Weights are from random-effects analysis. TRUE-HTA is a study name.
Weights are from random-effects analysis. TRUE-ESP and TRUE-HTA are study names.
Weights are from random-effects analysis.
eTable 1. Search Strategy MEDLINE
eTable 2. Excluded Studies Based on Full-Text Assessment With Reason (n = 28)
eTable 3. Study Characteristics Details of Blood Pressure Measurements for Included Studies (n = 31)
eFigure 1. Flowchart of Study Selection
eFigure 2. Forest Plot of the Mean Difference in Systolic BP Between Awake ABP (Reference) and Routine Office BP Measurement in Samples With Systolic AOBP ≥130 mmHg
eFigure 3. Forest Plot of the Mean Difference in Diastolic BP Between Awake ABP (Reference) and Routine Office BP Measurement in Samples With Systolic AOBP ≥130 mmHg
eFigure 4. Forest Plot of the Mean Difference in Systolic BP Between AOBP (Reference) and Awake ABP in Samples With Systolic AOBP <130 mmHg
eFigure 5. Forest Plot of the Mean Difference in Diastolic BP Between AOBP (Reference) and AABP in Samples With Systolic AOBP ≥130 mmHg
eFigure 6. Forest Plot of the Mean Difference Between AOBP (Reference) and Awake ABP in Samples With Systolic AOBP ≥130 mmHg, by Added Rest Period
eFigure 7. Forest Plot of the Mean Difference Between AOBP (Reference) and Awake ABP in Samples With Systolic AOBP ≥130 mmHg, by Interval Between AOBP Readings
eFigure 8. Forest Plot of the Mean Difference Between AOBP (Reference) and Awake ABP in Samples With Systolic AOBP ≥130 mmHg, by Device
eFigure 9. Forest Plot of the Mean Difference Between AOBP (Reference) and Awake ABP in Samples With Systolic AOBP ≥130 mmHg, by Source of the Sample
eFigure 10. Forest Plot of the Mean Difference Between AOBP (Reference) and Awake ABP in Samples With Systolic AOBP ≥130 mmHg, High Quality Studies
eFigure 11. Funnel Plot of Mean Differences Between AOBP and Awake ABP in Samples With Systolic AOBP ≥130 mmHg, SBP
eFigure 12. Funnel Plot of Mean Differences Between Diastolic AOBP and Awake ABP in Samples With Systolic AOBP ≥130 mmHg
eFigure 13. Pooled Estimates of Mean Differences in Systolic BP Leaving Out One Study at a Time in Samples With Systolic AOBP ≥130 mmHg
eFigure 14. Pooled Estimates of Mean Differences in Diastolic BP Leaving Out One Study at a Time in Samples With Systolic AOBP ≥130 mmHg
Customize your JAMA Network experience by selecting one or more topics from the list below.
Roerecke M, Kaczorowski J, Myers MG. Comparing Automated Office Blood Pressure Readings With Other Methods of Blood Pressure Measurement for Identifying Patients With Possible Hypertension: A Systematic Review and Meta-analysis. JAMA Intern Med. 2019;179(3):351–362. doi:10.1001/jamainternmed.2018.6551
Should automated office blood pressure (recording several blood pressure readings using a fully automated oscillometric sphygmomanometer with the patient resting alone in a quiet place) measurement replace readings recorded by nurses and physicians in routine clinical practice?
This systematic review and meta-analysis of 31 articles comprising 9279 participants compared automated office blood pressure with awake ambulatory blood pressure, a standard for predicting cardiovascular risk. Mean automated office blood pressure readings were similar to the awake ambulatory blood pressure readings and did not exhibit the “white coat effect” associated with routine office blood pressure measurement.
Automated office blood pressure measurement should replace the recording of blood pressure by nurses and physicians in routine clinical practice.
