Prevalence and clinical implications of the inter-arm blood ... - Nature

Journal of Human Hypertension (2006) 20, 923–931 & 2006 Nature Publishing Group All rights reserved 0950-9240/06 $30.00 www.nature.com/jhh

ORIGINAL ARTICLE

Prevalence and clinical implications of the inter-arm blood pressure difference: a systematic review CE Clark1, JL Campbell1, PH Evans1 and A Millward2 1

Primary Care Research Group, Institute of Health & Social Care Research, Peninsula Medical School, Exeter, Devon, UK and 2Institute of Biomedical and Clinical Science, Peninsula Medical School, Derriford, Plymouth, Devon, UK

A blood pressure (BP) difference between arms was first reported over 100 years ago. Knowledge of its prevalence and relevance to the accurate measurement of BP remains poor. Current hypertension guidelines do not emphasise it. The objectives of this study were to establish the best estimate of prevalence of the interarm difference (IAD) in the population, to consider its implications for accurate BP measurement and treatment, and to discuss its aetiology and potential as a risk marker for cardiovascular disease. Systematic literature review was carried out. The data sources were Medline EMBASE and CINAHL databases, and Index of Theses. Studies reporting prevalence rates of IAD were retrieved and considered for inclusion against explicit methodological criteria. Point prevalence rates were extracted and weighted mean prevalence rates calculated. The main outcome measures were weighted mean preva-

lences of systolic IADX10 and X20 mm Hg and of diastolic IADX10 mm Hg. Thirty-one studies were identified. Most had methodological weaknesses; only four met the inclusion criteria. Pooled prevalences of the IAD from these four studies were 19.6% systolic X10 mm Hg (95% CI 18.0–21.3%), 4.2% systolic X20 mm Hg (95% CI 3.4–5.1%) and 8.1% diastolic X10 mm Hg (95%CI 6.9– 9.2%). In conclusion, an IAD is present in a substantial number of patients and should be looked for whenever diagnosis and treatment depend on accurate measurements of BP. The importance of an IAD should be better emphasised in current hypertension management guidelines. There is evidence associating an IAD with peripheral vascular disease, raising the possibility that its presence may predict cardiovascular events. Journal of Human Hypertension (2006) 20, 923–931. doi:10.1038/sj.jhh.1002093; published online 12 October 2006

Keywords: blood pressure measurement; interarm difference

Introduction Patients are regularly encountered in the clinical setting with a different systolic or diastolic blood pressure (BP) in each arm. Comparisons of arm pressures have been made ever since the modern sphygmomanometer was introduced,1 yet the significance of this finding, which Cyriax named the ‘differential BP sign’,2 is still poorly appreciated today. Hypertension is one of the major causes of premature morbidity and mortality throughout the developed and developing worlds.3 The vast majority of patients with hypertension are managed in primary care4 and measurement of BP is the most common investigation performed in this setting.5 Correspondence: Dr CE Clark, Primary Care Research Group, Institute of Health & Social Care Research, Peninsula Medical School, Smeall Building, St Luke’s Campus, Magdalen Rd, Exeter, Devon EX1 2LU, UK. E-mail: [email protected] Received 10 May 2006; revised 21 August 2006; accepted 25 August 2006; published online 12 October 2006

The management of hypertension is known to be suboptimal;6 therefore, knowledge of the latest guidelines, and the many factors known to influence readings, is crucial.7 The prevalence of a systolic (sIAD) or a diastolic (dIAD) inter-arm difference, its importance for the management of hypertension, its significance as a potential marker of peripheral vascular disease (PVD) and predictor of cardiovascular disease (CVD) in hypertensive patients are the subjects of this review. Although recent ambulatory BP monitoring guidelines recommend routine bilateral assessment,8 the latest British guidelines for the management of hypertension merely state that ‘BP should initially be measured in both arms as patients may have large differences (410 mm Hg) between arms. The arm with the higher values should be used for subsequent measurements’.9 This appears to be guidance repeated almost verbatim through previous versions10–14 stretching back over 60 years.15 The guidelines do not consider the prevalence of the IAD and suggest that differences X20 mm Hg systolic and/or 10 mm Hg diastolic warrant specialist

The inter-arm blood pressure difference CE Clark et al 924

referral.16 As only one in three measurements of BP in practice are made according to guidelines,17,18 it is unlikely that even this passing reference to an IAD is acted upon. The latest Joint National Committee guidelines (JNC VII) merely suggest BP ‘verification in the contralateral arm’.19 Appreciation of the presence and implications of an IAD is clearly vital for accurate diagnosis and consistent management of hypertension. Following an earlier small study of a mixed normotensive and hypertensive cohort of primary care patients, we proposed that the IAD may be caused by PVD.20 As PVD is itself a strong predictor of CVD,21–26 the recognition of the IAD as a sign of PVD would suggest that the IAD may also have prognostic value as a marker for predicting cardiovascular events in a similar manner to a reduced ankle-brachial pressure index (ABPI). ABPI measurement has been proposed as a screening test,27 but is not routinely undertaken in the general practice assessment of hypertensive patients in the United Kingdom. It is time consuming and requires specialised equipment and a degree of training, whereas bilateral brachial BP measurements are easily taken and are in any case recommended in the assessment of new hypertensive patients.9 In symptomatic chronic upper limb ischaemia, where the IAD is marked,28 the predominant cause is atherosclerosis (Table 1), and this is associated with PVD.29 It therefore seems logical that the causes of an asymptomatic IAD should be similar. Consequently, the recognition of an IAD could have value in providing a simple screening test to apply in routine consultations to identify patients potentially at the highest risk of cardiovascular events and therefore in need of further assessment. This would facilitate more appropriate targeting of limited resources in primary prevention, as required by the National Service Framework for Coronary Heart Disease (CHD).30

