Unreliable oscillometric blood pressure measurement: prevalence, repeatability and characteristics of the phenomenon

(2009) 23, 794 800 & 2009 Macmillan Publishers Limited All rights reserved 0950-9240/09 $32.00 www.nature.com/jhh ORIGINAL ARTICLE Unreliable oscillometric blood pressure measurement: prevalence, repeatability and characteristics of the phenomenon GS Stergiou, P Lourida, D Tzamouranis, NM Baibas Hypertension Center, Third University Department of Medicine, Sotiria Hospital, Athens, Greece Oscillometric devices are being widely used for ambulatory, home and office blood pressure (BP) measurement. However, even successfully validated oscillometric devices fail to provide accurate measurements in some patients. This study investigated the prevalence, the reproducibility and the characteristics of the phenomenon of unreliable oscillometric BP (UOBP) measurement. A total of 5070 BP measurements were obtained simultaneously (Y connector) using a professional oscillometric device (BpTRU) and a mercury sphygmomanometer in 755 patients (1706 visits). UOBP readings were defined as those with 410 mm Hg difference (systolic or diastolic) between the two methods. UOBP was found in 15% of systolic and 6.4% of diastolic BP measurements. In all, 18% of the participants had their first but not their second visit, or the reverse. However, 49% of these participants had at least one more UOBP visit after their second visit within the study database. Patients with persistent UOBP were more likely to be female and had lower arm circumference. The systolic BP discrepancy between the two methods was associated with pulse pressure (r ¼ 0.41) and inversely with diastolic BP (r ¼ 0.40) and arm circumference (r ¼ 0.30), whereas the diastolic discrepancy with diastolic BP (r ¼ 0.61) and inversely with pulse pressure (r ¼ 0.32). There was a consistent significant trend for larger systolic BP discrepancy and smaller diastolic from the lower to the higher pulse pressure quintile (Po0.0001). A decreasing arm circumference was a significant predictor of persistent UOBP. These data suggest that the UOBP measurement is particularly common, not very reproducible and mainly affected by pulse pressure and arm circumference. (2009) 23, 794 800; doi:10.1038/jhh.2009.20; published online 26 March 2009 Keywords: blood pressure measurement; blood pressure monitoring; electronic devices; oscillometric devices Introduction Devices that measure blood pressure (BP) using the oscillometric technique are being widely used in clinical practice. This technique has almost completely dominated the area of ambulatory BP monitoring, 1,2 and home BP monitoring seems to follow the same direction. 3,4 More recently, professional oscillometric devices for office/clinic BP measurement appeared on the market and are expected to be increasingly used, particularly because of the recent ban of mercury sphygmomanometers in several countries. 5,6 Thus, any limitations of oscillometric BP measurement deserve close attention and thorough investigation. Correspondence: Professor GS Stergiou, Hypertension Center, Third University Department of Medicine, Sotiria Hospital, 152 Mesogion Avenue, Athens 11527, Greece. E-mail: gstergi@med.uoa.gr Received 14 November 2008; revised 20 February 2009; accepted 22 February 2009; published online 26 March 2009 For the assessment of the accuracy of oscillometric devices, validation protocols have been developed by the American Association for the Advancement of Medical Instrumentation, 7 the British Hypertension Society 8 and the European Society of Hypertension. 9 In the last two decades several oscillometric BP monitors have successfully fulfilled the validation criteria of one or more of these protocols. 10 However, it is recognized that, even the best oscillometric BP monitors that comfortably pass the requirements of established validation protocols fail to provide accurate BP measurements in some patients. 11,12 This drawback of oscillometric BP measurement is well recognized and, therefore, the validation protocols allow for a rather high level of inaccuracy for devices that successfully pass the protocol requirements. 11,12 For example, the European Society of Hypertension International Protocol (ESH-IP) 9 accepts successfully validated devices to have up to 24 out of 99 BP measurements (24%) to differ by 410 mm Hg from the reference auscultatory BP measurement and up to 11 of 33 patients

