Arm position and blood pressure: a risk factor for hypertension?

Transcription

1 (3) 17, & 3 Nature Publishing Group All rights reserved /03 $ ORIGINAL ARTICLE Arm position and blood pressure: a risk factor for hypertension? A Mourad, S Carney, A Gillies, B Jones, R Nanra and P Trevillian Department of Nephrology, John Hunter Hospital, Hunter Region Mail Centre, NSW, Australia The objective of this study was to re-evaluate the effect of arm position on blood pressure (BP) measurement with auscultatory and oscillometric methods including ambulatory blood pressure monitoring (ABPM). The setting was the hospital outpatient department and the subjects chosen were normotensive and hypertensive. The effect of lowering the arm from heart level on indirect systolic BP (SBP) and diastolic BP (DBP) measurement as well as the importance of supporting the horizontal arm were measured. In the sitting position, lowering the supported horizontal arm to the dependent position increased BP measured by a mercury device from / to / mmhg in normotensive subjects, a mean increase of 8/7 mmhg (Po0.01). In hypertensive subjects, a similar manoeuvre increased BP from / to / mmhg, an increase of 23/10 mmhg (Po0.01). Combined results from normotensive and hypertensive subjects demonstrate a direct and proportional association between BP (SBP and DBP) and the increase produced by arm dependency. Similar changes and associations were noted with oscillometric devices in the clinic situation. However, supporting the horizontal arm did not alter BP. Of particular interest, analysis of 13 hypertensive subjects who underwent ABPM on two occasions, once with the arm in the usual position and once with the arm held horizontally for BP measurement during waking hours, demonstrated changes comparable to the other devices. The mean 12-hour BP was / mmhg during the former period and significantly decreased to / mmhg during the latter period (Po0.01). Regression analysis of the change in SBP and DBP with arm position change again demonstrated a close correlation (r 2 ¼ and ; Po0.001) with the artefact being larger with higher systolic and diastolic pressures. In conclusion, arm movements lead to significant artefacts in BP measurement, which are greater, the higher the systolic or diastolic pressure. These systematic errors occur when using both auscultatory and oscillometric (clinic and ABPM) devices and might lead to an erroneous diagnosis of hypertension and unnecessary medication, particularly in individuals with high normal BP levels. Since clinical interpretations of heart level vary, the horizontal arm position should be the unambiguous standard for all sitting and standing BP auscultatory and oscillometric measurements. (3) 17, doi: /sj.jhh.1563 Keywords: blood pressure; indirect measurement; mercury and oscillometric device; ABPM Introduction Blood pressure (BP) measurement is the most common clinical measurement in hospitals, consulting rooms, and more recently home and the workplace. Most clinical practice audits reinforce the importance of the cuffed upper arm being at heart level; 1,2 yet published definitions and clinical interpretations of heart level vary, possibly owing to the heart extending from the sternal angle to the fifth intercostal space. 3 WHO guidelines recommend that the cubital fossa be placed at the fourth intercostal space. The American Heart Association defines Correspondence: Dr SL Carney, Department of Nephrology, John Hunter Hospital, Locked Bag 1, Hunter Region Mail Centre, NSW 2310, Australia. scarney@mail.newcastle.edu.au Received 22 July 2; revised 17 January 3; accepted 14 February 3 heart level as arbitrarily taken to the junction of the fourth intercostal space and the lower left sternal border. As a subsequent practical clinical guide, these guidelines state: When the patient is seated, placing the arm on a nearby tabletop a little above waist level will result in a satisfactory position. 5 The British Hypertension Society recommends that the arm be horizontal at the mid-sternal level. 6 This definition best approximates the right atrium, which is the reference level. 7 While the influence of arm position on BP measurement was first reported almost years ago, 8 Webster et al 9 more recently re-evaluated this issue and documented a mean increase of 11 (SBP)/ 12 (DBP) mmhg when the horizontal arm was made dependent in hypertensive subjects. Similar changes were noted in normotensives, and in addition up to 10/6 mmhg increases were measured if the horizontal arm was unsupported. These