Automated office blood pressure (AOBP) measurement involves recording several blood pressure (BP) readings using a fully automated oscillometric sphygmomanometer with the patient resting alone in a quiet place. Although several studies have shown AOBP measurement to be more accurate than routine office BP measurement and not subject to a “white coat effect,” the cumulative evidence has not yet been systematically reviewed.
To perform a systematic review and meta-analysis to examine the association between AOBP and office BP readings measured in routine clinical practice and in research studies, and ambulatory BP recorded during awake hours, as the latter is a standard for predicting future cardiovascular events.
The MEDLINE, Embase, and Cochrane Library were searched from 2003 to April 25, 2018.
Studies on systolic and diastolic BP measurement by AOBP in comparison with awake ambulatory BP, routine office BP, and research BP measurements were included if they contained 30 patients or more.
Data Extraction and Synthesis
Study characteristics were abstracted independently and random effects meta-analyses and meta-regressions were conducted.
Main Outcomes and Measures
Pooled mean differences (95% CI) of systolic and diastolic BP between types of BP measurement.
Data were compiled from 31 articles comprising 9279 participants (4736 men and 4543 women). In samples with systolic AOBP of 130 mm Hg or more, routine office and research systolic BP readings were substantially higher than AOBP readings, with a pooled mean difference of 14.5 mm Hg (95% CI, 11.8-17.2 mm Hg; n = 9; I2 = 94.3%; P < .001) for routine office systolic BP readings and 7.0 mm Hg (95% CI, 4.9-9.1 mm Hg; n = 9; I2 = 85.7%; P < .001) for research systolic BP readings. Systolic awake ambulatory BP and AOBP readings were similar, with a pooled mean difference of 0.3 mm Hg (95% CI, −1.1 to 1.7 mm Hg; n = 19; I2 = 90%; P < .001).
Conclusions and Relevance
Automated office blood pressure readings, only when recorded properly with the patient sitting alone in a quiet place, are more accurate than office BP readings in routine clinical practice and are similar to awake ambulatory BP readings, with mean AOBP being devoid of any white coat effect. There has been some reluctance among physicians to adopt this technique because of uncertainty about its advantages compared with more traditional methods of recording BP during an office visit. Based on the evidence, AOBP should now be the preferred method for recording BP in routine clinical practice.
Ambulatory blood pressure (ABP) monitoring is now recognized as the best method for predicting the risk of cardiovascular events in relation to an individual’s blood pressure (BP) level. The association between ABP and the risk of cardiovascular events is continuous, consistent, and independent of other risk factors. Although the accurate measurement of BP is the cornerstone for appropriate diagnosis and treatment of hypertension, recent guidelines have questioned the accuracy, and consequently the role, of manual BP measurement in routine clinical practice.1-3 Routine office BP measurement is not only more susceptible to a “white coat effect” (when BP measured in the office is higher than ABP), but is also less accurate, correlating relatively poorly with the awake ABP, and is more likely to be associated with digit preference (rounding off readings to the nearest zero value).4 The advent of oscillometric sphygmomanometers as a replacement for manual BP measurement may improve the quality of office BP readings by recording multiple measurements automatically and by eliminating some aspects of human error.5
During the past decade, oscillometric sphygmomanometers have been used in mostly 2 ways. One approach has been to adapt oscillometric sphygmomanometers designed for self-measurement of BP by patients in the home for use in the office. These oscillometric sphygmomanometers provide 1 or more BP readings, with most other aspects of the BP measurement process remaining the same, including office staff in close proximity to the patient. The net result is that oscillometric BP recordings in clinical practice are not much different from manual BP readings, with both methods being associated with a marked white coat effect.6,7