(33%) to have a 45 mm Hg BP difference from auscultation in two of their three comparisons. Thus, it is accepted that successfully validated and therefore recommended oscillometric devices fail to provide accurate BP measurements in a significant proportion of patients. There is no doubt that this limitation of the oscillometric technique has important implications in the application of such monitors in clinical practice. However, factors responsible for the failure of the oscillometric technique to provide accurate BP measurements remain largely unknown. This prospective study was designed to investigate the prevalence and the reproducibility of the phenomenon of unreliable oscillometric BP (UOBP) measurement and the characteristics of these patients. Patients and methods Consecutive patients referred to an outpatients hypertension clinic for elevated BP, treated or untreated, were recruited. Participants had simultaneous same arm BP measurements (Y connector) using a validated professional oscillometric device (BpTRU Medical Devices, Ltd, Coquitlam, BC, Canada) 13,14 and a standard mercury sphygmomanometer (Baumanometer; WA Baum Co. Inc., New York, NY, USA). This is a standard procedure for routine office BP measurement in the outpatients hypertension clinic since 2003. Three cuffs of different size were used to fit the arm of each individual (inflatable bladder to cover 80 100% of the individual s arm circumference). BP measurements were performed in the clinic by three trained observers experienced in BP monitoring research. Before the study initiation and twice during the study period the observers were re-tested for agreement in BP measurement according to the British Hypertension Society protocol, 8 and the components of the mercury sphygmomanometers were carefully checked. Triplicate BP measurements were taken per participant at each visit after at least 5 min sitting rest and with at least 30 s between measurements. Information about the presence of cardiovascular disease (stroke, transient ischaemic attack, coronary heart disease, heart failure and peripheral artery disease), left ventricular hypertrophy (yearly echocardiography and electrocardiogram), renal failure (serum creatinine 41.2 mg/dl) and smoking (current, never and ex 46 months) were obtained from the patients records. Arrhythmia was detected by the investigators during the auscultatory BP measurements. The protocol was approved by the hospital scientific committee, which did not require an informed consent for the analysis of the outpatients BP data and the retrospective collection of demographic and other data from the patients records. Unreliable oscillometric BP (UOBP) readings were defined as individual readings with 410 mm Hg difference (systolic and/or diastolic) between the oscillometric and the auscultatory measurement. UOBP visits were defined as those with at least two UOBP readings of the triplicate readings performed at each visit. Patients with UOBP were defined as those with at least one UOBP visit. The reproducibility of the UOBP was assessed in patients with two or more visits by comparing the first versus the second visit. Patients with persistent UOBP were defined as those with at least two UOBP visits (any visits within the study database). A case control comparison of the characteristics of these patients (systolic and diastolic BP, pulse pressure, body mass index, arm circumference, cuff size, diabetes mellitus, renal failure, cardiovascular disease, left ventricular hypertrophy, arrhythmia, smoking, time from hypertension diagnosis, number of antihypertensive drugs and hypolipidemic drug treatment) was carried out against the first consecutive patients, sex and age (within 5 years) matched, but without none of their two or more study visits. Unpaired t-tests were used to compare continuous parameters of different subgroups of participants and w 2 -tests for categorical parameters. Paired t-tests were used for the case control comparison of patients with persistent UOBP against their controls. The kappa statistic was applied to assess the agreement in the detection of different visits of the same patients (first versus second visit of patients with at least two visits during the study). Pearson s correlations were used to investigate the relationship between the oscillometric auscultatory BP measurement discrepancy and systolic and diastolic BP, pulse pressure and arm circumference. Conditional logistic regression models were used to explore the association of variables with cases of persistent UOBP versus age- and sex-matched controls with non-discrepant BP measurements, in a case control analysis. The variables explored were auscultatory systolic and diastolic BP, arm circumference, pulse rate and auscultatory pulse pressure. Separate models were fitted for systolic/diastolic BP and pulse pressure to avoid collinearity between the variables. One-way analysis of variance was used to compare the oscillometric auscultatory BP discrepancy across the quintiles of auscultatory pulse pressure. Statistical analysis was carried out using the MINITAB INC Statistical Software (release 13.31) (Stage College, PA, USA) and conditional logistic regression using the STATA statistical package (StataCorp LP, College Station, TX USA). A probability level Po0.05 was regarded as statistically significant. Results Patients and BP measurements A total of 5154 BP readings were collected within 14 months. Erroneous readings and visits with fewer 795

