Ambulatory Blood Pressure
Normality and Comparison With Other Measurements
Abstract—Previous studies have reported results on 24-hour ambulatory blood pressure (ABP) in Europe and Japan, but no data exists from South America. In this study, we conducted a population survey to identify reference values and to compare ambulatory blood pressure with clinic, home, and self-measured values. A random sample of 2650 adults was selected among 190 000 people covered by our prepaid healthcare institution. Clinic (physician and nurse) and home (nurse) blood pressure measurements were performed 3 times each, with semiautomatic electronic equipment. Self-measurements were performed by the subjects manually activating the ambulatory device. We analyzed 1573 individuals who were not receiving antihypertensive therapy from 1921 participants. Self-measurement was available in a subgroup of 577 participants younger than the whole sample. Normal ambulatory blood pressure limits were estimated as those that best correlated with 140/90 mm Hg at clinic. Estimated values were 125/80 mm Hg for 24-hour ambulatory (range: 122 to 128 and 77 to 83 mm Hg) and 129/84 mm Hg for daytime ambulatory (range: 127 to 132 and 81 to 86) blood pressure, depending on gender and age. Ambulatory and clinic blood pressures increased with age. The age-dependent increase in ABP was similar in women and men. Average blood pressure at clinic was 124/79 mm Hg by physician and 123/78 mm Hg by nurse. Nurse measurement at home was 125/78 mm Hg, daytime ambulatory was 121/77 mm Hg, and 24-hour ambulatory was 118/74 mm Hg. The values of the subgroup with self-measurement were physician 119/77 mm Hg; nurse at clinic 118/77 mm Hg; nurse at home 121/78 mm Hg; self-measured 115/72 mm Hg; daytime ambulatory 119/77 mm Hg; and 24-hour ambulatory 115/73 mm Hg. This study shows that a 24-hour ABP average value of 125/80 mm Hg and a daytime ABP average value of 129/84 mm Hg are suitable upper limits for normality. Higher limits would yield an artificially higher prevalence of white coat hypertension. Most subjects showed higher blood pressure levels when measurements were performed by healthcare personnel at a clinic or at home than when self-measured at home.
- blood pressure monitoring, ambulatory
- reference values
- hypertension, white coat
- blood pressure monitoring
We will not have definitive reference values for ambulatory blood pressure (ABP) until all the prospective studies evaluating its relation with subsequent cardiovascular events are completed.1 Meanwhile, reference values of ABP are needed for patient care. Several studies have proposed normal values for ABP populations in Japan2 and in Europe (Irish,3 Belgian,4 Italian PAMELA,5 and Danish6 studies), but there are not data that relate to the South America population.
Evidence of the risks of arterial hypertension and the benefits of its treatment are based on casual blood pressure (BP) measurements.7 8 Nevertheless, BP can be measured in different environments (clinic, home, and workplace), by different personnel (physician, nurse, or patient), and with a conventional sphygmomanometer or ABP devices. This introduces the problem of the correspondence among BP values measured in these different ways and situations, with its practical implications in the usual care of normotensive and hypertensive individuals. Correlation among ABP and the remaining BP measurement methods have been published in the above referred studies, but no one has compared the BP measured by a physician at clinic with BP measured by a nurse at clinic or at home, self-measured BP, and 24-hour ABP in the same sample. The purpose of this study was to estimate normal values of 24-hour ABP in an Uruguayan population and to perform a direct comparison of ABP results with clinic (physician and nurse), home, and self-measured BP values.
The Asociación Española Primera de Socorros Mutuos is a nonprofit mutual medical care institution located in Montevideo, Uruguay. The covered population is ≈190 000 individuals, roughly representative of the 1 345 000 Montevideo inhabitants. The study was planned and performed in the Department of Cardiology. The study protocol was approved by the Institutional Ethics Committee. In October 1995, we randomly selected within 6 age strata a sample of 2650 subjects >19 years of age with no upper age limit. The selected subjects were invited to participate in the survey by telephone and/or by letter. When the study was finished, in June 1998, we had recruited 2070 of them (78%). All the patients gave informed consent to participate in the study.
