The increasing awareness of hypertension as a serious

AJH 2002; 15:787 792 Evaluation of the Performance of a Wrist Blood Pressure Measuring Device With a Position Sensor Compared to Ambulatory 24-Hour Blood Pressure Measurements Sakir Uen, Burkard Weisser, Paul Wieneke, Hans Vetter, and Thomas Mengden This randomized, single-center, open, within-subject study evaluated the performance of the Braun PrecisionSensor 2000 blood pressure (BP) measurement wrist device (BP 2000, Braun GmbH, Germany) with and without the position sensor, and compared the results with data obtained from 24-h ambulatory BP measurement (ABPM, A&D TM 2430). In a crossover design, 43 subjects performed BP measurements with the BP 2000 during two 7-day periods, one with the position sensor and the other without the position sensor. The correlation coefficients between all wrist self-measurements and ABPM were 0.73 for systolic and 0.65 for diastolic BP (with position sensor) compared to 0.70 and 0.60 for readings without position sensor, respectively. The categorization of subjects as normotensive or hypertensive, using the wrist device in subjects without specific training, achieved a correct classification of 84% (with position sensor) and 81% (without position sensor) in comparison to ABPM. When the measurements were performed in the clinic by an experienced observer, the correct classification was 79% in comparison to ABPM. The crossover study design did not show any significant variation between the two devices with regard to reproducibility of readings at the wrist. In conclusion, self-bp measurements with BP 2000, by untrained subjects, produced results consistent with those found recently with self-bp measurements with upper arm devices, when both data sets are compared with ABPM. The rates of false classification of normotension/hypertension with the wrist device were small and at least as reliable as office measurements. Am J Hypertens 2002;15:787 792 2002 American Journal of Hypertension, Ltd. Key Words: Blood pressure self-measurement, wrist device, position sensor, arterial hypertension, ambulatory 24-h BP measurement. The increasing awareness of hypertension as a serious health hazard and the importance of its early detection and treatment have resulted in an increased availability of a variety of blood pressure (BP) measurement devices. These devices may be used, not only in supervised settings such as hospitals and clinics, but also unsupervised in the home by individuals who must interpret the measurements before deciding whether to seek professional advice. The effectiveness and reliability of unsupervised screening and interpretation depends on the correct operation of the particular BP measurement device and its reproducibility. The Braun PrecisionSensor 2000 (BP 2000) device under evaluation was developed for unsupervised use in the home and uses the well-known oscillometric approach for obtaining BP values. The hydrostatic pressure difference is an important factor in BP measurement at the wrist. Not holding the cuff at heart level will result in an offset of the BP values. The BP 2000 is equipped with a position sensor that allows the user to adopt the correct position of the forearm and wrist. The user is required to sit upright with arms crossed. The forearm with the device attached must be raised up to the angle indicated by the position sensor, as guided on the display. The objectives of the present study were, first, to evaluate the performance of the BP 2000 measurement device with the position sensor in comparison without the position sensor, and second, to compare the results of the BP 2000 measurements with and without position sensor with data obtained from one profile (24 h) ambulatory BP measurement (ABPM). Received March 11, 2002. First decision March 29, 2002. Accepted April 16, 2002. From the Department of Internal Medicine (SU, BW, HV, TM), University of Bonn, Bonn; and Braun GmbH (PW), Kronberg, Germany. Address correspondence and reprint requests to Dr. Thomas Mengden, University of Bonn, Department of Internal Medicine, Wilhelmstr.35, 53111 Bonn, Germany; e-mail: mengden@uni-bonn.de 2002 by the American Journalof Hypertension, Ltd. Published by Elsevier Science Inc. 0895-7061/02/$22.00 PII S0895-7061(02)02979-5

