Ambulatory blood pressure monitoring: which arm? - Nature

Journal of Human Hypertension (2000) 14, 227–230  2000 Macmillan Publishers Ltd All rights reserved 0950-9240/00 $15.00 www.nature.com/jhh

ORIGINAL ARTICLE

Ambulatory blood pressure monitoring: which arm? JC O’Shea and MB Murphy Department of Pharmacology and Therapeutics, University College and Mercy Hospital, Cork, Ireland

To determine the effects of routinely selecting the nondominant arm for ambulatory blood pressure monitoring (ABPM) on estimates of patients’ blood pressure (BP) and to evaluate the practise of using manual BP from one arm and ambulatory BP from the other on the estimation of white coat effect (WCE), an observational study was conducted in 10 volunteers, exhibiting an interarm resting clinic systolic BP (SBP) difference ⭓10 mm Hg. The main outcome measures were: (i) average ambulatory SBP measured on right and left arm simultaneously during 24 h, and (ii) estimate of WCE derived, by current practise, as the difference between the referral clinic BP (the higher of the manual readings from both arms) and ambulatory non-dominant arm BP, contrasted with the WCE calculated as the difference

between clinic and ambulatory readings from the same arm (the arm with the higher manual readings). The supine referral clinic SBP was 16 ⴞ 6 mm Hg higher in the right compared with the left arm. Average 24 h ambulatory SBP was 6 ⴞ 7 mm Hg higher in the right arm (range ⴙ17 to ⴚ3 mm Hg), P ⴝ 0.025. Diastolic BP measurements mirrored the systolic findings. One-third of the WCE, estimated by current practise, could be attributed to inconsistency in the choice of arm for BP measurement. Thus, inconsistency in the selection of arms for BP measurement, by different techniques, may confound estimation of patients’ cardiovascular morbidity risk. Journal of Human Hypertension (2000) 14, 227–230

Keywords: ambulatory blood pressure monitoring; right arm; left arm; white coat effect

Introduction Hypertension is an important risk factor for premature cardiovascular morbidity and mortality with a prevalence of about 20% in the adult population.1 Accurate and reproducible measurement of blood pressure (BP) is essential and, to this end, extensive guidelines have been published on the appropriate use of the sphygmomanometer.2,3 At the initial patient encounter, measurements must be made on both arms. The diagnosis must be established based on readings from the arm with the higher pressure, and the response to therapy monitored long-term using recordings from that arm.2,3 Up to threequarters of individuals exhibit a difference in BP between right and left arms exceeding 10 mm Hg while seated or supine. Pressure tends to be higher on the right.4–10 Non-invasive ambulatory BP monitoring (ABPM) is being used increasingly in the diagnosis and management of hypertension. Apart from the paucity of prospective studies quantifying the prognostic implications of different levels of ABP, with consequent uncertainty regarding the point at which drug treatment should be applied, several methodological questions remain unanswered. This study addresses the single question: does it matter which arm is monitored? Correspondence: Prof MB Murphy, Department of Clinical Pharmacology, Mercy Hospital, Grenville Place, Cork, Ireland Received 18 October 1999; revised and accepted 29 December 1999

ABPM requires the continuous wearing of a cuff, which may hinder the normal use of the arm. Conversely, arm movement may diminish the accuracy of recordings. Consequently, the practise of placing the cuff on the non-dominant arm (the left in 90% of cases) to minimise interference with daily activity and enhance the accuracy of the recording, has evolved for convenience. This study tested the hypothesis that the pragmatic selection of the nondominant arm in ABPM introduces a significant bias in estimating patients’ blood pressure. To this end, ABP was measured simultaneously in both left and right arms during 24 h in a cohort of patients. Two new concepts were popularised by ABPM: ‘white coat hypertension’ and ‘white coat effect’. White coat hypertension has been defined as a persistently elevated clinic BP with a normal BP at other times away from the medical setting.11 Many authorities eschew treatment of these patients. The white coat effect, defined as a transient elevation of BP occurring in the medical setting12 has, in practice, been quantified as the difference between office and ambulatory BP values. Clinic BP, however, is measured predominantly on the right arm while, as we have indicated, ABPM is conducted on the left. The extent to which this methodology influences the estimate of the white coat effect was also assessed in these subjects.