Methods

Table 1 Causes of chronic upper limb ischaemia

Results

K K K K

K K

Atherosclerosis Vasculitis e.g. polyarteritis nodosa, giant cell arteritis, Buerger’s disease Fibromuscular hyperplasia Connective tissue disorders e.g. scleroderma, mixed connective tissue disease, systemic lupus Radiation arteritis Thoracic outlet compression

Adapted with permission from Thompson and Kinsella.28

Identification of studies

The principal investigator searched Medline (1966– 2006), Old Medline (1950–1965), EMBASE (1974– 2006) and CINAHL (1982–2006) databases in May 2006 using the following MeSH, and full text terms: ‘blood pressure (restricted to analysis, instrumentation, classification or physiology) or hypertension or arm’ and ‘difference or differential’. The result set was limited to human and English language results. Further references were identified from specialist cardiovascular and hypertension journal collections. The Index of Theses (1970–2006) was searched and authors recently active in the field were contacted. Further citations were identified by hand searching of reference lists in retrieved papers. Only English language papers were reviewed, but some findings from foreign language papers quoted in a previous review31 have been cited in the Discussion. Selection and quality assessment of studies

Cross-sectional or cohort studies reporting a prevalence figure for sIAD or dIAD were retrieved for assessment. Reports meeting the predefined quality criteria based on sample selection, sample size and method of measurement of the IAD (Box 1) were included in the analysis. Data collection and synthesis

The principal investigator extracted data on prevalence rates, study population and methods of measurement from all studies reporting prevalences of the IAD. Where possible data were combined to derive pooled estimates of prevalences and magnitudes of sIAD and dIAD using weighted means and standard errors.32

Searches identified 357 citations, from which 18 articles were retrieved. Hand searching of references identified a further 19 older papers. After exclusion of papers without prevalence data, only four papers met the inclusion criteria (Figure 1). Description of studies

All potential studies for inclusion were crosssectional studies reporting point prevalences of

Box 1 Inclusion criteria for studies Sample size Selection of subjects

n4100 Defined primary or secondary care population not selected individuals

Method of measurement

Simultaneous using automated sphygmomanometers Crossover design, repeated measurements Controlled setting

Journal of Human Hypertension


Citations retrieved in search (n=357) Citations excluded from search data (n=341) Full text papers retrieved for potential inclusion (n=18) Additional papers identified from references (n=19) Full text papers considered for inclusion (n=37) Full text papers excluded: no prevalence data (n=6) Full text papers assessed against inclusion criteria (n=31) Papers not meeting inclusion criteria (n=27) Papers meeting inclusion criteria (n=4) Figure 1 QUORUM flow diagram (adapted from Moher et al 199980).

IADs. In total, 31 English language series published between 1915 and 2005 were assessed against the inclusion criteria. Many studies exhibited selection bias in reporting selective or opportunistic samples, or had unclear selection criteria,2,31,33–40 and older studies tended to report highly selected series.34,36,41–45 Others suffered from bias in the selection of results for analysis37,46 or reported small sample sizes (defined as o100 subjects).2,34,34–36,41,43,45–50 Studies reported one or more pairs of BP readings measured either simultaneously or sequentially. Simultaneous techniques used either two observers37,46 or a pair of automated sphygmomanometers41,51,52 to record BP in both arms at the same time. Sequential methods involved measurement of BP in one arm and then the other.2,20,29,33,35,36,38,39,42–44,49,53–59 Excluded studies, with reasons for exclusion, are summarised in Appendix A. Reported prevalences

Only four studies met the inclusion criteria37,51,52,60 (Table 2). The mean prevalences of sIAD in these studies were 19.6%X10 mm Hg (95% CI 18.0– 21.3%) and 4.2%X20 mm Hg (95% CI 3.4–5.1%) and for dIAD 8.1%X10 mm Hg (95% CI 6.9–9.2%). The mean prevalences for excluded studies were significantly higher (Table 3). One study presented data for normotensive and hypertensive subjects,60 the prevalences of sIADX20 mm Hg and

Table 2 Characteristics of studies included in the review Study (country of origin)

Study population

Method of measurement

Sample size

Prevalence of systolic differences

Prevalence of diastolic differences

Amsterdam and Amsterdam (1943)60 (USA)

Normotensives

Repeated averaged simultaneous measurements

1000

31.2%X10 mm Hg 8.4%420 mm Hg

9.3%410 mm Hg

272

51%420 mm Hg

27.2%410 mm Hg

447

26.6%X10 mm Hg 7.2%X20 mm Hg 5.3%X10 mm Hg 0.1%X20 mm Hg 10%X10 mm Hg indirect 6%X10 mm Hg direct (difference NS)

