AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 84:343-349 (1991)
Ambulatory Blood Pressure Monitoring: Assessing the Diurnal Variation of Blood Pressure GARY D. JAMES AND THOMAS G. PICKERING Cardiovascular Center, Cornell University Medical College-New Hospital, New York, New York 10021
KEY WORDS
York
Activity changes, Emotional differences, Mood,
Occupation
ABSTRACT The purposes of this paper are to examine the effects of activity, situation of measurement, mood, and occupation on the daily variation of blood pressure and to discuss the potential utility of ambulatory blood pressure monitoring in physical anthropological studies. The subjects of the blood pressure variability study are 125 men who were referred to the Hypertension Center at New York Hospital-Cornell Medical Center for evaluation of hypertension. There were 1,386 blood pressure measurements from these subjects available for study, which were taken using noninvasive ambulatory blood pressure monitoring techniques. Pressures were transformed to z scores using the subject’s daily mean pressure and standard deviation to assess the relative elevation during the experience of the various factors. The results show that activity and mood are the most significant sources of blood pressure variation (P < .005) and are additive. Occupation, which may be an indicator of social class in this population, also modified the mood effects. Because ambulatory blood pressure monitors obtain many readings over a day under a variety of circumstances, their use can improve epidemiologicaland human biological studies of the inheritance and variability of blood pressure. Ambulatory blood pressure monitoring is an important new tool in the study of human biological variation. Over the ast decade, noninvasive ambulatory bloo pressure monitors have been developed that can reliably record changes in pressure over periods of 24 hr or longer (Pickering et al., 1985). Physiological and medical studies employing these devices have demonstrated that there is an enormous intraindividual variability of pressure over 24 hr (Pickering et al., 1986; James et al., 1990a; Pickering, 1990),some of which can be attributed to cyclical sources such as respiratory oscillations (Dornhurst et al., 1952) and possible ultradian (Shimada and Marsh, 1979) and diurnal rhythms (Clark et al., 1987) but much of which cannot (James et al., 1988a). This noncyclic variation most likely arises as a consequence of adaptive processes, so that the individual can maintain homeostasis (James and Baker, 1990). That is, as behavior changes over the day, there are corresponding blood
f
@ 1991 WILEY-LISS, INC.
pressure changes that kee blood flow to the various body tissues re atively constant. Much of the behavioral variation arises in response t o numerous environmental stressors, some of which are physical in nature, but others may be psychological (Krantz and Manuck, 1984;James et al., 1989).This variation is important in that there is some suggestive evidence that it contributes to the development of hypertension and cardiovascular events (Alderman et al., in ress). Several studies have shown t at, during the day, pressure varies by activity (Clark et al., 1987; Van Egeren and Madarasmi, 1988; Sundberg et al., 1988>,the situation of
P
K
Address correspondence to Gary D. James, Ph.D Cardiovascular Center, Cornell University Medical College-NewYork Hospital, 525 E. 68th Street, New York, NY 10021. Received March 5,1990; accepted August 13,1990.
344
G.D. JAMES AND T.G. PICKERING
measurement (Pickering et al., 1982; Harshfield et al., 1982; James et al., 19861, and emotional state and intensity (James et al., 1986). The extent of these effects may also depend on the sex of the individual (James et al., 198813).However, little is known about how mood and activity interact, over the day to affect blood pressure variability or whether factors such as occupation that define aspects of the social environment also affect daily pressure variability. The purpose of this paper is to examine the effects of activity, situation of measurement, mood, and occupation on daily blood pressure variability and to discuss the implications of ambulatory monitoring on the study of blood pressure and cardiovascular disease in physical anthropology. MATERIALS AND METHODS
Design and subjects To assess the daily variability of blood pressure with regard to the factors listed, a cross tabular approach was used in which three moods, three places of measurement, seven activities, and two occupational groups were examined. The moods were happiness, anger, and anxiety. The places of measurement included work, home, and elsewhere (places other than work and home). Some 26 activities were collapsed into seven eneral categories, and occupation was dic otomized into professional and nonprofessional groups. These occupational classifications were arbitrary but likely reflect income and other social class differences in the population studied. The classifications of activity and occupation are shown in Table 1. The subjects of the study were 125 male atients from the Hypertension Center of f;ew York Hospital-Cornell Medical Center who were referred for evaluation of hypertension. The study was limited to men to eliminate the potential confounding effects of sex on the factors examined in the study (James et al., 1988b). Similar proportions of subjects in each occupational category were measured during winter and summer months, so that seasonal effects were similar inbothgroups (James et al., 1990b).Patients who were on antih pertensive medication had it discontinue for at least 2 weeks before ambulatory monitoring. Table 2 shows some of the biologxal and demographic characteristics of the study sample. As indicated, the two groups are quite similar with regard to these characteristics.
