Abstract

Background. Automated blood pressure (BP) devices are commonly used in doctor’s offices. How BP measured on these devices relates to ambulatory BP monitoring is not clear.

Objective. To assess how well office-based manual and automated BP predicts ambulatory BP.

Methods. Using data on 654 patients, we assessed how well sphygmomanometer measurements and measurements taken with an automated device (BpTRU) predicted results on ambulatory BP monitoring. We assess positive and negative predictive values and overall accuracy. We look at different cut-points for systolic (130, 135 and 140 mmHg) and diastolic (80, 85 and 90 mmHg) BP.

Results. A single automated office BP (AOBP) assessment provides superior predictive values and overall accuracy compared to three manual office BP assessments. For systolic BP, the predictive values are ≤69% for any of the cut-points while the positive predictive values for the single automated measurement is between 80.0% and 86.9% and the overall accuracy gets as high as 74% for the 130 mmHg cut-point. For diastolic BP, the automated readings are also more predictive but in this case, it is the negative predictive values that are better, as well as the overall accuracy.

Conclusions. Based on the results, we suggest that 135/85 mmHg continue to be used as the cut-point defining high BP with the BpTRU device. However, future research might suggests that values in a grey zone between 130–139 mmHg systolic and 80–89 mmHg diastolic be confirmed using ambulatory BP monitoring. As well, three AOBP assessments might produce much greater accuracy than the single AOBP assessment used in the study.

Introduction

The accuracy and consistency of measurement of biological functions, such as blood pressure (BP), are affected by the internal and external environment. BP may vary from day to day, even moment to moment, due to activity level, stress, illness and ingestion of food, alcohol, tobacco or other drugs, to mention a few. The mere presence of the person doing the measurement can affect the results. The care taken by the person conducting the measurement, ensuring that they are doing it in a manner that will provide accurate and consistent results, is also important; in a busy practice, this level of care may not always be applied. It is little wonder that the variance in manual office BP (MOBP) is large and its relationship to outcomes is less than other types of measurement, such as home BP (HBP) monitoring and ambulatory BP monitoring (ABPM).1–4

Because of the inherent difficulties with accurate BP measurement in the physician’s office, there has been a trend lately to suggest that it be replaced by HBP and ambulatory BP (ABP). However, the relatively recent development of automated office BP (AOBP) measurement devices may make it a little premature to forsake the use of office BP measurements. In Canada, there are three devices designed for AOBP measurement: the BpTRU device (BpTRU Medical Devices, Coquitlam, British Columbia, Canada), the OMRON HEM 907 and the MicroLife WatchBP Office. All have had validation studies completed confirming their accuracy.5–7 Of these, the BpTRU device has received a great deal of attention and is being studied in several sites across the country in relation to its use and utility in clinical practice. The exact prevalence of the use of AOBP measurement devices in the management of high BP is unknown. It is known that ∼10 000 units of the BpTRU device have been sold in Canada (Martin Myers, personal communication) but these are not all being used by family physicians in their offices. Presumably many would be used by specialists, hospitals, emergency rooms and for research.

Studies completed to-date suggest that, for the BpTRU device, the best estimate for the cut-point between normal and high BP is 135/85 mmHg, which is convenient because it is the same as the cut-point for HBP monitoring and awake ABPM. Beckett and Godwin8 in their 2005 study compared BpTRU with 24 hour ABPM and suggested that this cut-point be used as the boundary between normal and high BP, <135 mmHg systolic and <85 mmHg diastolic being considered normal.

This study is an attempt to further clarify the relationship between one of these AOBP measurement devices, the BpTRU, and what has become the gold standard, ABPM. It includes the data used in the Beckett and Godwin article and adds a further 173 patients, a different analysis, and discussion about how this device should probably be used in practice.

