Introduction

The world population is continuing to age rapidly [1]. The older population itself has been aging, with the oldest growing faster than younger demographics. Hypertension increases with age, and approximately 70–80% of oldest-old individuals have hypertension [2, 3]. High blood pressure (BP) is a major risk factor for cardiovascular complications in individuals of all ages [4, 5], and high BP is associated with declining physical function and premature disability in older adults [6,7,8]. While several studies have shown that treatment of hypertension is beneficial for elderly patients, including those aged over 80 years [9, 10], controlling BP in the elderly is challenging because of highly diverse comorbidities, as well as polypharmacy, frailty, cognitive impairment, and variable life expectancy [11].

Self-measurement of BP at home is recommended for older adults because it provides more reliable and reproducible estimations of BP, including minimization of the white-coat effect and the detection of masked hypertension [12]. For the accurate measurement of home BP, the use of validated devices that have been tested for clinical application according to validation standards is mandatory. However, in some special populations, automated devices are problematic, i.e., patients with arterial stiffness [13, 14] and the oldest old in whom vessel stiffness is most frequently encountered [15], and evidence on the accuracy of automated BP devices for the oldest old or elderly populations is very limited [16,17,18]. Therefore, we aimed to investigate the accuracy and reliability of the Omron HEM-7080IC home BP measurement device (Omron Healthcare Co., Ltd., Kyoto, Japan) in individuals older than 85 years according to the International Organization for Standardization (ISO) 81060-2:2013 protocol [19].

Methods

Participants

Outpatients with or without hypertension were recruited from the Department of Geriatric Medicine, Osaka University Hospital (Suita, Japan). We followed the ISO protocol designed for a special population because the oldest-old population was considered to be a possible special population. Since Omron HEM-7080IC uses the same measurement technology as Omron 705-IT, which was validated previously in the general population [20, 21], 35 special population patients were required. A total of 38 patients aged 85 years or older without chronic atrial fibrillation or dementia were screened. Of those, two participants whose BP was unstable and one who could not tolerate repeated measurements were excluded. The number of participants in the present study therefore totaled 35.

The study protocol was approved by the ethics committee of Osaka University Hospital (approval number: 17378). Written informed consent was obtained from each participant.

Device

The Omron HEM-7080IC is an upper-arm electronic oscillometric device with a capacitive pressure sensor designed to measure pressure in the range of 0–299 mmHg and pulse in the range of 40–180 beats per min. Systolic BP (SBP), diastolic BP (DBP), and pulse are shown on a liquid crystal digital display. Inflation is controlled by an automatic pumping system, and deflation is controlled by an automatic pressure release valve. The unit is powered by four alkaline batteries (1.5 V, type LR6), which enable approximately 300 measurements, or by an AC adapter.

A standard cuff can be used for individuals with arm circumferences in the range of 22–32 cm, while a small-sized cuff (arm circumference: 17–22 cm) and large-sized cuff (arm circumference: 32–42 cm) are also available. We measured the arm circumference in individuals so that an adequately sized cuff could be selected.

Procedure

All procedures were conducted in accordance with the ISO protocol [19]. The validation procedure was carried out using same-arm sequential measurements by a team of three medical professional observers who were trained and experienced in BP measurement. Two of the three observers simultaneously measured BP by the auscultatory technique using a standard mercury sphygmomanometer (ACOMA Medical Industry Co., Ltd, Tokyo, Japan), which had been calibrated before the study, and a dual-headed binaural stethoscope with a Y-shaped tube (TSUBASA INDUSTRY Co., Ltd, Tokyo, Japan). The cuffs used for observer measurements were selected such that the bladder dimensions were approximately 80–100% the length/40% the width of each patient’s arm circumference [19, 22]. The observers’ individual measurement values were averaged to determine the observer BP measurements. The third observer was a supervisor who checked the values obtained by the two observers and measured the BP with the test device. The observers were trained previously on how to reach a consensus regarding BP measurements, and each observer was blinded to the other observer’s measurements as well as to the test device measurements throughout the study.

After at least 5 min of rest, the participant sat in a quiet temperature-controlled examination room with his or her arm and back supported, and BP was measured sequentially at the level of the heart on the left arm. Two observers simultaneously measured BP using the mercury sphygmomanometer with the dual-headed binaural stethoscope, followed by measuring BP with the Omron HEM-7080IC as the test device. Neither of these running-in measurements was used for the calculation of the test device accuracy.

