- Email: [email protected]
Home blood pressure: accuracy is independent of monitoring schedules
AJH
2000;13:625– 631
Home Blood Pressure: Accuracy Is Independent of Monitoring Schedules Robert D. Brook
Long-term morbidity and mortality from hypertension are more closely related to home than to casual office blood pressure levels. There is no generally accepted recommendation on how to best schedule home blood pressure (HBP) recordings, perhaps because the effect of varying the home monitoring schedule on the HBP average is not well studied. The goals of this analysis are to describe the effects of HBP monitoring schedules on the accuracy of resultant HBP averages and to determine which monitoring schedule parameters correlate with HBP accuracy. Twelve published studies, each including home, office, and awake ambulatory blood pressure means were identified. Accuracy of office and HBP averages were determined by their agreement with corresponding awake ambulatory averages. Variations in HBP monitoring schedule parameters did not significantly affect the accuracy of the resultant HBP averages among the studies. In univariate analyses, no individual parameter correlated significantly with the final HBP average accuracy.
S
elf-monitoring of blood pressures by patients at home, or in the nonclinic setting, has steadily become more common during recent years.1 The advent of more accessible, affordable, and user-friendly automated blood pressure ma-
Received April 16, 1999. Accepted October 25, 1999. From the Division of Hypertension, Department of Internal Medicine, University of Michigan Hospital, Ann Arbor, Michigan. Address correspondence and reprint requests to Robert D. Brook, MD, 3918 Taubman Center, Division of Hypertension, Department of Internal Medicine, University of Michigan Hospital, Ann Arbor, MI 48109; e-mail: [email protected]
© 2000 by the American Journal of Hypertension, Ltd. Published by Elsevier Science, Inc.
As the total number of HBP readings obtained increased, or as other monitoring schedule parameters intensified, the superior accuracy of HBP levels as compared to that of casual office values also failed to significantly improve. No HBP accuracy differences were found among groups characterized by different HBP schedule parameter ranges. In conclusion, the accuracy of HBP measurements, as determined by their agreement with an awake ambulatory mean, is maintained regardless of substantial variations in HBP monitoring schedules. Therefore, the majority of the benefits derived from HBP monitoring will likely be achieved by obtaining only a few HBP measurements using a minimally complex monitoring schedule. Am J Hypertens 2000; 13:625– 631 © 2000 American Journal of Hypertension, Ltd.
KEY WORDS:
Hypertension, home blood pressure, ambulatory blood pressure.
chines has certainly helped to fuel this growth. Advantages of home blood pressure (HBP) levels over office values have been widely known for years,2–12 and have been the subject of detailed reviews on HBP monitoring.4,13 HBP readings are more reproducible2,3 and are better predictors of future blood pressure levels.8 They may improve patient compliance and blood pressure control,9 –12 lessen medical care costs,9 and eliminate the white coat effect.4 Perhaps most important, it has recently been demonstrated that HBP averages predict the risk of future cardiovascular disease and mortality, as well as target organ damage better than do traditional office levels.5–7 However, 0895-7061/00/$20.00 PII S0895-7061(99)00273-3
626
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
BROOK
TABLE 1. STUDIES USING HBP MONITORING Study 16
Mancia Sega17 Staessen-118 Staessen-219 Marloff20 Stergiou-115 Stergiou-221 Stergiou-321 Stergiou-422 Stewart23 Mengden24 Baily25 Minami26
Frequency 1 ⫻ am and pm 1 ⫻ am and pm 5 sequential readings 5 sequential readings 1 measurement am/mid-day/pm Duplicate readings am and pm 3 ⫻ week (above) (above) (above) 5 ⫻ day 1 ⫻ am and pm Triplicate readings am and pm Triplicate readings am and pm
Duration
Demographics
1 day 1 day 1 day 2 days 7 days 2 weeks 2 weeks 2 weeks 2 weeks 5 days 14 days 7 days 7 days
Sample, auto Sample, auto Sample, nurse Sample, nurse HTN, self HTN, combo HTN, auto HTN, self HTN, auto HTN, auto HTN, self HTN, self HTN, auto
Sample, patients from random population sample; HTN, patients from hypertensive population; Auto, HBP measured by automatic device; Nurse, HBP measured by nurse; Self, ausculatory HBP by self measurement.
despite these advantages, several important barriers, including the lack of consensus on HBP level definitions, fear of patient reporting bias,14 and the fact that prospective trial data are based on office blood pressures are hindering the universal acceptance of HBP measurements as important and useful adjuncts to office blood pressure values. One infrequently addressed shortcoming is the lack of a proven long-term HBP monitoring schedule that ensures both accuracy and feasibility. As a consequence, patients are measuring HBP with a highly variable frequency15 and physicians recommend very different monitoring schedules, if they recommend one at all. Even studies addressing the fundamental scientific aspects and the clinical significance of HBP measurements have used drastically different monitoring schedules (Table 1). Currently, no evidence exists to support the accuracy and feasibility of any particular HBP monitoring schedule over another. As reported by the American Society of Hypertension Ad Hoc Panel on home blood pressure monitoring,5 there have been three official position statements made in the United States about HBP monitoring. None state clear recommendations for proper usage, nor do they address the proper schedule (frequency, timing, number) of measurements. Only general recommendations were made by this panel, despite the fact that many factors involved in HBP monitoring schedules, including frequency of recordings, time of day, number of replications, season,26 work status,27 and recent activities28 could potentially affect the final levels obtained. Just as improper technique or the use of faulty monitors can negate the benefits derived from HBP monitoring,15 so too could improper monitoring schedules. Before definite recommendations on proper monitoring schedules can be made, the outcome of varying schedule parameters on HBP averages should be in-
vestigated. Therefore, the primary goal of this analysis is to determine whether variations in HBP monitoring schedules significantly affect the accuracy of resultant HBP averages. Secondary, which monitoring schedule factors, if any, that predict HBP accuracy will be determined. METHODS A literature search for English language-published studies that used all three home, clinic, and ambulatory monitoring of blood pressure was performed using the MEDLINE database. Studies were included if they presented summaries of office blood pressure averages, HBP averages, and a 12-h ambulatory blood pressure mean. Each study must have described in detail the particular HBP monitoring schedule used during the study period to be included. Two studies by Staessen et al18,19 did not include office blood pressures. Data was included from these two articles, however, because information on an important range of total HBPs was provided without any change in the analysis due to this lack of office values. Studies that specifically provided 12-h or daytime ambulatory blood pressure means were included. Those reporting only 24-h ambulatory values were not used in the analysis as the importance of the nocturnal blood pressure and its relationship with awake office or HBP levels are less clear. There were no inclusion or exclusion criteria based on patient demographics or hypertension status and any method of HBP, ambulatory, or office measurement was accepted. No other entry or exclusion criteria were used. For the purposes of this analysis, a daytime ambulatory blood pressure mean was used as the standard to represent the usual blood pressure of a patient and thus the best predictor of cardiovascular risk.29 –32 Therefore, a more accurate HBP is one that more