Methods

Participants and procedures

The data used in this study were collected during the course of enrolment of patients (baseline data) into two randomized controlled trials being conducted at the Centre for Studies in Primary Care at Queen’s University, Kingston, Ontario, Canada—the Home Monitoring of Blood Pressure Study (ISRCTN25105161) and the Intensive Scheduled Management of Hypertension Study (ISRCTN05874865), which were funded by the Heart and Stroke Foundation of Ontario. Baseline data from both intervention and control group subjects in both studies were used in this analysis. Both studies had the same eligibility criteria. Subjects were recruited from 55 family practices in eastern Ontario. All participants had a diagnosis of hypertension and were being treated with antihypertensive medications. Each patient’s chart was reviewed to abstract the BPs recorded (MOBP readings) at the last three office visits at which BP was measured. Only one recording was used from any single office visit. If there was more than one recording at a given visit, the last measurement recorded for that visit was used. These visits ranged from several weeks to several months apart, depending on the practice of the physician regarding follow-up of hypertensive patients. If the mean of these three readings, taken at different visits, was ≥140/90, the patient was labelled as ‘uncontrolled’ by office measurement, meaning that the patient had not achieved the treatment target (≥130/80 was used for diabetics) and then the patient was eligible for the study. Subjects were excluded if they were <18 years of age, pregnant or had a known secondary cause for their hypertension.

At the enrollment visit, each patient had their BP measured using the BpTRU device and then had an ABPM applied. The patient returned the next day to have the ambulatory monitor removed.

Devices

BpTRU.

The BpTRU device uses the oscillometric technique to determine BP. This is the same technique used by most ambulatory and HBP measuring devices. The BpTRU device meets all requirements of the Association of Advancement of Medical Instrumentation and achieved a Grade ‘A’ in the British Hypertension Society protocol.9 It is designed to take an initial reading while the clinician is present and then with the patient alone in the room, proceeds to take five or more measurements at intervals of 1–5 minutes after which it calculates and displays the average of these five readings. The specific steps we used were (i) the subjects were seated for at least 5 minutes, (ii) the BpTRU cuff was applied to the non-dominant arm by the research nurse, (iii) the initial BpTRU BP reading was taken and recorded, (iv) the nurse then left the room leaving the patient alone while the BpTRU device took five more readings at intervals of either 1 minute or 2 minutes and (v) these five readings were averaged and displayed by the device. The mean of the five measurements taken with the patient in the room alone is considered the patient’s BP as assessed by the BpTRU device. The BpTRU devices were calibrated at the beginning of the study. There is also a self-calibration process built into the device.

ABPM device.

The A&D TM-2430 device was used in these studies. It has been tested and met the criteria of the British Hypertension Society for BP measurement devices.10 The devices were placed on the participants by a trained research assistant on the morning of 1 day and removed 24 hours later.

Manual sphygmomanometers.

We did not assess the type, quality or calibration of the devices used in family physicians offices. We used the measurements they recorded in their charts from the devices they used on a daily basis. Since these MOBP measurements were taken before the study started and collected retrospectively from the charts, there was no standardization of the technique of data collection. This is what we wanted; of course, point-of-service BP measurements taken the way they are taken during regular practice.

Variables

The criterion variable was the mean of the BP measurements taken by the ABPM device while the patient was awake. The predictor variables were the mean of the three MOBP readings (taken on three different occasions weeks or months apart), and the mean of the five AOBP readings (BpTRU) taken at one setting with the patient alone in the room.

Analysis

The data were analysed using SPSS version 16. Descriptive statistics were used to describe the population. Sensitivity, specificity and predictive values were calculated for MOBP and AOBP (BpTRU) in relation to awake means on ABPM. Pearson correlations were used to compare mean scores of MOBP, AOBP and ABPM. Bland–Altman plots were used to visualize the ability of AOBP and MOBP to predict ABPM. The goal of the analysis was to clarify how well office sphygmomanometer measurements (MOBP readings) and BpTRU measurements (AOBP) predicted results on awake ABPM.

Normal for awake ABPM was defined as <135 mmHg for systolic BP and <85 mmHg for the diastolic BP. For the average of the last three MOBP measurements and for the BpTRU, we tested three different cut-points: 140, 135 and 130 mmHg for systolic BP and 90, 85 and 80 mmHg for diastolic BP to determine positive and negative predictive values for elevated BP on awake ABPM. Finally, we analysed the data in relation to a recent proposal algorithm regarding the use of AOBP in clinical practice.10 The algorithm contends that measurements on AOBP >140/90 mmHg can be considered elevated with a high degree of certainly without further assessment by ABPM, that measurements <130/80 mmHg can be considered to be normal with relative certainty and that BPs between 130–139 and 80–89 mmHg need to be assessed using ABPM to confirm whether or not hypertension is present. We assessed how well our data supported these assumptions.

Results

Table 1 describes the population of 654 primary care patients included in the analysis. These patients all had a diagnosis of hypertension, were being treated for hypertension with medications, but were not at target BP.