a
d
TABLE 1. Classification of activities and occupation Professional occupations
Nonprofessional occupations
Top executive Lawyer CPA MD/PhD/DDS Engineer
Technician Salesman Musician Interior decorator Union workers
Activities Talking Writing Physical activity Eating Relaxing Dressing Traveling
TABLE 2. Selected characteristics of the sample groups Item N Age (years) Ethnic g r o w (% wGte) Alcohol (% daily drinkers) Married (W) Weight (kgs) Height (cms) 24 Hr Avg Sys BP 24 Hr Avg Dias BP
Professional
Nonurofessional
82
43 5 0 f 10 95
16
19
77 79.5 9.5 177 7 134 rt 15 89 z t 1 2
81 78.5 f 9.2 177 f 7 136 5 15 87 10
49
* 10 89 +
*
*
Techniques The study subjects arrived at the Hypertension Center in the morning (usual1 between 8 and 11AM) and were fitted wit the Spacelabs ICR (5200 model) ambulatory blood pressure monitor. This unit has been previously described and validated (Pagny et al., 1987; Santucci et al., 1988). After the monitor was fitted, it was calibrated to a standard mercury column. In this rocedure, a “ T connector was utilized, whic attached the unit to the mercury column. The automatic pum in the monitoring unit was used to inflate t e cuff on the subject’s arm, with pressure being simultaneously registered in the machine and on the mercury column. The same trained listener, using a stetholaced just distal to the recorders’ scope one, read the pressure from the colmicrop umn. This listener was blinded from the machine’s readings and determined systolic ressure at the a pearance of sound (Korot[off phase I) ancf diastolic pressure at the disappearance of sound (Korotkoff phase V). The monitor was considered accurate when five consecutive readings from it agreed to within 5 mm H of the trained listener for both systolic an diastolic pressure. The automatic timer on the monitor was reset to 15 min intervals, and, after the cayibration procedure, the men were instructed to pro-
L
K
R
K
f
345
DAILY BLOOD PRESSURE VARIATION
ceed through a normal daily routine. Each time the monitor took a reading, they were asked to remain as motionless as possible through the entire inflation and deflation cycle and then to record in a diary his activity, situation (work, home, or elsewhere), and posture (sitting, standing, or reclining). He was also asked to record on a scale of 1to 10 his degree of happiness, sadness, anger, or anxiety if he felt that he was experiencing one of these states just prior to the reading. Artifactual readings were omitted from further analysis based on previously ublished criteria (Pickerin et al., 1982).Alt ough the atients were as ed to report sadness, too Few readings were taken while subjects experienced this emotion, so they were omitted from further analysis. In addition, because posture was found to be highly colinear with activity, it was removed as a factor from the analysis. With regard to the emotional classifications, the men were not given any specific instructions for what constituted ares onse. We relied on their own judgement o what they were experiencing, allowin them to use whatever criteria they chose. T is approach was necessary, given that the stressors and environmental conditions that an individual experiences over a day cannot be known a priori. Of the ap roximately 4,500 readings obtained for tYlese men, 1,386 had activity, location, and mood classifications. There were 63 men who had recordings during reported hap iness, 35 during reported anger, and 81 uring reported anxiety. Sixtythree had readings at work, 71 had readings at home, and 56 had readings elsewhere. Finally, the number of individuals ex eriencing each activity were as follows: tal ing 71, writing 46, hysical activity 58, eating 52, relaxing 66, $ressing 8, traveling 23. To examine the variability of the readings, they were transformed to z scores using the subjects overall mean daily ressure and standard deviation with the ollowing formula:
E
a
a
P
cp
K
F
zi
=
(BPI-X,)/S,
wkere BPI = a given reading from individual i, Xi = individual i’s mean daily pressure, sl= individual i’s daily pressure standard deviation. Adjusting the readings in this manner effectively eliminated the influence of between-individual differences in the level of
blood pressure by putting all the recordings on the same relative scale. The analysis of the z scores addressed the question of why a pressure was located at a particular position in the individual’s distribution of pressures. The null hypothesis tested was that the effect of the levels of a given parameter such as mood (an er, anxiety, or happiness) were the same, wit7l the average pressure during each effect being close to zero (the mean pressure). If the pressure z scores are substantially positive or negative for a given level of a factor (such as an er), this would suggest that experiencing t at particular effect either increased or decreased pressure (relative to the individual’s mean pressure). There are potential biases with this kind of anal sis articularly related to the different num er o measurements per person and the fact that not all individuals experience all activities, moods, or situations, which we have discussed at length (James et al., 1986, 1988b, 1990b). In general, the biases are toward type I error (rejecting real relationships), such that, for example, if differences between alternative moods or activities are found, they are likely to be larger than actually calculated (James et al., 1990b). The effects of activity, mood, situation of measurement, and occupation were examined by analysis of covariance, in which time of day (which is automatically recorded by the monitor on a continuous 24 hr clock)was included as a covariate. Only the main effects of activity were examined, because there were insufficient numbers of readings to assess interactions. The main and interactive effects,however, of all other fixed factors were included in the model.
a
35 ’r!
RESULTS
Table 3 shows the analysis of covarience results for systolic pressure. As indicated, the model parameters account for about 9% of the variance in the standardized pressures. Activity, mood, and the interactions of mood with situation of measurement and occupational classification were all significant sources of variation. Table 4 shows the covariance results for diastolic pressure. Again, activity and mood are additive sources of variation, and, additionally, the interaction of mood with occupational classification contributed to the standardized diastolic pressure variation. The model parameters account for 11%of the variation in the standardized diastolic pressures.
346
G.D. JAMES AND T.G. PICKERING TABLE 3. Analysis of covariance results for systolic pressure'
Source Occupation Place of measurement Activity Mood Occupation X place Occupation X mood Place X mood Occ. X place X mood Time Error
SSE
DF
P>F
0.63 8.83 88.02 13.23 1.81 12.72 7.46 12.28 8.18 1,516.26
1 2 6 2 2 2 4 4 1 1,361
NS .02 ,0001 ,003 NS .005 .03 NS .01
Regression coefficient
-
.02
lR2 = .09.
TABLE 4. Analysis of covariance results for diastolic pressure' Source Occupation Place of measurement Activity Mood Occupation X place Occupation X mood Place X mood Occ. X place X mood Time Error
SSE
DF
P>F
1.29 4.91 153.00 29.53 4.16 10.06 7.90 8.68 4.00 1.516.26
1
2 6 2 2 2 4 4 1 1.361
NS NS ,0001
Regression coefficient
.0001
NS .03 NS NS .09
-
.015
lR2 = .11.
Table 5 shows the average systolic and diastolic z scores for the moods experienced in each situation for the professional and non rofessional men. For systolic pressure, reagngs taken during happiness were lower than those taken during anger or anxiety. In addition, there were significant differences between professionals and nonprofessionals in the effects of happiness and anxiety experienced in situations not at home or work. For diastolic pressure, readings taken durin happiness were again lower than those ta en during anger or anxiety. However, in addition, happiness has a more accentuated effect on the pressures of the nonprofessionals. Table 6 presents the average effects of the activities on the standardized systolic and diastolic blood pressures. As indicated, relative to the daily mean, hysical activity tends to elevate pressure t e most, and relaxing tends to decrease blood pressure. Final1 , although Tables 5 and 6 have shown t e impact of the various factors on the standardized pressures, it is of interest to examine the effects in actual mm Hg.