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
HOME BLOOD PRESSURE ACCURACY
627
TABLE 2. TRIALS INCLUDING OFFICE, HBP, AND AWAKE AMBULATORY BLOOD PRESSURES Study
HBPs
Patients
Mancia16 Sega17 Staessen-118 Staessen-219 Marloff20 Stergiou-115 Stergiou-2(O)21 Stergiou-3(A)21 Stergiou-422 Stewart23 Mengden24 Baily25 Minami-1(S)26 Minami-2(W)26
2 2 5 10 21 24 24 24 24 25 28 42 42 42
1438 800 718 328 31 50 46 46 189 85 51 42 50 50
SBP (o-h)
DBP (o-h)
SBP (h-a)
DBP (h-a)
SBP (o-a)
DBP (o-a)
SBP (od-hd)
DBP (od-hd)
8.6 9.4
7.4 5.7
3.4 6.7
1 9.4
⫺0.6 5.7
5.6 3.1 5.6 4.9 5.3 4.5 4.5 4.2 1.8 2.4
⫺4 1 0.0 ⫺4.5 0.9 0.3 ⫺1.9 ⫺1.2 ⫺0.9 ⫺1.0 0.5 0.5 1.4 1.6
4.8 20
13.2 4.1 4.9 4.4 5.4 2.5 3.0 1.3 11.0 9.9
⫺3.8 10.6 1 ⫺3 0.1 4.2 0.5 1 1.5 5.0 ⫺3.5 2.3 ⫺3.2 ⫺0.8
13.3 8.3 5.4 5.4 6.9 7.5 3 3.6 7.8 9.1
6.5 3.4 3.7 3.7 4.4 3.5 4.5 4.7 3.2 4.0
13.2 4.1 4.9 4.4 5.4 2.5 ⫺0.5 1.3 4.6 8.3
5.6 2.2 1.8 2.5 3.5 2.5 4 4.2 1.8 2.4
HBPs, total number of HBP readings per trial; patients, total number of patients per trial; O, oscillometry; A, ausculatory; S, summer; W, winter; SBP, systolic blood pressure; DBP, diastolic blood pressure; (o-h), office-HBP; (h-a), HBP-ambulatory blood pressure; (o-a), office-ambulatory blood pressure; (od-hd), (o-a)-(h-a) [(h-a) in absolute values].
closely approximates an ambulatory mean blood pressure. The absolute difference between the ambulatory mean and the HBP average was used as the first measure of HBP accuracy. A second measure of accuracy was calculated by first subtracting ambulatory blood pressure means from both the office and HBP values (expressed in absolute values). The difference between these two calculated values represents the magnitude (in mm Hg) by which HBP readings more accurately approximate ambulatory mean values than do office blood pressures. This second measurement of HBP accuracy illustrates the superiority of HBP levels over office levels in predicting the usual blood pressure of a patient (ambulatory mean). Average HBP differences from office and ambulatory means were calculated using the Microsoft Excel spreadsheet. Graphs were also performed using Excel. All statistical analyses (Pearson correlations, partial correlations, Wilcoxon nonparametric tests, and a multiple regression model) were performed using the SPSS statistical package. RESULTS Study Characteristics Twelve publications were found that fit the entry criteria. A summary of the characteristics of the different studies, including the results of the calculated measures of HBP accuracy, are presented in Tables 1 and 2. Data from two studies were presented in two separate halves, resulting in a total of 14 studies listed in Table 2. The study by Minami et al26 included separate patient data from summer and winter. One of the publications by Stergiou et al21 was divided by patients that used auto-
mated machines versus those that used traditional ausculatory technique. Study parameters and designs varied among publications (Tables 1 and 2). There were an average of 319 patients included per study. In particular, the HBP monitoring schedules used differed greatly among publications (Table 1). The average duration of home monitoring and total number of home readings obtained among all studies were 7 days and 21 recordings, respectively. HBP Characteristics In every study, the resultant HBP averages were lower than their corresponding office values. This is illustrated in Table 2 in the Office ⫺ HBP columns. Regardless of the HBP monitoring schedule used, the resultant HBP values were on average 6.5 and 4.6 mm Hg lower than the office values for systolic (SBP) and diastolic (DBP) blood pressures, respectively. The results of the first measure of HBP accuracy are presented in Table 2 under the HBP ⫺ Ambulatory Mean columns and are illustrated visually as absolute values in Figure 1. Slightly more than half (nine SBP and eight DBP) HBP levels were higher than their corresponding ambulatory means. The average unweighted absolute differences between HBP readings and ambulatory readings were only 2.9 and 1.4 mm Hg for SBP and DBP, respectively; whereas the average difference between office and ambulatory means were much higher at 7.9 mm Hg for SBP and 4.3 mm Hg for DBP. Except for the one outlying systolic difference in the study by Sega et al,17 all HBP values, both SBP and DBP, were within 5 mm Hg of their corresponding ambulatory level.
628
BROOK
FIGURE 1. Bar graph representing the absolute difference between HBP and awake ambulatory blood pressure means for each trial.
All HBP averages were closer to their corresponding ambulatory mean levels than were office averages except for one SBP and one DBP average out of 24 total blood pressure averages (SBP and DBP) (Table 2). Figure 2 demonstrates this second measure of HBP accuracy visually, as all but two bars on the graph are positive. HBP values were on average 4.5 mm Hg for SBP and 2.8 mm Hg for DBP greater than office blood pressures in predicting the ambulatory mean values, excluding the two HBP readings that were less than office values. The two studies by Staessen et al18,19 did not include clinic blood pressure averages and therefore, were not included in this second measurement of HBP accuracy. Analysis of HBP Schedules In a univariate analysis across all studies there was no significant correlation between the total number of HBP readings obtained and the first measure of HBP accuracy (HBP ⫺ Ambulatory Mean). For SBP the correlation was slightly better (R ⫽ ⫺0.38, P ⫽ .18; Pearson two-tailed) than for DBP (R ⫽ ⫺0.34, P ⫽ .23; Pearson two-tailed); however, both were nonsignificant. If the study by Sega et al17 (with the outlying SBP difference) was excluded from analysis, both correlations became even weaker and remained nonsignificant. No other schedule parameter (number of measurements per day, number of replications per measurement session, or
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
total duration of HBP monitoring) predicted the first measure of HBP accuracy. All correlations coefficients were nonsignificant and were even weaker than that for the total number of HBP readings obtained. It has been demonstrated that the differences among HBP, office, and ambulatory levels are frequently magnified in hypertensive patients compared with normotensives.14 When partial correlations adjusting for patient hypertension status were performed, all correlations among HBP accuracy and monitoring schedule parameters remained nonsignificant and became even weaker. No significant correlations were found in univariate analysis across all included studies between total HBP readings obtained to the second measure of HBP accuracy (SBP R ⫽ ⫺0.13, P ⫽ .70; DBP R ⫽ ⫺0.02, P ⫽ .95; Pearson two-tailed). No difference occurred if the trial by Sega et al17 was excluded. There were no significant correlations between the second measure of HBP accuracy to changes in any other HBP monitoring schedule parameters, whether or not they were adjusted for patient hypertension status. A recently published study22 concluded that HBP values differ between the first and second days of measurement, but not significantly during additional days thereafter. A nonparametric Wilcoxon test was performed comparing both measures of HBP accuracy between groups of studies lasting 1 day to all studies of longer duration. No significant differences in either marker of HBP accuracy were found between the two groups (SBP[h-a] P ⫽ .09; DBP[h-a] P ⫽ .35; SBP[odhd] P ⫽ 1.00; DBP[od-hd] P ⫽ 1.00). No statistically significant group differences could be found in HBP accuracy by using any breakpoint for group analysis for all schedule parameters. A multivariate regression model simultaneously taking into account the effect of all schedule parameters discussed was constructed. No schedule parameter independently predicted either measure of HBP accuracy in this model. In fact, the entire model failed to reach statistical significance for its prediction of both measurements of HBP accuracy. DISCUSSION
FIGURE 2. Bar graph illustrating the second measurement of HBP accuracy for each trial (see text for definitions).
