant offsets, (2) this technique is best suited for assessing pressure changes, (3) physiological and pharmacological interventions do not consistently affect finger pressure accuracy, (4) many reports of finger blood pressure measuring devices are based on direct readings obtained with inadequate system response characteristics, and (5) the tested instrument falls short of the standard requirements (bias < or = 5 +/- 8 mm Hg) for devices that measure intermittently.