a
K
i
Table 7 shows the effects of the activities on systolic and diastolic pressure in mm Hg. These effects are calculated by multiplying the z scores presented in Table 6 by the average 24 hr standard deviation (the average intraindividual dispersion of blood pressure over the day; for systolic, the standard deviation equals 13; for diastolic, it equals 9). The specific effects of the moods and locations shown in Table 5 can also be calculated in the same manner. DISCUSSION
The results show that the proportional reduction in pressure variation associated with the model parameters (R2)was greater for diastolic (.11)than for systolic (.09) pressure. For both systolic and diastolic pressures the main effects of activity and mood were the most significant contributors to blood ressure variation. In addition, the time o day and place of measurement also affected the variation of systolic pressure as well as the interactions of mood with occupation and mood with place of measurement. These interactions were such that the pat-
F
347
DAILY BLOOD PRESSURE VARIATION
TABLE 5. Comparison of the mean effects of place of measurement and mood on the systolic and diastolic pressures in the two occupational groups Place Systolic pressure Work Home Elsewhere Diastolic pressure Work Home Elsewhere
Mood
Professional
f
Nonprofessional
Happy Angry Anxious Happy Angry Anxious Happy Angry2 Anxious
0.20 1.12 0.71 -0.10 0.57 0.16
NS NS
0.29 1.54 0.65 0.04 0.24 -0.07 1.05 -0.13 0.68
Happy Angry Anxious Happy Angry Anxious Happy Angry Anxious
0.31 1.37 0.80 -0.48 0.74 0.18 -0.01 1.01 0.94
NS
0.06
NS NS NS < ,001
1.15 1.09
< .05
-
NS
0.54 0.70 1.09 0.28 0.98 -0.13 0.48 -0.57 0.49
NS NS
< .001 NS NS < .05
-
< .05
‘Average z score of each category,indicating that the pressures associated with a given effect are, on average, the presented number of standard deviations away from an individual’s daily mean pressure. 20nly one observation among the nonprofessionals so that comparison could not be made.
diastolic pressure, the interactive effects of occupation and mood also contributed to variation in a similar pattern to systolic Activity Systolic Diastolic pressure. These findings are consistent with Talking .51 .59 other studies examining the singular effects Writing .33 67 of mood and activity on daily blood pressure Physical activity .86 .74 Eating .63 .70 variability (James et al., 1986, 1988b; Van Relaxing -.20 -.36 Egeren and Madarasmi, 1988; Clark et al., Dressing .75 .97 1987); however, the present study shows Traveling .65 .55 that these effects are additive. Furthermore, the results suggests that daily pressure variation is affected by occupational classificaTABLE 7. Mean effects of activity on blood tion, which may be an indicator of social class pressure (mm Hg)‘ in this study PO ulation. Specifically, the findings suggest t at the difference in social Activitv Svstolic Diastolic environment defined by occupation may in6.6 5.3 Talking fluence the blood pressure response to emoWriting 4.3 6.0 tional stimuli. Physical activity 11.2 6.7 Eating 8.2 6.3 There are also important implications of Relaxing -3.0 -3.2 these results in human biological and geDressing 9.8 8.7 netic epidemiological studies of blood presTraveling 8.5 5.0 sure. For example, single blood pressure ‘Calculated by multiplyingthe effects in Tahle 7 by the average measurements have been used as the esti24 hr standard deviation (systolic -13, diastolic -9). The values reflect the average number of mm Hg from the 24 hr mean mate of the “trait”blood ressure in genetic pressure e idemiologicalstudies. here is substantial p ysiological and medical evidence that a single measurement or even an average of two or three measurements taken under contern of the mood effects on pressure while trolled conditions does not accurately reflect elsewhere (not at home or work) were differ- the systemic pressure (Armitage and Rose, 1966; Beckman et al., 1981; James et al., 1988a; Pickering et al., 1988; James and Baker, 1990). Nonetheless, with single caTABLE 6. Mean effects of activity on the standardized blood pressures
R
R
f
348
G.D. JAMES AND T.G. PICKERING
sual measurements, genetic models have been developed that define both the biological and the “cultural” inheritance of blood pressure (Morton et al., 1980; Rao et al., 1976; Ward, 1990). “Cultural” inheritance includes a wide array of environmental factors. The inferences and estimates of genetic and environmental sources of blood pressure variance in these models are limited by the measurement methodology and could be significantly improved by employing ambulatory blood pressure monitoring techniques for two main reasons. First, the average ambulatory pressure has been shown to be a better reflection of systemic pressure and is more reproducible than casual measurements (James et al., 1988a).Thus the definition of the “trait” blood pressure is more precise using the ambulatory average. Second, it is also possible with ambulatory measurements to estimate the effects of a number of sources of environmental stress on pressure variability (as in this study). Thus the environmental variance can be better estimated. In addition to genetic studies, ambulatory blood pressure monitors may also be useful in human population biological studies of cardiovascular adaptation (either intra- or inter-individual); they are able to take measurements in a wide variet of naturalistic settings. By employing AN VA techniques, it is possible to estimate the effects of specific environmental stressors on blood ressure. Finally, many ambulatory blooLY pressure monitoring devices are currently available that have been validated against various standards as well as against each other (Harshfield et al., 1979, 1989; Pagny et al., 1987; Santucci et al., 1988, 1989; White, 1988; White et al., 1989; Cates et al., 1990). Because of the variety of devices and their continuous technological improvement, a protracted discussion of their operation is inappropriate for this report. Interested readers should consult the aformentioned references. However, overall, the currently available devices are quite portable (they enerally weigh about 1lb) and can be interaced with personal com uters to retrieve the data they automatical y collect and store. They run (in general) on either 4 “AA” or “C” cell batteries and are well tolerated by subjects. There are no studies in which these monitors have been used in extreme environments. However, they have been found to be fully functional across seasons when used in populations living in temperate climates
B
f
P
(Japan, western Europe, and the United States). There is clearly a great potential for these devices in enetic epidemiolo ‘cal and human biologica studies in physicaflanthropology.
