Original article 333 Optimum frequency of office blood pressure measurement using an automated sphygmomanometer Martin G. Myers a, Miguel Valdivieso a and Alexander Kiss b,c Objective To determine the optimum interval between serial blood pressure measurements using an automated BpTRU sphygmomanometer. Methods Two groups of 2 patients each had automated office measurements taken using the BpTRU device at either 1-min or 2-min intervals from the start of one reading to the start of the next reading with a 24-h ambulatory blood pressure (ABP) recording being performed. Another series of 5 patients had BpTRU readings taken at 1-min and 2-min intervals before and after 24-h ABP monitoring. The difference between the mean awake ABP and the mean automated office BP readings were compared for recordings taken at 1-min versus 2-min intervals. Results In the between-patient comparison (n = 4), mean awake ABP was similar to automated BP recordings in the examining room at either 1-min or 2-min intervals except for a slightly lower ( 4 mmhg) diastolic BP with the 1-min interval (P <.1 vs. ABP). In the within-patient comparison (n = 5), there was no consistent difference between automated BP readings taken in the examining room at 1-min versus 2-min intervals. Overall, the mean automated BP values tended to be slightly lower than the mean awake ABP. Conclusion Automated measurement of BP in the office setting with devices such as the BpTRU can be taken as frequently as every 1 min without affecting the accuracy of the reading. Small differences in BP between the 1 and 2-min settings and between the automated BpTRU and ABP readings were within accepted clinical standards for validation criteria. Blood Press Monit 13:333 338 c 28 Wolters Kluwer Health Lippincott Williams & Wilkins. Blood Pressure Monitoring 28, 13:333 338 Keywords: automated sphygmomanometers, blood pressure measurement a Division of Cardiology, Schulich Heart Centre, b Institute for Clinical and Evaluative Sciences, Sunnybrook Health Sciences Centre and c Department of Medicine, University of Toronto, Ontario, Canada Correspondence to Dr Martin G. Myers, MD, FRCPC, Sunnybrook Health Sciences Centre, A-22, 275 Bayview Avenue, M4N 3M5 Toronto, Ontario, Canada Tel: + 41 648 4749; fax: + 41 648 544; e-mail: martin.myers@sunnybrook.ca Received 3 April 28 Revised 26 June 28 Accepted 27 June 28 Introduction Automated blood pressure (BP) measurement is now recognized as the best predictor of cardiovascular events related to hypertension [1]. Out-of-office BP such as home BP recordings and 24-h ambulatory BP monitoring (ABPM) have been incorporated into algorithms [2] for making a diagnosis of hypertension with the latter considered to be the gold standard for classifying the BP status of patients. In routine clinical practice, it is much more practical to monitor an individual s BP status in the office or clinic than elsewhere, but conventional BP readings using manual methods are well known to be subject to different types of measurement error. The determination of BP using mercury or aneroid devices in the office may be affected by either observer or patient factors, or a combination of the two. As a consequence, office readings may be higher than during usual daily activities (white coat hypertension) and reproducibility of BP from visit-to-visit may be relatively poor, making it difficult to diagnose hypertension or monitor a patient s response to drug therapy. In 21, a validated automated sphygmomanometer, the BpTRU (BpTRU Medical Devices, Ltd., Coquitlam, British Columbia, Canada), became available for routine use in the office or clinic [3]. The BpTRU was specifically designed to minimize measurement error by eliminating observer patient interaction. This device can be set to take readings at 1 5 min intervals between the start of one reading and the start of the next reading with the patient sitting quietly alone in the examining room. The first reading is taken by the observer (either physician or other health professional) using the BpTRU to verify that the cuff has been placed in the correct position so that the reading is valid. Most often, the device is set to take measurements at either 1-min or 2-min intervals and the observer then leaves the room. After 5 or min, the BpTRU displays the mean of five readings, heart rate and individual BP readings for verification purposes. Currently, no data on the optimum frequency for recording BP with an automated device such as the BpTRU are observed. Both the 1-min and 2-min settings have been used in research studies 1359-5237 c 28 Wolters Kluwer Health Lippincott Williams & Wilkins DOI:.97/MBP.b13e32834247