15%X10 mm Hg

Hypertensives Probably hospital, not stated Harrison et al. (1960)37 Patients attending (USA) clinic Selected volunteers with differences on simultaneous indirect measurement Orme et al. (1999)52 (UK)

Staff, visitors and day case surgery patients at one DGH No CVD CVD

Lane et al. (2002)51 (UK)

Staff and patients in one general hospital

Single sequential pair of measurements Simultaneous three pairs of measurementsa Simultaneous direct intra-arterial

Automated simultaneous 2 2 crossover ¼ 4 pairs readings averaged

447 53

4%X10 mm Hg 8%X10 mm Hg indirect 1%X10 mm Hg direct (difference NS)

462

8.2%X10 mm Hg 0X20 mm Hg

6.5%X10 mm Hg 0.4%X20 mm Hg

364

2.7%X10 mm Hg 0X20 mm Hg 19.4%X10 mm Hg 0X20 mm Hg

6.3%X10 mm 0.3%X20 mm 7.1%X10 mm 1.0%X20 mm

20%410 mm Hg 3.5%420 mm Hg

11.3%410 mm Hg 3.8%420 mm Hg

98 Simultaneous automated, 400 2 pairs of measurements

Hg Hg Hg Hg

Abbreviation: CVD, cardiovascular disease. a Only this data set included in review. Journal of Human Hypertension


Table 3 Comparison of weighted mean prevalences of inter-arm differences for studies included and excluded Inter-arm difference

sIADX10 mm Hg sIADX20 mm Hg dIADX10 mm Hg

Included studies

Excluded studies

Mean (%)

95% CI

Mean (%)

95% CI

19.6 4.3 8.1

18.0–22.3% 3.4–5.1% 7.0–9.2%

34.5 7.5 25.3

32.0–37.1% 4.9–10.0% 22.8–27.9%

Significance

Po0.001 Po0.001 Po0.001

Abbreviations: CI, confidence interval; dIAD, diastolic inter-arm difference; sIAD, systolic inter-arm difference.

dIADX10 mm Hg were significantly higher in the presence of hypertension (sIADX20 mm Hg 8.4 vs 51%; OR 11.40, 95% CI 8.12–16.01; Po0.0001 and dIADX10 mm Hg 9.3 vs 27.2%; OR 3.64, 95% CI 2.55–5.20; Po0.0001). One study reported data for subjects with or without CVD,52 and showed a significantly higher rate of sIADX10 mm Hg in patients with CVD (19.4% with CVD vs 2.7% without CVD; OR 4.34, 95% CI 2.09–9.04; Po0.0001).

Discussion The prevalence rates presented demonstrate that the IAD is a common finding in clinical practice whenever it is looked for. The majority of reports and all included studies was from selected or secondary care populations, the true general population or primary care prevalence is unknown. Studies using sequential or single pair measurement techniques appear to overestimate prevalence compared with studies meeting our inclusion criteria. Strengths and weaknesses of the review

Published data on the prevalence of an IAD stretch back to 1900.61 Thorough hand searching appears to have yielded as full a set of papers as possible, but as only English language papers were retrieved some data may have been excluded. No foreign language citations were identified after 1929;62 therefore, it seems unlikely that our findings could have been substantially altered by their inclusion. Data extraction was only carried out by one investigator but was not a complex task, and the inclusion criteria were clear and objective. The choices of inclusion criteria were our own but based on a published critique32 of the methodology of one of the studies.46 Implications for BP measurement in practice

Failure to recognise the IAD and to standardise readings to the higher arm runs the risks of inadequate treatment of hypertensive patients, false diagnosis of hypotension due to over treatment (pseudohypotension36), a delay in the diagnosis of hypertension,63 or physician confusion creating the potential for ‘clinical inertia’.64 The Health Survey of England states that ‘if systolic BP were lowered by X10 mm Hg among the 71.5% of hypertensive subjects uncontrolled (in the survey), we estimate Journal of Human Hypertension

that 44000 fatal and nonfatal coronary events and 46000 fatal and nonfatal strokes could be prevented each year in England.’6 Given our estimate of the prevalence of a sIADX10 mm Hg of 19.6%, there is a one in 10 risk that a BP measurement could be underestimated by this amount if the IAD has not been excluded. We suggest that it is vital to undertake measurements of BP in a consistent manner in both arms for all subjects, to record the findings accurately, and to record measurements taken from the ‘higher arm’ to guide subsequent assessments. This approach should be more clearly emphasised in future hypertension guidelines, and given the prevalence figures that we have presented, the advice to refer subjects with a sIAD X20 mm Hg or dIAD X10 mm Hg for specialist assessment is seen to be impractical and should be revised. Pathophysiology of the IAD