788 CLINICAL EVALUATION OF A NEW WRIST DEVICE AJH September 2002 VOL. 15, NO. 9 Methods Study Population Normotensive or hypertensive (systolic BP 140 mm Hg, diastolic BP 90 mm Hg) male or female subjects (aged 18 years) were included. Exclusion criteria comprised previous experience with wrist BP measurement devices, wrist circumferences 13.5 cm and 19.5 cm, history of significant cardiac arrhythmia, hypertensive crisis, claudication or restricted blood flow, damage or injury of the left wrist, participation in any other clinical trial at the same time, pregnancy, and lactation. Approval was granted by an independent and applicable Ethics Review Committee, and written, informed consent was obtained from each subject before entry into the study. BP Measurements Arm circumferences were measured for both arms and wrists. The appropriate cuff size (for the auscultatory device) was selected for each subject. For the BP 2000 device, only one cuff size was available (wrist circumference, 13.5 to 19.5 cm). Subjects were assigned randomly into two groups in a crossover study design, with a duration of 16 days. From days 1 to 8, group 1 was assigned home measurements with BP 2000 with activated position sensor (BP 2000 P), and group 2 with BP 2000 without activated position sensor (BP 2000 P). A 24-h BP measurement device was used to record BP in all subjects between days 8 and 9. From days 9 to 16, group 1 was provided with BP 2000 P, and group 2 with BP 2000 P. A nonresponder test was performed on day 16. The mean value of three office measurements on the left arm were compared with the mean value of three supervised measurements with the activated position sensor (BP 2000 P). The ABPM was performed with a highly accurate device (A&D TM-2430) validated according to the Association for the Advancement of Medical Instrumentation (AAMI) and British Hypertension Society protocols (grade A/A) and recommended for clinical use. 1,2 The ABPM was performed during average working days and BP was measured at preset 15-min intervals from 7 AM to 10 PM (daytime readings), and at 20-min intervals at night (nighttime readings). We used mean whole-day ABPM (24 h) and daytime readings for statistical comparisons with office readings and self-measured BP values. Conventional sphygmomanometric office BP measurement was performed in the sitting position at 2-min intervals with an auscultatory BP measurement device according to the recommendations of the American Heart Association with at least four measurements/patient at entry of the study. Two office measurements per arm were performed at study entry. Home BP measurements were performed with the BP 2000, a new oscillometric wrist BP monitor (Braun GmbH, Kronberg, Germany). The cuff of the device was inflated and deflated automatically after the start of the measurement. Systolic and diastolic BP, heart rate, and the date and time of measurement were stored in the device memory automatically. All subjects were requested to measure BP at home according to the instructions for use, and to measure their BP twice in the morning and twice in the evening. For statistical analysis only stored data were used. Before the start of the study all BP measurement systems were calibrated according to the manufacturers instructions. All measurements (ABPM, office, and selfmeasured readings) were performed at the left arm/wrist. Before the study started, the investigators were trained for the appropriate use of any BP measurement device used in the study, and had to demonstrate proficiency with all devices according to the manufacturers instructions. Analysis of Data Statistical comparisons were made between the mean BP values resulting from BP 2000 measurements and those resulting from office readings and the 24-h ABPM readings. Subjects were classified as hypertensive if the following criteria were fulfilled: systolic BP 130 mm Hg or diastolic BP 80 mm Hg for the 24-h BP measurement; systolic BP 135 mm Hg or diastolic BP 85 mm Hg for the daytime values of the 24-h BP measurement; systolic BP 135 mm Hg or diastolic BP 85 mm Hg for self-bp measurement; and systolic BP 140 mm Hg or diastolic BP 90 mm Hg for office measurements. Variations between the BP values within each subject for both BP 2000 measurement devices were compared statistically. To evaluate the reproducibility of the results for each of the two measurement devices the standard deviation of the measurements within each subject were regarded as dependent variable. For both measurement devices, BP 2000 P and BP 2000 P, mean and standard deviation were statistically compared. A subject