Office Blood Pressure Measurement
The Weak Cornerstone of Hypertension Diagnosis
See related article, pp 848–857
For almost a century, office blood pressure (OBP) measurement has been regarded as the cornerstone for diagnosis and management of hypertension because the vast majority of the evidence on the risks associated with elevated BP and the benefits of treatment-induced BP lowering has been based on OBP measurements. However, it is now acknowledged that there are several sources of error with OBP measurement, not least being the recent recognition that the term has been used loosely to denote measurement by several different methods. In addition, OBP is confounded by the white coat and the masked hypertension phenomena, inaccurate devices, and observer-related factors, such as imperfect methodology, observer error and bias, and failure to standardize the circumstances of measurement. Because of these many shortcomings, the 2017 US guidelines now recommend OBP measurement solely as a screening method for the diagnosis of hypertension and out-of-office BP measurement (ambulatory or home) as diagnostic method.1
OBP in the SPRINT
The SPRINT (Systolic Blood Pressure Intervention Trial)2 is a landmark outcome trial which considerably influenced the 2017 US guidelines for hypertension.1 In this issue of the journal, new data examining the impact of differences in OBP measurement methodology on the SPRINT study findings are presented.3 The investigators reported that they have used both unattended and attended OBP measurements using automated devices, and there were no differences between the 2 methods in average follow-up OBP levels and in cardiovascular disease risk reduction in the intensively treated group.3
This article is very important but also problematic.3 It is important because it attempts to clarify the true relevance of the unattended OBP (no observer present during measurements), which has been suggested to question the validity and interpretation of the SPRINT study findings.4 On the other hand, it is problematic because it is largely based on a post hoc survey and clinic staff recall, and different subjects used each OBP measurement method.3 Being aware of these methodological limitations, the authors have worked meticulously to overcome them by performing exhaustive and complex statistical analyses and also by evaluating several end points (OBP levels, drug use, adverse events, and outcome).3 In the long history of OBP research, no previous outcome study has ever come under such intense scrutiny for its measurement methodology.
Types and Fallacies of OBP
The SPRINT study has brought to attention a subject that has been ignored for too long, namely the standardization of methods of BP measurement, in this case OBP. Traditionally, OBP has been loosely described as a method involving 3 essential components: a patient or subject, an observer (usually doctor or nurse), and a device to measure BP. Many factors can affect the interaction between these components and consequently the levels of OBP; these include (1) number of visits (familiarization—first versus repeated visits); (2) number of readings (single, 2, 3, 6, first one discarded); (3) sitting posture (arm position, legs crossed); (4) device type and measurement method (automated versus auscultatory); (5) observer error and bias (with auscultatory measurement); (6) observer presence (attended or unattended); and (7) talking (during rest period and measurements). However, the most fallible factor of all is the human observer using the auscultatory method.
Thus, there are at least 3 general types of OBP.
Clinical trial or research setting OBP, in which the methodology is carefully standardized and follows a protocol. Based initially on auscultatory measurement using mercury sphygmomanometers, such measurements are observer dependent and require careful training. In the past 2 to 3 decades, validated electronic (oscillometric) devices have been used in most outcome hypertension trials,5 which avoid the observer factor and require minimal training. Thus, OBP with electronic devices now is an evidence-based methodology.
General practice OBP, in which auscultatory or automated devices are used under poorly defined conditions of measurement, without reference to rest, position, number of readings, averaging method, etc. Several studies have shown that OBP in these circumstances is higher than research setting OBP by 6 to 14/4 to 10 mm Hg (systolic/diastolic),6 with this difference being variable because of the unstandardized methodology.
Automated OBP, in which the patient is resting alone in the examination room (unattended) and BP is measured using an automated device that gives the average of several measurements.6 Unattended OBP seems to provide lower values than research setting auscultatory OBP by ≈5 mm Hg (systolic and diastolic) and similar to daytime ambulatory BP.6 Unattended OBP has the advantage of removing many confounding factors, but requires additional resources and may not be applicable in all medical settings.
These data suggest that, without standardized methodology, OBP measurements vary considerably and, therefore, cannot provide reliable indication for further evaluation and should not be used for scientific analyses. As mercury devices have been abandoned and automated devices are being increasingly used,5 it is important that a standardized inexpensive and accurate OBP methodology, which is widely applicable in clinical practice, is agreed and universally recommended. A review of hypertension outcome studies of the last 3 decades shows considerable variation in the methodology and devices used for OBP measurement.5 This is an issue of major importance for research because evidence-based guidelines, that ultimately influence decision making in individual patients, rely on the results from clinical trials. It is mandatory, therefore, that a standardized methodology is used in all research studies.
Unattended Versus Attended Automated OBP
Several elements of the unattended OBP method might lead to lower levels than manual auscultatory OBP: (1) the observer error and bias is prevented; (2) multiple measurements are taken; and (3) talking of the patient is prevented (a neglected factor known to considerably increase BP). However, after eliminating all these factors, does the presence of the observer per se (unattended) affect the OBP levels?
Six studies, including the SPRINT, have provided comparative data on unattended versus attended OBP taken using automated devices,3,7–11 which allowed the evaluation of the net effect of the observer’s presence, given that with both methods OBP was taken using the same device and measurement protocol. These studies consistently showed that when several OBP measurements are taken using automated devices without talking to the patient, the presence of the observer has little or no effect on measured OBP (95% CIs in all studies excluded any clinically important difference; Table).
These data suggest that unattended OBP has no clear advantage over attended OBP, provided that the latter is taken with an automated device and there is no talking. The unattended OBP, however, provides a guarantee that these requirements will be followed. It should be noted that these data have been obtained at OBP levels close to the diagnostic thresholds (135–140/80–90 mm Hg; Table). For higher OBP levels, further investigation is required as there might be differences which, however, are less important as both methods will indicate the need to consider treatment.
At the present time, and for some time to come, it is likely that the diagnosis and management of hypertension in most people will be based on OBP measurement. However, even the most standardized OBP can be misleading in a considerable proportion of subjects, and out-of-office BP measurements should be used before making a diagnosis of hypertension and to guide treatment.1 In case of disagreement between OBP and out-of-office BP measurements (ambulatory or home), decisions should always be based on the latter.
A feasible methodology for OBP for both clinical research and clinical practice would be to take triplicate measurement using a validated automated device, with the subject resting and not talking. This methodology is devoid of most sources of error, has relatively low cost and requires little maintenance, and therefore can easily replace existing OBP methodologies and devices in most settings.
G. Stergiou, G. Parati, and E. O’Brien conducted validation studies for various manufacturers and advised manufacturers on device development. The other author reports no conflicts.
The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.
- © 2018 American Heart Association, Inc.
- Whelton PK,
- Carey RM,
- Aronow WS,
- et al
- Johnson KC,
- Whelton PK,
- Cushman WC,
- et al
- Kjeldsen SE,
- Mancia G
- Giorgini P,
- Weder AB,
- Jackson EA,
- Brook RD
- Myers MG
- Greiver M,
- White D,
- Kaplan DM,
- Katz K,
- Moineddin R,
- Dolabchian E
- Rinfret F,
- Cloutier L,
- Wistaff R,
- Birnbaum LM,
- Ng Cheong N,
- Laskine M,
- Roederer G,
- Van Nguyen P,
- Bertrand M,
- Rabasa-Lhoret R,
- Dufour R,
- Lamarre-Cliche M
- Bauer F,
- Seibert FS,
- Rohn B,
- Bauer KAR,
- Rolshoven E,
- Babel N,
- Westhoff TH