Although the ability of arteriosclerosis to affect BP has long been recognised65 and the presence of an IAD is used to help diagnose symptomatic upper limb ischaemia,28 in health it has historically been dismissed as a normal variant. Anatomical explanations have been proposed31,39,49,53,60,66 but there is now evidence for an association with PVD, suggesting a pathological rather than a physiological aetiology. Initially the IAD was described as a sign of aortic aneurysm67 or vascular disease,68,69 and the earliest report dismissed its magnitude in health as negligible, with differences of 10 mm Hg being unusual.61 The earliest series of 36 patients reported a pressure difference of 10 mm Hg or more in 20%, leading to the conclusion that ‘the difference between the right and left brachial arteries is by no means diagnostic of aneurysm’.70 Another early series of 100 healthy subjects found only 10% to have equal BPs71 and the argument against arteriosclerosis as a cause ever since has been that these differences are too frequent,37 and therefore represent a spectrum of normality. Consequently, anatomical explanations were advanced, which considered the angulation and anatomy of the aortic arch and its branches.31,39,49,53,60,66 The high prevalence of an IAD in war victims with unilateral injuries seemed to support this explanation.35 But normal anatomy alone cannot explain the higher left arm pressures noted.36,51 It is our contention that the IAD may be due


to occult PVD59 and one recent study of subclavian artery stenosis has shown such an association.29 If indeed such an observation was confirmed, one would predict that factors associated with PVD should have a higher prevalence in patients with an IAD. CHD is one common condition strongly associated with PVD, as defined by a reduced ABPI.21–25,72 No study has set out to compare the prevalence of an IAD in patients with or without CHD; however, data extracted from one study meeting our inclusion criteria showed a significantly higher prevalence of sIADX10 mm Hg in patients with CVD than without (19.4 vs 2.7%; OR 4.34, 95% CI 2.09–9.04; Po0.0001).52 Arteriosclerosis was also more common in subjects with an IAD in Kay and Gardner’s study,38 whereas although Singer and Hollander found no relationship between coronary risk factors and an IAD, they did find the mean sIAD to be significantly higher in patients with known coronary artery disease than without (14.5 vs 10.4 mm Hg; P ¼ 0.05).55 Another smaller study reported higher prevalence of IAD in patients with PVD compared to both CHD patients and controls.36 Thus, there is a body of indirect evidence to support our hypothesis. There is also direct evidence from vascular studies. In one angiographic series, 83% of patients undergoing vascular surgery with a difference in systolic BPs had evidence of innominate or subclavian artery stenosis on the side with the lower BP.73 IAD prevalence is also high (78–88%) in subclavian steal syndrome,74,75 and the severity of the stenosis has been associated with the size of the BP difference.76 In another large series of patients presenting for coronary angiography, the non-simultaneously measured IAD had a low sensitivity and positive predictive value, but high specificity (85% for 410 mm Hg and 94% for X20 mm Hg) and negative predictive value (99% for 410 mm Hg and 98% for X20 mm Hg) for predicting subclavian stenosis,77 and aortic arch angiography in a series of subjects with PVD showed a measurable stenosis of at least one brachiocephalic artery in 42% of 48 subjects studied.78 Thus, there is evidence of an association of the IAD with symptomatic PVD. We propose that there is a similar underlying aetiology of asymptomatic arterial disease causing the IAD reported in these studies. As the risk of cardiovascular events is similarly high for subjects with symptomatic and asymptomatic PVD in primary care,79 the recognition of the IAD, assuming our hypothesis to be correct, could become an important part of the risk assessment of hypertensive patients. This would facilitate the focusing of finite resources for primary prevention onto those at highest risk, as demanded by the National Service Framework for Coronary Heart Disease.30 To justify this approach, evidence from prospective survival studies is needed. To date, only the authors have reported a correlation of reduced event-free survival with the IAD in a general20 and hypertensive (Clark CE, MSc Dissertation, Peninsula Medical School, 2006) primary care

population. Larger studies using careful and standardised bilateral assessments of BP are needed to confirm this important observation, and to examine the additional contribution as a risk factor that recognition of the IAD can make to cardiovascular risk assessment. Thus, Cyriax’s words from 1921: ‘Neglect to exclude the presence of the differential blood-pressure sign may easily lead to totally erroneous conclusions as to the state of the circulation’33 remain just as relevant now.

Conclusions The IAD has been repeatedly shown to occur in substantial numbers of patients studied in primary and secondary care, and needs to be looked for in all patients whose diagnosis and treatment depend on accurate BP measurements. Many studies have overestimated prevalence due to poor design, and our estimate of prevalence from good quality studies is 19.6% sIADX10 mm Hg, 4.3%X20 mm Hg and 8.1%X10 mm Hg dIAD. The prevalence in coronary heart disease and hypertensive patients appears to be much higher. This fact, and the importance of measuring both arms during assessment, should be better emphasised in current hypertension management guidelines. There is a body of evidence associating the IAD with PVD, suggesting a pathological rather than physiological cause. This raises the possibility that the presence of the IAD may have a prognostic value in predicting cardiovascular events. There is preliminary evidence for this but further large, carefully controlled studies are needed to confirm it, and to establish what independent contribution to cardiovascular risk assessment the detection of an IAD may make.