796 than three valid readings were discarded. The analysis was on the basis of 5070 valid BP readings taken from 755 patients in 1706 clinic visits. In this dataset 40% of the patients had one clinic visit, 26% had two and 34% three or more visits. The characteristics of the study participants are presented in Table 1. Average auscultatory BP was 141.7±19.0/86.9±11.8 mm Hg (systolic/diastolic) and average oscillometric BP was 136.6±19.9/ 83.6±11.3 (mean difference systolic 5.0±6.2 mm Hg and diastolic 3.4±5.0 mm Hg). The proportion of UOBP readings as well as the proportion of UOBP visits and of participants with UOBP is presented in Figure 1 (separately for systolic and diastolic BP). UOBP reproducibility A total of 453 patients had at least two office visits during the study. Assessment of the UOBP reproducibility by comparing the first with the second visit of these 453 patients showed that 39 (9%) had UOBP in their first but not their second visit, 41 had the reverse (9%), 16 (4%) had both the first and the second visit and 357 (79%) in none (disagreement in UOBP detection between the two visits in 18% of patients; kappa statistic 0.18, s.e. 0.06, 95% confidence intervals (CI) 0.07, 0.30). Patients with UOPB in the first but not the second visit, or the reverse, (n ¼ 80) were compared with those with both visits (n ¼ 16), regarding age, gender, duration of hypertension, left ventricular hypertrophy, cholesterol, smoking, auscultatory BP, pulse rate, BP difference and antihypertensive treatment change between the two visits (initiation, addition or substitution). No differences between the two groups were found (data not shown), apart from a larger proportion of treatment changes between the two visits in the group with both visits (37 versus 14%, P ¼ 0.02). Thus, none of these factors seemed to explain the poor reproducibility of the UOBP phenomenon between the first and second visits. It is interesting that, of the 80 patients with disagreement regarding the UOBP diagnosis in the first versus the second visit, 39 (49%) had another visit after the second visit within the study database. Furthermore, as shown in Table 2, patients with at least one UOBP visit had a good chance to repeat this phenomenon if they had additional visits within the study database. For example, those with at least one UOBP visit and overall four or more visits within the database (on an average 4.8 visits) had 50% chance to have at least one more UOBP visit within the study database, 21% chance to have at least another two UOBP visits and 9% chance to have another three UOBP visits within the study database (Table 2). UOBP patients characteristics Fifty-nine patients had persistent UOBP defined as at least two study visits (any visits within the study database) with at least two UOBP readings per visit. A comparison of the characteristics of these patients against the rest of the study population showed that patients with persistent UOBP were more likely to be female and had lower arm circumference (Table 1). A case control comparison of these patients against the first consecutive 59 patients, sex and age matched, but without none of their two or more study visits, showed UOBP patients to have lower arm circumference (mean difference 1.2±3.1 cm, 95% CI 0.26, 2.1, P ¼ 0.01), lower body mass index (mean difference 3.0±4.4 kg/m 2, 95% CI 1.8, 4.1, Po0.0001) and a tendency for higher auscultatory systolic BP (mean difference 2.5±26.8 mm Hg, P ¼ NS) and higher pulse pressure (mean difference 2.7±22.4 mm Hg, P ¼ NS). No differences in other study parameters were found. BP Readings (n=5.070) Office Visits (n=1.706) Subjects (n=755) 4.8 6.4 7.7 10 15.2 Systolic BP Diastolic BP 18.4 0 5 10 15 20 Proportion (%) Figure 1 Prevalence of unreliable oscillometric blood pressure measurement in individual readings, visits or patients (for definitions see Patients and methods; BP, blood pressure). Table 1 Characteristics of subjects with persistent unreliable oscillometric blood pressure (UOBP) measurement compared with subjects without persistent UOBP and the total study population Persistent UOBP (n ¼ 59) Other non-uobp (n ¼ 696) Total database (n ¼ 755) Age (years) 58.2±15.8 58.1±12.6 58.1±12.9 Women 38 (64%) 319 (46%) * 367 (49%) Body mass index (kg/m 2 ) 26.0±4.0 28.1±4.3 27.9±4.3 Arm circumference (cm) 28.7±2.9 29.9±3.3 ** 29.7±3.3 Antihypertensive drugs 1.8±1.0 1.8±1.0 1.8±1.0 Diabetes mellitus 5 (9%) 64 (9%) 69 (9%) *P ¼ 0.01; **P ¼ 0.009 compared with subjects with persistent UOBP.

