Introduction

Accurate blood pressure (BP) measurements are needed in clinical practices, intervention studies, and health examination surveys. It is known that small inaccuracies in BP measurement can have considerable consequences at the population level. Decreases or increases in systolic and diastolic BP (SBP and DBP) by 10/5 mmHg result in quite large changes in hypertension prevalence (22.1% vs. 53.4%) and control rate of hypertension (21.2% vs. 3.6%), respectively [1]. Accurate BP measurements are based on the measurement environment, behavior of the subject, measurement procedures, devices used for the measurement, and the observer. A systematic review indicated that 29 factors may influence the accuracy of BP readings [2]. Among the various influencing factors, back support is an important factor. Cushman et al. [3] indicated that the auscultatory seated DBP measured without back support (on an examining table) was 6.5 mmHg higher than that measured with back support in 48 men with a history of essential hypertension, but the SBP was not significantly different between the two conditions. Later, two review papers were published by Handler et al. and Tolonen et al. Handler [4] pointed out that the SBP was 6–10 mmHg higher when the back was unsupported than when the back was supported. A systematic review by Tolonen et al. [5] indicated that SBP/DBP levels in patients with unsupported backs were higher by ~5–15/6 mmHg.

However, in 85 adult participants with a mean oscillometric SBP/DBP of 119.2 ± 16.4/68.2 ± 8.8 mmHg (range of 80–220/50–120 mmHg) with back support, Ringrose et al. recently demonstrated that BP levels with an unsupported back were higher but found that the differences in SBP/DBP were only 0.7 ± 4.9/1.8 ± 3.0 mmHg. Therefore, the authors concluded that the provision of back support has a small effect on the mean oscillometric BP, especially for SBP [6].

Because the impact of unsupported back BP is controversial [3,4,5,6], and oscillometric BP measurement is widely used in the clinic and home at present, it is necessary to evaluate the difference in BP readings between supported and unsupported back BP readings when using oscillometric BP devices. This study was designed to evaluate the impact of an unsupported back on oscillometric BP readings.

Method and subjects

This study was approved by the Ethics Committee of the Second Affiliated Hospital of Nanchang University. From June to July 2020, 224 outpatients of the cardiovascular department (18–86 years, 54.5 ± 15.5 years old, 105 males) with sinus rhythm were enrolled. Age, history of hypertension, and information on antihypertensive treatment were collected. Among them, 100 participants were treated with calcium channel blockers, 19 with angiotensin-converting enzyme inhibitors, 50 with angiotensin receptor blockers, 36 with beta blockers, and 33 with diuretics.

The exclusion criteria were acute coronary syndrome, acute stroke, acute heart failure and atrial fibrillation, known peripheral artery disease, and walking disabilities.

BP measurements

BP was measured by a doctor specifically trained on BP measurement according to the Chinese hypertension guideline standards and those of the American Heart Association [7, 8]. The participants were advised to avoid cigarette smoking and consumption of coffee, tea, and alcohol for at least 30 min before the BP measurement. After proper placement of the cuff, the arm BP was measured using an automatic oscillometric BP monitor (Omron HEM-1300) after a 5-min rest period. During the BP measurement, the feet of participants were always on the ground.

Two protocols were designed to attenuate the bias induced by BP measurement order: one was supported–unsupported–supported–unsupported and the other was unsupported–supported–unsupported–supported. The participants followed one protocol at random for BP measurement. After a BP measurement, the participants changed their back position for 1 min before the next BP measurement.

The mean SBP and DBP of two readings in the same position were recorded as the final BP value. The difference in SBP/DBP between the unsupported and supported back positions was calculated as delta SBP and delta DBP. Meanwhile, the percentage variation (PV) was calculated with the formula delta BP/BP using the unsupported back measurement.

Statistical analysis

Data are expressed as the means ± standard deviation for continuous variables. The paired sample t-test was used to determine the significance of delta BP between different conditions, such as age, sex, and hypertension history. Hypertension history was defined as a patient having ever been diagnosed with hypertension (SBP > 140 mmHg or DBP > 90 mmHg). For multivariable regression analysis, the dependent factor was PV of the SBP or DBP difference level, and the independent factors included age, sex (0 = woman; 1 = man), hypertension history (0 = no hypertension history, 1 = hypertension history), and supported BP level. A two-sided p value < 0.05 was considered statistically significant.

Results

The SBP/DBP ranges were 90–190/60–120 mmHg in these 224 adult participants. The SBP levels for the unsupported back position were 132.7 ± 19.5 mmHg, which was slightly but significantly higher by 2.3 mmHg than in the supported back position; the DBP levels in the unsupported position were 79.6 ± 12.9 mmHg, which were higher by only 1.0 mmHg (no significant difference) than in the supported back position. The HR levels were similar between the two positions (with a mean difference of 0.4 bpm) (Table 1).

Table 1 BP and HR between the back unsupported and supported positions
Full size table

Compared with the supported back measurements, SBP and DBP levels measured with the back unsupported were higher in all six subgroups regardless of age, sex, or hypertension situation.

The >65-year subgroup had higher SBP and lower DBP than the ≤65-year subgroup. Moreover, the >65-year subgroup also had significantly higher delta SBP (4.6 ± 13.9 mmHg vs. 1.5 ± 6.4 mmHg, p < 0.05).

Although the hypertension subgroup had higher SBP and DBP than the normotension subgroup, no significant difference in delta SBP or DBP was found between the two subgroups.