Acknowledgements CEC was supported by a Clinical Academic Fellowship from Plymouth tPCT and a grant from Mid Devon Research Group. We owe thanks to Ginny Newton and her colleagues at Exeter Medical Library for their persistence in obtaining so many historic papers and books for this review. Conflict of interests None to declare.

References 1 Riva-Rocci S. Un Sfigmomanometro nuovo. Gaz med Torino 1896; 47: 981–996, 1001–1017 (translated in Ruskin A. Classics in Arterial Hypertension. CC Thomas: Springfield, IL, 1956). 2 Cyriax EF. Unilateral alterations in blood-pressure caused by unilateral pathological conditions: the differential blood pressure sign. Q J Med 1920; 13: 148–164. 3 Ezzati M, Lopez AD, Rodgers A, Vander Hoom S, Murray CJL, the Comparative Risk Assessment Collaborating Group. Selected major risk factors and global and regional burden of disease. Lancet 2002; 360: 1347–1360. Journal of Human Hypertension


4 Williams B, Poulter NR, Brown MJ, Davis M, McInnes GT, Potter JF et al. British Hypertension Society guidelines for hypertension management 2004 (BHSIV): summary. BMJ 2004; 328(7440): 634–640. 5 O’Brien E. General principles of blood pressure measurement. In: O’Brien E, Beevers DG, Marshall HJ (eds). ABC of Hypertension. British Medical Journal Publishing Group: London, 1995, pp 3–8. 6 Primatesta P, Brookes M, Poulter NR. Improved hypertension management and control: results from the Health Survey for England 1998. Hypertension 2001; 38(4): 827–832. 7 Reeves RA. Does this patient have hypertension? How to measure blood pressure. JAMA 1995; 273(15): 1211–1218. 8 O’Brien E, Coats A, Owens P, Petrie J, Padfield PL, Littler WA et al. Use and interpretation of ambulatory blood pressure monitoring: recommendations of the British Hypertension Society. BMJ 2000; 320(7242): 1128–1134. 9 Williams B, Poulter NR, Brown MJ, Davis M, McInnes GT, Potter JF et al. Guidelines for management of hypertension: report of the fourth working party of the British Hypertension Society, 2004-BHS IV. J Hum Hypertens 2004; 18(3): 139–185. 10 Kirkendall WM, Burton AC, Epstein FH, Freis ED. Recommendations for human blood pressure determination by sphygmomanometers. Circulation 1967; 36: 980–988. 11 Petrie JC, O’Brien ET, Littler WA, de Swiet M. Recommendations on blood pressure measurement. BMJ 1986; 293(6547): 611–615. 12 Perloff D, Grim C, Flack J, Frohlich ED, Hill M, McDonald M et al. Human blood pressure determination by sphygmomanometry. Circulation 1993; 88(5): 2460–2470. 13 O’Brien ET, Petrie JC, Littler WA, de Swiet M, Padfield PL, Dillon MJ. Blood Pressure Measurement: Recommendations of the British Hypertension Society, 3rd edn. BMJ Publishing Group: London, 1997. 14 Ramsay L, Williams B, Johnston G, MacGregor G, Poston L, Potter J et al. Guidelines for management of hypertension: report of the third working party of the British Hypertension Society. J Hum Hypertens 1999; 13(9): 569–592. 15 American Heart Association, Cardiac Society of Great Britain and Ireland, Committee of Standardisation of Blood Pressure Readings. Standard methods for taking blood pressure readings. JAMA 1939; 113: 294. 16 Beevers G, Lip GYH, O’Brien E. ABC of hypertension: blood pressure measurement. BMJ 2001; 322(7292): 981–985. 17 Cushman WC. A century of indirect blood pressure measurement. Back to basics. Arch Intern Med 1996; 156(17): 1922–1923. 18 Wingfield D, Pierce M, Feher M. Blood pressure measurement in the community: do guidelines help? J Hum Hypertens 1996; 10(12): 805–809. 19 Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo Jr JL et al. Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension 2003; 42(6): 1206–1252. 20 Clark CE, Powell RJ. The differential blood pressure sign in general practice: prevalence and prognostic value. Family Pract 2002; 19(5): 439–441. Journal of Human Hypertension