Materials and methods The records of 39 consecutive right handed patients who attended our unit for ABPM between November

ABP monitoring: which arm? JC O’Shea and MB Murphy

228

1995 and January 1996 were reviewed. Thirteen exhibited a systolic BP (SBP) ⭓10 mm Hg higher in the right arm based on ‘casual’ BP readings made by the same observer consecutively measuring BP in one arm and then the other at the initial screening. These subjects were invited to undergo simultaneous right and left arm 24-h ABP monitoring. Of the 10 subjects who subsequently volunteered for the study, none were on cardiovascular drugs at the time of the study and none had known arterial disease. Immediately prior to monitoring, BP was measured by a nurse previously trained and assessed according to the British Hypertension Society (BHS) protocol using the BHS video film ‘Blood Pressure Measurement’.13 BP was recorded three times on each arm in random order, such that the right arm was measured first in six cases. Left and right arm circumferences were measured and a large cuff was used when the circumference exceeded 31 cm. ABPM was undertaken using the Tycos QuietTrak, a device that has satisfied the Association for the Advancement of Medical Instrumentation (AAMI) and the BHS accuracy standards.14,15 Monitors placed on both arms were programmed to measure BP simultaneously at 20-min intervals over the 24-h period. Since the quality of ABP recording is influenced by arm movement, the extent of divergence in movement was determined in each subject using electronic activity monitors (Gaewihler Electronics, Hombrechtikon, Switzerland) on right and left wrists during the study period. The battery-powered unit, slightly larger than a wristwatch, is a monaxial piezoelectric accelerometer, programmed to record activity every 10 sec. The monitor integrates motor activity (threshold 0.1 g) over a defined time period, and converts it into an activity value ranging from 0 to a maximum of 253 units. We have previously demonstrated that the average level of daytime physical activity, measured by this device, is a significant determinant of diurnal BP variability.16 Summary data are reported as mean ± standard deviation (s.d.). Comparisons were made between the data from the different arms by paired t-tests, with P values less than 0.05 deemed statistically significant. Statistical analysis was performed using the Astute Statistical Software programme (DDU software, Leeds University, UK).

Results Ten patients, six male, with an average age of 45 ± 17 years took part in the study. The average right and left arm circumferences were identical, 30 ± 3 cm; the large BP cuff was used on both arms in two of the subjects. SBP at clinic, measured by mercury sphygmomanometer, was 170 ± 22 mm Hg and 154 ± 18 mm Hg; P = 0.001 on right and left arms respectively (this order applies to all the data hereafter). The average of three SBP measurements immediately prior to ABP monitoring was 160 ± 24 mm Hg vs 152 ± 20 mm Hg; P = 0.025 (Table 1). The quality of ambulatory recordings was similar Journal of Human Hypertension

in both arms. The monitors’ analytical programme deemed 91 ± 5% of readings on the right and 92 ± 8% of those on the left to be acceptable. Analysis of the timing of corresponding right and left BP measurements showed that all were made within 1 min of each other. The average 24-h SBP was 6 ± 7 mm Hg higher in the right arm (range + 17 to −3 mm Hg), P = 0.025. Significant interarm differences in SBP were seen during daytime (6 ± 8 mm Hg, P = 0.04) and night-time (4 ± 4 mm Hg, P = 0.02) recording periods (Table 2). As a consequence of this observation, the white coat effect was clearly influenced by the methodology. The apparent white coat effect (estimated as clinic right arm SBP less the average daytime ambulatory SBP) was significantly higher when the left arm ambulatory recording was used contrasted with the right arm ambulatory data, 15 ± 10 mm Hg vs 11 ± 12 mm Hg respectively; P = 0.03 (Figure 1). Thus, 33 ± 52% (range 157% to −12%) of the difference between the manual and ABP recording resulted from the use of different arms to make the estimate. While subjects were selected on the basis of SBP differences only, diastolic BP (DBP) measurements mirrored the systolic findings. Screening supine clinic DBP was 102 ± 11 mm Hg vs 96 ± 11 mm Hg; P = 0.025 with an average DBP prior to ABPM of 98 ± 12 mm Hg vs 94 ± 14 mm Hg (right vs left arm); P = 0.05. Average 24-h DBP was 84 ± 12 mm Hg vs 81 ± 11 mm Hg; P = 0.30 (Table 1). Corresponding right and left arm daytime activity levels were 36 ± 9 vs 32 ± 11 (units), P = 0.052. Night-time activity scores were identical, 1.5 ± 1 (Table 1). These activity scores may be contrasted with those of a series of 120 routine clinic attenders studied during ABPM where the average daytime dominant arm activity score was 43 ± 13.15

