In order to evaluate the influence of cuff size on blood pressure (BP) measurement, we compared values obtained with a standard cuff of 12 x 23 cm with those obtained with a large cuff of 15 x 28 cm in a group of 554 consecutive outpatients. Higher values were obtained with the standard cuff (systolic/diastolic 133.9 +/- 22/80.5 +/- 12 mm Hg) compared to the large cuff (128 +/- 21/76.5 +/- 11.6 mmHg (m +/- SD). This difference reached an average of 10/7 mm Hg for the large arm circumferences (AC) and for high blood pressure values. With the standard cuff, hypertensive values were found twice as often as with the large cuff. Thus the choice of a cuff adapted to AC is of great importance in identifying and treating hypertensive patients. In 77 subjects with a range of AC from 16 to 48 cm, the values measured with the standard and the large cuffs were then compared to those obtained with a reference triple cuff fitting to AC of 8.5 x 26-12 x 36-15 x 46 cm. The measurement of BP was most accurate when the ratio between cuff breadth and AC was between 0.42 and 0.44. The error increased by 0.6-0.7% per centimeter of AC above and below this ratio for the short inflatable bladders (twice as long as broad) of the standard and large cuffs. However, the use of long bladders (3 times as long as broad) combined with the triple cuff appreciably decreased this error.(ABSTRACT TRUNCATED AT 250 WORDS)