Measuring patients’ blood pressures in the nonclinic setting, such as at home, is extremely useful in managing patients with hypertension. The many benefits of HBP measurements compared to office values have been discussed in the literature. Because no definite recommendations on an optimal HBP monitoring schedule exists, this analysis sought to determine the importance of monitoring schedules in predicting the accuracy of HBP averages. The primary finding of this study is that large and substantial HBP monitoring schedule variations among 12 previously published studies did not significantly affect the accuracy of the
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
final HBP averages. The lack of a relevant HBP monitoring schedule contribution to HBP accuracy is best exemplified by the fact that there were no significant correlations between total number of HBP readings obtained across studies to either measure of HBP accuracy. In fact, HBP accuracy was not significantly correlated with any single HBP monitoring schedule parameter (number of measurement sessions per day, replications per measurement session, or total duration of home monitoring). When comparing studies of only one day’s duration versus all other longer studies, no differences were found in both determinants of HBP accuracy. Therefore, this analysis would support the notion that HBP monitoring schedules as a whole, as well as monitoring schedule parameters individually, are not significant determinants of overall HBP accuracy. Previous studies of HBP monitoring have come to similar conclusions.2,22,33,34 In a recent publication, only eight total home recordings obtained over 2 day’s duration were needed to generate the majority of the strength in the correlation between home and ambulatory readings.22 Using test–retest correlation coefficients of 0.8 or higher between pairs of HBP measurements (recordings done at the beginning of the study to those done 2 months later) as indicative of HBP reliability, Garcia-Vera and Sanz34 found that the average of single HBP recordings performed in the evening for 3 consecutive days yields reliable HBP levels (total of three HBP recordings). Three recordings performed on the same day almost met criteria for HBP reliability. Although previous studies support the notion that relatively few HBP recordings are needed for effective HBP monitoring, the distinguishing feature of this present analysis is that it accounted for the effect of more than one schedule parameter, both individually and in combination, on the resultant HBP accuracy. Several other important differences also exist. The conclusion of Stergiou et al,22 that eight total HBP recordings taken on 2 separate days are needed to ensure HBP accuracy, is primarily due to the finding of higher HBP readings obtained on day 1 than on all other days for the remainder of the week. The four readings obtained on the second day were statistically adequate by themselves to assure HBP accuracy; however, a possible alerting response to measuring one’s HBP on day 1 necessitated recording HBP for a total of 2 days. On the contrary, this analysis found no significant alerting response, as there were no significant differences in HBP accuracy between studies of one day’s duration and ones of longer duration. This particular difference in results, as well as other minor differences between the conclusions of the present study and previous studies, can be attributed to the methodologic differences among studies by which
HOME BLOOD PRESSURE ACCURACY
629
HBP accuracy was defined or determined. Although each method, such as test–retest correlation, may have relative merits and shortcomings, comparing HBP averages to office and ambulatory means permitted the inclusion of multiple HBP schedule parameters in the assessment of HBP accuracy. The fundamental issue underlying the effect of monitoring schedule variations on the resultant HBP accuracy is whether the majority of the benefits derived from HBP values are due to them being intrinsically different from office levels and better reflections of the usual blood pressure level or simply due to the more numerous blood pressure readings obtained on more separate occasions. The former being the case, then very few HBP readings would be needed to differentiate them from office values and for them to better reflect the usual blood pressure. A well-defined schedule would be relatively unimportant to the value of HBP monitoring. The latter being the case, then a threshold number of readings would be required before HBP levels would become superior predictors of cardiovascular risk. Monitoring schedule factors would likely play a larger role in determining the value of the HBP averages. Furthermore, both of the above could be acting in consort to produce the benefits of HBP monitoring. The information gained from this analysis lends support for the former hypothesis, as even as few as two HBP recordings taken on 1 day are superior to office values in predicting ambulatory blood pressure averages. Acquiring additional HBP readings did not significantly add to the HBP accuracy. Several important shortcomings of this study need to be discussed. First, this analysis does not discount the hypothesis that an as of yet undiscovered optimal HBP monitoring schedule may exist that best predicts awake ambulatory mean blood pressures. As only 12 studies were found that fit the inclusion criteria, a limited number of HBP monitoring schedule variations could be analyzed. However, the variety of HBP monitoring schedules that were examined well encompassed the full spectrum of those used clinically. Implementing even more complex schedules, or adding to the duration of monitoring, would likely be impractical and thus clinically irrelevant. In addition, as the correlations among HBP monitoring schedule parameters to the two measures of HBP accuracy were so weak, it is doubtful that the addition of more schedule variations within this clinically feasible range would alter the outcome of this analysis. It is also conceivable that these multiple concomitant differences in schedule parameters may have disrupted the univariate analyses across studies of HBP accuracy along any one given schedule parameter. Despite this shortcoming, it is clear from observing the unanalyzed data (Table 2) that very little differences in
630
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
BROOK
HBP accuracy occurred among trials regardless of the combined effects of multiple schedule parameter variations. In addition, a multivariate regression model simultaneously accounting for the effects of all schedule parameters failed to significantly predict HBP accuracy. A final weakness of this analysis is that other unaccounted interstudy differences may have contributed to the HBP accuracy variations and thus disrupted the statistical analyses. Variations in patient HBP measurement techniques, accuracy of home devices used, time of year performed,26 as well as many other variables may have also contributed to interstudy HBP accuracy differences. Overall, it is important to stress that despite all these variables, which cannot be accurately accounted for, both measurements of HBP accuracy demonstrated very little variations among all the studies (Figures 1 and 2). The most relevant clinical message from this analysis is that the majority of the benefits derived from HBP monitoring can be achieved with very few home measurements and with minimal schedule complexity. It appears that HBP values differentiate themselves from casual office values and better predict ambulatory mean blood pressures after as little as two HBP readings. Therefore, a schedule of HBP monitoring of only a few measurements (if done properly) should provide statistically the same information as any greater amount of recordings obtained with any degree of additional monitoring schedule parameter intensity. From the results of this study, it is conceivable that as few as two HBP recordings obtained on 1 day once per month may prove to be adequate for long-term HBP monitoring. Previous studies agree with the conclusions that very few HBP readings are needed for acceptable HBP monitoring22,34; however, methodologic difference among studies prohibit direct comparisons among results. The finding of this study that only two HBP readings provide an accurate resultant HBP average, whereas additional recordings were needed in previous studies, is likely a result of these variations in study methodology and definitions of HBP accuracy. Importantly, several other benefits of HBP monitoring, such as aid in the diagnosis of orthostatic symptoms or characterization of blood pressure variability, may require more readings or additional monitoring days. Confirmation of this analysis by a study specifically designed to investigate possible HBP accuracy variations obtained by using differing HBP monitoring schedules would be helpful. However, even if variations in HBP accuracy are found, the results of this analysis would support the notion that any accuracy differences would likely be relatively small and of minor clinical importance.