A second approach followed the introduction of oscillometric recorders such as the Omron 907 (Omron Healthcare) and BpTRU (BpTRU Medical Devices Inc), which were capable of recording multiple BP readings automatically without the need to have a physician or nurse present with the patient. These devices have a built-in delay that allows time for office staff to initiate the readings and then leave the patient alone before the first reading starts. The most important innovation with this technique, subsequently called automated office BP (AOBP) measurement,8 was that conversation between the patient and office staff was no longer possible because the patient was alone, as talking was known to be a major cause of the white coat effect. Also, removing clinic staff likely reduces any anxiety caused by the presence of nurses or physicians. Automated office BP measurement may also be obtained if the readings are recorded with the patient sitting alone in the waiting area of the physician’s office or in a community pharmacy, provided that office or pharmacy staff are not interacting with the patient.9,10 Subsequent research into the use of AOBP measurement has confirmed the initial belief that reducing human involvement in the measurement of BP in the office would improve the quality of the readings.6 Despite encouraging results in clinical studies in different settings, physicians in the United States and Europe have been slow to adopt AOBP measurement into routine office practices, often claiming it was not feasible to do so.11 The Canadian experience suggests otherwise, with 1 recent survey reporting that more than 50% of physicians in primary care may be now using AOBP measurement in their practices.12
Until now, to our knowledge, there has not yet been a comprehensive evaluation of the literature on the comparability of AOBP measurement and conventional office BP measurement. In the present study we have performed a systematic review and meta-analysis of various aspects of the use of AOBP measurement, including a comparison with office BP measurement in routine clinical practice, office BP recorded in research studies, and awake ABP. Systolic BP is the primary focus of this review, because diastolic BP has been less than 80 mm Hg in most studies involving AOBP. Systolic BP is also more affected by the white coat response and is a more important determinant of an individual’s risk of experiencing a cardiovascular event.
Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses,13 we conducted a systematic literature search (eFigure 1 in the Supplement) using MEDLINE, Embase, and the Cochrane Library from 2003 through April 25, 2018, as well as our personal libraries, using key words relating to automated BP measurement (eTable 1 in the Supplement). In addition, we searched reference lists of identified articles and related meta-analyses and reviews. Inclusion criteria were as follows: (1) full-text article with unattended and fully automated AOBP assessment, (2) reported mean difference (MD) and SE (or enough data to calculate these) between AOBP and at least 1 of 3 other BP measurements (awake ABP, research BP, or routine office BP), (3) sample size of 30 or more individuals, (4) maximum time elapsed between 2 types of BP measurement 1 month or less, (5) maximum interval between AOBP measurements 2 minutes or less, (6) at least 3 readings for AOBP (except 2 readings for the Mobil-o-graph device [IEM Gmbh]), and (7) study conducted in or after 2003 (when AOBP was first introduced).
No other restrictions were applied and authors were contacted for missing or additional information, when required. Two of us independently excluded articles based on title and abstract on the first pass. Articles with unsure eligibility were obtained in full text and discussed until consensus was reached. Eligibility of articles retrieved in full text was determined by consensus of all 3 of us.
From all relevant articles, we extracted the first author’s name, year of publication, country, calendar year(s) the study was conducted, setting of the study, age of participants, body mass index (BMI) of participants, sex of participants (percentage of men and women), number treated for hypertension, number of participants, details of BP assessment (added rest period, number of measurements, interval between measurements, attendance by personnel, device, sequence, and randomization), inclusion and exclusion criteria, and subgroup results defined by hypertension status (systolic AOBP, ≥130 mm Hg).
We considered unattended BP measurement recorded with fully automated devices as valid measurements of AOBP if they did not require any involvement of the patient, such as activating the device. All but 1 study14 used either the BpTRU (5 readings at 1- to 2-minute intervals after an initial test reading without antecedent rest), Omron 907 (3 readings, usually at 1-minute intervals with 5 minutes of antecedent rest), or the WatchBP Office (Microlife AG; 1-minute delay then 3 readings at 1-minute intervals).
No additional rest was mandated before the initiation of the AOBP measurements, although several studies,15-18 which otherwise followed AOBP principles, did include an additional rest period. We used daytime or awake ABP as the standard for BP measurement. Automated office BP measurement was available from all 31 eligible studies and awake ABP from 22 studies.