Discussion In the absence of coarctation of the aorta or atheromatous disease, inter-arm BP difference is predominantly a physiological phenomenon. Arterial pressure, as measured by sphygmomanometry, is a vectorial summation of pressure waves, which are transmitted forward from the ascending aorta, and reflected back from the peripheral arterioles. The tubular, distributed nature of the arterial system leads to differences in absolute pressure along the arterial tree at the same point in time.17 These differences are exaggerated in hypertension and the elderly when arterial stiffening results in more rapid propagation of the pulse wave along the major arteries and the consequent early return of wave reflection from the periphery.18 Conventionally, where differences in inter-arm BP are detected by sphygmomanometer, therapeutic decisions are based on the higher arm BP, since the findings of the major outcome trials in hypertension have been based on the higher arm BP. In contrast, ABP monitoring is conducted uniformly on the non-dominant arm. In this study, a small group of subjects, selected consecutively from a clinic population on the basis


229

Table 1 Average right and left arm clinic and ambulatory blood pressure levels and activity scores from 10 subjects n = 10

Right arm

Screening SBP (mm Hg) Average pre-ABPM SBP (mm Hg) Average 24-h SBP (mm Hg) Average daytime SBP (mm Hg) Average sleep SBP (mm Hg) Daytime activity score Night-time activity score Screening DBP (mm Hg) Average pre-ABPM DBP (mm Hg) Average 24-h DBP (mm Hg) Average daytime DBP (mm Hg) Average sleep DBP (mm Hg)

170 160 145 149 135 36 1.5 102 98 84 87 78

± ± ± ± ± ± ± ± ± ± ± ±

Left arm

22 24 21 21 22 9 1 11 12 12 11 15

154 152 139 143 131 32 1.5 96 94 81 84 76

± ± ± ± ± ± ± ± ± ± ± ±

18 20 17 17 21 11 1 11 14 11 11 13

Diff.

P value

± ± ± ± ± ± 0 ± ± ± ± ±

0.0003 0.025 0.025 0.04 0.02 0.052 0.5 0.025 0.05 0.3 0.4 0.4

16 8 6 6 4 4 6 4 3 2 2

9 7 7 8 4 5 7 7 8 8 9

Diff., Difference between right and left arm; SBP, systolic blood pressure; DBP, diastolic blood pressure.

Table 2 Twenty-four hour average SBP (mm Hg) recorded simultaneously in right and left arms in 10 subjects Subject

JS MD RiD BG RaD MT CD MB AD UN

Right arm

Left arm

Interarm difference

167 160 187 136 122 152 131 136 130 128

150 144 178 129 116 148 128 134 133 131

+17 +16 +9 +7 +6 +4 +3 +2 −3 −3

SBP, systolic blood pressure.

Figure 1 Estimation of white coat effect: comparison of manual right arm SBP less average daytime right arm ambulatory SBP vs manual right arm SBP less average daytime left arm ambulatory SBP. • P = 0.03.

of a higher right arm SBP of ⭓10 mm Hg, underwent bilateral 24-h ABPM. Great care was taken to insure that measurements were made simultaneously in both arms to eliminate differences that might arise from the short-term effects of bursts of physical activity. Both daytime and night-time SBP were