21 Newman AB, Sutton-Tyrrell K, Vogt MT, Kuller LH. Morbidity and mortality in hypertensive adults with a low ankle/arm blood pressure index. JAMA 1993; 270(4): 487–489. 22 Newman AB, Shemanski L, Manolio TA, Cushman M, Mittelmark M, Polak JF et al. Ankle-arm index as a predictor of cardiovascular disease and mortality in the Cardiovascular Health Study. The Cardiovascular Health Study Group. Arterioscler Thromb Vasc Biol 1999; 19(3): 538–545. 23 Violi F, Criqui M, Longoni A, Castiglioni C. Relation between risk factors and cardiovascular complications in patients with peripheral vascular disease. Results from the A.D.E.P. study. Atherosclerosis 1996; 120(1–2): 25–35. 24 Vogt MT, Cauley JA, Newman AB, Kuller LH, Hulley SB. Decreased ankle/arm blood pressure index and mortality in elderly women. JAMA 1993; 270(4): 465–469. 25 Zheng ZJ, Sharrett AR, Chambless LE, Rosamond WD, Nieto FJ, Sheps DS et al. Associations of ankle-brachial index with clinical coronary heart disease, stroke and preclinical carotid and popliteal atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) Study. Atherosclerosis 1997; 131(1): 115–125. 26 McDermott MM, Feinglass J, Slavensky R, Pearce WH. The ankle-brachial index as a predictor of survival in patients with peripheral vascular disease. J Gen Intern Med 1994; 9(8): 445–449. 27 Applegate WB. Ankle/arm blood pressure index. A useful test for clinical practice? JAMA 1993; 270(4): 497–498. 28 Thompson JF, Kinsella DC. Vascular disorders of the upper limb. In: Beard JD, Gaines PA (eds). Vascular and Endovascular Surgery. WB Saunders: London, 2001, pp 199–216. 29 Shadman R, Criqui MH, Bundens WP, Fronek A, Denenberg JO, Gamst AC et al. Subclavian artery stenosis: prevalence, risk factors, and association with cardiovascular diseases. J Am Coll Cardiol 2004; 44(3): 618–623. 30 Coronary Heart Disease. National Service Framework. Department of Health: London, 2000. 31 Korns KM, Guinand PH. 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. 32 Altman DG. Practical Statistics for Medical Research. Chapman & Hall: London, 1991. 33 Cyriax EF. Unilateral alterations in blood pressure: the differential blood pressure sign (second communication). Q J Med 1921; 14: 309–313. 34 Phipps C. Blood Pressure. Boston M & S J 1915; 173: 476. 35 Cyriax EF. Blood pressure in war traumatisms. BMJ 1918; 3: 132. 36 Frank SM, Norris EJ, Christopherson R, Beattie C. Right and left arm blood pressure discrepancies in vascular surgery patients. Anesthesiology 1991; 75: 457–463. 37 Harrison EG, Roth GM, Hines EA. Bilateral indirect and direct arterial pressures. Circulation 1960; 22: 419–436. 38 Kay WE, Gardner KD. Comparative blood pressures in the two arms. Calif Western Med 1930; 33: 578–579. 39 Israel E. Differences in blood pressure in both arms. Acta Med Orientalia 1944; 3: 86–89. 40 MacLaren JP. Medical Insurance Examination: Modern Methods and Rating of Lives, for Medical Practitioners,


41 42 43 44

45 46 47 48

49 50 51

52

53

54 55 56 57

58

59 60

and Insurance Officials. London: Bailliere, Tindall and Cox, 1927. Fotherby MD, Panayiotou B, Potter JF. Age-related differences in simultaneous interarm blood pressure measurements. Postgrad Med J 1993; 69(809): 194–196. Hashimoto F, Hunt WC, Hardy L. Differences between right and left arm blood pressures in the elderly. West J Med 1984; 141: 189–192. Kristensen BO, Kornerup HJ. Which arm to measure the blood pressure? Acta Med Scand 1982; 670(Suppl): 69–73. Pesola GR, Pesola HR, Nelson MJ, Westfal RE. The normal difference in bilateral indirect blood pressure recordings in normotensive individuals. Am J Emerg Med 2001; 19(1): 43–45. Yagi S, Ichikawa S, Sakamaki T, Takayama Y, Murata K. Blood pressure in the paretic arms of patients with stroke. N Engl J Med 1986; 315(13): 836. Gould BA, Hornung RS, Kieso HA, Altman DG, Raftery EB. Is the blood pressure the same in both arms? Clin Cardiol 1985; 8(8): 423–426. Goldhill DR. Bilateral simultaneous indirect systolic blood pressure measurements. Cardiovasc Res 1986; 20(10): 774–777. Panayiotou BN, Harper GD, Fotherby MD, Potter JF, Castleden CM. Interarm blood pressure difference in acute hemiplegia. J Am Geriatr Soc 1993; 41(4): 422–423. O’Shea JC, Murphy MB. Ambulatory blood pressure monitoring: which arm? J Hum Hypertens 2000; 14(4): 227–230. Swallow RA. Hypertension: which arm? BMJ 1975; 3: 370. Lane D, Beevers M, Barnes N, Bourne J, John A, Malins S et al. Inter-arm differences in blood pressure: when are they clinically significant? J Hypertens 2002; 20(6): 1089–1095. Orme S, Ralph SG, Birchall A, Lawson-Matthew P, McLean K, Channer KS. The normal range for interarm differences in blood pressure. Age Ageing 1999; 28(6): 537–542. Southby R. Some clinical observations on blood pressure and their practical application, with special reference to variation of blood pressure readings in the two arms. M J Australia 1935; 2(569): 580. Rueger MJ. Blood pressure variations in the two arms. Ann Int Med 1951; 35: 1023. Singer AJ, Hollander JE. Blood pressure. Assessment of interarm differences. Arch Intern Med 1996; 156(17): 2005–2008. Cassidy P, Jones K. A study of inter-arm blood pressure differences in primary care. J Hum Hypertens 2001; 15(8): 519–522. Pesola GR, Pesola HR, Lin M, Nelson MJ, Westfal RE. The normal difference in bilateral indirect blood pressure recordings in hypertensive individuals. Acad Emerg Med 2002; 9(4): 342–345. Arnett DK, Tang W, Province MA, Oberman A, Ellison RC, Morgan D et al. Interarm differences in seated systolic and diastolic blood pressure: the Hypertension Genetic Epidemiology Network study. J Hypertens 2005; 23(6): 1141–1147. Clark CE. Difference in blood pressure between arms might reflect peripheral vascular disease. BMJ 2001; 323(7309): 399–400. Amsterdam B, Amsterdam AL. Disparity in blood pressures in both arms in normals and hypertensives

