Ambulatory Physical Activity as a Determinant of Diurnal Blood Pressure Variation
Abstract—There are reports that indicate that diurnal blood pressure (BP) variation, in addition to high BP per se, is related to target organ damage and the incidence of cardiovascular events. However, the determinants of diurnal BP variation are not adequately understood. We used actigraphy and ambulatory BP monitoring to study the diurnal variation of BP and physical activity in 160 adults. Within individuals, activity was more strongly related to pulse rate than to BP. The correlation between BP and activity was stronger during sleep than when awake, but the correlation between activity and pulse rate was higher during the awake period than during sleep. Between individuals, the sleep/awake ratio of systolic BP (SBP) was correlated with mean sleep activity (r=.17, P<0.05), mean awake activity (r=−0.16, P<0.05), and, especially, the ratio of sleep/awake activity (r=.24, P<0.01). Awake BP variability (SD of awake SBP) was positively correlated with awake activity (r=.16, P<0.05). In regard to the effect of position, the standing-supine SBP difference was negatively correlated with the sleep/awake SBP ratio (r=−0.39, P<0.01) and positively correlated with awake SBP variability (r=.33, P<0.01). When we divided the subjects into 3 groups, 19 extreme dippers (with a sleep SBP decrease of ≥20% of awake SBP), 102 dippers (with decreases of ≥10% to <20%), and 39 nondippers (with decreases of <10%), no significant differences existed in awake activity among the groups. However, the nondippers exhibited greater sleep activity than extreme dippers (P<0.05) and an increased sleep/awake activity ratio compared with extreme dippers and dippers (P<0.01). Extreme dipping may also be associated with increased BP variability (P=0.08). Individual SBP responses to activity (the within-person slope of awake SBP regressed on activity) did not differ significantly among the 3 subgroups. In conclusion, physical activity is one of the determinants of ambulatory BP and its diurnal variation. We hypothesize that the association of sleep activity to sleep BP and dipping reflects differences in sleep quality.
Ambulatory BP monitoring (ABPM) has been used to show that both high BP levels and abnormal patterns of diurnal BP variation carry increased risk for cardiovascular disease in hypertensive patients. It has been proposed that nondippers, who exhibit a diminished nocturnal BP fall, constitute a subgroup whose abnormal diurnal BP variation is associated with increased rates of damage to all target organs (brain, heart, and kidney) and poorer prognosis for cardiovascular events when compared with dippers who exhibit a normal nocturnal BP fall.1 2 3 4 We recently identified a second potential high risk group among dippers: the subgroup of extreme dippers who have an exaggerated nocturnal BP fall. We found that among elderly hypertensive patients, extreme dippers have more marked cerebrovascular damage than dippers.5
However, the mechanism of these abnormal diurnal BP variation patterns in hypertensive patients remains unclear.6 The nondipper pattern has been reported in secondary hypertensive patients with endocrine abnormalities and in those with autonomic nervous system dysfunction.7 Sodium sensitivity has also been reported to be linked to nondipping status of BP,8 although sodium restriction restores the normal dipping pattern in sodium-sensitive individuals. Studies with spectral analysis of heart rate variability have shown that abnormal diurnal BP patterns are closely related to the diurnal variation of autonomic nervous system activity.9 10
Body position and physical activity are 2 major determinants of the momentary fluctuations of BP. We have recently reported significant relationships between abnormal diurnal BP variation and orthostatic BP variation in elderly hypertensive patients, such that the change in BP associated with the upright position during the daytime might partly account for the abnormal diurnal BP variation patterns of extreme dippers and nondippers.11 Although several studies have investigated the effect of position on diurnal ABP variation,12 none has evaluated the effect of physical activity. In addition, quality of sleep might also affect nocturnal BP dipping status. Actigraphy is now widely used both for the quantitative assessment of physical activity during the day and for the assessment of sleep quality and duration.13
In this paper, we investigate the relationships of ambulatory physical activity assessed by actigraphy with diurnal BP variations in asymptomatic adults who were either normotensive or mildly (nonmedicated) hypertensive.
We studied 121 men and 39 women 33 to 66 years of age (mean±SD: 48±8.6 years) who participated in the third wave of the Cornell Work Site BP Study.14 At initial recruitment, subjects had a body mass index ≤32.5kg/m2, had a screening BP ≤160/105 mm Hg, and were free of clinically overt cardiovascular disease, renal disease, and diabetes mellitus. All subjects were fluent in English and worked full-time in 1 of 9 organizations in New York City. Further details on inclusion/exclusion criteria have been reported previously.15 Each wave consisted of a comprehensive evaluation that included 24-hour ABPM, an echocardiogram, psychosocial and demographic questionnaires, and other tests performed at 3-year intervals. Clinic BPs were measured by sphygmomanometer in the sitting position, according to the American Heart Association’s recommended protocol. Actigraphy was added at wave 3. Eighty-three percent of the original 373 participants were successfully followed during the third wave. All but 18 were currently employed. Excluded from this analysis were those who had cardiovascular events (n=10), died (n=5), or were taking antihypertensive medication at follow-up (n=40). We also excluded those who did not wear an actigraph or for whom the actigraph failed (n=90), those with <15 awake or 4 sleep BP readings (n=62), and 1 individual whose BP diary contained inadequate information regarding the time of sleep onset. All subjects enrolled in this study signed an informed consent statement approved by the Institutional Review Board at Cornell University Medical Center.
