Comparison of unattended, attended and routine clinic systolic blood pressure measurements and determinants of blood pressure difference between attended and unattended BP

Journal of Human Hypertension ARTICLE www.nature.com/jhh OPEN Comparison of unattended, attended and routine clinic systolic blood pressure measurements and determinants of blood pressure difference between attended and unattended BP Fraser Todd 1, Joseph Rushton1, Emily Hayashi1, Hannah Brown1, Cynthia Nguyen ✉ Gary Yip1,2 and Mahesan Anpalahan 1,2 2 , Emmeline Lai2, Emily Schembri2, 1234567890();,: © The Author(s) 2026 The study compared unattended(uAOBP), attended(aAOBP) and routine clinic systolic blood pressures (SBP) and investigated the role of BP measurement sequence and aAOBP threshold on SBP difference between aAOBP and uAOBP. Eligible patients were randomized to the order of aAOBP and uAOBP measurements during a single clinic visit. aAOBP and uAOBP measurements were obtained under standardised conditions using an Omron HEM907 BP-monitor. Routine clinic BP(CBP) from the same visit was also obtained. 85 patients (mean age 69.9 ± 16.4 years, 52.9% males) underwent aAOBP and uAOBP measurements. CBP was available in 78. uAOBP was significantly lower than both aAOBP (mean difference 4.12 ± SD7.91 mmHg, P < 0.001) and CBP (mean difference 6.67 ± SD6.67 mmHg, P < 0.001). When comparing randomised groups, mean BP difference was significant in only those who underwent aAOBP measurement first (6.30 + 7.29 mmHg; P < 0.001), but not in those who underwent uAOBP measurement first (1.88 + 7.98 mmHg; P = 0.134). In subgroup analysis by aAOBP ≥130 mmHg(vs.<130 mmHg), the mean BP difference was significant in only those with an aAOBPs ≥130 mmHg (6.13 ± 0.07 mmHg, P < 001), but not in those with an aAOBPs <130 mmHg (1.63 + 7.05 mmHg, P = 0.162). Correlation and Bland-Altman analyses revealed closer correlation between uAOBP and aAOBP(r = 0.925, P < 0.001), with lower bias (4.12 mmHg) and narrower 95% limits of agreement (LoA) (−11.38, 19.62 mmHg), compared to clinic SBP (r = 0.784, P < 0.001; Bias: 6.67 mmHg; LoA:−20.67, 34.01 mmHg. An aAOBP of 140 mmHg corresponded to an uAOBP of 135.3 mmHg. Both CBP and aAOBP significantly overestimate uAOBP. BP differences are partly attributable to the sequence of BP measurement and the level of aAOBP. Journal of Human Hypertension; https://doi.org/10.1038/s41371-026-01162-5 INTRODUCTION Accurate measurement of blood pressure for the diagnosis of hypertension and treatment monitoring remains a challenge. Blood pressure in routine practice is usually measured using an automated oscillometric device with a clinician present, and this may cause falsely elevated blood pressure due to the white coat effect [1]. In 2015, the Systolic Blood Pressure Intervention Trial (SPRINT) brought the concept of unattended automated blood pressure measurement (uAOBP) to the forefront of blood pressure assessment [2]. Automation mitigates the risk of observer bias and readings taken with the patient alone reduces the risk of white coat effect [3]. Despite these advantages, the suitability of uAOBP as the standard method for routine office BP assessment remains controversial, with recommendations in guidelines ranging from a clear preference [4] to admonishment [5]. This may be because of a lack of strong evidence for predicting outcomes for uAOBP compared to attended automated office BP (aAOBP) [6–8]. Furthermore, studies comparing uAOBP and aAOBP techniques have produced mixed results in blood pressure readings [3, 9, 10]. Thus, we aimed to compare unattended, attended and routine office or clinic systolic blood pressures in patients attending outpatient medical clinics. Additionally, we investigated the potential modifying effect of BP measurement sequence and the attended BP threshold <130 mmHg (vs. ≥130) on the mean BP difference between aAOBP and uAOBP measurements. We also investigated the potential associations of demographic and clinical variables with the mean BP difference between attended and unattended BPs. MATERIALS AND METHODS Design, setting and participants This was a prospective, observational study conducted at Eastern Health, a tertiary teaching hospital network in Melbourne, Australia, where patients were randomly assigned to the measurement sequence of aAOBP and uAOBP. Eligible participants included all patients aged 18 or over with a documented history of hypertension and/or cardiovascular comorbidities attending medical outpatient clinics at Eastern Health. Data collection occurred between November 2023 and February 2025. Patients were approached during their clinic visit and provided with written information about the study. After obtaining informed verbal consent participants were randomised in a 1:1 ratio to one of two 1 Department of General Medicine, Eastern Health, Melbourne, VIC, Australia. 2Department of Medicine, Monash University, Melbourne, VIC, Australia. ✉email: Received: 18 November 2025 Revised: 3 May 2026 Accepted: 14 May 2026 F. Todd et al. 2 sequences: (1) uAOBP followed by aAOBP, or (2) aAOBP followed by uAOBP. Blood pressure measurement Both aAOBP and uAOBP measurements were performed using the Omron HEM-907 automated BP monitor. Measurements were taken using an appropriately sized cuff, with the patient seated comfortably and resting for five minutes prior to the first reading, in accordance with current blood pressure guidelines [5]. For the uAOBP measurement, the device was programmed to take three BP measurements, the first after a five-minute rest followed by two additional readings at one-minute intervals. After activating the device, the investigator left the room and only returned after the three readings had been complete. For the aAOBP measurement, the investigator remained present both during the five-minute rest and measurement periods and manually activated each reading at one-minute intervals. The mean value of the three blood pressure readings was used for analysis. During the aAOBP measurement no interaction or conversation was permitted between the patient and investigator. The patient was also not allowed to use cell phone. The investigator remained in the room but did not engage with the patient and quietly worked on the computer, collecting relevant study data from the electronic medical records. Routine clinic blood pressures taken by the treating clinician in the same visit were obtained retrospectively from medical records. Routine clinic blood pressures were obtained using either manual auscultatory or automated monitors at the discretion of the clinic physician and depending on device availability. The clinics were staffed by consultant physicians and trainee physicians working under supervision. Investigators were not present during routine clinic BP measurements and deliberately minimised any interactions with the clinic staff to avoid any potential influence of the study on usual clinical practice. As investigators were not present during clinic BP measurements, it is difficult to comment on the specific conditions under which routine clinic BPs were obtained, but they were presumably mea (...truncated)