Blood Pressure Measurements
The clinic BP measurements by physician and nurse and the home BP measurement by nurse were performed from Monday to Friday, after 8:00 am and before 7:00 pm, according to standard recommendations,9 10 with a previously validated semiautomatic blood pressure measuring machine (Omron HEM-705CP).11 Subjects refrained from smoking, drinking coffee or tea, or using other stimulants ≥30 minutes before arriving at the outpatient clinic. In the clinic, in a quiet warm office, the individual was seated with the back resting against the chair and the arm bared and supported at heart level. After 5 minutes of rest, the first BP reading was performed. During blood pressure measurement, the individual was instructed not to speak. A nurse or physician performed the measurement 3 times each, with ≥2 minutes between each measurement. Pulse frequency, BP, and the time of each measurement were registered. After a few minutes, BP was measured by a physician if the first set of measurements were performed by a nurse or vice versa. Clinic BP measurements were performed first by a physician in 41.4% of the visits and by a nurse in 58.6% of the visits, with ≈35 minutes between both sets of measurements. During the clinic visit, the date and time of the home visit was arranged. At home, BP was measured by a nurse 3 times after sitting at rest for 5 minutes, with ≥2 minutes between each measurement. ABP monitoring was performed with Spacelab 90207 devices. The devices were attached before individuals left the clinic and were programmed to measure BP at 20-minute intervals between 7:00 am to 11:00 pm and at 40-minutes interval during the night. Individuals were instructed to keep a diary of their asleep and awake times. The ABP study was accepted if ≥20 valid measurements were recorded, otherwise it was repeated. Self-measurement was obtained twice at home in a subset of 577 individuals (from among the last 1716 recruited subjects) by manual self-activated ABP machine in the early morning and in the evening after ≥5 minutes of rest in the sitting position. This measurement was incorporated to our study when we knew the design of the PAMELA study.
Body weight and height were registered, and a 12-lead ECG was performed. A set of standard questions related to the knowledge, habits, and attitudes of the individuals and their physicians was asked by nurses. Clinical records were screened by physicians searching for previous BP values and possible drug treatments. Some of this data will be reported elsewhere.
Age was stratified into 10-year span strata. The first stratum covered individuals 20 to 29 years of age, and the last one comprised individuals >69 years of age. The casual BP (clinic physician, clinic nurse, and home nurse) values used for calculations are averages of 3 successive measurements. The self-measured values are averages of a morning and an evening measurement. Conventional descriptive statistics for continuous variables and proportions were calculated when appropriate and used throughout the study for results presentation. Student’s t test for independent samples was applied to assess differences between strata. Student’s t test for differences was applied to compare results of different measurement methods, and bivariate Pearson correlation matrices were calculated to obtain the correlation coefficients among them. Bivariate linear correlation-regression was performed to assess the relationship between the results of diverse ABP measurement results and the physician at clinic BP values that were taken as reference ones. This technique was also used to estimate the age-related variations of BP. Diagnostic break points in correspondence with 140/90 mm Hg physician at clinic BP were calculated for the various outputs of ABP measurement as the predicted values, which were obtained by the following equations: where aiS and aiD are the intercepts and biS and biD are the slopes of the best fit straight lines found in the regression analysis performed on the different ABP outputs (identified by the subscript i) versus physician at clinic BP for SBP and DBP, respectively (Figure⇓). SEM of the estimate for the true regression line is shown as a reference on the incertitude of each predicted value. The referred procedure was executed on the whole sample and in the self-measurement subset on the whole set and in subsets defined by age and gender .
Casual BP (clinic physician, clinic nurse, and home nurse) and ABP were available from 1921 participants, 72% of the randomized sample. Three hundred forty-eight subjects of 1921 participants were taking antihypertensives. These participants were excluded from the analysis to avoid potential confounding effects. Then, the analyzed sample included 1573 subjects. The demographics of the population were as follows: age 47.4 years (SD 15.2, 20 to 88), female gender 58.0%, BMI 26.5 kg/m2 (SD 4.6, 17 to 48), years of education 11.6 (SD 4.5, 0 to 21), smokers 24.6%, former smokers 15.1%, diabetes 5.5%, and hyperlipidemia 17.6%.
In a subgroup of 577 subjects, self-measured BP was also available. Compared with the whole sample, this subgroup was younger (40.5 years), had fewer women (52.8%), and had similar BMI (26.5 kg/m2).