The techniques used to record a separate manual or oscillometric BP in research studies varied somewhat. We defined a research quality office BP as a measurement performed according to standard guidelines, such as those of the American Heart Association.3 A routine office BP measurement was defined as a manual or electronic BP reading taken in usual clinical practice and not as part of a research study. These readings were obtained retrospectively after office staff, who were unaware that the measurements would be used in a research study, had recorded them. In order for the BPs to be considered routine and to avoid observer bias, they could not be measured prospectively as part of a study.
Quality score use in meta-analyses remains controversial.19,20 We restricted our inclusion to studies in which AOBP measurement was performed in the absence of research staff (unattended) and recorded with a fully automated device in at least 30 participants. In the case of duplicate publications involving the same individuals, we used the most comprehensive data available for each analysis.
For studies with a mean systolic AOBP of 130 mm Hg or more, the influence of age, BMI, sex, source of the sample, interval between BP measurements, additional rest periods, proportion of patients being treated for hypertension, and the type of automated device being used for BP measurement were evaluated in meta-regression and subgroup analyses. Furthermore, we rated studies as high quality for a comparison between AOBP and awake ABP based on the following criteria (in addition to our inclusion criteria): (1) no additional rest period, (2) not based on a retrospective review of the medical record, (3) patients were alone during AOBP measurement, and (4) at least 3 AOBP readings were used to calculate a mean. Studies that fulfilled all criteria were classified as high methodological quality.
We used the reported MD with 95% CI to calculate the difference between each BP measurement type. When such data were not available, we calculated the MD and 95% CI based on reported means, 95% CIs were calculated based on reported SEs or SDs, and the correlation coefficient or P value for the MD of paired data.21 All P values were from 2-tailed tests and results were deemed statistically significant at P < .05. Interquartile ranges were used to estimate the SDs by the method of Wan et al.22 In cases in which only the mean and SD for 2 types of BP measurement were available (n = 3 studies for a comparison of AOBP and awake ABP), we used the weighted mean correlation coefficient from studies that provided such data.21
Mean differences were pooled with inverse-variance weighting using DerSimonian-Laird random-effect models to allow for between-study heterogeneity23 Small-study effects were examined using the regression-based test of Egger et al.24 Variation in the effect size because of between-study heterogeneity was quantified using the I2 statistic.25 Applying a random-effects meta-regression26 significance level of P < .10, we conducted analyses for the effect of: (1) mean age (continuous, years), (2) mean BMI (continuous, calculated as weight in kilograms divided by height in meters squared), (3) hypertension status (systolic BP ≥130 mm Hg based on AOBP), (4) rest period (standard procedure vs added rest), (5) interval between AOBP measurements, (6) setting (referral or specialist clinic vs general practice setting or population-based), (7) device (BpTRU, Omron 907, WatchBP Office, or Mobil-o-graph), and (8) percentage of treated patients. In a sensitivity analysis, we used awake systolic ABP (≥130 mm Hg) as the definition for hypertension status. All meta-analyses were conducted with Stata statistical software, version 14.2 (Stata Corp).
Of 2359 initial unique articles, 59 were reviewed in full text (eFigure 1 in the Supplement). Twenty-eight reports were excluded (reasons for exclusion are shown in eTable 2 in the Supplement). In total, we used data from 31 articles (Table 1;10,14-18,27-51 eTable 3 in the Supplement) including 9279 participants, with data from 4736 men and 4543 women. The sample size ranged from 43 to 2145 participants. The weighted mean age of participants was 55.9 years (range, 39-69.5 years). In approximately half of the studies, participants had a mean systolic AOBP of 130 mm Hg or more (n = 4892). Studies were conducted in a range of high-income countries (Table 1). Most studies (n = 18) were conducted in Canada10,15,16,27-41; 3 in the United States42-44; 3 in the Czech Republic17,45,46; 2 in Greece47,48; 1 each in Germany,49 Ireland,50 and Spain51; and 2 studies included multiple countries,14,18 including Australia and Poland. Two reports of the CAMBO (Conventional vs Automated Measurement of Blood Pressure in the Office) trial38,39 were used because the BP assessments were performed 2 years apart. Five studies randomized the order of BP measurement.18,28,29,31,37 All but 3 of the studies that reported awake ABP measurements17,37,47 performed ambulatory blood pressure monitoring (ABPM) on the same day after AOBP measurements.