REFERENCES 1.
Yarows SA: Home blood pressure monitoring in primary care. Blood Pressure Monitoring 1998;3(suppl 1): S11–S17.
2.
James GD, Pickering TG, Yees LS: The reproducibility of average ambulatory, home and clinic pressures. Hypertension 1988;11:545–549.
3.
Sakuma M, Imai Y, Nagai K, et al: Reproducibility of home blood pressure measurements over a one-year period. Am J Hypertens 1997;10:798 – 803.
4.
Pickering T, American Society of Hypertension Ad Hoc Panel: Recommendations for the use of home (self) and ambulatory blood pressure monitoring. Am J Hypertens 1995;9:1–11.
5.
Ohkubo T, Imai Y, Tsuji I, et al: 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.
6.
Abe H, Yokouchi M, Nagata S, et al: Relation of office and home blood pressure to left ventricular hypertrophy and performance in patients with hypertension. High Blood Press 1992;1:279 –285.
7.
Kleinert HD, Harshfield GA, Pickering TG, et al: What is the value of home blood pressure measurements in patients with mild hypertension? Hypertension 1984;6: 574 –578.
8.
Nesbitt SD, Amerena JV, Grant E, et al: Home blood pressure as a predictor of future blood pressure stability in borderline hypertension. The Tecumseh study. Am J Hypertens 1997;10:1270 –1280.
9.
Soghikan K, Casper SM, Fireman BH, et al: Home blood pressure monitoring: effect on use of medical services and medical care costs. Medical Care 1992;30: 855– 865.
10.
Edmonds D, Foerster E, Groth H, et al: Does selfmeasurement of blood pressure improve patient compliance in hypertension? J Hypertens 1985;3(suppl): S31–S34.
11.
Carnahan JE, Nugent CA: The effects of self-monitoring by patients on the control of hypertension. Am J Med Sci 1975;269:69 –73.
12.
Zarnke K, Feagan B, Mahon J, et al: A randomized study comparing a patient-directed hypertension management strategy with usual office-based care. Am J Hypertens 1997;10:58 – 67.
13.
Pickering TG, James GD: Some implications of the differences between home, clinic and ambulatory blood pressure in normotensive and hypertensive patients. J Hypertens 1989;7(suppl 3):S65–S72.
14.
Myers M: Self-measurement of blood pressure at home: the potential for reporting bias. Blood Press Monit 1998; 3(suppl 1):S12–S22.
15.
Stergiou GS, Malakos JS, Voutsa AV, et al: Home blood pressure monitoring: limited value in general practice. J Hum Hypertens 1996;10:219 –223.
16.
Mancia G, Sega R, Bravi C, et al: Ambulatory blood pressure normality: results from the PAMELA study. J Hypertens 1995;13:1377–1390.
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
HOME BLOOD PRESSURE ACCURACY
631
17.
Sega R, Cesana G, Milesi C, et al: Ambulatory and home blood pressure normality in the elderly Data from the PAMELA population. Hypertension 1997;30(part 1):1– 6.
26.
Minami J, Kawano Y, Ishimitsu T, et al: Seasonal variations in office, home, and 24 h ambulatory blood pressure in patients with essential hypertension. J Hypertens 1996;14:1421–1425.
18.
Staessen J, Fagard R, Lijen P, et al: Ambulatory blood pressure and blood pressure measured at home: progress report on a population study. J Cardiovasc Pharm 1994;23(suppl 5):S5–S11.
27.
Enstrom I, Pennert K: Does it matter whether ambulatory blood pressure is recorded during a work or a non-work day? J Hypertens 1996;14:565–569.
28.
19.
Staessen J, Bulpitt CJ, Fagard R, et al: Reference values for the ambulatory blood pressure and the blood pressure measured at home: a population study. J Hum Hypertens 1991;5:335–361.
Freestone S, Ramsey LE: Effect of coffee and cigarette smoking on the blood pressure of untreated and diuretic-treated hypertensive patients. Am J Med 1982;73: 348 –353.
29.
20.
Marloff AP, Hany S, Ba¨ttig B, et al: Comparison of casual, ambulatory and self-determined blood pressure measurement. Nephron 1987;47(suppl 1):142–145.
Omboni S, Ravogli A, Parati G, et al: The prognostic value of ambulatory blood pressure monitoring. J Hypertens 1991;9(suppl 3):S25–S28.
30.
21.
Stergiou GS, Voutsa AV, Achimastos AD, et al: Home self-monitoring of blood pressure: Is fully automated oscillometric technique as good as conventional stethoscope technique? Am J Hypertens 1997;10:428 – 433.
Giacono S, Levanti C, Fommei E, et al: Microalbuminuria and casual and ambulatory blood pressure monitoring in normotensives and in patients with borderline and mild essential hypertension. Am J Hypertens 1989; 2:259 –261.
22.
Stergiou GS, Skeva II, Zourbaki AS, et al: Self-monitoring of blood pressure at home: how many measurements are needed? J Hypertension 1998;16:725–731.
31.
23.
Stewart MJ, Gough K, Reid M, et al: White coat hypertension; a comparison of detection using ambulatory blood pressure monitoring or home monitoring of blood pressure (abst). J Hypertens 1996;12:1507.
Ohkubot T, Imai Y, Tsuji I, et al: Prediction of mortality by ambulatory blood pressure monitoring versus screening blood pressure measurements: a pilot study in Ohasama. J Hypertens 1997;15:357–364.
32.
Mancia G, Zanchetti A, Agabiti-Rosei E, et al: Ambulatory blood pressure is superior to clinic blood pressure in predicting treatment-induced regression of left ventricular hypertrophy. Circulation 1997;95:1464 –1470.
33.
Chatellier G, Dutrey-Dupagne C, Vaur L, et al: Home self blood pressure measurement in general practice. The SMART study. Am J Hypertens 1996;9:644 – 652.
34.
Garcia-Vera MP, Sanz J: How many self-measured blood pressure readings are needed to estimate hypertensive patients’ true blood pressure? J Behavioral Medicine 1999;22:93–113.
24.
25.
Mengden T, Binswanger B, Weisser B, et al: 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. Baily G, Ruddy MC, Malka ES, et al: Comparison of office, home, and 24-hour ambulatory blood pressures in borderline and mild hypertension. Angiology 1988; 39:752–760.