334 Blood Pressure Monitoring 28, Vol 13 No 6 and routine clinical practice. The 1-min setting is popular with clinicians because it minimizes the time required to measure BP but could theoretically increase BP by causing more arm discomfort. In this study, we have compared readings taken at either 1 or 2-min intervals to the mean awake ambulatory BP (ABP) to determine if recording BP as frequently as every minute has any effect on the measurement of BP. Methods Patient population BP readings taken at 1-min and 2-min intervals were evaluated in two series of patients using either a between-patient comparison (n = 4) or a withinpatient comparison (n = 5). All patients were referred to an ABPM centre with each patient having office BP recorded using the automated BpTRU device at 1-min or 2-min intervals followed by 24-h ABPM. The 5 patients who participated in the within-patient part of the study had BpTRU readings taken at both 1-min and 2-min intervals either before or after 24 h ABPM, with the order of the 1-min and 2-min recordings being randomized. Before initiating the ABP recording, patients were seated in a quiet examining room and the BP cuff of the BpTRU device was attached to the upper arm. An initial reading was obtained and then automatically discarded. For 2 consecutive patients, this device was set to take readings at 1-min intervals. For the next 2 patients, readings were performed at 2-min intervals. The timing of each setting is measured from the start of one reading to the start of the next reading. For the 1-min setting, there is 15 3 s between readings when the BP is lower and virtually no time between readings in patients with higher BP levels. Upon completion of five BpTRU readings, the mean BP and heart rate value were noted and the patient then underwent 24 h ABPM using a SpaceLabs model 927 (Spacelabs Healthcare Ltd., Issaquah, Washington, USA) device. Patients were asked to engage in usual daily activities and to return the ABPM recorder the next morning. The ABP monitor was set to take readings at 15 min intervals during awake hours (6: 22: h) and at 3 min intervals during the night. Mean awake ABP for data analysis was determined using individual diaries, which gave the precise time spent awake. The mean awake ABP was compared with the BpTRU readings taken during the visit to the ABPM unit. The primary outcome measure for the between-patient comparison was the difference between the automated office BP and the mean awake ABP with readings taken at 1-min versus 2-min intervals with differences being displayed using a Bland Atman-type plot [4] showing the individual data points. These differences were categorized as systolic BP of less than, to less than 2 and greater than or equal to 2 mmhg, and diastolic BP of less than 5, 5 to less than and greater than or equal to 2 mmhg. Differences between automated office BP versus ABP were also computed for hypertensive patients defined as a systolic BP greater than or equal to 135 mmhg and/or a diastolic BP greater than or equal to 85 mmhg using readings obtained with the BpTRU device. Mean BP values were also compared for the office versus ambulatory readings and Pearson s coefficients of correlation (r) values were calculated to examine the relationship between the office BP at 1-min and 2-min intervals versus the mean awake ABP. Differences in mean BP between the automated readings taken at 1-min versus 2- min intervals in the 5 patients who had readings at both settings were compared using repeated measures analysis of variance. Results Between-patient comparison No significant differences were observed in clinical characteristics in the between-patient analysis between the 1-min versus 2-min subgroups: mean age 63 ± 14 versus 63 ± 15, number of males/ number of females 94/6 versus 9/1, number of treated patients/ number of untreated patients 129/71 versus 8/92, respectively. In the two groups of 2 patients, mean BpTRU values were within 4 mmhg of the mean awake ABP regardless of whether readings were taken at one or two minute intervals (Table 1). Differences between the BpTRU versus mean ABP were identical for systolic BP ( 2.2 mmhg) for both the 1-min and 2-min groups. Diastolic BP taken with the BpTRU at 1-min intervals were significantly (P <.1) lower ( 3.8 ± 8.4 mmhg) versus the ABP. Diastolic BP taken at 2-min intervals were similar (.9 ± 8.1 mmhg) to the ABP. Table 1 Mean ( ± SD) BP (mmhg) readings taken in the ABPM unit are compared with the mean awake ABP in two separate groups of 2 patients and for patients with hypertension (mean awake ambulatory systolic BP Z 135 and/or diastolic BP Z 85 mmhg) Systolic BP Diastolic BP Interval between BP readings Office ABP Office ABP All patients (n) One minute (2) 133 ± 2 135 ± 13 72 ± 11* 76 ± Two minutes (2) 132 ± 19 134 ± 13 76 ± 12 77 ± 11 Hypertensive patients (n) One minute (4) 139 ± 22** 144 ±11 76 ± 12*** 82 ± Two minutes () 14 ± 19* 14 ± 12 81 ± 12* 83 ± *P <.5; **P <.1; ***P <.1 versus APB. ABMP, ambulatory blood pressure monitoring; ABP, ambulatory blood pressure; BP, blood pressure.