2 390 important observations have been replicated using an oscillometric device 10 and confirmed by intraarterial BP measurements. 11 Despite studies confirming the importance of arm position on BP measurement, it is likely that arm position varies markedly in clinical practice and also in published studies. Resting an arm on a desk or arm of a chair as often recommended could vary the BP significantly, depending not just on the dimensions of the furniture but also on the height of the individual and also their posture. It was the intention of this study to repeat part of the Webster et al 9 protocol to decide if arm position errors are more or less apparent in hypertensive subjects when compared to normotensives, an issue that has not been evaluated. In contrast to that study, the Hawksley random zero sphygmomanometer was not used owing to concerns about its accuracy. 12,13 However, to minimise potential observer bias, an automatic oscillometric device was also used. Confirmation of the effect of arm position on BP with oscillometric devices is important, given their increasing use particularly in hospitals and the home. Evaluating the effect of arm position on ambulatory blood pressure monitoring (ABPM) was an additional study objective, given the increasingly important role of ABPM in hypertension management. It was postulated that the usually dependent arm during daytime ABPM significantly overestimates BP and exaggerates nocturnal dipping. Subjects and methods Arm position and clinic BP A total of 51 subjects participated, 26 being hypertensive, two of whom were not on medication. BP was measured with a standard mercury sphygmomanometer and validated 14,15 oscillometric device (OMRON HEM 705CP, Japan) in random order. The cuff was on the nondominant arm. With both devices, the BP was measured with the patient sitting and standing and with the arm dependent, horizontal unsupported and horizontal supported. These manoeuvres were performed in random order. The manufacturer s instructions for using OMRON oscillometric devices are comparable to the American Heart Association recommendations with the arm being flexed and resting on a desk or table. BP measurement was carried out in a comfortable environment by a trained observer (AM) after the subject had been sitting or standing for 5 min. A minute was allowed between each arm position change. Conversation was avoided during the procedure. Phase 1 Korotkoff sounds represented systolic BP (SBP) and Phase V the diastolic BP (DBP). BP technique followed the British Hypertension Society recommendations 6 including the horizontal position of the arm using oscillometric and mercury devices. Arm position and ABPM In total, 15 subjects participated: 10 were on antihypertensive medications and 5 were untreated while being evaluated following clinic diagnosis of hypertension. They were randomly allocated to usual or horizontal arm position during the first of two monitoring periods. Both monitoring periods were performed on days of similar activity. During the horizontal ABPM period, patients were asked to hold the arm in the horizontal position during each BP measurement when not in bed. They were allowed to rest the arm on nearby furniture or hold it without support. They were not advised on the possible effects of arm position on BP. Successful completion of greater than 90% of BP readings was required for each ABPM period using a spacelabs oscillometric device (Space Labs Inc., Redmond, WA, USA), with the cuff on the nondominant arm. Volunteers in both studies were in sinus rhythm, did not have cardiac failure, autonomic neuropathy or a recent history of postural hypotension, and were not taking tricyclic antidepressants or a-adrenoceptor blockers. Statistical analysis Student s paired t-test was conducted to compare systolic and diastolic pressures. Analyses compared dependent position with horizontal supported, horizontal supported vs horizontal unsupported while sitting and standing with both devices. Simple linear regression was used to compare the relations between horizontal and dependent arm positions and the level of SBP and DBP. Differences between usual and horizontal ABPM periods were examined by analysis of variance for repeated measurements. Differences were tested using Students paired t-test. Linear relations between patient BP were estimated using