was classified as a nonresponder if the following absolute differences in BP between both types of measurement were fulfilled: [mean BP (systolic AUSC) mean (systolic BP 2000 P)] 15 mm Hg or [mean BP (diastolic AUSC) mean BP (diastolic BP 2000 P)] 10 mm Hg, where AUSC was the auscultatory BP measurement performed by the investigator, and BP 2000 P was the BP measurement supervised by the investigator. Results A total of 46 subjects were randomized into the study of which 43 completed the study with at least one self-bp measurement in each measurement period (intention-totreat [ITT] population). From the ITT population, 19 subjects performed 24 BP self-measurements in each measurement period. This group representing the per-protocol population (PP). For the PP data set all subjects performed 24 BP measurements. For the ITT data set the mean values of the number of performed BP measurements were 22.2

AJH September 2002 VOL. 15, NO. 9 CLINICAL EVALUATION OF A NEW WRIST DEVICE 789 Table 1. Characteristics of the study population (N 43) Range Mean SD Age (y) 23 81 51 14.3 Height (cm) 156 197 171.5 8.8 Weight (kg) 49 110 76.1 16.3 Body mass index (kg/m 2 ) 18.2 34.6 25.7 4.1 Left arm circumference (cm) 23.0 37.0 29.3 3.1 Right arm circumference (cm) 24.0 36.0 29.3 2.7 Left wrist circumference (cm) 15.0 19.5 17.7 1.2 Right wrist circumference (cm) 15.5 20.2 17.9 1.3 Left arm systolic BP (mm Hg) 103 185 131.9 17.3 Right arm systolic BP (mm Hg) 100 173 130.1 15.5 Left arm diastolic BP (mm Hg) 63 110 85.3 12.5 Right arm diastolic BP (mm Hg) 60 113 85.2 12.3 BP blood pressure. (SD 2.6) and 23.3 (SD 1.5) for measurement periods 1 and 2, respectively. There were 24 male (55.8%) and 19 female (44.2%) subjects in the study. The demographic and baseline characteristics of the ITT population are shown in Table 1. A crossover design was used in this study in which each subject performed BP measurements with the BP 2000 during two 7-day periods (period 1 and 2), one with the position sensor (BP 2000 P) and the other without the position sensor (BP 2000 P). There was no statistical evidence for unequal effects between the measurement devices for systolic or diastolic BP. When all BP selfmeasurements were combined for both time periods systolic BP values were on average 1.5 mm Hg (BP 2000 P) and 0.5 mm Hg (BP 2000 P) lower than the 24-h BP values (Table 2). Comparison with the daytime ABPM measurements, however, resulted in larger differences. The diastolic BP values were on average 5.2 mm Hg (BP 2000 P) and 6.0 mm Hg (BP 2000 P) higher than the 24-h ABPM values (Table 2). Comparison with the daytime ABPM measurements showed smaller differences for diastolic BP. Correlation coefficients between systolic BP values resulting from BP 2000 measurements and 24-h ABPM were stronger than for daytime ABPM measurements (0.70 BP 2000 P, 0.73 BP 2000 P, 24-h ABPM; 0.64 BP 2000 P, 0.67 BP 2000 P, daytime ABPM). The corresponding correlation coefficients for diastolic BP values were also stronger for 24-h ABPM than for daytime ABPM measurements (0.6 BP 2000 P, 0.65 BP 2000 P, 24-h ABPM; 0.48 BP 2000 P, 0.49 BP 2000 P, daytime ABPM). All correlations were highly significant (P.001). The PP analysis yielded findings similar to those of the ITT analysis, with no evidence for unequal crossover effects between BP 2000 measurement devices, and only minor differences in mean and standard deviation values. The potential influence of prior experience with BP 2000 P in time period 1 on subsequent measurements in period 2 with BP 2000 P was evaluated. Prior experience with BP 2000 P slightly reduced both the systolic and diastolic BP level measured with BP 2000 P, whereas prior experience with BP 2000 P increased the BP level measured with the BP 2000 P (Table 3). These effects were of approximately equal size for both BP 2000 measurement devices. The reproducibility of the results for each of the BP 2000 measurement devices was assessed within each subject. There was no evidence for unequal crossover effects between the BP 2000 measurement devices for systolic or Table 2. Differences between BP 2000 and 24-h measurements systolic and diastolic BP (mm Hg) (N 43) 24-h ABPM daytime ABPM 24-h ABPM daytime ABPM Systolic Mean 1.5 5.1 0.5 4.1 SD 13.7 15.1 15.0 16.2 P.4669.0338.8115.1073 Diastolic Mean 5.2 3.1 6.0 3.8 SD 7.9 9.6 8.9 10.2 P.0001.0433.0001.0186 ABPM ambulatory blood pressure monitoring; other abbreviation as in Table 1.

