Assessment of the white-coat effect

Background: A limitation of blood pressure measurements made in the physician's office is the transient elevation in pressure seen in many patients that does not appear to be linked to target organ damage or prognosis. This has been labeled the 'white-coat effect' (WCE), computed as the difference between blood pressure measurements taken by the physician and the ambulatory level or resting measures. It is unclear, however, which resting measure is most appropriate. The awake ambulatory blood pressure is the most widely used. However, while arguably the most useful measure for prediction of clinical outcomes, it is less appropriate for use as a resting measure, because it is influenced by many factors, including posture and physical activity level. Resting levels taken in the clinic may also be elevated, and will therefore underestimate the WCE. Methods: We addressed this question by taking resting measures in a non-medical setting on the day before patients were seen at a Hypertension Clinic (day 1), and comparing these with resting measures taken on the following day, in the clinic before the patient saw the physician. Results: As predicted, the day 1 resting levels were lower than those taken in the clinic prior to seeing the physician (P < 0.05 and P < 0.001 for systolic and diastolic pressures, respectively) in both normotensive and hypertensive patients. Using the day 1 resting levels, the estimated WCE for hypertensive patients was 5.3/6.9 mmHg (systolic/ diastolic blood pressures), compared with estimates, using the clinic resting levels, of 0.3/0.5 mm Hg. The pattern of changes was different in normotensive patients and hypertensive patients, with the physician pressures being slightly lower than day 1 pressures in the former, and substantially higher in the latter. Heart rate changes were similar and modest in both groups. Conclusion: The WCE may not just be limited to that narrow interval in which the patient actually sees the physician, but may generalize to the clinic setting, rendering a clinic 'resting' level invalid. While it is strongly positive in most hypertensive patients, it is frequently negative in normotensive patients. Our results suggest that improved methods of measuring blood pressure in the clinic setting are unlikely to resolve the confounding influence of the WCE, and that greater reliance will need to be placed on out-of-office monitoring.