The male subgroup had similar SBP and higher DBP levels than the unsupported back subgroup, and no significant difference in delta SBP or DBP was found between the two subgroups.

Furthermore, no significant difference in HR was found between the unsupported and supported back conditions regardless of age, sex, or hypertension situation (Table 2).

Table 2 The impact of age, gender, and hypertension on the delta BP and HR
Full size table

Multivariable regression analysis

The results showed that the supported BP level was a negative factor and that age was a positive factor for the PV of SBP differences. Age and supported DBP were two positive factors, but hypertension history was a negative factor for the PV of DBP differences (Table 3). However, the regression coefficient was small.

Table 3 The independent factors for the percentage variation from multivariable regression analysis
Full size table

Discussion

Our results showed that in the population with a mean oscillometric SBP/DBP of 132.7/79.6 mmHg when BP was measured with the back unsupported, the values were only 2.3/1.0 mmHg higher than those measured with the back supported. These values were generally close to the values of 0.7/1.8 mmHg reported by Ringrose et al. However, these values were obviously lower than those reported by previous studies [3,4,5].

Cushman et al. [3] performed a study in 48 male hypertensives and pointed out that the DBP was 6.5 mmHg higher when the participants sat on an examining table (back unsupported) than when their backs were supported. However, the Cushman study used a mercury sphygmomanometer rather than an oscillometric BP device, which was used in the current study and that of Ringrose. The different BP devices may be an important reason for the varying values, as using automated office BP measurements may reduce the white coat response [9]. In other words, the larger BP difference between the unsupported and supported back measurements in Cushman’s study may be attributed to the white coat effect of BP to some extent. The second reason may be the different studied populations. In addition to Cushman’s study, we also noted that there were two review papers dealing with the BP difference between unsupported and supported back positions. Handler [4] reported that the difference in SBP was 6–10 mmHg, and the differences in SBP/DBP reported by Tolonen et al. [5] were 5–15/6 mmHg. However, these values may mainly represent the experience of the authors, as the value in Handler’s paper was from Cushman’s study, and Handler’s paper was cited by Tolonen et al. As Ringrose’s study and our study were performed in a relatively large population (85 patients in the former and 224 subjects in the latter) and the repeated BP measurement protocol was assigned strictly in random order, the results may be reasonable, at least for oscillometric BP measurement.

Second, the multivariable regression analysis indicated that age was a positive factor for the PV of SBP and DBP differences. These results mean that the unsupported back position may induce a higher rise in SBP and DBP in aging subjects, even after correlation with baseline BP levels. The comparison of the delta SBP levels between the >65-year and ≤65-year subgroups also demonstrated this point (4.6 vs. 1.5 mmHg); therefore, a supported back position is more important for elderly subjects to obtain an accurate BP reading, especially for SBP.

Our results also showed that supported SBP level was a negative factor for VP of SBP, but supported DBP was a positive factor for VP of DBP. Furthermore, we noted that hypertension history was a negative factor for VP of DBP.

The exact mechanism for the difference in BP between the supported and unsupported back conditions has not been deeply investigated until now. However, some studies have investigated the impact of back support on physiological position. For example, a study found that both the myoelectric activity and disc pressure decrease when the back is supported compared with when the back is unsupported [10]. Another study also found that sitting with a fitted backrest reduced peak pressure under the ischia, reduced muscular activity, maintained total and segmental lumbar lordosis, and increased lumbar intervertebral disc heights [11]. By measuring electromyographic activity, Watanabe et al. [12] indicated that sitting reclining against the backrest of a chair could bring about the correct lumbar curvature, effectively stabilize the lumbopelvic region, and decrease focal stress on passive structures.

Recently, the impact of different sitting positions on physiological position has attracted increasing attention [13]. A study indicated that 40-min periods of unsupported sitting had broad impacts on discomfort, posture, spine load, and localized muscle fatigue [14]. Compared with other sitting postures, slouched sitting for 20 min had the lowest bilateral obliquus internus/transversus abdominis activity [15]. With these data, we recognize that the unsupported back position could increase spine load, disc pressure, localized muscle fatigue and myoelectric activity, discomfort, and even psychosocial stress. These changes may activate the sympathetic nervous system and then respond to the increased BP in the unsupported back position.

Clinical implications

Although the delta SBP difference was only 2.3 mmHg, it was statistically significant. Thus, physicians should allow subjects to sit in a back-supported position for BP measurement, especially elderly individuals. Previously, several studies demonstrated the impact of a 2 mmHg reduction in population BP on hypertension prevalence and hypertension control rate [16].

At present, BP kiosks are used in Canada and the USA, as well as in China. However, some BP kiosks are not equipped with backrest chairs. Even with backrest chairs, the backs of some people may not be supported when measuring BP. It is better to equip backrest chairs in the BP kiosks and provide a language indication during the BP measurement. Fortunately, the BP difference between the supported and unsupported back conditions is not large.

Limitations

There are several limitations in this study. The sample size of this study was not large. Moreover, we did not evaluate the impact of antihypertensive treatment on the BP difference, as various antihypertensive therapies were used.

Conclusion

In a population with a mean SBP/DBP of 132.7/79.6 mmHg when measured in the unsupported back position, the mean SBP/DBP levels measured with an oscillometric BP device with the back in an unsupported position were 2.3/1.0 mmHg higher than those taken with the back in a supported position. Age was a positive factor for large percentage increases in SBP and DBP differences.