2000;13:625– 631
Home Blood Pressure: Accuracy Is Independent of Monitoring Schedules Robert D. Brook
Long-term morbidity and mortality from hypertension are more closely related to home than to casual office blood pressure levels. There is no generally accepted recommendation on how to best schedule home blood pressure (HBP) recordings, perhaps because the effect of varying the home monitoring schedule on the HBP average is not well studied. The goals of this analysis are to describe the effects of HBP monitoring schedules on the accuracy of resultant HBP averages and to determine which monitoring schedule parameters correlate with HBP accuracy. Twelve published studies, each including home, office, and awake ambulatory blood pressure means were identified. Accuracy of office and HBP averages were determined by their agreement with corresponding awake ambulatory averages. Variations in HBP monitoring schedule parameters did not significantly affect the accuracy of the resultant HBP averages among the studies. In univariate analyses, no individual parameter correlated significantly with the final HBP average accuracy.
S
elf-monitoring of blood pressures by patients at home, or in the nonclinic setting, has steadily become more common during recent years.1 The advent of more accessible, affordable, and user-friendly automated blood pressure ma-
Received April 16, 1999. Accepted October 25, 1999. From the Division of Hypertension, Department of Internal Medicine, University of Michigan Hospital, Ann Arbor, Michigan. Address correspondence and reprint requests to Robert D. Brook, MD, 3918 Taubman Center, Division of Hypertension, Department of Internal Medicine, University of Michigan Hospital, Ann Arbor, MI 48109; e-mail: [email protected]
© 2000 by the American Journal of Hypertension, Ltd. Published by Elsevier Science, Inc.
As the total number of HBP readings obtained increased, or as other monitoring schedule parameters intensified, the superior accuracy of HBP levels as compared to that of casual office values also failed to significantly improve. No HBP accuracy differences were found among groups characterized by different HBP schedule parameter ranges. In conclusion, the accuracy of HBP measurements, as determined by their agreement with an awake ambulatory mean, is maintained regardless of substantial variations in HBP monitoring schedules. Therefore, the majority of the benefits derived from HBP monitoring will likely be achieved by obtaining only a few HBP measurements using a minimally complex monitoring schedule. Am J Hypertens 2000; 13:625– 631 © 2000 American Journal of Hypertension, Ltd.
KEY WORDS:
Hypertension, home blood pressure, ambulatory blood pressure.
chines has certainly helped to fuel this growth. Advantages of home blood pressure (HBP) levels over office values have been widely known for years,2–12 and have been the subject of detailed reviews on HBP monitoring.4,13 HBP readings are more reproducible2,3 and are better predictors of future blood pressure levels.8 They may improve patient compliance and blood pressure control,9 –12 lessen medical care costs,9 and eliminate the white coat effect.4 Perhaps most important, it has recently been demonstrated that HBP averages predict the risk of future cardiovascular disease and mortality, as well as target organ damage better than do traditional office levels.5–7 However, 0895-7061/00/$20.00 PII S0895-7061(99)00273-3
626
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
BROOK
TABLE 1. STUDIES USING HBP MONITORING Study 16
Mancia Sega17 Staessen-118 Staessen-219 Marloff20 Stergiou-115 Stergiou-221 Stergiou-321 Stergiou-422 Stewart23 Mengden24 Baily25 Minami26
Frequency 1 ⫻ am and pm 1 ⫻ am and pm 5 sequential readings 5 sequential readings 1 measurement am/mid-day/pm Duplicate readings am and pm 3 ⫻ week (above) (above) (above) 5 ⫻ day 1 ⫻ am and pm Triplicate readings am and pm Triplicate readings am and pm
Duration
Demographics
1 day 1 day 1 day 2 days 7 days 2 weeks 2 weeks 2 weeks 2 weeks 5 days 14 days 7 days 7 days
Sample, auto Sample, auto Sample, nurse Sample, nurse HTN, self HTN, combo HTN, auto HTN, self HTN, auto HTN, auto HTN, self HTN, self HTN, auto
Sample, patients from random population sample; HTN, patients from hypertensive population; Auto, HBP measured by automatic device; Nurse, HBP measured by nurse; Self, ausculatory HBP by self measurement.
despite these advantages, several important barriers, including the lack of consensus on HBP level definitions, fear of patient reporting bias,14 and the fact that prospective trial data are based on office blood pressures are hindering the universal acceptance of HBP measurements as important and useful adjuncts to office blood pressure values. One infrequently addressed shortcoming is the lack of a proven long-term HBP monitoring schedule that ensures both accuracy and feasibility. As a consequence, patients are measuring HBP with a highly variable frequency15 and physicians recommend very different monitoring schedules, if they recommend one at all. Even studies addressing the fundamental scientific aspects and the clinical significance of HBP measurements have used drastically different monitoring schedules (Table 1). Currently, no evidence exists to support the accuracy and feasibility of any particular HBP monitoring schedule over another. As reported by the American Society of Hypertension Ad Hoc Panel on home blood pressure monitoring,5 there have been three official position statements made in the United States about HBP monitoring. None state clear recommendations for proper usage, nor do they address the proper schedule (frequency, timing, number) of measurements. Only general recommendations were made by this panel, despite the fact that many factors involved in HBP monitoring schedules, including frequency of recordings, time of day, number of replications, season,26 work status,27 and recent activities28 could potentially affect the final levels obtained. Just as improper technique or the use of faulty monitors can negate the benefits derived from HBP monitoring,15 so too could improper monitoring schedules. Before definite recommendations on proper monitoring schedules can be made, the outcome of varying schedule parameters on HBP averages should be in-
vestigated. Therefore, the primary goal of this analysis is to determine whether variations in HBP monitoring schedules significantly affect the accuracy of resultant HBP averages. Secondary, which monitoring schedule factors, if any, that predict HBP accuracy will be determined. METHODS A literature search for English language-published studies that used all three home, clinic, and ambulatory monitoring of blood pressure was performed using the MEDLINE database. Studies were included if they presented summaries of office blood pressure averages, HBP averages, and a 12-h ambulatory blood pressure mean. Each study must have described in detail the particular HBP monitoring schedule used during the study period to be included. Two studies by Staessen et al18,19 did not include office blood pressures. Data was included from these two articles, however, because information on an important range of total HBPs was provided without any change in the analysis due to this lack of office values. Studies that specifically provided 12-h or daytime ambulatory blood pressure means were included. Those reporting only 24-h ambulatory values were not used in the analysis as the importance of the nocturnal blood pressure and its relationship with awake office or HBP levels are less clear. There were no inclusion or exclusion criteria based on patient demographics or hypertension status and any method of HBP, ambulatory, or office measurement was accepted. No other entry or exclusion criteria were used. For the purposes of this analysis, a daytime ambulatory blood pressure mean was used as the standard to represent the usual blood pressure of a patient and thus the best predictor of cardiovascular risk.29 –32 Therefore, a more accurate HBP is one that more
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