Blood Pressure Findings
Casual Blood Pressure
Casual systolic blood pressure (SBP) showed a sustained increase through increasing age strata independent of gender. Casual diastolic blood pressure (DBP) increased in men 59 years of age (0.51 mm Hg/y, 95% confidence limits [CLs]: 0.42 to 0.6) and in women 69 years of age (0.32 mm Hg/year, 95% CLs 0.27 to 0.37), P<0.0001 in both cases. We observed a slight decline in DBP in older individuals, which was steeper in men (−0.42 mm Hg/y, 95% CLs: −0.65 to 0.19; P<0.0002 versus −0.04 mm Hg/y, P=NS in women). At all age strata, casual BP was higher in men than women (P<0.05), except those individuals >69 years (P=NS). Physician and nurse at clinic BPs were significantly different, but the difference was clinically unimportant. When the physician took the first set of readings, the average BP values were 126/78 mm Hg by physician and 121/76 mm Hg (lower) by nurse (P<0.0001). When the nurse took the first set of readings, the average values were 125/79 mm Hg by nurse and 123/78 mm Hg (lower) by physician (P<0.0001). We found small although significant (P<0.05) differences between clinic BP (nurse or physician) and nurse at home BP.
Ambulatory Blood Pressure
When systolic and diastolic ABP were incremented as a function of age, this increase was not as steep as that observed with casual BP; 0.36 (95% confidence limits: 0.32 to 0.40) versus 0.75 mm Hg/y (95% confidence limits: 0.69 to 0.81) SBP and 0.17 (95% confidence limits: 0.14 to 0.20) versus 0.24 mm Hg/y DBP (95% confidence limits: 0.20 to 0.28), P<0.005 in both. In all cases, the Pearson bivariate correlation between clinic physician BP, clinic nurse BP, home nurse BP, 24-hour ABP, and daytime ABP was >0.75 for SBP and >0.65 for DBP (P<0.0001 in all). Nighttime ABP also significantly correlated with r values >0.64 for SBP and >0.54 for DBP. In the group as a whole, casual SBP and DBP were significantly higher than 24-hour ABP, with a higher difference observed in men (8 to 10 mm Hg) than in women (5 to 7 mm Hg), P<0.05. Daytime ABP was also lower than casual BP (P<0.05). In women, this effect was not observed or, if it was, it was clinically unimportant. When we analyzed the differences between casual and ABP by gender and age strata, we observed similar systolic values (or clinically unimportant differences) in women <50 years of age and in men <40 years of age and similar diastolic values in women <40 years of age and in men <30 years of age. Larger differences were observed in women >49 years of age and in men >39 years of age, with the differences higher in men (P<0.05). Table 3⇓ shows the predicted values of ABP that correspond to 140/90 mm Hg of physician BP as estimated by first-degree linear regression analysis when the individuals were classified by age and gender. The normal upper limits of mean 24-hour ABP in the whole sample (the predicted values that corresponded to the traditional break points of 140/90 mm Hg) were 125/80 mm Hg (95% CLs: 124.5 to 125.5/79.6 to 80.4 mm Hg), with variations between 122 to 128 mm Hg for SBP and 77 to 83 mm Hg for DBP related to gender and age. The normal upper limits of daytime ABP in the whole sample were 129/84 mm Hg (95% CLs: 128.4 to 129.6/83.3 to 84.7 mm Hg), with age and gender-related variations between 127 to 132 mm Hg for SBP and 81 to 86 mm Hg for DBP.
Self-Measurement of Blood Pressure
The observed BP values in the subgroup with self-measurement are shown in Tables 4 (global) and 5 (by gender and age). Self-measured BP and 24-hour ABP were very close (except in individuals >69 years of age, in which only 5 individuals performed the self-measurement). Differences between casual BP, self-measured BP, and 24-hour ABP were smaller in women than in men (P<0.05) and smaller than the lower BP values (P<0.05). In all cases, the Pearson bivariate correlation between clinic physician BP, clinic nurse BP, home nurse BP, self-measured BP, 24-hour ABP, and daytime ABP was >0.60 for SBP and >0.50 for DBP (P<0.001 in all). In addition, nighttime ABP significantly correlated with r values >0.64 for SBP and >0.59 for DBP. Self-measured BP and 24-hour ABP were significantly correlated with r>0.79 for SBP and r>0.69 for DBP (P<0.001 in both). Table 6⇓ compares our results with those from other ABP population studies.
Physician measured BP≥140/90 mm Hg was present in 26.1% of the sample (as stated, subjects under antihypertensive therapy were excluded in this study). Prevalence rose to 33.2% if calculated on the 2070 surveyed subjects (BP≥140/90 mm Hg or antihypertensive therapy).