Frequency of BP measurement Myers et al. 335 Comparative BP values for patients with hypertension based on a mean awake ABP greater than or equal to 135/85 are shown in Table 1. Systolic and diastolic BP values taken with the BpTRU device were significantly lower than the awake ABP for both the 1-min and 2-min settings, with differences ranging from less than 1 mmhg up to 6 mmhg. Differences between the individual BpTRU readings taken at 1 or 2-min intervals and the corresponding mean awake ABP were plotted against the mean ABP (Figs 1 and 2). For the 1-min versus 2-min groups, the number of patients with systolic BP differences of less than, to less than 2 and greater than or equal to 2 mmhg were 161, 27, 12 versus 165, 19, 16, respectively. Similarly, for diastolic BP, differences between the BpTRU and ABP for values less than 5, 5 to less than and greater than or equal to mmhg for 1-min versus 2-min groups were 166, 27, 7 versus 15, 33, 17, respectively. Pearson s coefficients of correlation between the BpTRU and mean awake ABP readings for systolic/diastolic BP were r =.64/ r =.693 and r =.631/r =.757 for the 1-min and 2-min groups, respectively. None of the differences in coefficients of correlation between groups were statistically significant. Within-patient comparison Mean automated office BP measurements taken at 1-min intervals were similar to the values obtained at 2-min intervals with the systolic BP being 2 mmhg higher for the 1-min setting and the diastolic 1 mmhg higher for readings taken every 2 min (Table 2). The automated systolic and diastolic office BP readings recorded at either 1-min or 2-min intervals were 2 5 mmhg lower than the mean awake ABP. For the 1-min versus the 2-min readings, the number of patients with systolic BP differences between the BpTRU and ABP values of less than, 19 and greater than or equal to 2 mmhg were 2, 22 and 8 versus 24, 21 and 5, respectively. Similarly, for diastolic BP, differences between the BpTRU and ABP values of less than 5, 5 to less than and greater than or equal to mmhg for 1-min versus 2-min readings were 21, 24 and 5 versus 3, 12 and 8, respectively. Pearson s coefficients of correlation between the mean awake systolic/diastolic ABP and automated readings taken at either 1-min intervals (r =.695/r =.689) were similar to the correlation between the ABP and automated readings taken at 2-min intervals (r =.717/ r =.82). Discussion Automated BP taken in the office or clinic may be performed at either 1-min or 2-min intervals without affecting the relationship between these measurements and the mean awake ABP. In the between-patient comparison, differences in systolic BP between the office and ABP were similar regardless of whether readings were taken at 1 or 2-min intervals using the automated BpTRU device. Diastolic BP readings with the automated office BP were slightly lower when taken at 1-minute intervals. Percentages of patients with hypertension were similar in the 1 and 2-min groups. No consistent differences were observed between the automated readings taken with the BpTRU in the same