HOME BLOOD PRESSURE ACCURACY
627
TABLE 2. TRIALS INCLUDING OFFICE, HBP, AND AWAKE AMBULATORY BLOOD PRESSURES Study
HBPs
Patients
Mancia16 Sega17 Staessen-118 Staessen-219 Marloff20 Stergiou-115 Stergiou-2(O)21 Stergiou-3(A)21 Stergiou-422 Stewart23 Mengden24 Baily25 Minami-1(S)26 Minami-2(W)26
2 2 5 10 21 24 24 24 24 25 28 42 42 42
1438 800 718 328 31 50 46 46 189 85 51 42 50 50
SBP (o-h)
DBP (o-h)
SBP (h-a)
DBP (h-a)
SBP (o-a)
DBP (o-a)
SBP (od-hd)
DBP (od-hd)
8.6 9.4
7.4 5.7
3.4 6.7
1 9.4
⫺0.6 5.7
5.6 3.1 5.6 4.9 5.3 4.5 4.5 4.2 1.8 2.4
⫺4 1 0.0 ⫺4.5 0.9 0.3 ⫺1.9 ⫺1.2 ⫺0.9 ⫺1.0 0.5 0.5 1.4 1.6
4.8 20
13.2 4.1 4.9 4.4 5.4 2.5 3.0 1.3 11.0 9.9
⫺3.8 10.6 1 ⫺3 0.1 4.2 0.5 1 1.5 5.0 ⫺3.5 2.3 ⫺3.2 ⫺0.8
13.3 8.3 5.4 5.4 6.9 7.5 3 3.6 7.8 9.1
6.5 3.4 3.7 3.7 4.4 3.5 4.5 4.7 3.2 4.0
13.2 4.1 4.9 4.4 5.4 2.5 ⫺0.5 1.3 4.6 8.3
5.6 2.2 1.8 2.5 3.5 2.5 4 4.2 1.8 2.4
HBPs, total number of HBP readings per trial; patients, total number of patients per trial; O, oscillometry; A, ausculatory; S, summer; W, winter; SBP, systolic blood pressure; DBP, diastolic blood pressure; (o-h), office-HBP; (h-a), HBP-ambulatory blood pressure; (o-a), office-ambulatory blood pressure; (od-hd), (o-a)-(h-a) [(h-a) in absolute values].
closely approximates an ambulatory mean blood pressure. The absolute difference between the ambulatory mean and the HBP average was used as the first measure of HBP accuracy. A second measure of accuracy was calculated by first subtracting ambulatory blood pressure means from both the office and HBP values (expressed in absolute values). The difference between these two calculated values represents the magnitude (in mm Hg) by which HBP readings more accurately approximate ambulatory mean values than do office blood pressures. This second measurement of HBP accuracy illustrates the superiority of HBP levels over office levels in predicting the usual blood pressure of a patient (ambulatory mean). Average HBP differences from office and ambulatory means were calculated using the Microsoft Excel spreadsheet. Graphs were also performed using Excel. All statistical analyses (Pearson correlations, partial correlations, Wilcoxon nonparametric tests, and a multiple regression model) were performed using the SPSS statistical package. RESULTS Study Characteristics Twelve publications were found that fit the entry criteria. A summary of the characteristics of the different studies, including the results of the calculated measures of HBP accuracy, are presented in Tables 1 and 2. Data from two studies were presented in two separate halves, resulting in a total of 14 studies listed in Table 2. The study by Minami et al26 included separate patient data from summer and winter. One of the publications by Stergiou et al21 was divided by patients that used auto-
mated machines versus those that used traditional ausculatory technique. Study parameters and designs varied among publications (Tables 1 and 2). There were an average of 319 patients included per study. In particular, the HBP monitoring schedules used differed greatly among publications (Table 1). The average duration of home monitoring and total number of home readings obtained among all studies were 7 days and 21 recordings, respectively. HBP Characteristics In every study, the resultant HBP averages were lower than their corresponding office values. This is illustrated in Table 2 in the Office ⫺ HBP columns. Regardless of the HBP monitoring schedule used, the resultant HBP values were on average 6.5 and 4.6 mm Hg lower than the office values for systolic (SBP) and diastolic (DBP) blood pressures, respectively. The results of the first measure of HBP accuracy are presented in Table 2 under the HBP ⫺ Ambulatory Mean columns and are illustrated visually as absolute values in Figure 1. Slightly more than half (nine SBP and eight DBP) HBP levels were higher than their corresponding ambulatory means. The average unweighted absolute differences between HBP readings and ambulatory readings were only 2.9 and 1.4 mm Hg for SBP and DBP, respectively; whereas the average difference between office and ambulatory means were much higher at 7.9 mm Hg for SBP and 4.3 mm Hg for DBP. Except for the one outlying systolic difference in the study by Sega et al,17 all HBP values, both SBP and DBP, were within 5 mm Hg of their corresponding ambulatory level.
628
BROOK
FIGURE 1. Bar graph representing the absolute difference between HBP and awake ambulatory blood pressure means for each trial.
All HBP averages were closer to their corresponding ambulatory mean levels than were office averages except for one SBP and one DBP average out of 24 total blood pressure averages (SBP and DBP) (Table 2). Figure 2 demonstrates this second measure of HBP accuracy visually, as all but two bars on the graph are positive. HBP values were on average 4.5 mm Hg for SBP and 2.8 mm Hg for DBP greater than office blood pressures in predicting the ambulatory mean values, excluding the two HBP readings that were less than office values. The two studies by Staessen et al18,19 did not include clinic blood pressure averages and therefore, were not included in this second measurement of HBP accuracy. Analysis of HBP Schedules In a univariate analysis across all studies there was no significant correlation between the total number of HBP readings obtained and the first measure of HBP accuracy (HBP ⫺ Ambulatory Mean). For SBP the correlation was slightly better (R ⫽ ⫺0.38, P ⫽ .18; Pearson two-tailed) than for DBP (R ⫽ ⫺0.34, P ⫽ .23; Pearson two-tailed); however, both were nonsignificant. If the study by Sega et al17 (with the outlying SBP difference) was excluded from analysis, both correlations became even weaker and remained nonsignificant. No other schedule parameter (number of measurements per day, number of replications per measurement session, or
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
total duration of HBP monitoring) predicted the first measure of HBP accuracy. All correlations coefficients were nonsignificant and were even weaker than that for the total number of HBP readings obtained. It has been demonstrated that the differences among HBP, office, and ambulatory levels are frequently magnified in hypertensive patients compared with normotensives.14 When partial correlations adjusting for patient hypertension status were performed, all correlations among HBP accuracy and monitoring schedule parameters remained nonsignificant and became even weaker. No significant correlations were found in univariate analysis across all included studies between total HBP readings obtained to the second measure of HBP accuracy (SBP R ⫽ ⫺0.13, P ⫽ .70; DBP R ⫽ ⫺0.02, P ⫽ .95; Pearson two-tailed). No difference occurred if the trial by Sega et al17 was excluded. There were no significant correlations between the second measure of HBP accuracy to changes in any other HBP monitoring schedule parameters, whether or not they were adjusted for patient hypertension status. A recently published study22 concluded that HBP values differ between the first and second days of measurement, but not significantly during additional days thereafter. A nonparametric Wilcoxon test was performed comparing both measures of HBP accuracy between groups of studies lasting 1 day to all studies of longer duration. No significant differences in either marker of HBP accuracy were found between the two groups (SBP[h-a] P ⫽ .09; DBP[h-a] P ⫽ .35; SBP[odhd] P ⫽ 1.00; DBP[od-hd] P ⫽ 1.00). No statistically significant group differences could be found in HBP accuracy by using any breakpoint for group analysis for all schedule parameters. A multivariate regression model simultaneously taking into account the effect of all schedule parameters discussed was constructed. No schedule parameter independently predicted either measure of HBP accuracy in this model. In fact, the entire model failed to reach statistical significance for its prediction of both measurements of HBP accuracy. DISCUSSION
FIGURE 2. Bar graph illustrating the second measurement of HBP accuracy for each trial (see text for definitions).