Table 2 Persistence of the unreliable oscillometric blood pressure measurement (UOBP) phenomenon in repeated visits within the study dataset in subjects with at least one UOBP visit 797 Number of visits in the study dataset Average number of visits Xone visit Xtwo visits Number of subjects (%) Xthree visits Xfour visits X2 2.2 131 41 (31) X3 3.9 90 34 (38) 13 (14) X4 4.8 56 28 (50) 12 (21) 5 (9) X5 5.5 29 17 (59) 7 (24) 4 (14) UOBP associations In patients with persistent UOBP the systolic BP discrepancy between the oscillometric and the auscultatory measurement was strongly correlated with auscultatory pulse pressure (r ¼ 0.41, P ¼ 0.001) and inversely correlated with auscultatory diastolic BP (r ¼ 0.40, P ¼ 0.002) and arm circumference (r ¼ 0.30, P ¼ 0.05), whereas a weak association was found with auscultatory systolic BP (r ¼ 0.17, P ¼ NS). On the other hand, the diastolic BP discrepancy was strongly correlated with auscultatory diastolic BP (r ¼ 0.61, Po0.0001) and inversely with auscultatory pulse pressure (r ¼ 0.32, P ¼ 0.01), whereas no relationship was found with arm circumference (r ¼ 0.11) or auscultatory systolic BP (r ¼ 0.04). Table 3 Conditional logistic regression models investigating independent predictors of persistent unreliable oscillometric blood pressure (BP) measurement cases versus age- and sexmatched controls Variable Odds ratio P (95% CI) Model 1 Arm circumference 0.69 0.02 (0.51, 0.95) Auscultatory systolic BP 1.01 0.64 (0.98, 1.04) Auscultatory diastolic BP 0.97 0.43 (0.91, 1.04) Pulse rate 1.03 0.29 (0.97, 1.09) Model 2 Arm circumference 0.72 0.02 (0.55, 0.96) Pulse pressure 1.005 0.74 (0.98, 1.03) Pulse rate 1.03 0.04 (0.97, 1.08) Abbreviation: CI, confidence intervals. UOBP predictors In conditional logistic regression analysis of patients with persistent UOBP measurements (n ¼ 59) versus age- and sex-matched controls with non-discrepant BP measurements, arm circumference was the only variable that showed a significant inverse association with persistently discrepant UOBP cases in both models that were fitted (Table 3). Pulse pressure Q1 Q2 Q3 Q4 Q5 P<0.0001 3.6 4.8 6.0 7.2 Systolic blood pressure discrepancy (mmhg) BP discrepancy and pulse pressure The levels of systolic and diastolic BP discrepancy in quintiles of auscultatory pulse pressure in the total 5070 readings are shown in Figure 2 (1014 BP readings per quintile; Q1: mean pulse pressure 33.4 mm Hg, range 12 40; Q2: mean 43.8, range 40 48; Q3: mean 52.0, range 48 56; Q4: mean 62.2, range 56 68; Q5: mean 82.6, range 68 158). There was a consistent trend for larger systolic BP discrepancy and smaller diastolic across the pulse pressure quintiles from Q1 to Q5 (Po0.0001 for both). Discussion This study investigated the phenomenon of UOBP measurement in patients attending an outpatients BP linic. The main findings are, first, that this Pulse pressure Q1 Q2 Q3 Q4 Q5 P<0.0001 1.5 3.0 4.5 6.0 Diastolic blood pressure discrepancy (mmhg) Figure 2 Systolic and diastolic blood pressure discrepancy between the oscillometric and the auscultatory measurement across the pulse pressure quintiles (Q1 Q5; 1014 readings per quintile; mean difference with 95% confidence intervals). phenomenon is particularly common, second that it is poorly reproducible and third that seems to be affected by pulse pressure.