The pooled MD in systolic BP between routine BP and awake ABP measurements14,27,33,34,36,38 was 13.4 mm Hg (95% CI, 9.5-17.5; I2 = 96%; P < .001) (eFigure 2 in the Supplement). For diastolic BP, the pooled MD was 5.9 mm Hg (eFigure 3 in the Supplement). Based on 5 studies17,35,37,44,47 the pooled MD between awake ABP and AOBP measurements in studies with systolic AOBP less than 130 mm Hg was 5.4 mm Hg (95% CI, 1.7-9.1; I2 = 96%; P < .001) (eFigure 4 in the Supplement), showing that awake ABP was higher than AOBP in this normotensive BP range. The difference in MD between patients with normal BP and hypertension based on AOBP was statistically significant. Nineteen samples from 18 articles10,14-16,18,27,30,32-36,38-41,47,51 reported comparisons between awake ABP, research BP, and routine BP measurements and mean systolic AOBP of 130 mm Hg or more. There was no difference in systolic BP between awake ABP and AOBP (pooled MD, 0.3; 95% CI, −1.1 to 1.7; n = 19; I2 = 90%, P < .001) (Figure 1). Blood pressure measurements obtained under research conditions15,18,28,32,33,35,49,51 were substantially higher compared with systolic AOBP (pooled MD, 7.0; 95% CI, 4.9-9.1; n = 9 samples; I2 = 85.7%; P < .001) (Figure 2). The largest difference between routine BP measurements and systolic AOBP measurements was observed in 9 studies14,27,29,33,34,36,38,42,45 (pooled MD, 14.5; 95% CI, 11.8-17.2; I2 = 94.3%; P < .001) (Figure 3). This difference was comparable with the difference between awake ABP and routine office BP measurements (eFigure 2 in the Supplement). Repeating the analysis with diastolic BP readings showed similar results (eFigure 5 in the Supplement).
We conducted several meta-regression analyses to investigate the statistical heterogeneity observed in the comparison between AOBP and awake ABP measurements in studies with AOBP of 130 mm Hg or more. Age (−0.3 mm Hg per increase in years; n = 19; P = .22), BMI (−0.2 mm Hg per 1-unit increase in BMI; n = 7; P = .93), number of treated patients (0.001–mm Hg increase per 1% increase in proportion of treated patients; n = 16; P = .76), and sex (−0.2 mm Hg per 1% increase in the proportion of males; n = 19; P = .45) did not explain the statistical heterogeneity. Four studies15,16,18,47 used added rest before AOBP measurement, which did not result in a significant difference compared with standard AOBP measurement in comparison with awake ABP measurement for each device (−2.4 mm Hg; n = 19; P = .31) (eFigure 6 in the Supplement). Similarly, the interval between AOBP measurements (1 or 2 minutes) was not significantly associated with the difference between AOBP and awake ABP (−1.9 mm Hg; n = 14; P = .25) (eFigure 7 in the Supplement).