f
LITERATURE CITED Alderman MH, Ooi WL, Madhavan S, and Cohen H Blood Pressure reactivity and subsequent myocardial infarction among treated hypertensive patients. J. Clin. Epidemiol. (in press). Armitage P, and Rose GA (1966) The variability of measurements of casual blood pressure. I. A laboratory study. Clin Sci 30:325-335. Beckman M, Panfilov V, Sivertsson R, Sannerstedt R, and Anderson 0 (1981)Blood pressure and heart rate recordings at home and at the clinic. Acta Med. Scand. 21@97-102. Cates EM, Schlussel YR, James GD, and Pickering TG (1990) A validation study of the Spacelabs 90207 ambulatory blood pressure monitor. J . Ambulatory Monitoring 3:149-154. Clark LA, Denby L, Pregibon D, Harshfield GA, Pickermg TG, Blank S, and Laragh JH (1987)A quantitative analysis of the effects of activity and time of day on the diurnal variations of blood pressure. J. Chron. Dis. 40:671-687. Dornhurst AC, Howard P, and Leathart GL (1952)Respiratory variations in blood pressure. Circulation 6:553-558. Harshfield GA, Hwan C, Blank SG, and Pickering TG (1989) Research tectniques for ambulatory monitorin . In N Schneidennan, SM Weiss, and PG Kaufman e s.):Handbook of ResearchMethods in Cardiovascular Behavioral Medicine. New York: Plenum Press, pp. 293309, Harshfield GA, Pickering TG, Blank S, Lindahl C, Stroud L, and Laragh J H (1979) A validation study of the Del Mar Avionics ambulatory blood pressure system. Ambulatory Electrocardiol. 1 :7-12. Harshfield GA, Pickering TG, Kleinert HD, Blank S, and Laragh J H (1982) Situational variation of blood pressure in ambulatory hypertensive patients. Psychosom. Med. 44:237-245. James GD, and Baker PT (1990) Human PO ulation biology and blood pressure: Evolutionary anaecological aspects of blood pressure. In JH Laragh and BM Brenner (eds.): Hypertension: Pathophysiology, Diagnosis, and Management. New York: Raven Press, Ltd., pp. 137-145. James GD, Crews DE, and Pearson J (1989) Catecholamines and stress. In MA Little and JD Haas (eds.): Human Population Biology: A Transdisciplinary Science. Oxford: Oxford University Press, pp. 280-295. James GD, Pickering TG, Schlussel YR, Clark LA, Denby L, and Pre ‘bon D (1990a) Measures of re ro ducibility of bloofpressure variability measuref by noninvasive ambulatory blood pressure monitors. J Ambulatory Monitoring 3:139-147. James GD, Pickering TG, Yee LS, Harshfield GA, Riva S, and Laragh JH (1988a) The reproducibility of average ambulatory, home, and clinic pressures. Hypertension 11:545-549. James GD, Yee LS, Harshfield GA, Blank SG, and Pickering TG (1986)The influence of happiness, anger and anxiety of the blood pressure of borderline hypertensives. Psychsom. Med. 48:502-508. James GD, Yee LS, Harshfield GA, and Pickering TG (1988b) Sex differences in factors affecting the daily
(2
DAILY BLOOD PRESSURE VARIATION variation of blood pressure. SOC.Sci. Med. 26:10191023. James GD, Yee LS, and Pickering TG (1990b) Wintersummer differences in the effects of emotion, osture and place of measurement on blood pressure. %c. Sci. Med. 31:1213-121 7. Krantz DS, and Manuck SB (1984) Acute sychophysio logic reactivity and risk of cardiovascurar disease: A review and methodologic critique. Psychol. Bull. 96:435464. Morton NE, Gulbrandsen CI, Rao DC, Rhoads CG, and Kagan A (1980) Determinants of blood pressure in Japanese-American families. Hum. Genet. 53:261266. Pagny J , Chatellier G, Devries C, Janod J, Corvol, and Menard J (1987) Evaluation of the Spacelabs ambulatory blood pressure recorder: Comparison with the Remler M2000. Cardiol. Rev. Rep. 8:31-36. Pickering TG (1990) Diurnal rhythms and other sources of blood pressure variabilit in normal and hypertensive subjects. In J H Larag$ and BM Brenner (eds.): Hypertension: Patho hysiology, Diagnosis, and Management. New York $awn Press, pp. 1397-1405. Pickering TG, Harshfield GA, Blank S, James GD, Laragh JH, Clark L, Denby L, and Pregibon D (1986) Behavioral determinants of 24-hour blood pressure patterns in borderline hypertension. J. Cardiol. Pharmacol. 8:S89-S92. Pickering TG, Harshfield GA, Devereux RB, and Laragh JH (1985) What is the role of ambulatory blood pressure monitoring in the management of hypertensive patients? Hypertension 7:171-177. Pickering TG, Harshfield GA, Kleinert HD, Blank S, and Laragh J H (1982) Blood pressure during normal daily activities, sleep, and exercise. Comparison of values in normal and hypertensive subjects. JAMA 247:992996.
349
Pickering TG, James GD, Boddie C , Harshfield GA, Blank SG, and Laragh J H (1988) How common is white coat hypertension? JAMA 259:225-228. Rao DC, Morton NE, and Yee S (1976) Resolution of cultural and biological inheritance by path analysis. Am. J . Hum. Genet. 28:228-242. Santucci S, Cates EM, James GD, Schlussel YR, Steiner D, and Pickering TG (1989) A comparison of two ambulatory blood pressure recorders, The Del Mar Avionics Pressurometer N and the Spacelabs 90202. Am. J. Hypertens. 2:797-799. Santucci S, Steiner D, Zimbler M, James GD, and Pickering TG (1988) A validation study of the Spacelabs 90202 and 5200 ambulatory blood pressure monitors. J Ambulatory Monitoring 1 :211-216. Shimada SG, and Marsh DJ (1979) Oscillations in mean arterial blood pressure in concious dogs. Circ. Res. 44:692-700. Sundburg S, Kohvakka A, and Gordin A (1988) Rapid reversal of circadian blood pressure rhythm in shift workers. J. Hypertens. 6:393-396. Van Egeren LF, and Madarasmi S (1988) A computer assisted diary (CAD) for ambulatory blood pressure monitoring. Am. J. Hypertens. I:179%-185S. Ward R (1990) Familial aggregation and genetic epidemiology of blood pressure. In J H Laragh and BM Brenner (eds.): Hypertension: Pathophysiology, Diagnosis, and Management. New York: Raven Press, pp. 81-100. White WB (1988) Ambulatory blood pressure monitoring: recorders and clinical applications. Comp. Ther. 14:3-9. White WB, Lund-Johansen P, and McCabe E J (1989) Clinical evaluation of the Colin ABPM 630 at rest and during exercise: An ambulatory blood pressure monitor with gas-powered cuff inflation. J. Hypertens. 7:477-483.