798 This study has neither the elegance nor the complexity of a formal validation study of an oscillometric BP monitor. In contrast with validation studies (a) BP measurements were taken at an outpatients clinic instead of a research setting, (b) simultaneous instead of sequential BP measurements were obtained to compare the oscillometric versus the reference auscultatory BP measurements, and (c) a single instead of two observers performed each set of measurements. This simplified protocol was necessary to shorten the validation procedure in order to be applicable in a large number of patients. On the other hand, in order to increase the accuracy of the study in detecting the UOBP phenomenon, carefully standardized observers obtained three pairs of measurements and the threshold for UOBP was set at 410 mm Hg difference between the oscillometric and the auscultatory BP measurement instead of the conventional 5 mm Hg threshold. 7 9 The latter aimed to eliminate the expected underestimation of auscultatory systolic and overestimation of diastolic BP because of the faster deflation rate of the electronic device (3.5 4.0 mm/s) compared with that recommended for auscultatory BP measurement (2 3 mm/s). 15 Furthermore, at least two of the triplicate measurements were required to define an UOBP visit, and two such visits were required to define persistent UOBP that was used to investigate predictors of this phenomenon. Oscillometric BP measurements in this study were obtained using a professional device (BPM- 100 or BpTRU) that has fulfilled the requirements of American Association for the Advancement of Medical Instrumentation validation protocol and achieved an A grade of the British Hypertension Society validation protocol for both systolic and diastolic BP measurement. 13,14 Although not formally investigated, the average oscillometric auscultatory BP discrepancy found in this study seems to fulfill the American Association for the Advancement of Medical Instrumentation validation criterion of o5 mm Hg with s.d. o8 mm Hg. 7 Furthermore, the 85% of the total of BP readings lying within 10 mm Hg from the reference method (Figure 1) also complies with the ESH-IP phase 2.1 criteria. 9 Despite the use of a successfully validated professional oscillometric device, the prevalence of the UOBP phenomenon in this study is striking. Irrespective of how this phenomenon was approached (in terms of single BP readings, single visits or individual patients), in about 10 18% of the cases for systolic BP and 5 8% for diastolic the oscillometric measurement failed to provide an accurate assessment of BP (Figure 1). These data confirm the findings of multiple validation studies of oscillometric BP monitors using the ESH-IP (phase 2.2), showing about 20% of each study s participants having a 45 mm Hg difference between the oscillometric and the reference BP method in two or more of their triplicate BP readings. 11 Although a looser threshold (10 mm Hg) was applied in this study, as mentioned above this was deemed necessary to improve the accuracy of the detection of the UOBP phenomenon in the conditions of a clinic rather than a research setting. The most puzzling finding of this study is the poor reproducibility of the UOBP phenomenon. About 9% of the participants with their first study visit were not classified as such in their second visit, whereas another 9% did the reverse. Thus, in 18% of the participants there was disagreement between the first and the second visit in the UOBP yes/no classification, with kappa statistic suggesting only slight agreement. 16 It is interesting that, although the appearance of a UOBP visit was often sporadic, as suggested by the poor reproducibility of the UOBP phenomenon when only the first and second visits were assessed, it seems that once the phenomenon appears, it is very likely that it will be found again if repeated visits are performed (Table 2). This study failed to identify factors related with the poor reproducibility of the UOBP phenomenon. The imperfect reproducibility of the BP measurement in the clinic probably is one of the factors responsible. 17 It should be noted that no earlier information on the reproducibility of the UOBP is available because in both the validation studies, and in other studies that assessed the UOBP phenomenon, all BP measurements of each participant were obtained in a single occasion. Several different approaches have been used in this study to identify factors related with the UOBP phenomenon (results of UOBP patients characteristics, associations, predictors and BP discrepancy and pulse pressure). The most interesting finding of this analysis is that pulse pressure, which is a reliable marker of arterial stiffness, seemed to be related with the oscillometric auscultatory BP discrepancy. Although not significant in all the study analyses, the trend was always the same, with increasing pulse pressure being associated with larger oscillometric versus auscultatory BP discrepancy. As shown in Figure 1, with