patients at 1-min versus 2-min intervals. The automated office systolic was slightly lower using the 1-min setting whereas the diastolic BP was slightly higher. Overall, in the between-patient analysis, differences between the office readings and ABP tended to be small with 81 and 83% of 1-min and 2-min interval patients having differences in systolic BP less than mmhg and 83 and 75% having differences in diastolic BP less than 5 mmhg. Subgroup analysis did not show any influence of BP status (hypertension vs. normal BP) on the findings. In particular, readings in hypertensive patients were not increased by the shorter 1-min office BP interval. If anything, BpTRU readings using the 1-min setting tended to be slightly lower. This finding is noteworthy in that the 1-min interval between the start of one reading and the start of the next in hypertensive patients leaves virtually no period of rest between deflation of the cuff and inflation for the next reading, especially if the large adult cuff size is used. Any arm discomfort caused by the more frequent BP measurements using the 1-min setting does not seem to affect the BP reading. A similar finding was noted in a preliminary communication by Eguchi et al. [5] which reported no differences in the mean of three consecutive readings taken at intervals of either s or of 1 min from the end of one reading to the start of the next. The authors concluded that a s interval between multiple BP readings is as accurate as a 1-min interval for evaluating a patient s blood pressure. Culleton et al. [6] examined BpTRU readings taken at 5 min intervals and observed significantly lower systolic ( 9.7 mmhg) and diastolic BP ( 3.9 mmhg) values compared with mean awake ABP. Use of the 5-min interval raised concern that BP recordings over a 25 min period might be more appropriately classified as basal and not comparable with the casual BP, generally used to assess BP status. Furthermore, a 25 min period for determining BP in routine clinical practice would be impractical. The only other study [7] comparing office BP and ABP in a large population showed a close relationship between
336 Blood Pressure Monitoring 28, Vol 13 No 6 Fig. 1 (a) 6 5 BpTRU minus mean awake ambulatory systolic BP 4 3 2 2 3 4 5 1 12 13 14 15 16 17 18 19 Mean awake ambulatory systolic BP (mmhg) (b) 35 3 BpTRU minus mean awake ambulatory diastolic BP 25 2 15 5 5 15 2 25 5 6 7 8 9 1 Mean awake ambulatory diastolic BP (mmhg) Individual data for mean awake ambulatory systolic (panel a) and diastolic (panel b) in 2 patients are plotted against the difference between the BpTRU readings and the ambulatory BP with the readings taken at 1-min intervals. the automated office BP (14/8 mmhg) and the mean awake ABP (142/8 mmhg). In this study, readings were taken at either 1 or 2-min intervals, each after a 5-min period of rest. The authors did not report separate results for readings taken at 1-min versus 2-min intervals. It is noteworthy that the mean diastolic BP readings taken with the BpTRU device and the mean awake ambulatory diastolic BP were identical.