Measuring patients’ blood pressures in the nonclinic setting, such as at home, is extremely useful in managing patients with hypertension. The many benefits of HBP measurements compared to office values have been discussed in the literature. Because no definite recommendations on an optimal HBP monitoring schedule exists, this analysis sought to determine the importance of monitoring schedules in predicting the accuracy of HBP averages. The primary finding of this study is that large and substantial HBP monitoring schedule variations among 12 previously published studies did not significantly affect the accuracy of the
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
final HBP averages. The lack of a relevant HBP monitoring schedule contribution to HBP accuracy is best exemplified by the fact that there were no significant correlations between total number of HBP readings obtained across studies to either measure of HBP accuracy. In fact, HBP accuracy was not significantly correlated with any single HBP monitoring schedule parameter (number of measurement sessions per day, replications per measurement session, or total duration of home monitoring). When comparing studies of only one day’s duration versus all other longer studies, no differences were found in both determinants of HBP accuracy. Therefore, this analysis would support the notion that HBP monitoring schedules as a whole, as well as monitoring schedule parameters individually, are not significant determinants of overall HBP accuracy. Previous studies of HBP monitoring have come to similar conclusions.2,22,33,34 In a recent publication, only eight total home recordings obtained over 2 day’s duration were needed to generate the majority of the strength in the correlation between home and ambulatory readings.22 Using test–retest correlation coefficients of 0.8 or higher between pairs of HBP measurements (recordings done at the beginning of the study to those done 2 months later) as indicative of HBP reliability, Garcia-Vera and Sanz34 found that the average of single HBP recordings performed in the evening for 3 consecutive days yields reliable HBP levels (total of three HBP recordings). Three recordings performed on the same day almost met criteria for HBP reliability. Although previous studies support the notion that relatively few HBP recordings are needed for effective HBP monitoring, the distinguishing feature of this present analysis is that it accounted for the effect of more than one schedule parameter, both individually and in combination, on the resultant HBP accuracy. Several other important differences also exist. The conclusion of Stergiou et al,22 that eight total HBP recordings taken on 2 separate days are needed to ensure HBP accuracy, is primarily due to the finding of higher HBP readings obtained on day 1 than on all other days for the remainder of the week. The four readings obtained on the second day were statistically adequate by themselves to assure HBP accuracy; however, a possible alerting response to measuring one’s HBP on day 1 necessitated recording HBP for a total of 2 days. On the contrary, this analysis found no significant alerting response, as there were no significant differences in HBP accuracy between studies of one day’s duration and ones of longer duration. This particular difference in results, as well as other minor differences between the conclusions of the present study and previous studies, can be attributed to the methodologic differences among studies by which
HOME BLOOD PRESSURE ACCURACY
629
HBP accuracy was defined or determined. Although each method, such as test–retest correlation, may have relative merits and shortcomings, comparing HBP averages to office and ambulatory means permitted the inclusion of multiple HBP schedule parameters in the assessment of HBP accuracy. The fundamental issue underlying the effect of monitoring schedule variations on the resultant HBP accuracy is whether the majority of the benefits derived from HBP values are due to them being intrinsically different from office levels and better reflections of the usual blood pressure level or simply due to the more numerous blood pressure readings obtained on more separate occasions. The former being the case, then very few HBP readings would be needed to differentiate them from office values and for them to better reflect the usual blood pressure. A well-defined schedule would be relatively unimportant to the value of HBP monitoring. The latter being the case, then a threshold number of readings would be required before HBP levels would become superior predictors of cardiovascular risk. Monitoring schedule factors would likely play a larger role in determining the value of the HBP averages. Furthermore, both of the above could be acting in consort to produce the benefits of HBP monitoring. The information gained from this analysis lends support for the former hypothesis, as even as few as two HBP recordings taken on 1 day are superior to office values in predicting ambulatory blood pressure averages. Acquiring additional HBP readings did not significantly add to the HBP accuracy. Several important shortcomings of this study need to be discussed. First, this analysis does not discount the hypothesis that an as of yet undiscovered optimal HBP monitoring schedule may exist that best predicts awake ambulatory mean blood pressures. As only 12 studies were found that fit the inclusion criteria, a limited number of HBP monitoring schedule variations could be analyzed. However, the variety of HBP monitoring schedules that were examined well encompassed the full spectrum of those used clinically. Implementing even more complex schedules, or adding to the duration of monitoring, would likely be impractical and thus clinically irrelevant. In addition, as the correlations among HBP monitoring schedule parameters to the two measures of HBP accuracy were so weak, it is doubtful that the addition of more schedule variations within this clinically feasible range would alter the outcome of this analysis. It is also conceivable that these multiple concomitant differences in schedule parameters may have disrupted the univariate analyses across studies of HBP accuracy along any one given schedule parameter. Despite this shortcoming, it is clear from observing the unanalyzed data (Table 2) that very little differences in
630
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
BROOK
HBP accuracy occurred among trials regardless of the combined effects of multiple schedule parameter variations. In addition, a multivariate regression model simultaneously accounting for the effects of all schedule parameters failed to significantly predict HBP accuracy. A final weakness of this analysis is that other unaccounted interstudy differences may have contributed to the HBP accuracy variations and thus disrupted the statistical analyses. Variations in patient HBP measurement techniques, accuracy of home devices used, time of year performed,26 as well as many other variables may have also contributed to interstudy HBP accuracy differences. Overall, it is important to stress that despite all these variables, which cannot be accurately accounted for, both measurements of HBP accuracy demonstrated very little variations among all the studies (Figures 1 and 2). The most relevant clinical message from this analysis is that the majority of the benefits derived from HBP monitoring can be achieved with very few home measurements and with minimal schedule complexity. It appears that HBP values differentiate themselves from casual office values and better predict ambulatory mean blood pressures after as little as two HBP readings. Therefore, a schedule of HBP monitoring of only a few measurements (if done properly) should provide statistically the same information as any greater amount of recordings obtained with any degree of additional monitoring schedule parameter intensity. From the results of this study, it is conceivable that as few as two HBP recordings obtained on 1 day once per month may prove to be adequate for long-term HBP monitoring. Previous studies agree with the conclusions that very few HBP readings are needed for acceptable HBP monitoring22,34; however, methodologic difference among studies prohibit direct comparisons among results. The finding of this study that only two HBP readings provide an accurate resultant HBP average, whereas additional recordings were needed in previous studies, is likely a result of these variations in study methodology and definitions of HBP accuracy. Importantly, several other benefits of HBP monitoring, such as aid in the diagnosis of orthostatic symptoms or characterization of blood pressure variability, may require more readings or additional monitoring days. Confirmation of this analysis by a study specifically designed to investigate possible HBP accuracy variations obtained by using differing HBP monitoring schedules would be helpful. However, even if variations in HBP accuracy are found, the results of this analysis would support the notion that any accuracy differences would likely be relatively small and of minor clinical importance.
