Janus faces of the white coat effect: blood pressure not only rises, it may also fall

Abstract

For reliable control of hypertension, it is essential to know the prevalence of the white coat effect (WCE), which is defined as either a difference of 10 mmHg or more in systolic blood pressure (BP) and/or a difference of 5 mmHg or more in diastolic BP measured by a general practitioner (GP), a nurse, or the patient him/ herself. The objective was to assess the WCE in hypertensive patients visiting GPs in primary healthcare centres. A blinded, randomised study in six primary healthcare centres was conducted, where eight GPs, nine nurses and 210 patients (92 men and 118 women) on a hypertension control regimen participated. Heart rate and BP in GPs', nurses' and self-service rooms were measured twice using a sphygmomanometer and an automatic device. Altogether, 148 patients (70%) performed the self-measurements of BP and heart rate perfectly successfully in all the three rooms. These were included in the final analyses. One out of three patients (33%) showed a marked alerting WCE in the GP's room (systolic BP rose by at least 10 mmHg and/or diastolic BP by at least 5 mmHg). On the contrary, one out of 10 (10%) showed a marked relaxing WCE in the GP's room (systolic BP decreased by at least 10 mmHg and/or diastolic BP by at least 5 mmHg). It can be concluded that the WCE in general practice has two faces: an alerting reaction and a relaxing reaction. This should be taken into account in hypertension control.

Doctors record higher blood pressures than nurses: systematic review and meta-analysis

“…32,35 Lack of clear blinding of doctors and nurses to readings by each other was a concern for all but three studies. 29,32,33 Summary risk of bias judgments were low for seven studies, [25][26][27]30,31,33,34 and high for the remainder (Table 1). 28,29,32,[35][36][37][38][39] Differences in blood pressure measurements Differences in blood pressures measured by doctors and nurses could be pooled for 16 cohorts from 14 studies.…”

“…Eleven studies recruited hypertensive subjects from primary health care 25,26 or secondary care clinic settings; [27][28][29][30][31][32][33][34][35] others recruited a combination of normotensive and hypertensive subjects. [36][37][38][39] Studies excluded after assessment are available from the authors on request.…”

“…32,34,35 and one study averaged mercury and automated readings. 25 Methods were not stated for one study, 36 and two adopted different measurement methods for doctors and nurses. 32,35 Lack of clear blinding of doctors and nurses to readings by each other was a concern for all but three studies.…”

“…Overall systolic measurements by nurses were -7.0 (95% CI = -4.7 to -9.2) mmHg lower than those by doctors; diastolic measurements were -3.8 (95% CI = -2.2 to -5.4) mmHg lower. Pooled differences for studies at low risk of bias 25,27,30,31,33,34 were significantly lower than for studies at high risk of bias 28,29,32,[35][36][37][38][39] for systolic readings: -4.6 (95% CI = -1.9 to -7.3) versus -8.4 (95% CI = -6.3 to -10.4; P = 0.03; Figure 2), and for diastolic readings: -1.7 (95% CI = -0.1 to -3.2) versus -5.0 (95% CI = -3.4 to -6.5; P = 0.003; Figure 3). Considerable heterogeneity existed between studies (I 2 statistic >50%) that was not explained by individual outlying studies.…”

“…28,36,37 Pooled systolic blood pressures were -10.1 (95% CI = -8.2 to -12.0) mmHg lower for nurses compared with doctors. Eleven cohorts (nine studies) 25,27,[29][30][31][32][33][34][35] recruited hypertensive cohorts and pooled systolic blood pressure differences were -5.4 (95% CI = -3.2 to -7.6) mmHg lower for nurses compared with doctors, with no difference between the high and low risk of bias subgroups. The difference for the normotensive subgroup was greater than for the hypertensive subgroups; sensitivity analysis according to risk of bias did not alter this finding ( Figure 4).…”