61 62 63

64 65 66 67 68 69 70 71 72

73 74

75 76

77

78

79

80

and its clinical significance. NY State J Med 1943; 43: 2294–2300. Hensen H. Beitrage zur Physiologie und pathologie des Blutdrucks. Deutsches Arch f klin Med 1900; 67: 436. Van Balen GF. Bloeddrukmetingen aan beide armen. Nederlandsch Tijdschr v Geneesk 1929; 73: 932. Banks MJ, Erb N, George P, Pace A, Kitas GD. Hypertension is not a disease of the left arm: a difficult diagnosis of hypertension in Takayasu’s arteritis. J Hum Hypertens 2001; 15(8): 573–575. Phillips LS, Branch WT, Cook CB, Doyle JP, El Kebbi IM, Gallina DL et al. Clinical inertia. Ann Intern Med 2001; 135(9): 825–834. Janeway T. The Clinical Study of Blood Pressure; A Guide to the Use of the Sphygmomanometer. D. Appleton & Co: New York & London, 1904. Shock NW, Ogden E. The differences between blood pressure measurements obtained simultaneously on the two arms. Q J Exp Physiol 1940; 30: 155. Osler W. Modern Medicine. Lea & Febiger: Philadelphia, 1915. Geisbock F. Die Bedeutung der Blutdruckmessung fur die praxis. Deutsches Arch f klin Med 1905; 83: 363 (quoted from Korns and Guinand (32)). Norris GW. Blood Pressure: Its Clinical Applications 1917. Bing HJ. Ueber die Blutdruckrnessung bei Menschen. Berl Klin Wchnschr 1906; 43: 1650. Bodenstab WH. Blood pressure – difference of readings in the two arms. Journal Lancet 1925; 45(15): 360–361. Sutton-Tyrrell K, Wildman R, Newman A, Kuller LH. Extent of cardiovascular risk reduction associated with treatment of isolated systolic hypertension. Arch Intern Med 2003; 163(22): 2728–2731. Moll F, Six J, Mutsaerts D. Misleading upper extremity blood pressure measurements in vascular occlusive disease. Bruit 1983; 8: 18–19. Hennerici M, Klemm C, Rautenberg W. The subclavian steal phenomenon: a common vascular disorder with rare neurologic deficits. Neurology 1988; 38(5): 669–673. Lawson JD, Petracek MR, Buckspan GS, Dean RH. Subclavian steal: review of the clinical manifestations. South Med J 1979; 72(11): 1369–1373. Tan TY, Schminke U, Lien LM, Tegeler CH. Subclavian steal syndrome: can the blood pressure difference between arms predict the severity of steal? J Neuroimaging 2002; 12(2): 131–135. English JA, Carell ES, Guidera SA, Tripp HF. Angiographic prevalence and clinical predictors of left subclavian stenosis in patients undergoing diagnostic cardiac catheterization. Catheterization Cardiovasc Interv 2001; 54(1): 8–11. Gutierrez GR, Mahrer P, Aharonian V, Mansukhani P, Bruss J. Prevalence of subclavian artery stenosis in patients with peripheral vascular disease. Angiology 2001; 52(3): 189–194. Leng GC, Lee AJ, Fowkes FG, Whiteman M, Dunbar J, Housley E et al. Incidence, natural history and cardiovascular events in symptomatic and asymptomatic peripheral arterial disease in the general population. Int J Epidemiol 1996; 25(6): 1172–1181. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of reporting of metaanalyses. Lancet 1999; 354(9193): 1896–1900. Journal of Human Hypertension

930


Appendix A Summary of excluded studies Study population

Reason for exclusion

Phipps (1915)34 (USA)

Mixed in-patients

Sequential measurements

Cyriax (1918)35 (UK)

War wounded

Cyriax (1920)2 (UK)

Cyriax (1921)33 (UK)

Sample size

Prevalence of systolic differences

Prevalence of diastolic differences

36

20%X10 mm Hg

Not stated


36

72%45 mm Hg

Not stated

Post-Surgical in-patients Many war trauma. Mainly with ‘unilateral’ wounds


73

83%X10 mm Hg 20%X20 mm Hg


Private practice (93) and in-patients (35): Unilateral or bilaterally unequal surgical cases Bilaterally equal and constitutional disorders


128 56

‘Confirmed previous findings’

Kay and Gardner (1930)38 (USA)

General medical practice, opportunistic sample

Korns and Guinand (193l)31 (USA)