increasing pulse pressure the systolic BP discrepancy was increased, whereas the diastolic discrepancy reduced. Clinically important (45 mm Hg) discrepancy between the oscillometric and the auscultatory BP measurement appeared at pulse pressure quintiles Q4 Q5 (pulse pressure 456 mm Hg) for systolic BP and Q1 for diastolic (pulse pressure o40 mm Hg). Thus it seems that the algorithm of the tested oscillometric device is accurate within a given range of pulse pressure (arterial stiffness). When pulse pressure is higher than this range the systolic BP measurement accuracy is reduced and when it is lower the diastolic. Four earlier studies, each one with entirely different design, support the view that arterial stiffness is related with UOBP measurement. 18 21 First is a population-based cohort study of 1808

healthy patients aged 455 years. 18 BP was measured twice using an oscillometric device (Dinamap, GE Healthcare, Chalfont St Gilles, UK) in the supine position by a physician (always first), and again 15 mins later using a Hawksley random-zero sphygmomanometer (Hawksley and Sons, Lancing, Sussex, UK) in the sitting position by a nurse. As expected, because of the measurement protocol the oscillometric BP was by 11/4 mm Hg (systolic/ diastolic) higher. Yet, the BP discrepancy was related to arterial stiffness assessed by carotidfemoral pulse wave velocity, independently of age, gender and BP level. 18 Second is a study in 305 patients with BP measured three times using an oscillometric device (Dinamap) and a mercury device in random order, which showed that the BP discrepancy between the two devices tended to increase in very old participants and in those with high pulse pressure. 19 Third is a retrospective analysis of six simultaneous BP measurements (SpaceLabs Inc, Redmont, Washington, WA, USA oscillometric versus mercury device) obtained in 192 young adults (diabetics type-1 and healthy controls) that showed the BP discrepancy between the two devices to differ in the two groups, and to be influenced mainly by the presence of diabetes and less so by age and diabetes duration. 20 Last is a validation study of an oscillometric home BP monitor in patients with end-stage renal disease (Microlife AG, Widnau, Switzerland), who are known to have advanced arterial stiffness. 21 Although the device fulfilled the ESH-IP validation criteria, the degree of the arterial stiffness assessed by measuring the augmentation index affected the accuracy of oscillometric BP measurement. 21 In this study, a relationship of an increasing oscillometric auscultatory BP discrepancy with small arm circumference was consistently found in three different analyses (results of UOBP patients characteristics, associations and predictors). An earlier validation study of another oscillometric device (Omron Corp., Kyoto, Japan) in 197 patients aged 6 16 years showed better accuracy for diastolic BP measurement at arm circumference 22 32 cm compared with 14 21 cm. 22 More research is needed to show the effect of arm/cuff size on the accuracy of oscillometric BP measurement. Taken together the above findings suggest an important role of pulse pressure and arterial stiffness on the phenomenon of UOBP. This study confirms this concept supported through different approaches applied in earlier studies, and in addition provides a real-life, prospectively collected and large database that allowed the investigation of the true prevalence, the reproducibility and the characteristics of this clinically important and common phenomenon. The conclusion is that oscillometric BP measurement technology is still imperfect and caution is needed when such monitors are used in patients with arterial stiffness. What is known about this topic K Oscillometric devices are being widely used for ambulatory, home and office blood pressure (BP) measurement. K Even successfully validated oscillometric devices fail to provide accurate measurements in some patients. Factors related to unreliable oscillometric BP (UOBP) measurement remain largely unknown. What this study adds K The prevalence, the reproducibility and the characteristics of the UOBP phenomenon were investigated in a prospective study involving 45000 BP measurements. K The findings suggest that in clinical practice the UOBP phenomenon is particularly common, poorly reproducible and mainly affected by pulse pressure and arm circumference. References 1 O Brien E, Asmar R, Beilin L, Imai Y, Mallion JM, Mancia G et al. European Society of Hypertension Working Group on Blood Pressure Monitoring. European Society of Hypertension recommendations for conventional, ambulatory and home blood pressure measurement. J Hypertens 2003; 21: 821 848. 2 Pickering TG, Hall JE, Appel LJ, Falkner BE, Graves J, Hill MN et al. Subcommittee of Professional and Public Education of the American Heart Association Council on High Blood Pressure Research. Recommendations for blood pressure measurement in humans and experimental animals: Part 1: blood pressure measurement in humans: a statement for professionals from the Subcommittee of Professional and Public Education of the American Heart Association Council on High Blood Pressure Research. Hypertension 2005; 45: 142 161. 