Then, BP was measured first by the observers and then by the test device and observers alternately, for a total of seven sequential measurements with at least 1 min between each measurement. Each test-device reading was compared with the mean of the observer measurements taken immediately before and after the test-device measurement; therefore, three pairs of the test device and observer BP measurements were obtained for each participant. According to the ISO protocol [19], we allowed an observer-to-observer difference within 4 mmHg; otherwise, another pair of observer-test BP measurements was taken, with a maximum of eight pairs.

Statistical analysis

Data were analyzed according to criteria 1 and 2 required by the ISO standard [19]. For criterion 1, each test device measurement value was compared with the corresponding observer measurement. The differences between the test device and observer SBP and DBP measurements were calculated using the mean and standard deviation (SD). Bland–Altman plots were used to show deviations in the data [23]. For criterion 2, the mean and SD of each paired BP measurement (test device and observer) was calculated individually for each participant and compared.

We used a linear mixed model that included sex, age, cuff size, upper arm circumference, corresponding observer BP, and pulse pressure to identify confounders causing BP differences. The data were analyzed using SPSS 22.0 for Windows (IBM Japan, Tokyo, Japan). P-values < 0.05 on two-sided tests were considered significant.

Results

The characteristics of the 35 participants are presented in Table 1. Of the participants, 65.7% were women, and the age ranged from 85 to 96 years (87.7 ± 2.6 years). Six participants used a small-sized cuff (a cuff for children: 90-mm-wide cuff bladder was used for BP measurements with a mercury sphygmomanometer), but no one required a large-sized cuff. Three valid determination pairs of the test device and observer BP were obtained, and the ranges of SBP and DBP were 89–180 and 53–97 mmHg for the observer measurements and 77–184 and 44–89 mmHg for the device measurements, respectively. The observer-to-observer differences were 0.0 ± 1.7 mmHg for SBP and 0.0 ± 2.0 mmHg for DBP.

Table 1 Characteristics of the study participants
Full size table

Bland–Altman plots of BP differences between the test device and observer measurements with a total of 105 separate SBP and DBP pairs are presented in Fig. 1. The summary statistics of the differences between observer and device measurements are shown in Table 2. The mean and SD values of the differences between the test device and observer SBP and DBP measurements fulfilled both criterion 1 (−0.7 ± 7.1 mmHg for SBP and −1.1 ± 4.5 mmHg for DBP) and criterion 2 (−0.7 ± 5.8 mmHg for SBP and −1.1 ± 4.1 mmHg for DBP) of the ISO protocol.

Fig. 1
figure 1

Bland–Altman plots of the comparison between the Omron HEM-7080IC device and observer measurements against the mean pressure of the device and observer measurement of the systolic (a) and diastolic (b) blood pressures. Mean differences and 95% limits of agreement between paired measurements are represented by dashed and 2-dot-dashed lines, respectively

Full size image
Table 2 Difference between the device and observer blood pressure measurements
Full size table

In the linear mixed model, the SBP difference between the test device and observer measurements was significantly correlated with the cuff size (P = 0.002, Table 3). The mean device-observer differences were −7.3 ± 6.9 mmHg for the small cuff and 0.7 ± 6.4 mmHg for the standard cuff. However, the DBP difference was associated with pulse pressure (P < 0.001). When divided into three equal subgroups by pulse pressure values, the BP differences of the study participants were all within 5 ± 8 mmHg.

Table 3 Factors associated with BP differences between test device and observer measurements
Full size table

Discussion

To the best of our knowledge, no one has assessed whether home BP can be appropriately measured in an oldest-old population. We assessed the accuracy of the Omron HEM-7080IC device in oldest-old individuals aged 85 years or older according to the ISO 81060–2:2013 protocol. We demonstrated that the device fulfills criterion 1 as well as criterion 2 of the ISO requirements for both SBP and DBP, indicating that the device was accurate and reliable in the oldest-old population.

The increased BP fluctuation in older adults is well known, and BP is easily affected by measurement conditions [24, 25]. Measurement of home BP repeatedly and regularly with unbiased methods is essential regardless of the characteristics of people, including older patients. No study, however, has validated electronic oscillometric monitors used for home BP measurement among an oldest-old population. The Omron HEM-7080IC device uses the same technology as the Omron 705-IT. Coleman A et al. [20] and El Assaad M et al. [21] reported the accuracy of the Omron 705-IT according to the 1993 British Hypertension Society Revised Protocol [26] and the 2002 European Society of Hypertension International Protocol (ESH-IP) [27] among 85 patients aged 47 ± 15 years and among 33 patients aged 54 ± 13 years, respectively. To validate the device in an oldest-old population, we included 35 participants aged 85 years or older in this study according to the requirement of the ISO protocol as a special population. The interdevice differences observed in our study were similar to those observed in the two previous studies of adult populations for both SBP and DBP [20, 21].

In the present study, approximately half of the participants had systolic hypertension, whereas no patients had a DBP of 100 mmHg or higher. Stiffening of the arterial wall and decreasing vessel wall compliance become conspicuous with aging [15]. Such structural changes of the arterial tree increase SBP and decrease DBP, leading to a wider pulse pressure and isolated systolic hypertension as the predominant form of hypertension in older people [28]. The BP distribution in this study would therefore be typical in common oldest-old individuals. Moreover, some previous validation studies excluded patients with isolated systolic hypertension [16, 17], which would be dominantly observed in older populations [28]. The previous ESH-IP [27] specifies three BP categories: low (SBP 90–129/DBP 40–79 mmHg), medium (130–160/80–100 mmHg), and high (161–180/101–130 mmHg). However, the protocol does not clarify, in which BP category patients with isolated systolic hypertension should be placed. The oscillometric method is based on recording arterial wall oscillations, and a greater pulse pressure is associated with a larger oscillometric versus auscultatory BP discrepancy [13, 14]. The probability of incorrect measurement might increase in elderly patients who often show a wider pulse pressure range. In the present study, the differences between the test device and observer measurements were correlated with pulse pressure for DBP measurements. Nevertheless, the BP differences remained within acceptable levels based on the validation criteria. Our findings offer new information on the accuracy of an electronic automated oscillometric BP-measurement device for an oldest-old population in a real-world setting, including those with isolated systolic hypertension.

The multivariate analysis suggested that the cuff size affected the differences in BP measurements between the test device and observer measurements, while the upper arm circumference did not. It has been shown that miscuffing leads to BP under/over estimation [29]. We selected appropriate cuffs for both the test device and reference device following the manufacturer’s instructions and international protocol [19, 22]. Nevertheless, the SBP difference was large among six patients measured with the small cuff; the mean difference was −7.3 ± 6.9 mmHg for the small cuff and 0.7 ± 6.4 mmHg for the standard cuff. Although the measurements satisfied the criteria of the protocol, Stergiou et al. [30] showed reduced accuracy for DBP using a small-sized cuff compared with a standard cuff in a validation study among children and adolescents, which was followed by another validation study in adults [31]. Although a definite conclusion cannot be drawn based on the small number of patients in our study, caution may be required when a small cuff is applied in an oldest-old population. To date, few previous validation studies targeting adults have included participants with slender arms who require a small cuff [31, 32]. Further studies are needed to clarify the accuracy of the oscillometric devices in individuals who require a small cuff. Furthermore, we did not have a patient with an arm thickness of 32 cm or more who required a large-sized cuff. Japanese patients generally have slender arms. Japanese anthropometric reference data reported that the median (10th−90th percentile) arm circumference in individuals over 85 years old is 24.0 (19.4–28.4) cm in men and 22.6 (18.2–27.9) cm in women [33]. While there are some countries or regions where it may be rare to use a large-sized cuff [34, 35], in other countries such as the United States of America, there are a number of patients with thick arms who need a large-sized cuff [36, 37]. For example, the 75th percentile value of arm circumference in those aged 80 years or older in the U.S.A. is 31.9 cm for females and 33.0 cm for males; a quarter of older people need a large-sized cuff [36].

Participants with chronic atrial fibrillation or any sustained arrhythmia were not included in this study. Frequent ectopic beats could affect the accuracy of both auscultatory and oscillometric techniques as a result of increased beat-to-beat BP variability. Although atrial fibrillation is more prevalent in older people than younger adults, there is currently no agreed-upon procedure for BP monitor validation in patients with such conditions [22].

Self-measurement of BP at home is recommended for older adults. We showed that the Omron HEM-7080IC can be recommended for self/home use in the oldest-old population, although caution may be required when a small cuff is applied. However, oldest-old individuals have heterogeneous patterns of comorbidities and severity of conditions [11]. In a clinical setting, medical professionals need to take into account the patient’s condition, such as cognitive function and activities of daily living, when applying self-measurement of home BP to oldest-old individuals. The issue regarding which of the measured home BP values should be used for clinical evaluation also needs to be considered [38, 39].

In conclusion, this study is the first to report that automated electronic devices for BP monitoring are able to accurately and reliably measure BP in a special population of oldest-old individuals. We can expand the application of self-measurement of home BP to oldest-old populations, although some considerations are required.