790 CLINICAL EVALUATION OF A NEW WRIST DEVICE AJH September 2002 VOL. 15, NO. 9 Table 3. Influence of prior experience with BP 2000 P upon BP 2000 P systolic and diastolic blood pressure measurements (mm Hg) Period 1 Period 2 BP 2000 P BP 2000 P BP 2000 P BP 2000 P M1 2 M1 M1 2 M1 M1 2 M1 M1 2 M1 N 20 20 23 23 23 23 20 20 Systolic Mean 128.8 130.7 135.4135.9 132.3 134.0 127.3 128.3 SD 16.8 17.5 23.2 23.5 22.9 24.7 17.7 18.4 Diastolic Mean 84.2 85.3 88.4 88.8 86.0 86.9 83.2 84.1 SD 9.2 9.411.1 11.3 11.3 12.0 10.6 11.6 Abbreviation as in Tables 1 and 2. diastolic BP. No evidence of statistically significant differences for reproducibility was found between the two measurement devices. As previously, the PP analysis yielded findings similar to those of the ITT analysis. As discussed previously, the potential influence of BP measurements performed with BP 2000 P in time period 1 in comparison with measurements made in period 2 with BP 2000 P was evaluated in respect of reproducibility. Both measurement devices showed a deterioration in reproducibility in period 2 relative to period 1. Thus, prior experience with BP 2000 P had influenced reproducibility of BP measurements with BP 2000 P, and vice-versa. An important aim of BP measurement is to decide whether a subject is hypertensive or not. The 24(12)-h ambulatory BP measurement is considered as a better parameter for the diagnosis of hypertension in comparison to office readings by many researchers and may be regarded as the gold standard for diagnosis of hypertension. This was used as the reference method by which to compare other methods for correctly classifying subjects as normotensive or hypertensive. Auscultatory measurements made by an experienced investigator, and both BP 2000 devices were evaluated with respect to false-positive and false-negative decisions (Table 4). Twenty-seven hypertensive and 16 normotensive subjects were identified by 24-h ABPM (cut-off value 130/80 mm Hg). The categorization of subjects as normotensive or hypertensive, using the wrist device in subjects without specific training, achieved a correct classification of 84% (with position sensor) and 81% (without position sensor) in comparison to the ambulatory 24-h measurement as the gold standard technique. When the measurements were performed in the clinic by an experienced observer (office readings), the correct classification was 79% in comparison to ambulatory 24-h measurement. A nonresponder test was performed on day 16. There was a strong correlation between the absolute differences for systolic and diastolic BP (correlation coefficient 0.56, P.0001). In total there were five nonresponders (11.6%), all of whom failed to meet the criterion for systolic pressure measurement, and three of whom also failed the diastolic pressure criterion. If only responders (38 subjects) were included in the analysis of the correct classification of normotension/hypertension in comparison with 24-h ABPM only five subjects were misclassified according to self-measurement with BP 2000 P (13%) compared to nine subjects (24%) according to office readings. Discussion Self-measurement of BP for diagnostic and therapeutic purposes has been recommended by the World Health Organization and the International Society for High Blood Pressure, 3 and is also used increasingly in epidemiologic and pharmacologic studies. 4 7 Self-measurement has a Table 4. Evaluation of false decisions with respect to 24-h blood pressure measurements (hypertension 130 mm Hg systolic or 80 mm Hg diastolic) Office BP BP 2000 P BP 2000 P False positive decision? Yes 2 13% 0 0% 0 0% No 1487% 16 100% 16 100% False negative decision? Yes 7 26% 7 26% 8 30% No 20 74% 20 74% 19 70% Abbreviation as in Tables 1 3.