78


‘Confirmed previous findings’ 45%X10 mm Hg 4%X20 mm Hg


125

12% 420 mm Hg

13% 410 mm Hg

Healthy University students, 73% male, mean age 20

Simultaneous but single pair of measurements

1000

22.2%X10 mm Hg

21.7%X10 mm Hg

Southby (1935)53 (Australia)

Patients attending one general practice


516

Differences above 20 mm systolic or 10 mm diastolic in 60% of cases

Israel (1944)39

Hypertensives, population not defined


125

18.4% 420 mm Hg

2.4% 410 mm Hg

Office patients


755

50%410 mm Hg 13.6% 420 mm Hg

33.7% 410 mm Hg

Swallow (1975)(letter)50 (UK)

General practice, 50–54 year old males

Measurement technique not stated

33

76%X10 mm Hg

Kristensen and Kornerup (1982)43 (Denmark)

Highly selected normotensive patients without signs of CVD Unselected normotensive in-patients


55

34.5%X10 mm Hg 3.4%X20 mm Hg 60.9%X10 mm Hg 26.1%X20 mm Hg 49.1%X10 mm Hg 15.8%X20 mm Hg 5.2%X30 mm Hg 59.7%X10 mm Hg 25.8%X20 mm Hg 8.1%X30 mm Hg

14.5%X10 mm Hg

20%X10 mm Hg 3.4%X20 mm Hg

9.7%X10 mm Hg

Rueger (1951)

54

(USA)

23

Hypertensive out-patients

57

Hypertensive out-patients

62

Hashimoto et al. (1984)42 (USA)

Elderly residential home residents attending Sequential measurements for BP checks. 50% were hypertensive

Gould et al. (1985)46 (UK)

Hypertensives

Simultaneous random zero measurements with two observers

174 91

Dismissed differences over 20 mm as erroneous, found no apparent mean differences over 10 mm systolic or diastolic. Individual measured

43.5%X10 mm Hg 8.7%X20 mm Hg 29.8%X10 mm Hg 5.2%X20 mm Hg 1.8%X30 mm Hg 37.1%X10 mm Hg 12.9%X20 mm Hg 1.6%X30 mm Hg

The inter-arm blood pressure difference CE Clark et al

Reference (country of origin)

differences were 8% readings 410 mm Hg systolic and 3% diastolic. Concluded that no bias is introduced by making measurements in different arms Goldhill (1986)47 (USA)

Not stated

One observer listening to bilateral dopplers Review of recordings of dopplers and pressures

52 68

37%X6 mm Hg 13%X11 mm Hg 6%X6 mm Hg None49 mm Hg

Yagi et al. (1986)45 (Japan)

Convalescent stroke in-patients

Simultaneous automated measurements

47

Mean 131/83 in paretic arms, 129/78 in intact arms, Po0.01/Po0.001

Frank et al. (1991)36 (USA)

Peripheral vascular disease in-patients


58

21%X20 mm Hg 41%X10 mm Hg 16%X10 mm Hg 3%X20 mm Hg 13%X10 mm Hg 0%X20 mm Hg

31%X10 mm Hg 4%X15 mm Hg 5%X10 mm Hg 3%X15 mm Hg 5%X10 mm Hg 3%X15 mm Hg None

Coronary heart disease in-patients Controls: orthopaedic & urology in-patients

38 38

Fotherby et al. (1993)41 (UK)

Elderly in & out-patients Young in & out-patients


40 40

10%410 mm Hg None

Panayiotou et al. (1993)48 (UK)

Hemiplegic in-patients following acute stroke


15

No correlation with side of paresis, mean inter-arm difference 4.4/4.7 mm Hg

Singer and Hollander (1996)55 (USA)

Ambulant patients over 5 years old attending university hospital emergency department

Sequential measurements Single (not repeated) simultaneous automated indirect

300 310

O’Shea and Murphy (2000)49 (Ireland)

Patients attending for ambulatory BP monitoring


39

33%X10 mm Hg

Clark (2001)59 (UK)

General practice


205


13%X10 mm Hg

Cassidy and Jones (2001)56 (UK)

General practice


237

23%X20 mm Hg

40%X10 mm Hg

Normotensive patients attending ER and staff accepting BP screening


100

15%X10 mm Hg

Clark and Powell (2002)20 (UK)

General practice


280

40.7%X10 mm Hg 13.6%X20 mm Hg

Pesola et al. (2002)57 (USA)

Staff and visitors to an ER with history of hypertension willing to be screened


100

18%X10 mm Hg

Shadman et al. (2004)29 (USA)

General population Vascular patients


2975 1248

1.9%415 mm Hg 7.1%415 mm Hg

Not stated Not stated

Arnett et al. (2005)58 (USA)

Random population sample Hypertensive siblings


824

9.2%410 mm Hg 1.1%420 mm Hg 14.2%410 mm Hg 1.8%420 mm Hg

1.6%410 mm Hg 0420 mm Hg 2.8%410 mm Hg 0.1%420 mm Hg

(USA)

23.2%X10 mm Hg

931


2195

28%410 mm Hg 20.6% 410 mm Hg

The inter-arm blood pressure difference CE Clark et al

Pesola et al. (2001)

44

13.3%420 mm Hg 11.6% 420 mm Hg