TABLE 1

Study population characteristics

Female (%) 370 (56.6) 
Age in years, mean (SD) 63 (18) 
Diagnosis of hypertension on chart (%) 649 (99.2) 
On antihypertension medications (%)a 652 (99.7) 
Diagnosis of diabetes (%) 117 (18) 
Years with hypertension 
    <1 (%) 53 (8.1) 
    1–5 (%) 249 (38.1) 
    6–10 (%) 122 (18.7) 
    >10 (%) 230 (35.1) 
Body mass index, mean (SD) 30.7 (5.2) 
Average of last three manual BP measurements on the chart in mmHg, mean (SD) 
    Systolic 148.5 (10.9) 
    Diastolic 82.5 (8.4) 
BP using BpTRU device in mmHg, mean (SD) 
    Systolic 139.2 (17.9) 
    Diastolic 79.8 (10.9) 
Average of awake BP readings on ABPM in mmHg, mean (SD) 
    Systolic 140.9 (13.3) 
    Diastolic 79.6 (7.8) 
Female (%) 370 (56.6) 
Age in years, mean (SD) 63 (18) 
Diagnosis of hypertension on chart (%) 649 (99.2) 
On antihypertension medications (%)a 652 (99.7) 
Diagnosis of diabetes (%) 117 (18) 
Years with hypertension 
    <1 (%) 53 (8.1) 
    1–5 (%) 249 (38.1) 
    6–10 (%) 122 (18.7) 
    >10 (%) 230 (35.1) 
Body mass index, mean (SD) 30.7 (5.2) 
Average of last three manual BP measurements on the chart in mmHg, mean (SD) 
    Systolic 148.5 (10.9) 
    Diastolic 82.5 (8.4) 
BP using BpTRU device in mmHg, mean (SD) 
    Systolic 139.2 (17.9) 
    Diastolic 79.8 (10.9) 
Average of awake BP readings on ABPM in mmHg, mean (SD) 
    Systolic 140.9 (13.3) 
    Diastolic 79.6 (7.8) 
a

This is not 100% despite it being an eligibility criterion because after enrollment, two people informed us that they were not taking medication.

Tables 2 and 3 show the results of our analysis. Tables 2 displays the positive predictive values, negative predictive values and overall accuracy for the three different systolic cut-points for both the MOBP and the AOBP. Positive predictive values measure the degree to which a test or device accurately predicts the true value on the criterion or gold standard measurement device when the value on the test or device is positive. Negative predictive values measure the degree to which a test or device accurately predicts the true value on the criterion or gold standard measurement device when the value on the test or device is negative. Table 3 displays the same results for diastolic pressure.

TABLE 2

Predictive values and overall accuracy of three systolic MOBP assessments and a single systolic automated (BpTRU) office assessment

  Positive predictive value, N (%)a Negative predictive value, N (%)b Overall accuracy, N (%)c 
MOBP Cut-point 140 mmHg 373 (69.2) 52 (46.4) 425 (65.2) 
Cut-point 135 mmHg 422 (67.8) 18 (56.3) 440 (67.3) 
Cut-point 130 mmHg 429 (67.8) 7 (50.0) 436 (66.6) 
AOBP Cut-point 140 mmHg 246 (86.9) 181 (48.8) 427 (65.3) 
Cut-point 135 mmHg 302 (81.8) 151 (53.0) 453 (69.3) 
Cut-point 130 mmHg 357 (80.0) 128 (58.7) 485 (74.2) 
  Positive predictive value, N (%)a Negative predictive value, N (%)b Overall accuracy, N (%)c 
MOBP Cut-point 140 mmHg 373 (69.2) 52 (46.4) 425 (65.2) 
Cut-point 135 mmHg 422 (67.8) 18 (56.3) 440 (67.3) 
Cut-point 130 mmHg 429 (67.8) 7 (50.0) 436 (66.6) 
AOBP Cut-point 140 mmHg 246 (86.9) 181 (48.8) 427 (65.3) 
Cut-point 135 mmHg 302 (81.8) 151 (53.0) 453 (69.3) 
Cut-point 130 mmHg 357 (80.0) 128 (58.7) 485 (74.2) 
a

Proportion of the time a measurement at cut-point or above predicts a mean ABP result of ≥135 mmHg.

b

Proportion of the time a measurement below cut-point predicts a mean ABP result of <135 mmHg.

c

Combination of positive and negative predictive values.

TABLE 3

Predictive values and overall accuracy of three diastolic MOBP assessments and a single diastolic automated (BpTRU) office assessment

  Positive predictive value, N (%)a Negative predictive value, N (%)b Overall accuracy, N (%)c 
MOBP Cut-point 90 mmHg 53 (39.8) 413 (79.7) 466 (71.3) 
Cut-point 85 mmHg 91 (33.1) 311 (82.1) 402 (61.5) 
Cut-point 80 mmHg 129 (29.8) 191 (86.4) 320 (48.9) 
AOBP Cut-point 90 mmHg 70 (58.3) 445 (83.3) 515 (78.7) 
Cut-point 85 mmHg 104 (51.0) 395 (87.8) 499 (76.3) 
Cut-point 80 mmHg 135 (40.7) 298 (92.5) 433 (66.2) 
  Positive predictive value, N (%)a Negative predictive value, N (%)b Overall accuracy, N (%)c 
MOBP Cut-point 90 mmHg 53 (39.8) 413 (79.7) 466 (71.3) 
Cut-point 85 mmHg 91 (33.1) 311 (82.1) 402 (61.5) 
Cut-point 80 mmHg 129 (29.8) 191 (86.4) 320 (48.9) 
AOBP Cut-point 90 mmHg 70 (58.3) 445 (83.3) 515 (78.7) 
Cut-point 85 mmHg 104 (51.0) 395 (87.8) 499 (76.3) 
Cut-point 80 mmHg 135 (40.7) 298 (92.5) 433 (66.2) 
a

Proportion of the time a measurement at cut-point or above predicts a mean ABP result of ≥135 mmHg.

b

Proportion of the time a measurement below cut-point predicts a mean ABP result of <135 mmHg.

c

Combination of positive and negative predictive values.

It should be remembered that we are comparing a single assessment by AOBP (using the BpTRU device) with the average of three different BP assessments using MOBP. A quick scan of the tables reveals that the single AOBP assessment is superior in predictive values and overall accuracy to the three MOBP assessments. Looking at systolic BP, the predictive values and overall accuracy of the manual readings never gets >69% for any of the cut-points while the positive predictive values for the single automated measurement is between 80.0% and 86.9% depending on the cut-point used and the overall accuracy gets as high as 74% for the 130 mmHg cut-point. For diastolic BP, the automated readings are also more predictive but in this case, it is the negative predictive values that are better, as well as the overall accuracy.

We also calculated Pearson correlation coefficients and did Bland–Altman graphs. It is apparent from these that the single AOBP correlates better than the three MOBP with the awake BP on ABPM. The correlation coefficients are in Table 4; they are similar to those reported in the previously quoted Beckett and Godwin paper.8

TABLE 4

Means and Pearson correlations of awake ABPM versus manual and automated BP measurements

 Awake ABPM, Mean (95% CI) AOBP, Mean (95% CI) MOBP, Mean (95% CI) ABPM versus AOBP, Pearson's r ABPM versus MOBP, Pearson's r 
Systolic BP 140.9 (139.8–141.9) 139.1 (137.8–140.5) 148.5 (147.7–149.4) 0.591 0.173 
Diastolic BP 79.5 (78.9–80.1) 79.8 (78.9–80.7) 82.5 (81.8–83.1) 0.587 0.306 
 Awake ABPM, Mean (95% CI) AOBP, Mean (95% CI) MOBP, Mean (95% CI) ABPM versus AOBP, Pearson's r ABPM versus MOBP, Pearson's r 
Systolic BP 140.9 (139.8–141.9) 139.1 (137.8–140.5) 148.5 (147.7–149.4) 0.591 0.173 
Diastolic BP 79.5 (78.9–80.1) 79.8 (78.9–80.7) 82.5 (81.8–83.1) 0.587 0.306 

The Bland–Altman plots are also useful in comparing AOBP and MOBP and how they relate to awake measurements on ABP. Figures 1 and 2 plot systolic MOBP versus ABP and AOBP versus ABP, respectively. While far from showing ideal agreement, the AOBP versus ABP shows better agreement than MOBP versus ABP. The Bland–Altman plots for diastolic BP (Figs 3 and 4) are both better than those for systolic BP; however, again the AOBP versus ABP shows superior agreement.

FIGURE 1

Bland-Altman plot comparing manual office systolic BP with awaks ambulatory systolic BP

FIGURE 1

Bland-Altman plot comparing manual office systolic BP with awaks ambulatory systolic BP

FIGURE 2

Bland-Altman plot comparing automated office systolic BP with awaks ambulatory systolic BP

FIGURE 2

Bland-Altman plot comparing automated office systolic BP with awaks ambulatory systolic BP

FIGURE 3

Bland-Altman plot comparing manual office diastolic BP with awaks ambulatory diastolic BP

FIGURE 3

Bland-Altman plot comparing manual office diastolic BP with awaks ambulatory diastolic BP

FIGURE 4

Bland-Altman plot comparing automated office diastolic BP with awaks ambulatory systolic BP

FIGURE 4

Bland-Altman plot comparing automated office diastolic BP with awaks ambulatory systolic BP

Finally, in Figure 5, we present the results in relation to the algorithm proposed by Myers11 as described under the analysis section above … recalling still that we are looking at the accuracy of a single AOBP assessment. When systolic AOBP was ≥140 mmHg, there was a 74.2% likelihood that the ABPM would agree it is ≥140 mmHg but perhaps most importantly there is only a small chance that the individual would be labelled as normal (<130 mmHg) on ABPM. If systolic AOBP was <130 mmHg, then ABPM agreed only 42.5% of the time but it labelled the person as hypertensive on 21.7% of the time. The majority (52.4%) of individuals with AOBP 130–139 mmHg were hypertensive ≥140 mmHg on ABPM. For diastolic BP, only 15.8% of people labelled hypertensive (≥90 mmHg) by AOBP were considered normal by ABPM; however, there were more found to be borderline than actually hypertensive. When diastolic AOBP was <80 mmHg, <1% were found to be hypertensive on ABPM.

FIGURE 5

Assessing the proposed algorithm (Myers8) for management of BP measurement by AOBP

FIGURE 5

Assessing the proposed algorithm (Myers8) for management of BP measurement by AOBP

Discussion

This paper is about trying to determine which of the two methods of measuring BP in the physician’s office, MOBP versus AOBP, is best at predicting whether clinically important elevated BP is likely to exist. We have defined the awake mean BP on ABPM as the gold standard. The MOBP was given an advantage because it was the mean of three BP assessments done on separate occasions while the AOBP was based on a single assessment by the BpTRU device.

In real life, we would never make a diagnosis of the presence or absence of hypertension based on one measurement unless the BP was extremely high. The three manual BP readings that we used in this study would be closer to what is normally done. Yet, it appears that these readings are not even as accurate as a single automated measurement. This lack of accuracy on the part of MOBP recordings by physicians may be due to the technique being used and the care being taken in a busy practice.

While a single AOBP assessment is superior to repeated MOBP assessments, it is still not as good as one would like in order to be able to say with certainty whether or not hypertension exists in an individual patient. It is likely that repeated AOBP assessments, even two or three of them over a period of several weeks, as is the norm in practice, would greatly improve its already superior predictability. This is the study that critically needs to be done. AOBP devices like the BpTRU are proliferating and it is imperative that we understand how to interpret its results.

The assessment of the AOBP measurements and awake ABP measurements in relation to the proposed algorithm by Myers11 is important to consider. Remembering still that we are talking about a single AOBP assessment of BP, it appears that the systolic threshold of ≥140 mmHg works reasonably well at ‘ruling in’ hypertension but most importantly has a low false positive rate (4.9%). However, AOBP is less accurate when labelling people as having normal systolic pressure (<130 mmHg) in that 21.7% of these people had elevated systolic pressure on ABPM. For diastolic pressure, AOBP of ≥90 mmHg is confirmed by ABPM only 36.7% of the time but most of the rest are borderline and only 15.8% would be normal on ABPM and hence falsely labelled as hypertensive by AOBP. This is reasonable but not as good as for systolic pressure. Where AOBP is most accurate is when it says the diastolic BP is <80 mmHg; in that situation, there is <1% likelihood that the individual will have an elevated diastolic BP (≥90 mmHg) on ABPM. While these data are supportive of the use of AOBP and are much better than MOBP, the results would likely be more accurate if repeated AOBP measurements over several weeks were used. There is a need for a study that assesses the accuracy of repeated measurements of AOBP.

Limitations

The most obvious limitation of this study is the fact that it was done on people with an existing diagnosis of hypertension. These data therefore can only reliably be applied to the situation where decisions are being made about whether a person with an existing diagnosis of hypertension remains hypertensive or has achieved their BP target. The authors believe that that these data do reasonably reflect the situation for initial diagnosis of hypertension; however, the study mentioned above where the predictability and accuracy of multiple AOBP assessments is determined, should be conducted on a range of individuals, both normotensive and hypertensive individuals. These data are from baseline enrolment in two different randomised controlled trials (RCTs). However, both RCTs had exactly the same eligibility criteria so this should not lead to different populations. Both studies recruited participants from practices in the same city.

Conclusions

A single AOBP assessment (BpTRU) predicts the results on awake ABP better than the average of three MOBP measurements. Until we have better data using multiple AOBP assessments over time, the current recommendation of 135/85 mmHg as the AOBP cut-point for defining elevated BP should probably still be used. However, a case for a cut-point of 130/80 mmHg could be made and future studies may suggest this. It is also not unreasonable to consider a graded approach, as is being proposed by Myers.11 This approach has the cut-point for definitely elevated BP set at 140/90 mmHg; a BP <130/80 mmHg is considered normal and the 130–139/80–89 mmHg range would warrant assessment with ABPM and individualization of management based on individual circumstances, such as lifestyle, risk factors and target organ damage. More research is needed to clarify whether multiple AOBP assessments will provide data to better justify this recommendation. Data from this study on a single AOBP assessment provides support for this recommendation.

Declaration

Funding: Heart and Stroke Foundation of Ontario (4884 and 4882).

Ethical approval: Research Ethics Board of Queen's University, Kingston, Ontario, Canada.

Conflict of interest: none.

The data for this paper were collected during the conduct of two studies funded by the Heart and Stroke Foundation of Ontario.

References

1
Sega
R
Facchetti
R
Bombelli
M
, et al.  . 
Prognostic value of ambulatory and home blood pressures compared with office blood pressure in the general population: follow-up results from the Pressioni Arteriose Monitorate e Loro Associazioni (PAMELA) study
Circulation
 , 
2005
, vol. 
111
 (pg. 
1777
-
83
)
2
Verdecchia
P
Porcellati
C
Schillaci
G
, et al.  . 
Ambulatory blood pressure—an independent predictor of prognosis in essential hypertension
Hypertension
 , 
1994
, vol. 
24
 (pg. 
793
-
801
)
3
Clement
DL
De Bacquet
DA
Duprez
DA
, et al.  . 
Prognostic value of ambulatory blood-pressure recordings in patients with treated hypertension
N Engl J Med
 , 
2003
, vol. 
348
 (pg. 
2407
-
15
)
4
Ohkubo
T
Imai
Y
Tsui
I
, et al.  . 
Home blood pressure measurement has a stronger predictive power for mortality than does screening blood pressure measurement: a population based observation in Ohasama, Japan
J Hypertens
 , 
1998
, vol. 
16
 (pg. 
971
-
5
)
5
Wright
JM
Mattu
GS
Perry
TL
Jr
, et al.  . 
Validation of new algorithm for the BPM-100 electronic oscillometric office blood pressure monitor
Blood Press Monit
 , 
2001
, vol. 
6
 (pg. 
161
-
5
)
6
White
WG
Anwar
YA
Evaluation of the overall efficacy of the Omron office digital blood pressure HEM-907 monitor in adults
Blood Press Monit
 , 
2001
, vol. 
6
 (pg. 
107
-
10
)
7
Stergiou
GS
Tzamouranis
D
Protogerou
A
Nasothimiou
E
Kapralos
C
Validation of the Microlife WatchBP Office professional device for office blood pressure measurement according to the International Protocol
J Hypertens
 , 
2008
suppl 1
pg. 
S481
 
8
Beckett
L
Godwin
M
The BpTRU automatic blood pressure monitor compared to 24 h ambulatory blood pressure monitoring in the assessment of blood pressure in patients with hypertension
BMC Cardiovasc Disord
 , 
2005
, vol. 
5
 pg. 
18
 
9
Mattu
GS
Heran
BS
Wright
JM
Overall accuracy of the BpTRU—an automated electronic blood pressure device
Blood Press Monit
 , 
2004
, vol. 
9
 (pg. 
47
-
52
)
10
Palatini
P
Frigo
G
Bertolo
O
, et al.  . 
Validation of the A&D TM-2430 device for ambulatory blood pressure monitoring and evaluation of performance according to subjects characteristics
Blood Press Monit
 , 
1998
, vol. 
3
 (pg. 
255
-
60
)
11
Myers
MG
A proposed algorithm for diagnosing hypertension using automated office blood pressure measurement
J Hypertens
 , 
2010
, vol. 
28
 (pg. 
703
-
708
)