AJH September 2002 VOL. 15, NO. 9 CLINICAL EVALUATION OF A NEW WRIST DEVICE 791 similar value for the prognosis as casual screening BP measurement made on an ambulatory basis. 8 When the international consensus conference recommended selfmeasurement of BP, this only referred to studies that were made with upper arm devices. 9 Moreover, where selfmeasurement of BP has been performed in clinical studies, subjects have received training and instruction on the use of devices in specialized tertiary care centers. The majority of patients who purchase self-measurement BP devices do so at their own initiative with only a minority receiving specific training in the use of the devices. 10 The international consensus conference recommended appropriate training for patients measuring their own BP. 11 In reality, most patients are unlikely to receive training, and for the wrist sector especially, one cannot always be certain that the measuring point is correctly aligned, and that reliable and reproducible data are generated. The Braun PrecisionSensor wrist device contains an active positioning system that ensures that the measuring point of the wrist sleeve is correctly aligned. The device has been validated using the AAMI protocol SP10 and has satisfied the criteria for accuracy both for systolic and diastolic BP. 12 Apart from evaluating accuracy under controlled laboratory conditions (AAMI or British Hypertension Society protocol), wrist devices need to provide reliable measurements under routine clinical practices. However, clinical studies using wrist devices have revealed significant discrepancies in comparison to the conventional BP measurements made in medical practices, and to ambulatory 24-h BP measurement. This has resulted in the incorrect classification in up to 46% of patients. 13 The objective of the present study was to validate the Braun PrecisionSensor wrist device under routine clinical conditions using subjects with no training in self-bp measurement. The use of 24-h BP measurements on an ambulatory basis with an oscillometric upper arm device provided not only the gold standard for clinical classification of BP, but also the basis for comparing measurements made with the wrist device with and without the active positioning system. The protocol adopted in the present study incorporated recommendations made for patients measuring their own BP. 9,11 Furthermore, the device used permits automatic data storage, which improves the precision of data reporting. 7 The correlation coefficients between self-measurement at the wrist and ambulatory 24-h BP measurements were 0.73 for systolic and 0.65 for diastolic BP. These values were similar to those reported previously when comparing upper arm devices with ABPM (0.7, systolic; 0.67, diastolic 6 ). The accuracy of the wrist device without an active position system tended to be slightly poorer than the device with the position sensor. The categorization of subjects as normotensive or hypertensive, using the BP 2000 without specific training in the use of the device, achieved a correct classification of 84% in comparison to ABPM. Office readings in the clinic by a highly trained physician produced a correct classification of 79% in comparison with the ABPM. This minor difference increased if subjects with clinical relevant differences between auscultatory readings and readings with the wrist device (nonresponders) were excluded from analysis (87% v 76% correct diagnosis for BP 2000 P and office readings, respectively). Thus, the measurement of BP with the Braun PrecisionSensor wrist device by untrained subjects was considered to be equally as good as that by an experienced observer. A second aim of the study was to compare both BP 2000 devices (BP 2000 P, BP 2000 P) with respect to their reproducibility. Recently, it has been shown that the use of the active positioning system improves the reproducibility of multiple self-measurements. 14,15 However, our crossover study design did not reveal any significant variation between the two devices. We could not demonstrate a significant difference with regard to reproducibility of the two devices and this may be explained by the different study design, study population, and smaller sample size in our study. The number of subjects in the present study was too small statistically to confirm (or refute) the findings from a recent study with the BP 2000. 14 Furthermore, the inclusion of subjects in a clinical study with the need of written, informed consent may introduce considerable bias with regard to correct measurement performance. We cannot rule out that by selection bias mainly subjects with high compliance were included and this influenced the quality of readings with both BP 2000 P and BP 2000 P devices. In conclusion, self-bp measurements with the Braun PrecisionSensor wrist device, by untrained subjects, produced results consistent with those found recently with self-bp measurements with upper arm devices, when both data sets are compared with ambulatory 24-h measurements. The rates of false classification of normotension/ hypertension with the wrist device were small and as reliable as those self-measurements made in the clinic supervised by an experienced observer. The use of a position sensor slightly increased the accuracy of the device, but, in this study, the difference did not achieve statistical significance. Acknowledgment We wish to thank Braun GmbH Germany for an educational grant for the study. References 1. Palatini P, Frigo G, Bertolo O, Roman E, Corta R, Winnicki M: Validation of the A&D T-2430 device for ambulatory blood pressure monitoring and evaluation of performance according to subjects characteristics. Blood Press Monit 1998;3:255 260. 2. O Brien E, Waeber B, Parati G, Staessen J, Myers MG, on behalf of the ESH Working Group on Blood Pressure Monitoring: European Society of Hypertension recommendations on blood pressure measuring devices. BMJ 2001;322:531 536.