There was consistency among the recording devices, which showed no significant differences between AOBP and awake ABP (BpTRU, 0.8 mm Hg; 95% CI, −1.0 to 2.5 mm Hg; n = 14; WatchBP, −1.3 mm Hg; 95% CI, −4.0 to 1.4 mm Hg; n = 4; Mobil-o-graph, 0.00 mm Hg; 95% CI, −2.2 to 2.2 mm Hg; n = 1) (eFigure 8 in the Supplement). In addition, there was no difference in AOBP and awake ABP for studies that included patients from specialist clinics or those referred for 24-hour ABPM (pooled MD, 0.6 mm Hg; 95% CI, −1.4 to 2.6 mm Hg; n = 14; I2 = 91%; P < .001) and for unselected samples of patients (pooled MD, −0.3 mm Hg; 95% CI, −2.1 to 1.6 mm Hg; n = 5; I2 = 83%; P < .001) (eFigure 9 in the Supplement). Studies that were classified as high quality, which were more clinically homogeneous, showed similar results as our main analysis (pooled MD, 0.2 mm Hg; 95% CI, −1.3 to 1.6 mm Hg; I2 = 85%; P < .001) (eFigure 10 in the Supplement). Table 2 summarizes the results of the main and subgroup analyses.
We found no evidence for small-study effects or publication bias for the main analysis of a comparison between AOBP and awake ABP for individuals with hypertension based on funnel plots (eFigures 11 and 12 in the Supplement) or formal tests (Egger test, P = .93 for systolic BP and P = .67 for diastolic BP). There was little evidence of possible bias in the form of financial support by manufacturers of devices for the studies. Only 2 of 31 studies included in the meta-analysis declared partial support from a manufacturer.18,28 The results from 5 studies that randomized the order of BP measurement were generally in the same range as analyses using all available studies.
Leaving each trial out of the analysis one at a time revealed no meaningful differences in MD (eFigures 13 and 14 in the Supplement). A sensitivity analysis using awake systolic ABP of 130 mm Hg or more as the definition of hypertension showed similar results (pooled MD, 1.1 mm Hg; 95% CI; −0.7 to 2.8 mm Hg; n = 21; I2 = 94%; P < .001) for a comparison between AOBP and awake ABP.
This study is the first comprehensive systematic review and meta-analysis, to our knowledge, to compare AOBP with other techniques of BP measurement in identifying patients with possible hypertension to be confirmed subsequently by ABPM or home BP. The meta-analysis comparing AOBP with awake ABP in patients with hypertension showed that the mean systolic BP reading derived from 19 samples from 18 articles10,14-16,18,27,30,32-36,38-41,47,51 was virtually identical for both techniques (Figure 1). Systolic AOBP differed from awake ABP by more than 5 mm Hg in only 2 studies,41,51 1 of which41 was a retrospective medical record review from an ABPM center.
Otherwise, despite the large statistical heterogeneity, which is expected considering the relatively large sample size in each study, the results exhibited a minimal amount of clinical heterogeneity, with results being confirmed in our analysis of high-methodological quality studies. Although most of the studies did not randomize the sequence of BP measurements, leaving the possibility of an order effect, the studies that did randomize the sequence of BP measurements reported similar results to those that did not. With regard to timing of the measurements, almost all studies measured AOBP and awake ABP on the same day, thus minimizing differences due to timing. Furthermore, no study had an overly large influence, pooled results were consistent across subgroup and meta-regression analyses, and there was no evidence of publication bias. In addition, we examined the association between AOBP and awake ABP in patients with normal BP.17,35,37,44,48 With 1 exception,48 the mean AOBP was lower than the mean awake ABP by approximately 5 mm Hg. These findings are consistent with other comparisons between awake ABP and both manual and oscillometric office BP readings in patient populations with normal BP.52
Among the studies on AOBP included in our analyses, we also identified 6 studies14,27,33,34,36,38 comparing office BP readings recorded in routine clinical practice with awake ABP (eFigure 2 in the Supplement). The MD for routine office BP was significantly higher (13.5 mm Hg; 95% CI, 9.5-17.5 mm Hg) and similar in magnitude to the difference observed in other studies. Previous reviews of office BP recorded in research studies according to guidelines have equated an office BP of 140/90 mm Hg with an awake ABP of 135/85 mm Hg. The results of our meta-analyses clearly show that the routine office BP is substantially higher, suggesting that proper BP measurement techniques in clinical practice were rarely followed. This white coat effect was also present when office BP in primary care was recorded in duplicate using an oscillometric sphygmomanometer. In 27 211 patients with hypertension in the Spanish ABPM Registry, the mean oscillometric office BP was 160/89 mm Hg, compared with a mean awake ABP of 135/78 mm Hg.7 Thus, simply replacing manual BP with an oscillometric device did not eliminate the white coat effect.