REFERENCES 1.
Yarows SA: Home blood pressure monitoring in primary care. Blood Pressure Monitoring 1998;3(suppl 1): S11–S17.
2.
James GD, Pickering TG, Yees LS: The reproducibility of average ambulatory, home and clinic pressures. Hypertension 1988;11:545–549.
3.
Sakuma M, Imai Y, Nagai K, et al: Reproducibility of home blood pressure measurements over a one-year period. Am J Hypertens 1997;10:798 – 803.
4.
Pickering T, American Society of Hypertension Ad Hoc Panel: Recommendations for the use of home (self) and ambulatory blood pressure monitoring. Am J Hypertens 1995;9:1–11.
5.
Ohkubo T, Imai Y, Tsuji I, et al: 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.
6.
Abe H, Yokouchi M, Nagata S, et al: Relation of office and home blood pressure to left ventricular hypertrophy and performance in patients with hypertension. High Blood Press 1992;1:279 –285.
7.
Kleinert HD, Harshfield GA, Pickering TG, et al: What is the value of home blood pressure measurements in patients with mild hypertension? Hypertension 1984;6: 574 –578.
8.
Nesbitt SD, Amerena JV, Grant E, et al: Home blood pressure as a predictor of future blood pressure stability in borderline hypertension. The Tecumseh study. Am J Hypertens 1997;10:1270 –1280.
9.
Soghikan K, Casper SM, Fireman BH, et al: Home blood pressure monitoring: effect on use of medical services and medical care costs. Medical Care 1992;30: 855– 865.
10.
Edmonds D, Foerster E, Groth H, et al: Does selfmeasurement of blood pressure improve patient compliance in hypertension? J Hypertens 1985;3(suppl): S31–S34.
11.
Carnahan JE, Nugent CA: The effects of self-monitoring by patients on the control of hypertension. Am J Med Sci 1975;269:69 –73.
12.
Zarnke K, Feagan B, Mahon J, et al: A randomized study comparing a patient-directed hypertension management strategy with usual office-based care. Am J Hypertens 1997;10:58 – 67.
13.
Pickering TG, James GD: Some implications of the differences between home, clinic and ambulatory blood pressure in normotensive and hypertensive patients. J Hypertens 1989;7(suppl 3):S65–S72.
14.
Myers M: Self-measurement of blood pressure at home: the potential for reporting bias. Blood Press Monit 1998; 3(suppl 1):S12–S22.
15.
Stergiou GS, Malakos JS, Voutsa AV, et al: Home blood pressure monitoring: limited value in general practice. J Hum Hypertens 1996;10:219 –223.
16.
Mancia G, Sega R, Bravi C, et al: Ambulatory blood pressure normality: results from the PAMELA study. J Hypertens 1995;13:1377–1390.
AJH–JUNE 2000 –VOL. 13, NO. 6, PART 1
HOME BLOOD PRESSURE ACCURACY
631
17.
Sega R, Cesana G, Milesi C, et al: Ambulatory and home blood pressure normality in the elderly Data from the PAMELA population. Hypertension 1997;30(part 1):1– 6.
26.
Minami J, Kawano Y, Ishimitsu T, et al: Seasonal variations in office, home, and 24 h ambulatory blood pressure in patients with essential hypertension. J Hypertens 1996;14:1421–1425.
18.
Staessen J, Fagard R, Lijen P, et al: Ambulatory blood pressure and blood pressure measured at home: progress report on a population study. J Cardiovasc Pharm 1994;23(suppl 5):S5–S11.
27.
Enstrom I, Pennert K: Does it matter whether ambulatory blood pressure is recorded during a work or a non-work day? J Hypertens 1996;14:565–569.
28.
19.
Staessen J, Bulpitt CJ, Fagard R, et al: Reference values for the ambulatory blood pressure and the blood pressure measured at home: a population study. J Hum Hypertens 1991;5:335–361.
Freestone S, Ramsey LE: Effect of coffee and cigarette smoking on the blood pressure of untreated and diuretic-treated hypertensive patients. Am J Med 1982;73: 348 –353.
29.
20.
Marloff AP, Hany S, Ba¨ttig B, et al: Comparison of casual, ambulatory and self-determined blood pressure measurement. Nephron 1987;47(suppl 1):142–145.
Omboni S, Ravogli A, Parati G, et al: The prognostic value of ambulatory blood pressure monitoring. J Hypertens 1991;9(suppl 3):S25–S28.
30.
21.
Stergiou GS, Voutsa AV, Achimastos AD, et al: Home self-monitoring of blood pressure: Is fully automated oscillometric technique as good as conventional stethoscope technique? Am J Hypertens 1997;10:428 – 433.
Giacono S, Levanti C, Fommei E, et al: Microalbuminuria and casual and ambulatory blood pressure monitoring in normotensives and in patients with borderline and mild essential hypertension. Am J Hypertens 1989; 2:259 –261.
22.
Stergiou GS, Skeva II, Zourbaki AS, et al: Self-monitoring of blood pressure at home: how many measurements are needed? J Hypertension 1998;16:725–731.
31.
23.
Stewart MJ, Gough K, Reid M, et al: White coat hypertension; a comparison of detection using ambulatory blood pressure monitoring or home monitoring of blood pressure (abst). J Hypertens 1996;12:1507.
Ohkubot T, Imai Y, Tsuji I, et al: Prediction of mortality by ambulatory blood pressure monitoring versus screening blood pressure measurements: a pilot study in Ohasama. J Hypertens 1997;15:357–364.
32.
Mancia G, Zanchetti A, Agabiti-Rosei E, et al: Ambulatory blood pressure is superior to clinic blood pressure in predicting treatment-induced regression of left ventricular hypertrophy. Circulation 1997;95:1464 –1470.
33.
Chatellier G, Dutrey-Dupagne C, Vaur L, et al: Home self blood pressure measurement in general practice. The SMART study. Am J Hypertens 1996;9:644 – 652.
34.
Garcia-Vera MP, Sanz J: How many self-measured blood pressure readings are needed to estimate hypertensive patients’ true blood pressure? J Behavioral Medicine 1999;22:93–113.
24.
25.
Mengden T, Binswanger B, Weisser B, et al: 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. Baily G, Ruddy MC, Malka ES, et al: Comparison of office, home, and 24-hour ambulatory blood pressures in borderline and mild hypertension. Angiology 1988; 39:752–760.