Variation in cuff blood pressure in untreated outpatients with mild hypertension--implications for initiating antihypertensive treatment.

Thirty-two patients with mildly elevated blood pressure (BP), but without target organ damage, attended a BP measuring clinic where duplicate BP measurements were made on 12 visits. During visits 1-3, BP showed a systematic decrease which varied from patient to patient. During visits 4-12, no further systematic changes in BP were observed. During the latter period, between-visit variation in BP was substantial, the standard deviation of the difference in BP from one visit to another being 10.4 mmHg for systolic, 6.8 mmHg for diastolic (phase IV) and 7.0 mmHg for diastolic (phase V). These values were used to determine the chance that the BP estimated after a number of visits differed from the average stable BP. After visit 4, the chance of a difference of 5 mmHg or more was 0.50 systolic blood pressure (SBP) and 0.32 diastolic blood pressure (DBP; phase V). Increasing the number of visits to six or more reduced the chance of error. Before initiating lifelong treatment in mild hypertensives free of target organ damage, BP should be recorded in duplicate on a minimum of six visits.