3 Pickering TG, Miller NH, Ogedegbe G, Krakoff LR, Artinian NT, Goff D. American Heart Association; American Society of Hypertension; Preventive Cardiovascular Nurses Association. Call to action on use and reimbursement for home blood pressure monitoring: a joint scientific statement from the American Heart Association, American Society Of Hypertension, and Preventive Cardiovascular Nurses Association. Hypertension 2008; 52: 10 29. 4 Parati G, Stergiou GS, Asmar R, Bilo G, de Leeuw P, Imai Y et al. ESH Working Group on Blood Pressure Monitoring. European Society of Hypertension guidelines for blood pressure monitoring at home: a summary report of the Second International Consensus Conference on Home Blood Pressure Monitoring. J Hypertens 2008; 26: 1505 1526. 5 Pickering TG. What will replace the mercury sphygmomanometer? Blood Press Monit 2003; 8: 23 25. 6 Stergiou GS. Office blood pressure measurement with electronic devices: has the time come? Am J Hypertens 2008; 21: 246. 7 White WB, Berson AS, Robbins C, Jamieson MJ, Prisant LM, Roccella E et al. National standard for measurement of resting and ambulatory blood pressures with automated sphygmomanometers. Hypertension 1993; 21: 504 509. 8 O Brien E, Petrie J, Littler W, de Swiet M, Padfield PL, O Malley K et al. The British Hypertension Society protocol for the evaluation of automated and semi- 799

800 automated blood pressure measuring devices with special reference to ambulatory systems. J Hypertens 1990; 8: 607 619. 9 O Brien E, Pickering T, Asmar R, Myers M, Parati G, Staessen J, et al., On behalf of the Working Group on Blood Pressure Monitoring of the European Society of Hypertension Working Group on Blood Pressure Monitoring of the European Society of Hypertension. International Protocol for validation of blood pressure measuring devices in adults. Blood Press Monit 2002; 7: 3 17. 10 dabl s Educational Trust. Devices for blood pressure measurement. http://www.dableducational.org. Assessed 2 November 2008. 11 O Brien E, Atkins N. Validation and reliability of blood pressure monitors. In: W White (ed). Blood Press Monit Cardiovasc Med Ther. Humana Press Inc.: Totowa, NJ, USA, 2007, pp 97 132. 12 Gerin W, Schwartz AR, Schwartz JE, Pickering TG, Davidson KW, Bress J et al. Limitations of current validation protocols for home blood pressure monitors for individual patients. Blood Press Monit 2002; 7: 313 318. 13 Wright JM, Mattu GS, Perry Jr TL, Gelferc ME, Strange KD, Zorn A et al. Validation of a new algorithm for the BPM-100 electronic oscillometric office blood pressure monitor. Blood Press Monit 2001; 6: 161 165. 14 Mattu GS, Perry Jr TL, Wright JM. Comparison of the oscillometric blood pressure monitor (BPM-100 (Beta)) with the auscultatory mercury sphygmomanometer. Blood Press Monit 2001; 6: 153 159. 15 Pickering TG, Hall JE, Appel LJ, Falkner BE, Graves J, Hill MN et al. Subcommittee of Professional and Public Education of the American Heart Association Council on High Blood Pressure Research. Recommendations for blood pressure measurement in humans and experimental animals: Part 1: blood pressure measurement in humans: a statement for professionals from the Subcommittee of Professional and Public Education of the American Heart Association Council on High Blood Pressure Research. Hypertension 2005; 45: 142 161. 16 Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statistic. Fam Med 2005; 37: 360 363. 17 Stergiou GS, Alamara CV, Salgami EV, Vaindirlis IN, Dacou-Voutetakis C, Mountokalakis TD. Reproducibility of home and ambulatory blood pressure in children and adolescents. Blood Press Monit 2005; 10: 143 147. 18 Van Popele NM, Bos WJ, de Beer NA, van Der Kuip DA, Hofman A, Grobbee DE et al. Arterial stiffness as underlying mechanism of disagreement between an oscillometric blood pressure monitor and a sphygmomanometer. Hypertension 2000; 36: 484 488. 19 Ni H, Wu C, Prineas R, Shea S, Liu K, Kronmal R et al. Comparison of Dinamap PRO-100 and mercury sphygmomanometer blood pressure measurements in a population-based study. Am J Hypertens 2006; 19: 353 360. 20 Van Ittersum FJ, Wijering RM, Lambert J, Donker AJ, Stehouwer CD. Determinants of the limits of agreement between the sphygmomanometer and the SpaceLabs 90207 device for blood pressure measurement in healthy volunteers and insulin-dependent diabetic patients. J Hypertens 1998; 16: 1125 1130. 21 Thompson AM, Eguchi K, Reznik ME, Shah SS, Pickering TG. Validation of an oscillometric home blood pressure monitor in an end-stage renal disease population and the effect of arterial stiffness on its accuracy. Blood Press Monit 2007; 12: 227 232. 22 Stergiou GS, Yiannes NG, Rarra VC. Validation of the Omron 705 IT oscillometric device for home blood pressure measurement in children and adolescents: the Arsakion School Study. Blood Press Monit 2006; 11: 229 234.