Smokers were defined as current smokers. Body mass index was calculated as weight (kg)/height (m)2, which was measured by a nurse/technician at the time of a physical examination and/or echocardiography. Left ventricular mass index was calculated as described previously.16
24-Hour Ambulatory BP Monitoring
Noninvasive ABPM was performed on a normal weekday with an automatic ambulatory BP monitor (SpaceLabs 90207), which recorded BP and heart rate every 15 minutes during the awake period and every 60 minutes during sleep. The monitor was fitted while the subject was either at work or during a research visit to the Hypertension Center at the New York Presbyterian Hospital/Weill Medical College of Cornell University and calibrated by comparison of 5 successive SBP and diastolic BP (DBP) readings against simultaneously determined auscultatory readings performed by a trained technician with a mercury column with the patient in the sitting position before ABPM. The mean of the monitor’s and technician’s readings had to agree to within 5 mm Hg for the 24-hour monitoring to proceed. Each subject was asked to remain as motionless as possible each time the monitor took a reading during waking hours and then to record his or her position, location, and activity in a diary.
The average (±SD) number of ambulatory BP readings used for the determination of awake and sleep BPs were 57.6±7.4 and 7.2±1.6, respectively. Each subject’s average BP and pulse rate (PR) for each position (supine, sitting, and standing as recorded in the diary) were calculated whenever the subject had ≥4 readings in that position (n=160 for sitting, n=157 for standing, and n=62 for supine [while awake]). The standing-supine SBP difference (n=60) was used as a measure of orthostatic BP variability. The SD of BP and PR during the awake and sleep periods were calculated and used as indicators of variability.
Definition of Extreme Dippers, Dippers, and Nondippers
Sleep BP was defined as the mean BP from the time at which the patient went to bed until the time of awakening, and the awake BP was defined as the mean BP during the remaining portion of the day. The proportional nocturnal SBP fall was calculated as (awake SBP−sleep SBP)/awake SBP. The 160 subjects who were subclassified on the basis of this nocturnal SBP reduction were classified as follows: 19 extreme dippers with ≥20% nocturnal reduction of SBP; 102 dippers with ≥10% but <20% reduction; and 39 nondippers with <10% reduction.
Assessment of Physical Activity by Actigraphy
Subjects wore an actigraph (Ambulatory Monitoring, Inc) on their waist while awake and on a wrist during sleep. The internal clocks of the ABPM device and the actigraph were set by the same computer. Activity was assessed continuously and recorded in 10-second epochs throughout the 24-hour period. The units of activity are arbitrary; it measures the number of times (max=10 times/s) that the pointer of the acceleration sensor in the actigraph crosses zero during a 10-second epoch (max=100 times/10 s epoch).
On the basis of preliminary analyses of the within-person relationship of activity to awake ABP readings, a weighted average of activity during the 6-minute period before and including each BP measurement was used as the summary measure of current/recent physical activity that best predicts within-person fluctuations in ambulatory BP (as described in the results). Mean awake (sleep) activity was defined as the mean activity score of the 6-minute periods that corresponded to all awake BP assessments. Little evidence exists of a ceiling effect, because the awake mean is ≈20 and only 1% of the 10-second epoch scores are >63 and <0.1% are >75. The square root transformation is applied to all activity measures to reduce the positive skew of the distribution and make it more normal. This transformation also makes the within-person relationship between the weighted 6-minute activity scores and the ABP readings more linear.
BP Reactivity Index
For each subject, we calculated the slope of awake SBP readings regressed against activity, as illustrated in Figure⇓. This slope measures the individual’s BP reactivity (to activity). On the basis of the quartiles of this index for the SBP slopes, we divided the sample into 4 subgroups: extreme reactive group=top quartile; reactive group=second quartile; low reactive group=third quartile; and nonreactive group=lowest quartile.
Data are expressed as the mean±SD or percentage. One-way ANOVA was performed to detect differences among groups, and Tukey’s honestly significant differences (HSD) test was used for multiple pairwise comparisons of means among groups. The χ2 test was used to evaluate differences among groups in the proportions of men, smokers, and blacks. Pearson correlation coefficients were used to measure the relationships among continuous measures. Multiple linear regression analysis was performed to estimate and test the independent effects on nocturnal BP reduction of various measures of activity. Differences/associations with P<0.05 were considered to be statistically significant.