Given that the AOBP in patients with hypertension is comparable with the awake ABP, a routine office BP should be substantially higher than the corresponding AOBP reading. Our meta-analysis of 9 studies14,27,29,33,34,36,38,42,45 involving patients with hypertension showed the mean routine office systolic BP to be almost 15 mm Hg higher than the mean AOBP (Figure 3). This difference was similar in magnitude to what was seen in the comparison between the routine office BP and the awake ABP and was consistent. If a white coat effect is defined on the basis of the difference between mean office BP and mean awake ABP, then AOBP appeared to eliminate the overall white coat effect usually associated with office BP.
Because AOBP is comparable to the awake ABP in hypertensive patients, one would also expect the AOBP to be lower than office BP readings recorded in research studies according to guidelines. Results from the meta-analysis of 9 samples from 8 articles15,18,28,32,33,35,49,51 involving patients with hypertension (Figure 2) confirmed this hypothesis, with the mean systolic AOBP being 7 mm Hg lower than the corresponding mean research office BP. Thus, even if a research quality BP were to be performed in routine clinical practice, AOBP would still have the advantage of having the same threshold as awake ABP (135/85 mm Hg) for diagnosing hypertension, whereas the research office BP threshold would still be 140/90 mm Hg. The 3 key aspects of AOBP must always be followed if higher readings are to be avoided: multiple BP readings, recorded with a fully automated device, with the patient resting quietly and alone. Only 1 study, by Bauer et al,49 has reported a mean research quality systolic BP similar to the AOBP. In contrast, the findings in the other 8 samples from 7 articles comparing research BP with AOBP showed the research BP to be consistently higher (Figure 2). It is questionable if the very strict BP measurement conditions in the study of Bauer et al,49 especially no conversation or other interaction with the patient, can be realistically reproduced in routine clinical practice.
If AOBP is to be used in clinical practice, readings must closely adhere to the procedures used in the AOBP studies in this meta-analysis, including multiple BP readings recorded with a fully automated oscillometric sphygmomanometer while the patient rests alone in a quiet place. Using this approach, AOBP has been the preferred technique for office BP measurement in the evidence-based Hypertension Canada guidelines2 since 2016 and is now routinely used by many Canadian primary care physicians.12 The existing evidence supports the use of AOBP to screen patients for possible hypertension in clinical practice, especially if one takes into account the white coat effect associated with current manual or oscillometric techniques for office BP measurement. The use of AOBP in clinical practice has also been recognized elsewhere, such as in a forthcoming statement on blood pressure measurement from the American Heart Association and in the recent guidelines from the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines.3 Automated office BP should now be the preferred method for recording BP in routine clinical practice to identify patients with possible hypertension, with the diagnosis to be confirmed by 24-hour ABPM or home BP.
Accepted for Publication: September 29, 2018.
Corresponding Author: Martin G. Myers, MD, FRCPC, Schulich Heart Program, Division of Cardiology, Sunnybrook Health Sciences Centre, Department of Medicine, University of Toronto, 2075 Bayview Ave, Toronto, ON M4N 3M5, Canada (firstname.lastname@example.org).
Published Online: February 4, 2019. doi:10.1001/jamainternmed.2018.6551
Author Contributions: Drs Roerecke and Myers had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Roerecke.
Administrative, technical, or material support: All authors.
Supervision: All authors.
Conflict of Interest Disclosures: None reported.