792 CLINICAL EVALUATION OF A NEW WRIST DEVICE AJH September 2002 VOL. 15, NO. 9 3. Guidelines Subcommittee: 1999 World Health Organization International Society of Hypertension Guidelines for the management of hypertension. J Hypertens 1999;17:151 183. 4. Jula A, Puukka P, Karanko H: Multiple clinic and home blood pressure measurements versus ambulatory blood pressure monitoring. Hypertension 1999;34:261 266. 5. Mengden T, Binswanger B, Weisser B, Vetter W: An evaluation of self-measured blood pressure in a study with a calcium-channel antagonist versus a -blocker. Am J Hypertens 1992;5:154 160. 6. Mengden T, Schwartzkopff B, Strauer BE: What is the value of home (self) blood pressure monitoring in patients with hypertensive heart disease? Am J Hypertens 1998;11:813 819. 7. Mengden T, Medina RMH, Alvarez E, Kraft K, Vetter H: Reliability of reporting self-measured blood pressure values by hypertensive patients. Am J Hypertens 1998;11:1413 1417. 8. Okhubo T, Imai Y, Tsuji I, Nagai K, Kato J, Kikuchi N, Nishiyama A, Aihara A, Sekino M, Kikuya M, Ito S, Satoh H, Hisamichi S: Home blood pressure measurement has a stronger predictive power for mortality than does screening blood pressure measurement: A population-based observation in Ohasama, Japan. J Hypertens 1998; 16:971 975. 9. Asmar R, Zanchetti A: Guidelines for the use of self-blood pressure monitoring: A summary report of the first international consensus conference. J Hypertens 2000;18:493 508. 10. Krecke H-J, Lutkes P, Maiwald M: Patient assessment of selfmeasurement of blood pressure: Results of a telephone survey in Germany. J Hypertens 1996;14:323 326. 11. Mengden T, Chamontin B, Phong Chau N, Luis Palma Gamiz J, Chanudet X: User procedure for self-measurement of blood pressure. Blood Press Monit 2000;5:111 129. 12. Wessig K, Hollinger S, Schmalzhaf I, Lenz T: Clinical evaluation of the efficacy of the Braun PrecisionSensor oscillometric wrist blood pressure monitor for use on adults versus auscultation as defined by ANSI/AAMI SP10 1992. Blood Press Monit 2000; 5:239 245. 13. Zweiker R, Schumacher M, Fruhwald FM, Watzinger N, Klein W: Comparison of wrist blood pressure measurement with conventional sphygmomanometry at a cardiology outpatient clinic. J Hypertens 2000;18:1013 1018. 14. Freund D, Wieneke P: Clinical evaluation of the performance of the BP2000 oscillometric blood pressure measurement device with and without position sensor. Biomedizinische Technik 2000;45(Suppl 1):349 350. 15. Grassi G, Foglia G, Dell Oror R, Seravalle G, Boari S, Mancia G: Reproducibility of home blood pressure monitoring by a new oscillometric wrist automatic device. J Hypertens 2001;19(Suppl 2): S22 S23.