Effect of Activity During and Before BP Measurement
To identify the time period of activity that most strongly related to the individual BP measurements, we compared the correlations of BP with (a) the activity scores for the 6 consecutive 10-second periods before and the six 10-second periods during each BP measurement, (b) the average activity during eleven 1-minute periods (the minute of the BP assessment back to 10 minutes before), and (c) several weighted averages of the activity scores for this 11-minute period. The correlations between BP and activity were stronger during sleep than when awake, but this difference was restricted to the 1 minute during and the 5 minutes before each BP measurement. The correlation between activity and PR was consistently higher when awake than during sleep, but for both states, the largest correlations were found with activity during the 1 minute of and the 1 minute before the measurement (as opposed to more distant measures of activity). Because the strongest pooled within-person correlation was obtained between each BP value and the square root of a weighted average of the activity during the 5-minute period just before and the 1 minute during each BP measurement (current minute of activity is weighted 4 times as heavily as each of the previous 5 minutes), we used this as the measure of physical activity linked to each BP reading. Similar to the BPs, these activity scores were averaged separately for the awake and sleep assessments. The mean awake and sleep activity scores (on the basis of the 6 minutes at and before the ABP readings) were strongly correlated with total activity during the entire awake and sleep periods, respectively (awake activity: r=.93, P<0.0001; sleep activity: r=.74, P<0.0001). Thus, we consider the mean of the weighted average of activity for the 6 minutes during and just before each BP measurement closely related to the average activity of the entire awake and sleep periods.
Characteristics of Groups With Different Patterns of Nocturnal BP Fall
Table 1⇓ shows the characteristics of the subjects classified according to the magnitude of nocturnal BP fall (extreme dippers, dippers, and nondippers). Apart from a modest difference in age, no significant differences existed among the groups in demographic or clinical characteristics, including gender, race/ethnicity, smoking status, body mass index, and casual/clinic BP.
Characteristics of 24-Hour Ambulatory BP Patterns
The 24-hour SBP was significantly higher in nondippers than in dippers, whereas no difference existed in DBP and PR among the 3 groups (Table 2⇓). Compared with the dippers, the sleep SBP and DBP were significantly higher in the nondippers and significantly lower in the extreme dippers, whereas no significant differences existed in awake SBP and DBP among the 3 groups. Standing SBP tended to be higher in the extreme dippers compared with the dippers (P=0.06, Tukey’s HSD) and nondippers (P=0.10). The standing-supine difference tended to be lower in nondippers than in dippers and extreme dippers, although this was not statistically significant because, we suspect, of the small number of individuals (n=62) with the minimum required number of supine BP readings. As a measure of short-term BP variability, the SD of the awake SBP tended to be higher in the extreme dippers than the dippers (P=0.08). The large group differences in the sleep/awake ratio of SBP were tautological and closely paralleled by group differences in the awake-sleep difference in SBP and DBP.
24-Hour Ambulatory Physical Activity
Sleep activity was greater in the nondippers than either the extreme dippers (P<0.05) or dippers (P=0.07), whereas awake activity was not significantly different among the 3 groups (Table 3⇓). The awake-sleep difference in activity was lower and the sleep/awake activity ratio was higher in nondippers than in extreme dippers and dippers. The average variability (SD) of sleep activity was also significantly higher in nondippers than in extreme dippers.
No significant differences existed in BP reactivity to activity among the 3 groups. However, the BP reactivity index was positively correlated with the variability of awake BP (SBP: r=.43, P<0.001; DBP: r=.30, P<0.001) but not with the variability of sleep BPs (Table 4⇓).
Correlations of Awake and Sleep BPs and PR With Activity and Position
Mean awake and sleep activity were positively correlated with awake PR (Table 4⇑). The correlation of sleep DBP with sleep activity was also significant, but awake BP and awake activity were essentially uncorrelated. The sleep/awake ratios of SBP and DBP were significantly and positively correlated with mean sleep activity and negatively correlated with awake activity (all, P<0.05). The sleep/awake ratios of SBP and DBP were also correlated positively with the sleep/awake activity ratio and correlated negatively with the standing-supine difference in SBP. A stepwise linear regression analysis that predicted the nocturnal drop in SBP (the sleep/awake ratio of SBP) revealed that after the sleep/awake activity ratio entered the equation (first step, P<0.001), none of the other measures of activity had a significant independent effect. Awake SBP variability (SD) was positively correlated with mean awake activity, the BP reactivity index, and the standing-supine difference in SBP.