found to be significantly higher in the right arm, confirming that differences in right and left arm blood pressures detected in the clinic persist during daily activity and are detectable by ABPM. Reduced vascular compliance during the active waking state, resulting in more rapid reflection of pulse waves from the periphery, may explain the greater interarm SBP difference seen during daytime hours. What is the clinical significance of these findings? In essence the cardiovascular risk of patients could be underestimated. The risk for heart attack or stroke increases continuously with rising SBP. Data from prospective observational studies indicate that sustained elevation in SBP of 9 mm Hg confers about a one-third higher stroke risk.19 Extrapolating from this, the estimated stroke risk in the 10 subjects in this study could be 20% higher by selecting the results of the right arm ABPM rather than the left. Additionally, more patients will be labelled as having white coat hypertension by current clinical practice than would be the case if choice of arm for BP measurement were consistent for both methodologies. The nature and biological implications of the white coat effect are controversial. This study has illustrated a flaw in its estimation, consequent on methodological inconsistencies. The use of the nondominant arm in ABP recording results in a significantly higher estimate of the white coat effect than would be obtained using the ipsilateral (right arm) ABP measurement, given that clinic BP is usually recorded on the right arm. We have alluded earlier to the rationale for the current practice of non-dominant arm ABP monitoring, ie, ABP interfering with physical activity during the recording and activity confounding accurate ABP recording. However, our data do not support these concerns. Activity levels of participants were within 1 s.d. of the average activity of our clinic population15 despite the fact that these subjects were wearing two ABPM units simultaneously. Furthermore, rejection rates were not significantly different between right and left arm recordings, while activity levels of both were almost identical. This project was solely designed to demonstrate the potential of our clinical practise for confounding estimation of a patient’s BP and consequent risk of Journal of Human Hypertension


230

premature cardiovascular morbidity. Establishing the prevalence and extent of the problem will require further study.

References 1 Kaplan N. Hypertension: prevalence, risks and effect of therapy. Ann Intern Med 1983; 98: 705–709. 2 Kirkendall W, Burton A, Epstein F, Freis E. Recommendations for human blood pressure determination by sphygmomanometers. Circulation 1967; 36: 980–988. 3 O’Brien E, O’Malley K. ABC of blood pressure measurement. Br Med J 1979; 2: 775–776. 4 Shock N, Ogden E. The difference between blood pressure measurements obtained simultaneously on the two arms. Q J Exp Physiol 1940; 30: 155–162. 5 Korns H, Guinand P. Inequality of blood pressure in the brachial arteries, with especial reference to disease of the arch of the aorta. J Clin Invest 1933; 12: 143–154. 6 Amsterdam B, Amsterdam A. Disparity in blood pressures in both arms in normals and hypertensives and its clinical significance: a study of 1000 normals and 272 hypertensives. N Y State J Med 1943; 43: 2294 – 2300. 7 Swallow R. Hypertension: which arm? Br Med J 1975; 3: 370. 8 Kay W, Gardner K. Comparative blood pressures in the two arms. Calif Med 1930; 33: 578–579. 9 Rueger M. Blood pressure variations in the two arms. Ann Intern Med 1951; 35: 1023–1027.

Journal of Human Hypertension

10 Kristensen B, Kornerup H. Which arm to measure the blood pressure? Acta Med Scand 1982; 670 (Suppl): 69–73. 11 Pickering T. White coat hypertension. In: Pickering TG (ed). Ambulatory Monitoring and Blood Pressure Variability. Science Press: London, 1991. 12 Gosse P et al. Clinical significance of white-coat hypertension. J Hypertens 1994; 12 (Suppl): S43–S47. 13 Jamieson M et al. Blood Pressure Measurement. Video for the British Hypertension Society. British Medical Journal Publications: London, 1980. 14 White W et al. Multicentre assessment of the QuietTrak ambulatory blood pressure recorder according to the 1992 AAMI guidelines. Am J Hypertens 1994; 7: 509–514. 15 O’Shea J, Murphy M. Factors confounding assessment of ambulatory blood pressure monitors, studied during formal evaluation of the Tycos QuietTrak. Am J Hypertens 1997; 10: 175–180. 16 Gretler D, Carlson G, Montano A, Murphy M. Diurnal blood pressure variability and physical activity measured electronically and by diary. Am J Hypertens 1993; 6: 127–133. 17 O’Rourke M. Second workshop on structure and function of large arteries: Part 1. Mechanical Principles in Arterial Disease. Hypertension 1995; 26: 2–9. 18 O’Rourke M. Arterial stiffness, systolic blood pressure and logical treatment of hypertension. Hypertension 1990; 15: 339– 347. 19 MacMahon S, Rodgers A. Blood pressure, antihypertensive treatment and stroke risk. J Hypertens 1994; 12 (Suppl): S5–S14.