the least-squares method. Results are expressed as mean 7 s.d. Results Arm position and clinic BP Hypertensive subjects were older than the normotensive group, although mean body weight and arm circumference were similar between the groups (Table 1a). BP measured with a mercury sphygmomanometer was comparable in the sitting and standing positions in both normotensive and hypertensive groups (Table 2a). However, lowering the horizontal supported arm to the dependent position increased mean BP by 8/7 mmhg in sitting normotensive subjects and by 7/5 mmhg when standing (Po0.01 for both SBP and DBP). In hypertensive subjects, the same arm manoeuvre increased mean BP by 23/10 mmhg sitting and 21/10 mmhg standing (Po0.01 for both SBP and DBP). When comparing the supported arm with the unsupported arm,

3 Table 1 Patient characteristics 391 (a) Arm position and clinic BP Normotensive (n=25) Hypertensive (n=26) Mean 7 s.d. (range) Mean 7 s.d. (range) Sex (M/F) 10/15 10/16 Age (years) (17 58) * (18 81) Body weight (kg) ( 131) (38 140) Body height (cm) ( 183) ( ) Arm circumference (cm) (20 40) (20 46) (b) Arm position and ABPM Mean 7 s.d. (range) Sex (M/F) 9/4 (18 80) Age (years) Arm horizontal day 1 7 Antihypertensive treatment 9 *Po0.05. Table 2 Clinic and ABPM results (a) Auscultatory and oscillometric blood pressure (mmhg) measurements in normotensive and hypertensive subjects Posture and arm position Normotensive Hypertensive Systolic Diastolic Systolic Diastolic Auscultatory Sitting Horizontal supported 103 (10) 60 (7) 143 (21) 78 (17) Horizontal unsupported 103 (10) 60 (7) 142 (21) 77 (18) Dependent 111 (14)** 67 (10)** 166 (29)** 88 (20)** Standing Horizontal supported 103 (11) 63 (9) 142 (19) 77 (15) Horizontal unsupported 102 (12) 62 (10) 143 (20) 77 (15) Dependent 110 (13)** 68 (10)** 163 (25)** 87 (17)** Oscillometric Sitting Horizontal supported 103 (11) 62 (7) 145 (24) 79 (18) Horizontal unsupported 102 (11) 63 (8) 144 (25) 81 (18) Dependent 113 (13)** 72 (8)** 163 (29)** 88 (19)** Standing Horizontal supported 102 (12) 63 (9) 139 (24) 77 (13) Horizontal unsupported 101 (10) 63 (7) 139 (25) 78 (13) Dependent 114 (14)** 73 (10)** 157 (29)** 88 (16)** (b) ABPM daytime results Mean 12-h SBP Mean 12-h DBP Arm horizontal Arm usual position ** ** **Po0.01. mean BP was not significantly different. A regression analysis of the differences between sitting and standing BP in normotensive and hypertensive subjects using both mercury and oscillometric devices was performed. Figures 1 and 2 represent the sitting auscultatory SBP (y ¼ x ) and DBP (y ¼ x ) results, and sitting oscillometric and standing auscultatory and oscillometric results were similar as were the regression equations for each analysis. In particular the difference between the supported horizontal and dependent arm increased with the increase in SBP and DBP. For example, an SBP of 155 mmhg in the horizontal arm would become 180 mmhg in the dependent position, an increase of 25 mmhg. A DBP of 85 mmhg could be incorrectly measured as

4 r 2 = P <0.001 dependent (mmhg) unsupported (mmhg) r 2 = P< horizontal (mmhg) Figure 1 Sitting auscultatory SBP: supported horizontal compared to dependent arm position supported (mmhg) Figure 3 Sitting auscultatory SBP: supported compared to unsupported horizontal arm position. 2 r 2 = P< dependent (mmhg) mmhg BPM horizontal (mmhg) Figure 2 Sitting auscultatory DBP: supported horizontal compared to dependent arm position. 96 mmhg, an increase of 11 mmhg following arm dependence. Evaluation of the regression analyses for systolic pressure (horizontal supported compared to dependent) demonstrated a slope that was significantly different from the diastolic pressure relation, regardless of whether sitting or standing. As demonstrated in Figure 3, supporting the arm did not significantly alter sitting SBP or DBP when compared to horizontal unsupported auscultatory readings (y ¼ x ). This observation persisted in standing subjects for SBP and DBP and was also noted with both oscillometric measurements. Arm position and ABPM Adequate measurements were achieved in 13 of 15 subjects, nine of whom were male and their age ranged from 18 to 80 years (Table 1b). Nine patients were taking antihypertensive medications and the period of assessment varied between 2 and 4 days TIME Figure 4 Analysis of the mean hourly SBP, DBP and pulse rate, during the 12-h awake period. Values are expressed as mean 7 s.d: + Po0.05, ++ Po0.01, and +++ Po0.001; (K) horizontal arm; (&) usual arm position. Results were analysed over a 12-h period between and h when all subjects were awake and out of bed. During this period, the mean SBP/DBP was / mmhg with the arm horizontal and / mmhg with the arm in the usual position for ABPM. Both systolic and diastolic pressures were significantly different (Po0.01) when the two 12-hourly evaluation periods were compared (Table 2b). An analysis of the mean hourly SBP (Figure 4) also demonstrated a persistently lower SBP in patients when their arm was horizontal during the 12-h awake period, and these differences were also statistically significant at 09:00 h (Po0.01), h (Po0.01), h (Po0.05), h (Po0.05) and h (Po0.001). While each mean hourly DBP was also consistently lower over this 12-h period, these differences did not reach statistical significance. The pulse rate was

5 a Arm Position (mmhg) b Horizontal (mmhg) r 2 = P< r 2 = P < comparable during each period (Figure 4). A comparison of mean horizontal vs usual arm position over the 12-h awake period in each of the 13 subjects revealed a significant correlation for both SBP (r 2 ¼ ) and DBP (r 2 ¼ ) (Figure 5), the subsequent regression lines being significantly different from zero and unity (Po0.001). Thus, the higher the SBP or DBP, the greater the artefact because of arm position. There was no difference in night-time sleeping SBP or DBP between the two periods. Discussion usual arm position (mmhg) Figure 5 SBP (a) and DBP (b) during 12-h awake period: horizontal compared to usual arm postion. While this study confirms the significant increase in SBP and DBP by making the arm dependent noted in earlier studies, including recent ones using a Hawksley random-zero sphygmomanometer, 9,10 it also demonstrates that oscillometric devices including those used for ABPM are subject to similar artefacts. The absolute mean changes found during arm movement with a mercury device were similar to those found by Webster et al, 4 and Netea et al. 10 In sitting hypertensives, Netea et al 10 noted a mean increase of 10/11 mmhg in subjects from a mean SBP/DBP of 133/78 mmhg. While the mean SBP/DBP of our hypertensive subjects was higher (143/ 78 mmhg), using our calculations we would have predicted an increase of 18/10 mmhg in their patients. In the earlier data from Webster et al, 9 their sitting hypertensive group mean SBP/DBP was 140/90 mmhg, which increased by 18/14 mmhg with dependency. Our prediction of 20/12 mmhg is comparable. Netea et al 10 also measured the effect of arm position using an oscillometric device and demonstrated a similar but slightly lower effect of arm movement. We observed similar effects of arm position with mercury and both oscillometric devices. Despite the widespread belief that the arm should be supported and confirmatory results from Webster et al, 9 we were unable to demonstrate any significant differences in sitting or standing subjects. We suspect that the additional cardiac work required to hold the arm up in healthy people is relatively trivial; however, on balance it is reasonable to maintain this standard during BP measurement. The 13 mmhg SBP and 8 mmhg DBP mean difference between the two ABPM periods is also consistent with our clinic study with mercury and oscillometric devices. Again the higher the SBP or DBP, the greater the effect of arm position, potentially a significant problem in patient management. During this study, we could not ensure that all pressure measurements during the arm horizontal day were in fact horizontal, and while complete compliance was reported, BP differences may have been reduced by impaired compliance. While the two situations are not strictly comparable, it is of interest to compare the ABPM and clinic oscillometric device results. Applying the clinic regression analysis of dependent and horizontal pressures to the 13 ABPM subjects, a BP of 141/74 mmhg with the arm horizontal (mean 12-h BP) should increase to 161/84 mmhg with the arm dependent. This prediction is close to the measured mean of 154/ 82 mmhg. The regression analyses of SBP and DBP in horizontal vs usual arm position in the 13 ABPM subjects are also similar to the clinic OMRON findings. Minor differences between clinic and ABPM studies are probably because of a variation in the arm position during the usual ABPM monitoring period. Although the role of ABPM in the diagnosis and treatment of hypertension now appears well defined, individual reproducibility remains an important issue. ABPM is often repeated as part of the diagnostic or treatment process, and while ABPM reproducibility appears to be superior to clinic BP measurement, particularly in large groups, 19 a significant variation has been detected in some individuals. 20 For example, some studies have demonstrated that BP recorded in hypertensives and normotensives during a work day can be higher than during a nonworking day, 21,22 although a large individual variation was observed. Seasonal variations can also alter BP in normotensive and hypertensive individuals. 23 The ability of arm 393