In this study of 1573 randomly selected individuals from Uruguay, who were not receiving antihypertensive therapy, we observed lower 24-hour ABPs than casual clinic BPs as measured by physician and by nurse and at home by nurse. We found that the 24-hour ABP average value of 125/80 mm Hg (95% CLs: 124.5 to 125.5/79.6 to 80.4 mm Hg) is the upper limit for discrimination between normal and hypertensive individuals, which was 140/90 mm Hg on casual BP measurements. Because this study had a large enough sample size in the highest age stratum, we were able to extend our findings to include elderly individuals. Within a subset of 577 individuals who performed self-measurement of BP at home, these values were similar to 24-hour ABP. Daytime ABP was intermediate between self-measured and casual BP.
Casual SBP increased with age independent of gender. Casual DBP increased at the ages 69 and 59 years in women and men, respectively; after these ages, we observed a decline. In agreement with other studies, men had significantly higher casual BPs than women,6 except in individuals >69 years of age. We observed no difference in clinic BP whether it was measured by physician (124/79 mm Hg) or by nurse (123/78 mm Hg), but we observed BP differences related to the order in which measurements were performed. When BP was first measured by a physician, the mean BP values were higher than those obtained by the nurse (126/78 versus 121/76 mm Hg); when the first measurement was performed by a nurse, the mean BP values were slightly higher than those read by physician (125/79 versus 123/78 mm Hg). Thus, the effect of the order in which the BP measurements are performed could be more important that the role of the person who performs them. These results are different than those observed in other studies. This could be explained by different environmental conditions (early evidence on the white coat phenomenon was obtained through measurements performed at a hospital),12 social prestige of physicians and nurses in different countries, and nontypical ambulatory medical practice conditions in this survey. BP measurements at home by nurses showed slightly higher values than those obtained at clinic by physician and by nurse, mainly SBP in young people, but these differences are not clinically important. ABP and home values were similar to those observed in the Belgian study.
We found that the 24-hour ABP average value of 125/80 mm Hg (95% CLs: 124.5 to 125.5/79.6 to 80.4) is the upper limit for discrimination between normal and hypertensive individuals that best correlates with 140/90 mm Hg on casual BP measurements. These ABP normality values that ranged from 122 mm Hg (younger women) to 128 mm Hg (older men) for SBP and 77 mm Hg (20 to 29 years) to 83 mm Hg (50 to 59 years men) for DBP are shown in Table 3⇑. Daytime limits were 127 to 131 mm Hg for SBP and 81 to 86 mm Hg for DBP. To assess the relationship between casual BP and ABP, we choose 140/90 mm Hg as a well-established upper limit for clinic BP normality, which was used in the PAMELA study. With the use of these upper limits of normal ABP, white coat hypertension was present in 23.3% of our sample. If 135/85 mm Hg was used as the normal upper limit, the prevalence of white coat hypertension would rise to 50.2%.
We observed ABP values identical to those of the PAMELA study, and ABP values almost identical to those shown in other ABP population studies (Table 6⇑), except for the Danish study. In the Danish study, a higher 24-hour SBP was observed, which is possibly related to the cold climate and older age of participants. In the Japanese study, the individuals, although older, showed a lower BMI than that found in the subjects of other studies including ours. Other small differences could be explained by different ages or gender proportions in the studies. In the PAMELA study, the clinic BP was slightly higher than our observations; this could be because in the Italian study, BP was only measured in the clinic in morning when, according to known circadian patterns, BP reaches high values. This small difference could also be due to the different proportion of individuals receiving antihypertensive therapy among the samples, which was higher in our population (348/1921) than in the PAMELA study (213/1651), P<0.0001. We found that 24-hour ABP values were lower than casual BP values in subjects, except in younger subjects, in whom similar values, or clinical unimportant differences, were observed. Daytime and nighttime periods for ABP analyses were defined according to the patient’s diary, instead of arbitrarily established periods, and this could explain the small differences observed with the other population studies. In contrast with results observed in the PAMELA study, in which differences between 24-hour ABP and self-measured increased with age, we did not observed such increases, but our data could be biased by the underrepresentation of older individuals in the self-measurement subgroup.