Ambulatory Blood Pressure Monitoring: Assessing the Diurnal Variation of Blood Pressure GARY D. JAMES AND THOMAS G. PICKERING Cardiovascular Center, Cornell University Medical College-New Hospital, New York, New York 10021
KEY WORDS
York
Activity changes, Emotional differences, Mood,
Occupation
ABSTRACT The purposes of this paper are to examine the effects of activity, situation of measurement, mood, and occupation on the daily variation of blood pressure and to discuss the potential utility of ambulatory blood pressure monitoring in physical anthropological studies. The subjects of the blood pressure variability study are 125 men who were referred to the Hypertension Center at New York Hospital-Cornell Medical Center for evaluation of hypertension. There were 1,386 blood pressure measurements from these subjects available for study, which were taken using noninvasive ambulatory blood pressure monitoring techniques. Pressures were transformed to z scores using the subject’s daily mean pressure and standard deviation to assess the relative elevation during the experience of the various factors. The results show that activity and mood are the most significant sources of blood pressure variation (P < .005) and are additive. Occupation, which may be an indicator of social class in this population, also modified the mood effects. Because ambulatory blood pressure monitors obtain many readings over a day under a variety of circumstances, their use can improve epidemiologicaland human biological studies of the inheritance and variability of blood pressure. Ambulatory blood pressure monitoring is an important new tool in the study of human biological variation. Over the ast decade, noninvasive ambulatory bloo pressure monitors have been developed that can reliably record changes in pressure over periods of 24 hr or longer (Pickering et al., 1985). Physiological and medical studies employing these devices have demonstrated that there is an enormous intraindividual variability of pressure over 24 hr (Pickering et al., 1986; James et al., 1990a; Pickering, 1990),some of which can be attributed to cyclical sources such as respiratory oscillations (Dornhurst et al., 1952) and possible ultradian (Shimada and Marsh, 1979) and diurnal rhythms (Clark et al., 1987) but much of which cannot (James et al., 1988a). This noncyclic variation most likely arises as a consequence of adaptive processes, so that the individual can maintain homeostasis (James and Baker, 1990). That is, as behavior changes over the day, there are corresponding blood
f
@ 1991 WILEY-LISS, INC.
pressure changes that kee blood flow to the various body tissues re atively constant. Much of the behavioral variation arises in response t o numerous environmental stressors, some of which are physical in nature, but others may be psychological (Krantz and Manuck, 1984;James et al., 1989).This variation is important in that there is some suggestive evidence that it contributes to the development of hypertension and cardiovascular events (Alderman et al., in ress). Several studies have shown t at, during the day, pressure varies by activity (Clark et al., 1987; Van Egeren and Madarasmi, 1988; Sundberg et al., 1988>,the situation of
P
K
Address correspondence to Gary D. James, Ph.D Cardiovascular Center, Cornell University Medical College-NewYork Hospital, 525 E. 68th Street, New York, NY 10021. Received March 5,1990; accepted August 13,1990.
344
G.D. JAMES AND T.G. PICKERING
measurement (Pickering et al., 1982; Harshfield et al., 1982; James et al., 19861, and emotional state and intensity (James et al., 1986). The extent of these effects may also depend on the sex of the individual (James et al., 198813).However, little is known about how mood and activity interact, over the day to affect blood pressure variability or whether factors such as occupation that define aspects of the social environment also affect daily pressure variability. The purpose of this paper is to examine the effects of activity, situation of measurement, mood, and occupation on daily blood pressure variability and to discuss the implications of ambulatory monitoring on the study of blood pressure and cardiovascular disease in physical anthropology. MATERIALS AND METHODS
Design and subjects To assess the daily variability of blood pressure with regard to the factors listed, a cross tabular approach was used in which three moods, three places of measurement, seven activities, and two occupational groups were examined. The moods were happiness, anger, and anxiety. The places of measurement included work, home, and elsewhere (places other than work and home). Some 26 activities were collapsed into seven eneral categories, and occupation was dic otomized into professional and nonprofessional groups. These occupational classifications were arbitrary but likely reflect income and other social class differences in the population studied. The classifications of activity and occupation are shown in Table 1. The subjects of the study were 125 male atients from the Hypertension Center of f;ew York Hospital-Cornell Medical Center who were referred for evaluation of hypertension. The study was limited to men to eliminate the potential confounding effects of sex on the factors examined in the study (James et al., 1988b). Similar proportions of subjects in each occupational category were measured during winter and summer months, so that seasonal effects were similar inbothgroups (James et al., 1990b).Patients who were on antih pertensive medication had it discontinue for at least 2 weeks before ambulatory monitoring. Table 2 shows some of the biologxal and demographic characteristics of the study sample. As indicated, the two groups are quite similar with regard to these characteristics.