6 394 position to alter ABPM significantly is another confounding factor in the interpretation of ABPM in individuals and this probably contributes to reported reproducibility problems, particularly when comparing work to nonwork days rather than the implied effects of work exertion and stress. The recent report by van der Steen et al 24 that body and arm positions can significantly influence sleeping ABPM supports this hypothesis. Guidelines for ABPM use recommend only that patients hold their arm still during pressure recordings, 25 and unless the subject is seated or lying, the arm will probably be dependent. The correlation between the level of BP and arm position has not been evaluated before and is an important clinical observation. While changes in arm position in subjects with low normal BP, particularly a normal SBP, would have few clinical implications, in subjects with borderline high pressures lowering the arm even if not to the dependent position could clearly produce a spurious diagnosis of hypertension. Some publications recommend correction factors for arm movement, with the American Heart Association suggesting a 0.8 mmhg correction for every 1 cm displacement of the sensing site with respect to heart level. 5 However, in a borderline or hypertensive population such corrections underestimate the potential error. It is believed that the difference in BP values produced by arm movement is largely accounted for by hydrostatic pressure 5 and this is supported by Netea et al, 11 although the differences demonstrated in regression analysis between SBP and DBP in our study cast doubt on such a simplistic view. While the British Hypertension Society recommends the method of differences as its preferred technique when comparing methods for measuring BP, 26 least-products regression analysis may be superior. 27 Nevertheless, ordinary least-squares regression analysis was used in this study despite its limitations because it is readily comprehensible to clinicians and, in particular, it clearly defines the magnification of the error produced by arm dependency as both SBP and DBP increases. Identifying fixed and proportional bias does not illuminate the important clinical observations of this study. This study confirms the significant effect of arm position on auscultatory BP. Such an effect occurs in the sitting and standing positions as well as with an oscillometric device. In addition, it is now clear that the higher the BP, the greater the error produced, particularly in the measurement of SBP. Given the current focus on the diagnosis and treatment of systolic hypertension, the potential for spurious readings should cause concern. The indirect measurement of BP is prone to numerous errors because of poor technique and even the presence of back support when sitting may influence BP. 28 Nevertheless, the marked effect of arm position on BP has been relatively ignored probably because the accepted heart level position is unclear and open to misinterpretation. Even a relatively small downward arm movement in a patient with borderline or high BP could significantly influence diagnosis and treatment. To avoid significant and probable widespread errors in current BP measurement, we commend all national and international organisations to clearly and unambiguously state that the arm must be supported in the horizontal position when BP is being measured, whether sitting or standing and whether auscultatory or oscillometric devices are being used. Furthermore, clinical research manuscripts should state arm position to allow comparison, although in the recently published comparison between BP and mortality from the seven countries study research group, 29 BP was measured in the supine position, which possibly serendipitously minimised arm position artefacts between several centres. The