Self-measured BP was lower than home BP measured by nurse and very close to 24-hour ABP, which was similar to results reported in the PAMELA study. The self-measurement subgroup was younger than the entire population; this fact could explain the lower BP values observed and the differences between casual, self-measured, and 24-hour ABP. The self-measurement was performed when the participants activated the ambulatory device. This procedure has the advantage of eliminating the interdevice sources of error. Another potential bias source is related to inter-arm differences in BP, which was also eliminated in our study. Nevertheless, the procedure used for self-measurement of BP had the disadvantage of being under used by older subjects.
This is the first ABP population study that performed a direct comparison between physician at clinic BP, nurse at clinic BP, nurse at home BP, 24-hour ABP, daytime ABP, and self-measured BP (the last one in a subgroup). This direct comparison in a large study shows that clinic BP is similar when measured by nurse or by physician, and the differences between them are related to the sequence of measurements and not to the person that performs the measurement. This comparison also shows that home BP measured by nurse is similar to clinic BP and higher than self-measured BP. These results are clinically relevant because they mean that nurse or physician BP readings are equivalent for usual practice. Normal BP values of 140/90 mm Hg are adequate for nurse readings at home but too high for self-measured BP. Our data show that in an adult randomly selected population, in which individuals receiving antihypertensive therapy are excluded, a 24-hour ABP upper limit of normality is 125/80 mm Hg, which is similar to self-measured BP and lower than physician or nurse measured BP.
In conclusion, our study shows that 24-hour ABP average values of 125/80 mm Hg and daytime ABP average values of 129/84 mm Hg are suitable upper limits for normality. Higher limits would yield an artificially higher prevalence of white coat hypertension. Most subjects showed higher blood pressure levels when measurements were performed by healthcare personnel at clinic or at home than when self-measured at home. Our data on the relationships between casual BP and ABP readings confirm most results published to date and show the marked similarity between groups from all over the world. This data should be considered in everyday practice when several BP measurement methods are available. The present population is now being further followed to assess normal ABP values, taking into account the incidence of cardiovascular events.
We thank the nurses and staff from the Department of Cardiology for their valuable collaboration and Dr. Emma Schwedt for her critical revision of the manuscript. We thank the Board of Regents of the Asociación Española Primera de Socorros Mutuos for the encouragement and financial support for this study.
- Received May 8, 1999.
- Revision received June 15, 1999.
- Accepted July 26, 1999.
Ohkubo T, Imai Y, Tsuji I, Nagai K, Ito S, Satoh H, Hisamichi S. Reference values for 24-hour ambulatory blood pressure monitoring based on a prognostic criterion: the Ohasama Study. Hypertension. 1998;32:255–259.
Imai Y, Nagai K, Sakuma M, Sakuma H, Nakatsuka H, Satoh H, Minami N, Munakata M, Hashimoto J, Yamagishi T, Watanabae N, Yabe T, Nishiyama A, Abe K. Ambulatory blood pressure of adults in Ohasama, Japan. Hypertension. 1993;22:900–912.
Staessen J, Fagard Rh, Lijnen PJ, Thijs L, Van Hulle S, Vyncke G, Amery A. Ambulatory blood pressure and blood pressure measurement at home: progress report on a population study. J Cardiovasc Pharmacol. 1994;23(suppl 5):S5–S11.
MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, Godwin J, Dyer A, Stamler J. Blood pressure, stroke, and coronary heart disease, I: prolonged differences in blood pressure: prospective observational studies corrected for the regression dilution bias. Lancet. 1990;335:765–774.
Collins R, Peto R, MacMahon S, Hebert P, Fiebach NH, Eberlein KA, Godwin J, Qizilbash N, Taylor JO, Hennekens CH. Blood pressure, stroke, and coronary heart disease, II: short-term reductions in blood pressure: overview of randomized drug trials in their epidemiological context. Lancet. 1990;335:827–838.
Perloff D, Grim C, Flack J, Frohlich ED, Hill M, McDonald M, Morgenstern BZ. Human blood pressure determination by sphygmomanometry. Circulation. 1993;88(pt 1):2460–2470.
Grupo de Hipertensión. Asociación Española. Validación de un equipo electrónico para medir la presión arterial. Rev Urug Cardiol. 1996;11:90–93.
Mancia G, Parati G, Pomidosi G, Grassi G, Casadei R, Zanchetti A. Alerting reaction and rise in blood pressure during measurement by physician and nurse. Hypertension. 1987;9:209–215.