a
d
TABLE 1. Classification of activities and occupation Professional occupations
Nonprofessional occupations
Top executive Lawyer CPA MD/PhD/DDS Engineer
Technician Salesman Musician Interior decorator Union workers
Activities Talking Writing Physical activity Eating Relaxing Dressing Traveling
TABLE 2. Selected characteristics of the sample groups Item N Age (years) Ethnic g r o w (% wGte) Alcohol (% daily drinkers) Married (W) Weight (kgs) Height (cms) 24 Hr Avg Sys BP 24 Hr Avg Dias BP
Professional
Nonurofessional
82
43 5 0 f 10 95
16
19
77 79.5 9.5 177 7 134 rt 15 89 z t 1 2
81 78.5 f 9.2 177 f 7 136 5 15 87 10
49
* 10 89 +
*
*
Techniques The study subjects arrived at the Hypertension Center in the morning (usual1 between 8 and 11AM) and were fitted wit the Spacelabs ICR (5200 model) ambulatory blood pressure monitor. This unit has been previously described and validated (Pagny et al., 1987; Santucci et al., 1988). After the monitor was fitted, it was calibrated to a standard mercury column. In this rocedure, a “ T connector was utilized, whic attached the unit to the mercury column. The automatic pum in the monitoring unit was used to inflate t e cuff on the subject’s arm, with pressure being simultaneously registered in the machine and on the mercury column. The same trained listener, using a stetholaced just distal to the recorders’ scope one, read the pressure from the colmicrop umn. This listener was blinded from the machine’s readings and determined systolic ressure at the a pearance of sound (Korot[off phase I) ancf diastolic pressure at the disappearance of sound (Korotkoff phase V). The monitor was considered accurate when five consecutive readings from it agreed to within 5 mm H of the trained listener for both systolic an diastolic pressure. The automatic timer on the monitor was reset to 15 min intervals, and, after the cayibration procedure, the men were instructed to pro-
L
K
R
K
f
345
DAILY BLOOD PRESSURE VARIATION
ceed through a normal daily routine. Each time the monitor took a reading, they were asked to remain as motionless as possible through the entire inflation and deflation cycle and then to record in a diary his activity, situation (work, home, or elsewhere), and posture (sitting, standing, or reclining). He was also asked to record on a scale of 1to 10 his degree of happiness, sadness, anger, or anxiety if he felt that he was experiencing one of these states just prior to the reading. Artifactual readings were omitted from further analysis based on previously ublished criteria (Pickerin et al., 1982).Alt ough the atients were as ed to report sadness, too Few readings were taken while subjects experienced this emotion, so they were omitted from further analysis. In addition, because posture was found to be highly colinear with activity, it was removed as a factor from the analysis. With regard to the emotional classifications, the men were not given any specific instructions for what constituted ares onse. We relied on their own judgement o what they were experiencing, allowin them to use whatever criteria they chose. T is approach was necessary, given that the stressors and environmental conditions that an individual experiences over a day cannot be known a priori. Of the ap roximately 4,500 readings obtained for tYlese men, 1,386 had activity, location, and mood classifications. There were 63 men who had recordings during reported hap iness, 35 during reported anger, and 81 uring reported anxiety. Sixtythree had readings at work, 71 had readings at home, and 56 had readings elsewhere. Finally, the number of individuals ex eriencing each activity were as follows: tal ing 71, writing 46, hysical activity 58, eating 52, relaxing 66, $ressing 8, traveling 23. To examine the variability of the readings, they were transformed to z scores using the subjects overall mean daily ressure and standard deviation with the ollowing formula:
E
a
a
P
cp
K
F
zi
=
(BPI-X,)/S,
wkere BPI = a given reading from individual i, Xi = individual i’s mean daily pressure, sl= individual i’s daily pressure standard deviation. Adjusting the readings in this manner effectively eliminated the influence of between-individual differences in the level of
blood pressure by putting all the recordings on the same relative scale. The analysis of the z scores addressed the question of why a pressure was located at a particular position in the individual’s distribution of pressures. The null hypothesis tested was that the effect of the levels of a given parameter such as mood (an er, anxiety, or happiness) were the same, wit7l the average pressure during each effect being close to zero (the mean pressure). If the pressure z scores are substantially positive or negative for a given level of a factor (such as an er), this would suggest that experiencing t at particular effect either increased or decreased pressure (relative to the individual’s mean pressure). There are potential biases with this kind of anal sis articularly related to the different num er o measurements per person and the fact that not all individuals experience all activities, moods, or situations, which we have discussed at length (James et al., 1986, 1988b, 1990b). In general, the biases are toward type I error (rejecting real relationships), such that, for example, if differences between alternative moods or activities are found, they are likely to be larger than actually calculated (James et al., 1990b). The effects of activity, mood, situation of measurement, and occupation were examined by analysis of covariance, in which time of day (which is automatically recorded by the monitor on a continuous 24 hr clock)was included as a covariate. Only the main effects of activity were examined, because there were insufficient numbers of readings to assess interactions. The main and interactive effects,however, of all other fixed factors were included in the model.
a
35 ’r!
RESULTS
Table 3 shows the analysis of covarience results for systolic pressure. As indicated, the model parameters account for about 9% of the variance in the standardized pressures. Activity, mood, and the interactions of mood with situation of measurement and occupational classification were all significant sources of variation. Table 4 shows the covariance results for diastolic pressure. Again, activity and mood are additive sources of variation, and, additionally, the interaction of mood with occupational classification contributed to the standardized diastolic pressure variation. The model parameters account for 11%of the variation in the standardized diastolic pressures.
346
G.D. JAMES AND T.G. PICKERING TABLE 3. Analysis of covariance results for systolic pressure'
Source Occupation Place of measurement Activity Mood Occupation X place Occupation X mood Place X mood Occ. X place X mood Time Error
SSE
DF
P>F
0.63 8.83 88.02 13.23 1.81 12.72 7.46 12.28 8.18 1,516.26
1 2 6 2 2 2 4 4 1 1,361
NS .02 ,0001 ,003 NS .005 .03 NS .01
Regression coefficient
-
.02
lR2 = .09.
TABLE 4. Analysis of covariance results for diastolic pressure' Source Occupation Place of measurement Activity Mood Occupation X place Occupation X mood Place X mood Occ. X place X mood Time Error
SSE
DF
P>F
1.29 4.91 153.00 29.53 4.16 10.06 7.90 8.68 4.00 1.516.26
1
2 6 2 2 2 4 4 1 1.361
NS NS ,0001
Regression coefficient
.0001
NS .03 NS NS .09
-
.015
lR2 = .11.
Table 5 shows the average systolic and diastolic z scores for the moods experienced in each situation for the professional and non rofessional men. For systolic pressure, reagngs taken during happiness were lower than those taken during anger or anxiety. In addition, there were significant differences between professionals and nonprofessionals in the effects of happiness and anxiety experienced in situations not at home or work. For diastolic pressure, readings taken durin happiness were again lower than those ta en during anger or anxiety. However, in addition, happiness has a more accentuated effect on the pressures of the nonprofessionals. Table 6 presents the average effects of the activities on the standardized systolic and diastolic blood pressures. As indicated, relative to the daily mean, hysical activity tends to elevate pressure t e most, and relaxing tends to decrease blood pressure. Final1 , although Tables 5 and 6 have shown t e impact of the various factors on the standardized pressures, it is of interest to examine the effects in actual mm Hg.
a
K
i
Table 7 shows the effects of the activities on systolic and diastolic pressure in mm Hg. These effects are calculated by multiplying the z scores presented in Table 6 by the average 24 hr standard deviation (the average intraindividual dispersion of blood pressure over the day; for systolic, the standard deviation equals 13; for diastolic, it equals 9). The specific effects of the moods and locations shown in Table 5 can also be calculated in the same manner. DISCUSSION
The results show that the proportional reduction in pressure variation associated with the model parameters (R2)was greater for diastolic (.11)than for systolic (.09) pressure. For both systolic and diastolic pressures the main effects of activity and mood were the most significant contributors to blood ressure variation. In addition, the time o day and place of measurement also affected the variation of systolic pressure as well as the interactions of mood with occupation and mood with place of measurement. These interactions were such that the pat-
F
347
DAILY BLOOD PRESSURE VARIATION
TABLE 5. Comparison of the mean effects of place of measurement and mood on the systolic and diastolic pressures in the two occupational groups Place Systolic pressure Work Home Elsewhere Diastolic pressure Work Home Elsewhere
Mood
Professional
f
Nonprofessional
Happy Angry Anxious Happy Angry Anxious Happy Angry2 Anxious
0.20 1.12 0.71 -0.10 0.57 0.16
NS NS
0.29 1.54 0.65 0.04 0.24 -0.07 1.05 -0.13 0.68
Happy Angry Anxious Happy Angry Anxious Happy Angry Anxious
0.31 1.37 0.80 -0.48 0.74 0.18 -0.01 1.01 0.94
NS
0.06
NS NS NS < ,001
1.15 1.09
< .05
-
NS
0.54 0.70 1.09 0.28 0.98 -0.13 0.48 -0.57 0.49
NS NS
< .001 NS NS < .05
-
< .05
‘Average z score of each category,indicating that the pressures associated with a given effect are, on average, the presented number of standard deviations away from an individual’s daily mean pressure. 20nly one observation among the nonprofessionals so that comparison could not be made.
diastolic pressure, the interactive effects of occupation and mood also contributed to variation in a similar pattern to systolic Activity Systolic Diastolic pressure. These findings are consistent with Talking .51 .59 other studies examining the singular effects Writing .33 67 of mood and activity on daily blood pressure Physical activity .86 .74 Eating .63 .70 variability (James et al., 1986, 1988b; Van Relaxing -.20 -.36 Egeren and Madarasmi, 1988; Clark et al., Dressing .75 .97 1987); however, the present study shows Traveling .65 .55 that these effects are additive. Furthermore, the results suggests that daily pressure variation is affected by occupational classificaTABLE 7. Mean effects of activity on blood tion, which may be an indicator of social class pressure (mm Hg)‘ in this study PO ulation. Specifically, the findings suggest t at the difference in social Activitv Svstolic Diastolic environment defined by occupation may in6.6 5.3 Talking fluence the blood pressure response to emoWriting 4.3 6.0 tional stimuli. Physical activity 11.2 6.7 Eating 8.2 6.3 There are also important implications of Relaxing -3.0 -3.2 these results in human biological and geDressing 9.8 8.7 netic epidemiological studies of blood presTraveling 8.5 5.0 sure. For example, single blood pressure ‘Calculated by multiplyingthe effects in Tahle 7 by the average measurements have been used as the esti24 hr standard deviation (systolic -13, diastolic -9). The values reflect the average number of mm Hg from the 24 hr mean mate of the “trait”blood ressure in genetic pressure e idemiologicalstudies. here is substantial p ysiological and medical evidence that a single measurement or even an average of two or three measurements taken under contern of the mood effects on pressure while trolled conditions does not accurately reflect elsewhere (not at home or work) were differ- the systemic pressure (Armitage and Rose, 1966; Beckman et al., 1981; James et al., 1988a; Pickering et al., 1988; James and Baker, 1990). Nonetheless, with single caTABLE 6. Mean effects of activity on the standardized blood pressures
R
R
f
348
G.D. JAMES AND T.G. PICKERING
sual measurements, genetic models have been developed that define both the biological and the “cultural” inheritance of blood pressure (Morton et al., 1980; Rao et al., 1976; Ward, 1990). “Cultural” inheritance includes a wide array of environmental factors. The inferences and estimates of genetic and environmental sources of blood pressure variance in these models are limited by the measurement methodology and could be significantly improved by employing ambulatory blood pressure monitoring techniques for two main reasons. First, the average ambulatory pressure has been shown to be a better reflection of systemic pressure and is more reproducible than casual measurements (James et al., 1988a).Thus the definition of the “trait” blood pressure is more precise using the ambulatory average. Second, it is also possible with ambulatory measurements to estimate the effects of a number of sources of environmental stress on pressure variability (as in this study). Thus the environmental variance can be better estimated. In addition to genetic studies, ambulatory blood pressure monitors may also be useful in human population biological studies of cardiovascular adaptation (either intra- or inter-individual); they are able to take measurements in a wide variet of naturalistic settings. By employing AN VA techniques, it is possible to estimate the effects of specific environmental stressors on blood ressure. Finally, many ambulatory blooLY pressure monitoring devices are currently available that have been validated against various standards as well as against each other (Harshfield et al., 1979, 1989; Pagny et al., 1987; Santucci et al., 1988, 1989; White, 1988; White et al., 1989; Cates et al., 1990). Because of the variety of devices and their continuous technological improvement, a protracted discussion of their operation is inappropriate for this report. Interested readers should consult the aformentioned references. However, overall, the currently available devices are quite portable (they enerally weigh about 1lb) and can be interaced with personal com uters to retrieve the data they automatical y collect and store. They run (in general) on either 4 “AA” or “C” cell batteries and are well tolerated by subjects. There are no studies in which these monitors have been used in extreme environments. However, they have been found to be fully functional across seasons when used in populations living in temperate climates
B
f
P
(Japan, western Europe, and the United States). There is clearly a great potential for these devices in enetic epidemiolo ‘cal and human biologica studies in physicaflanthropology.