Frequency of BP measurement Myers et al. 337 Fig. 2 (a) 6 5 BpTRU minus mean awake ambulatory systolic BP 4 3 2 2 3 4 5 1 12 13 14 15 16 17 18 19 Mean awake ambulatory systolic BP (mmhg) (b) 35 3 BpTRU minus mean awake ambulatory diastolic BP 25 2 15 5 5 15 2 25 5 6 7 8 9 1 Mean awake ambulatory systolic BP (mmhg) Individual data for mean awake ambulatory systolic (panel a) and diastolic (panel b) in 2 patients are plotted against the difference between the BpTRU readings and the ambulatory BP with the readings taken at 2-min intervals. In a recently completed BP survey involving 238 residents of the Province of Ontario [8], BP readings were taken in a quiet examining room with both the BpTRU and standard mercury sphygmomanometer. For the BpTRU set at 1-min intervals, mean BP taken with patients while alone were 3/3 mmhg lower than with the mercury device. This difference is consistent with a reduction in the white coat response when the BpTRU
338 Blood Pressure Monitoring 28, Vol 13 No 6 Table 2 Mean ( ± SD) awake (mmhg) ABP readings are compared with BpTRU readings (mmhg) taken in the same 5 patients at 1-min and 2-min intervals either before or after performing ABPM BP measurement Systolic BP Diastolic BP One minute interval 135 ± 16 74 ± 11*** Two minute interval 133 ± 17** 75± 12* Awake ABP 138 ± 13 77 ± 12 Versus awake ABP: *P =.5; **P =.1; ***P =.2. ABMP, ambulatory blood pressure monitoring; ABP, ambulatory blood pressure; BP, blood pressure. is used. For a 3-mmHg fall in systolic BP, the diastolic BP would be expected to be about 2 mmhg lower than with the mercury sphygmomanometer. The 3-mmHg difference in the diastolic BP seen with the BpTRU does not support a clinically important lower diastolic BP when readings are taken with the BpTRU at 1-min intervals. Protocols developed to validate the accuracy of automated BP recording devices [9,] such as the BpTRU and SpaceLabs 927 used in this study consider differences of up to 5 mmhg between the test device and the standard mercury sphygmomanometer to be within acceptable limits, given the recognized variability in BP measurement. Thus, automated devices considered to be equivalent for the recording of BP in clinical practice are not expected to produce identical BP readings when used in the same patient, even under standardized conditions. The differences between readings obtained with the BPTRU and SpaceLabs recorders in this study fall within the 5-mmHg range, although the size of the population sample resulted in some of the values being statistically significant. In summary, automated systolic BP readings taken at 1 or 2-min intervals in the office using the BpTRU device were similar to each other and to the mean awake ABP. The demonstration that repeated measurement of BP over a 5-min period does not adversely affect the accuracy of readings should encourage the use of automated BP measurement in routine clinical practice. Acknowledgements This study was supported in part by grant #ESA 5745 from the Heart and Stroke Foundation of Ontario. The authors thank Jacqueline Headley for secretarial assistance and Marko Katic for performing the data analysis References 1 Myers MG. Ambulatory blood pressure monitoring for routine clinical practice. Hypertension 25; 45:483 484. 2 Hemmelgarn BR, McAlister FA, Myers MG, et al; for the Canadian Hypertension Education Program: the 25 Canadian hypertension Education Program (CHEP) recommendations for the management of hypertension. Part 1: blood pressure measurement, diagnosis and assessment of risk. Can J Cardiol 25; 21:645 656. 3 Wright JM, Mattu GS, Perry TL Jr, Gelfer ME, Strange KD, Zorn A, Chen Y. Validation of new algorithm for the BPM- electronic oscillometric office blood pressure monitor. Blood Press Monit 21; 6:161 165. 4 Bland JM, Altman DG. Statistics Notes: some examples of regression towards the mean. Br Med J 1994; 39:78 789. 5 Eguchi K, Kuruvilla S, Ogedegbe G, Gerin W, Schwartz JE. Very short measurement interval between multiple readings is useful during home blood pressure (BP) measurement. J Hypertension 28; 26 (Suppl 1):S9. 6 Culleton BF, McKay DW, Campbell NR. Performance of the automated BpTRU measurement device in the assessment of white-coat hypertension and white-coat effect. Blood Press Monit 26; 11:37 42. 7 Beckett L, Godwin M. The BpTRU automatic blood pressure monitor compared to 24 h ambulatory blood pressure monitoring in the assessment of blood pressure in patients with hypertension. BMC Cardiovasc Disord 25; 5:18. 8 Myers MG, McInnis NH, Fodor GJ, Leenen FHH. Comparison between an automated and manual sphygmomanometer in a population survey. Am J Hypertens 28; 21:28 282. 9 O Brien E, Pickering T, Asmar R, Myers M, Parati G, Staessen J, et al. With the statistical assistance of Atkins N and Gerin W on behalf of the 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 22; 7:3 17. Association for the Advancement of Medical Instrumentation. American national standard: electronic or automated sphygmomanometers. Arlington, VA: AAMI; 1993.