application of such an unambiguous standard using ABPM is, however, less clear. Many community studies have facilitated the development of useful clinical criteria for hypertension diagnosis and assessment and to discount this information would be difficult, particularly since it is probably convenient for individuals to leave their arm at their side during cuff inflation. However, daytime ABPM is usually much lower than clinic pressures even if it is still artificially high owing to the dependent arm. On the other hand, night-time dipping is in part due to a daytime dependent arm. References 1 Feher M, Harris-StJohn K, Lant A. Blood pressure measurement by junior doctorsfa gap in medical education? Health Trends 1992; 24: Carney S et al. Hospital blood pressure measurement: staff and device assessment. J Qual Clin Pract 1999; 19: Moore KL. Clinically Oriented Anatomy, 3rd edn. Williams & Wilkins: Baltimore, MD, WHO. Guidelines for the management of mild hypertension. Memorandum from a WHO/ISH meeting. Hypertension 1993; 22: American Heart Association. Recommendations for human blood pressure determination by sphygmomanometers: report of a special taskforce appointed by the Steering committee. Hypertension 1988; 11: 210A 222A. 6 Petrie JC, O Brien ET, Littler WA, de Swiet M. British Hypertension Society Recommendations on blood pressure measurement. BMJ 1986; 293: Guyton A. Textbook of Medical Physiology. WB Saunders Company: Philadelphia; 1986, p Hill L. Measurement of systolic blood pressure in men. Heart 1909; 1: Webster J, Newnham D, Petrie JC, Lovell HG. Influence of arm position on measurement of blood pressure. BMJ 1984; 288: Netea RT, Lenders JWM, Smits P, Thien T. Arm position is important for blood pressure measurements. J Hum Hypertens 1999; 13:

7 11 Netea RT et al. Influence of arm position on intraarterial blood pressure measurement. J Hum Hypertens 1999; 13: Parker D et al. A comparison of the random zero and standard mercury sphygmomanometer. Hypertension 1988; 11: O Brien E, Mee F, Atkins N, O Mally K. Inaccuracy of the Hawksley random zero sphygmomanometer. Lancet 1990; 336: Mann S. Inaccuracy of electronic sphygmomanometers. Clin Exp Pharmacol Physiol 1992; 19: O Brien E, Mee F, Atkins N, Thomas M. Evaluation of three devices for self-measurement of blood pressure according to the revised British Hypertension Society Protocol: the Omron HEM-705CP, Phillips HP5332, and Nissei DS-175. Blood Press Monitor 1996; Fotherby MC, Potter JF. Reproducibility of ambulatory and clinic blood pressure measurements in elderly hypertensive subjects. J Hypertens 1993; 11: Mansoor GA, McCabe EJ, White WB. Long-term reproducibility of ambulatory blood pressure. J Hypertens 1994; 12: Palatini P, Mormino P, Caneli C. Factors affecting ambulatory blood pressure reproducibility Results of the HARVEST trial. Hypertension 1994; 23: Engfeldt P et al. 24-Hour ambulatory blood pressure monitoring in elderly normotensive individuals and its reproducibility after one year. J Hum Hypertens 1994; 8: Prisant LM et al. Repeatability of automated ambulatory blood pressure measurements. J Fam Pract 1992; 11: Peiper C, Warren K, Pickering TG. A comparison of ambulatory blood pressure and heart rate at home and work on work day and non-work days. J Hypertens 1993; 11: Enstrom I, Pennert K. Does it matter whether ambulatory blood pressure is recorded duirng a work day or non-work day? J Hypertens 1996; 14: Sega R et al. Seasonal variations in home and ambulatory blood pressure in the PAMELA population. J Hypertens 1998; 16: Van der Steen MS, Pleijers AMLJ, Lenders JCUM, Thein T. Influence of different supine body positions on blood pressure: consequences for night blood pressure/dipper-status. J Hypertens 0; 18: Prasad N, Isles C. Ambulatory blood pressure monitoring: a guide for general practitioners. BMJ 1996; 313: O Brien E et al. The British Hypertension Society Protocol for the evaluation of automated and semiautomated blood pressure measuring devices with special reference to ambulatory systems. J Hypertens 1990; 8: Ludbrook J. Comparing methods of measurement. Clin Exp Pharmacol Physiol 1997; 24: Cushman WC, Cooper KM, Horne RA, Meydrech EF. Effect of back support and stethoscope on seated blood pressure determinations. Am J Hypertens 1990; 3: Van den Hoogen PCW et al. The relation between blood pressure and mortality due to coronary heart disease among men in different parts of the world. N Engl J Med 0; 342: