Blood pressure measurement: Should technique define targets?

Abstract Accurate assessment of blood pressure (BP) is the cornerstone of hypertension management. The objectives of this study were to quantify the effect of medical personnel presence during BP measurement by automated oscillometric BP (AOBP) and to compare resting office BP by AOBP to daytime average BP by 24‐h ambulatory BP monitoring (ABPM). This study is a prospective randomized cross‐over trial, conducted in a referral population. Patients underwent measurements of casual and resting office BP by AOBP. Resting BP was measured as either unattended (patient alone in the room during resting and measurements) or as partially attended (nurse present in the room during measurements) immediately prior to and after 24‐h ABPM. The primary outcome was the effect of unattended 5‐min rest preceding AOBP assessment as the difference between casual and resting BP measured by the Omron HEM 907XL. Ninety patients consented and 78 completed the study. The mean difference between the casual and Omron unattended systolic BP was 7.0 mm Hg (95% confidence interval [CI] 4.5, 9.5). There was no significant difference between partially attended and unattended resting office systolic BP. Resting office BP (attended and partially attended) underestimated daytime systolic BP load from 24‐h ABPM. The presence or absence of medical personnel does not impact casual office BP which is higher than resting office AOBP. The requirement for unattended rest may be dropped if logistically challenging. Casual and resting office BP readings by AOBP do not capture the complexity of information provided by the 24‐h ABPM.

coat effect. 4 Hence, casual office BP is not recommended, neither for the diagnosis nor for the management of hypertension. Resting office BP is recommended by all professional societies, with variation in the definition of what this entails, ranging from a clear preference for automated oscillometric BP (AOBP) by Hypertension, Canada to a clear admonishment against using unattended AOBP from the European Society of Hypertension/European Society of Cardiology as well as the International Society of Hypertension. [5][6][7][8] In the largest randomized controlled trial (RCT) to inform clinical practice in this area, namely the Systolic blood PRessure InterveNtion Trial (SPRINT), AOBP was used for BP measurement. 9 While the methodology of BP assessment in this trial was standardized concerning the model and manufacturer of the automated oscillometric BP device, resting time, and number of recordings, it turned out that measurements in some centers were performed as completely unattended while in others with medical personnel present for part of the measurements. 10 The requirement for completely unattended rest may be logistically challenging for many clinical settings, and may entail the need to purchase new devices if the existing AOBP device does not allow for this. Lastly, though 24 h ambulatory BP monitoring (ABPM) provides the most comprehensive data on BP load during the day and night, diurnal patterns of BP, and BP variability, this was not performed at the same time as the office BP measurements in SPRINT, making comparisons between reported BP readings difficult. 10 To address these issues, we conducted a prospective crossover trial to examine the effect of complete absence and partial absence of medical personnel on resting BP. In addition to that, we compared results of resting automated office BP (in the complete and partial absence of medical personnel) to corresponding office visit casual BP and average of the daytime BP from 24-h ABPM administered at the time of office visit.

Study design
This was a prospective, randomized, crossover study to compare different BP measurement methods (see Figure S1 for study flow).

Inclusion criteria
All patients being followed in a tertiary care, referral hypertension clinic were eligible for enrollment.

Exclusion criteria
Patients for whom oscillometric measurements may be difficult were excluded, namely with inability to do oscillometric measurements (eg, arrhythmias, pain, device reporting error) or lack of consent from the patient.  Figure S1). Measurements for ABPM were done every 20 min during the awake period and every 30 min during sleep. 11 We only accepted reports comprising at minimum 75% recordings as being valid. Awake and sleep periods were defined according to the patients' diaries. The order of AOBP measurement methods was randomized. Randomization was done by generating a randomization list and allocation was concealed using sealed envelopes.

Outcomes
The primary outcome was the effect of unattended 5-min rest pre-

Analysis
Sample

RESULTS
Ninety patients consented for this study, and 78 patients completed the study as designed, over the period of study from January 2018 to January 2019 ( see details in  Table 1. All data as mean + standard deviation or number (percentage) unless otherwise specified.

TA B L E 1 Baseline characteristics
Abbreviations: BMI, body mass index; BP, blood pressure; IQR, interquartile range.

Casual BP compared to resting AOBP
The mean difference between the casual and Omron unattended sys-  Table 2 for more details).

Difference between the resting AOBP methods
The mean difference between the Omron unattended and partially attended was 2.7 mm Hg (95 % CI -4.0, 1.4 mm Hg). Similarly, there was not a significant difference between Omron unattended and BpTRU (mean 1.4, 95% CI -5.6, 2.8 mm Hg) or the Omron partially attended and BpTRU (mean -4.5, 95 5 CI 11.5, -20.5 mm Hg). For diastolic BP, the mean difference was not significant between the two Omron methods at 0.5 mm Hg, but the Omron unattended reported a lower diastolic BP compared to the BpTru (mean -3.4, 95% CI 0.5, 6.3 mm Hg), with more details in Table 3.

Comparison of resting AOBPs with daytime average from 24 h ABPM
The mean systolic BP with AOBP as measured by Omron unattended was lower than the daytime ABPM average (difference -7.8 mm Hg, 95 %CI -5.0, -10.8). Similarly, the mean systolic BP with other AOBP methods was also lower than the daytime ABPM (Omron partially TA B L E 2 Differences between casual and resting blood pressure readings

Comparison of casual, resting AOBP (pooled), and daytime ABPM
Casual office BP readings were significantly higher as compared to resting office BP by AOBP (see details in Table 5). The AOBP systolic and diastolic BP was significantly lower than the daytime ABPM by 10.3 mm Hg and 7.9 mm Hg, respectively.

Limits of agreement
The detailed limits of agreement using the Bland-Altman method are presented in Table 6. These were 26.7 to -12.7 mm Hg for casual compared to Omron unattended systolic BP ( Figure 2) and 15.0 to -32.8 for pooled AOBP compared to daytime ABPM ( Figure S2). Our study has several limitations. Firstly, at the time this study was conceived and executed we were not aware of differences between systolic AOBP and daytime systolic BP from 24-h ABPM described by Roerecke and coworkers and observed in the SPRINT ambulatory BP substudy. 29,31 Hence our data are relevant primarily to patients with systolic AOBP < 130 mm Hg. Secondly, due to protocol violations, the numbers in each group were not exactly 30 as planned. Despite this, the relative numbers in each group were robust enough and the differences in numbers between the groups small with regards to analysis of the primary outcome. Third, when interpreting this data, it is crucial to understand that even casual office BP in this study is of research quality with regards to execution and may not be generalizable to a casually performed BP in a busy primary care practice. Fourth, our data indicating no effect on presence of medical personnel during resting on AOBP should be interpreted with caution as "presence" does not equal presence of noise from moving or working and changes to lighting conditions etc.
Casual office BP readings are significantly higher compared to office BP readings after 5-min rest. As treatment thresholds and BP targets are based on "SPRINT-like" resting office BP, it is important to acknowledge that unattended resting may impose logistical issues for practitioners with no clinically meaningful impact on resting office BP. Casual and resting office BP readings by AOBP naturally cannot capture complexity of BP behavior and overall daytime and nocturnal BP load provided by 24-h ABPM and as such, they should not be the sole diagnostic tool for diagnosis and management of hypertension.