f
LITERATURE CITED Alderman MH, Ooi WL, Madhavan S, and Cohen H Blood Pressure reactivity and subsequent myocardial infarction among treated hypertensive patients. J. Clin. Epidemiol. (in press). Armitage P, and Rose GA (1966) The variability of measurements of casual blood pressure. I. A laboratory study. Clin Sci 30:325-335. Beckman M, Panfilov V, Sivertsson R, Sannerstedt R, and Anderson 0 (1981)Blood pressure and heart rate recordings at home and at the clinic. Acta Med. Scand. 21@97-102. Cates EM, Schlussel YR, James GD, and Pickering TG (1990) A validation study of the Spacelabs 90207 ambulatory blood pressure monitor. J . Ambulatory Monitoring 3:149-154. Clark LA, Denby L, Pregibon D, Harshfield GA, Pickermg TG, Blank S, and Laragh JH (1987)A quantitative analysis of the effects of activity and time of day on the diurnal variations of blood pressure. J. Chron. Dis. 40:671-687. Dornhurst AC, Howard P, and Leathart GL (1952)Respiratory variations in blood pressure. Circulation 6:553-558. Harshfield GA, Hwan C, Blank SG, and Pickering TG (1989) Research tectniques for ambulatory monitorin . In N Schneidennan, SM Weiss, and PG Kaufman e s.):Handbook of ResearchMethods in Cardiovascular Behavioral Medicine. New York: Plenum Press, pp. 293309, Harshfield GA, Pickering TG, Blank S, Lindahl C, Stroud L, and Laragh J H (1979) A validation study of the Del Mar Avionics ambulatory blood pressure system. Ambulatory Electrocardiol. 1 :7-12. Harshfield GA, Pickering TG, Kleinert HD, Blank S, and Laragh J H (1982) Situational variation of blood pressure in ambulatory hypertensive patients. Psychosom. Med. 44:237-245. James GD, and Baker PT (1990) Human PO ulation biology and blood pressure: Evolutionary anaecological aspects of blood pressure. In JH Laragh and BM Brenner (eds.): Hypertension: Pathophysiology, Diagnosis, and Management. New York: Raven Press, Ltd., pp. 137-145. James GD, Crews DE, and Pearson J (1989) Catecholamines and stress. In MA Little and JD Haas (eds.): Human Population Biology: A Transdisciplinary Science. Oxford: Oxford University Press, pp. 280-295. James GD, Pickering TG, Schlussel YR, Clark LA, Denby L, and Pre ‘bon D (1990a) Measures of re ro ducibility of bloofpressure variability measuref by noninvasive ambulatory blood pressure monitors. J Ambulatory Monitoring 3:139-147. James GD, Pickering TG, Yee LS, Harshfield GA, Riva S, and Laragh JH (1988a) The reproducibility of average ambulatory, home, and clinic pressures. Hypertension 11:545-549. James GD, Yee LS, Harshfield GA, Blank SG, and Pickering TG (1986)The influence of happiness, anger and anxiety of the blood pressure of borderline hypertensives. Psychsom. Med. 48:502-508. James GD, Yee LS, Harshfield GA, and Pickering TG (1988b) Sex differences in factors affecting the daily
(2
DAILY BLOOD PRESSURE VARIATION variation of blood pressure. SOC.Sci. Med. 26:10191023. James GD, Yee LS, and Pickering TG (1990b) Wintersummer differences in the effects of emotion, osture and place of measurement on blood pressure. %c. Sci. Med. 31:1213-121 7. Krantz DS, and Manuck SB (1984) Acute sychophysio logic reactivity and risk of cardiovascurar disease: A review and methodologic critique. Psychol. Bull. 96:435464. Morton NE, Gulbrandsen CI, Rao DC, Rhoads CG, and Kagan A (1980) Determinants of blood pressure in Japanese-American families. Hum. Genet. 53:261266. Pagny J , Chatellier G, Devries C, Janod J, Corvol, and Menard J (1987) Evaluation of the Spacelabs ambulatory blood pressure recorder: Comparison with the Remler M2000. Cardiol. Rev. Rep. 8:31-36. Pickering TG (1990) Diurnal rhythms and other sources of blood pressure variabilit in normal and hypertensive subjects. In J H Larag$ and BM Brenner (eds.): Hypertension: Patho hysiology, Diagnosis, and Management. New York $awn Press, pp. 1397-1405. Pickering TG, Harshfield GA, Blank S, James GD, Laragh JH, Clark L, Denby L, and Pregibon D (1986) Behavioral determinants of 24-hour blood pressure patterns in borderline hypertension. J. Cardiol. Pharmacol. 8:S89-S92. Pickering TG, Harshfield GA, Devereux RB, and Laragh JH (1985) What is the role of ambulatory blood pressure monitoring in the management of hypertensive patients? Hypertension 7:171-177. Pickering TG, Harshfield GA, Kleinert HD, Blank S, and Laragh J H (1982) Blood pressure during normal daily activities, sleep, and exercise. Comparison of values in normal and hypertensive subjects. JAMA 247:992996.
349
Pickering TG, James GD, Boddie C , Harshfield GA, Blank SG, and Laragh J H (1988) How common is white coat hypertension? JAMA 259:225-228. Rao DC, Morton NE, and Yee S (1976) Resolution of cultural and biological inheritance by path analysis. Am. J . Hum. Genet. 28:228-242. Santucci S, Cates EM, James GD, Schlussel YR, Steiner D, and Pickering TG (1989) A comparison of two ambulatory blood pressure recorders, The Del Mar Avionics Pressurometer N and the Spacelabs 90202. Am. J. Hypertens. 2:797-799. Santucci S, Steiner D, Zimbler M, James GD, and Pickering TG (1988) A validation study of the Spacelabs 90202 and 5200 ambulatory blood pressure monitors. J Ambulatory Monitoring 1 :211-216. Shimada SG, and Marsh DJ (1979) Oscillations in mean arterial blood pressure in concious dogs. Circ. Res. 44:692-700. Sundburg S, Kohvakka A, and Gordin A (1988) Rapid reversal of circadian blood pressure rhythm in shift workers. J. Hypertens. 6:393-396. Van Egeren LF, and Madarasmi S (1988) A computer assisted diary (CAD) for ambulatory blood pressure monitoring. Am. J. Hypertens. I:179%-185S. Ward R (1990) Familial aggregation and genetic epidemiology of blood pressure. In J H Laragh and BM Brenner (eds.): Hypertension: Pathophysiology, Diagnosis, and Management. New York: Raven Press, pp. 81-100. White WB (1988) Ambulatory blood pressure monitoring: recorders and clinical applications. Comp. Ther. 14:3-9. White WB, Lund-Johansen P, and McCabe E J (1989) Clinical evaluation of the Colin ABPM 630 at rest and during exercise: An ambulatory blood pressure monitor with gas-powered cuff inflation. J. Hypertens. 7:477-483.