Various factors seem to contribute to the individual diurnal BP pattern. In the present study, we first documented the relationship of physical activity (assessed by actigraphy) with BP levels and diurnal BP variation in study participants’ natural settings. Increased physical activity during sleep, which may be related to impaired sleep quality, partially determines nondipping status, whereas awake activity also partially determines awake BP variability and degree of dipping. The difference in ambulatory BP associated with changes in position (standing versus supine) may also be related to nondipping status and to BP variability.
In regard to the within-person relationships between activity and BP, our analysis of the individual BP readings and activity measurements obtained from the 160 subjects found that activity during the BP measurement and the preceding 5 minutes significantly predict BP during both the sleep and awake periods. The correlation between BP and activity was stronger during sleep than when awake. In addition, activity is a stronger predictor of PR than of BP and the correlation between activity and PR was consistently higher when awake than during sleep
We found that sleep activity was related to nocturnal BP dipping status. Recently, activity during the night has been accepted as an indicator of sleep quality. The board of American Sleep Disorders Association has approved the use of the actigraph for the clinical evaluation of sleep disorders.13 In individuals with impaired sleep quality, frequent arousals could cause increased physical movement, which would be detected by actigraphy. Some specific sleep disorders such as sleep apnea syndrome17 18 or fatal familial insomnia19 are frequently characterized by nondipping BP patterns. With the use of actigraphy, these specific disorders have also been reported to be associated with increased activity during sleep and a diminished sleep/awake activity ratio.20 21 We are only aware of one other study that has simultaneously monitored ambulatory BP and actigraphy in healthy adults.22 Although these investigators did not present results by dipping status group, they found a small-to-moderate positive association between awake-sleep differences in activity and BP, which was consistent with those reported in Table 4⇑. In our study, we also found a positive relationship between nondipping status and increased sleep activity. One small previous study with polysomnographic recording in 19 hypertensives found that the 9 nondippers had more frequent microarousals and less stage 4 sleep than the 10 dippers.23 Thus, the increased sleep activity in nondippers might reflect impaired sleep quality. In addition, the SD of asleep activity in nondippers was also increased compared with extreme dippers, which was consistent with the hypothesized fluctuation of sleep quality in this group. Impaired sleep quality might alter the sympathovagal balance toward increased sympathetic activity to produce nondipping status, as suggested in previous studies including ours.9 10
However, to address the relationship between dipping status and sleep quality directly, a study that assesses sleep quality by polysomnography in the ambulatory setting (at home) is necessary, because physical activity during sleep assessed by actigraphy is a nonspecific measure of sleep quality. Actigraphy also captures nocturnal behaviors (going to the bathroom or getting a drink) and nocturnal involuntary movements.
Although statistically significant, sleep activity or the sleep/awake ratio of activity accounted for 3% to 5% of the variation in the systolic and diastolic sleep/awake ratios. In addition, other factors are certainly important determinants of dipping. As stated above, endocrine abnormalities, autonomic nervous system dysfunction, and sodium sensitivity also affect nocturnal dipping. In addition, target organ damage to the brain (eg, silent cerebrovascular disease) may be both a cause and a consequence of nondipping. However, these mechanisms are probably not important in this study because almost all subjects were healthy adults with normotension or, at most, mild hypertension.
Awake activity was significantly though weakly correlated with awake SBP variability (SD). Additional correlates of awake BP variability were the standing-supine difference in BP and the SBP reactivity to activity index that was calculated to assess each individual’s BP response to activity while awake. The present study of a middle-aged population provides modest support (P=0.08) for our previous finding in elderly hypertensive patients that extreme dippers exhibit greater awake BP variability (SD) than dippers.5 In addition to an exaggerated diurnal BP difference, increased awake BP variability may also advance target organ damage in extreme dippers.
In regard to the BP response to changes in position during ABPM, we found that the standing-supine difference in SBP during the awake period was negatively correlated with the sleep/awake ratio of SBP. We previously reported that the sitting-supine difference in awake SBP was strongly correlated with the awake-sleep difference in SBP.12 In another study, we found that the orthostatic BP change assessed by the 70-degree head-up tilting was closely related to the diurnal BP variation: the greater the orthostatic change of SBP, the greater the diurnal BP variation (decrease in the sleep/awake SBP ratio).11 In the present study, this relationship was confirmed in the ambulatory setting. Thus, the effect of position on BP may also contribute to the different diurnal BP variation patterns of extreme dippers, dippers, and nondippers.
In conclusion, physical activity is probably a determinant of diurnal BP variation. In particular, both the degree of dipping and dipping status are related to the ratio of sleep activity and awake activity.
This study was supported in part by grant HL-47540 from the National Heart, Lung and Blood Institute and by the Sasagawa Health Science Foundation (Tokyo, Japan).
- Received February 8, 1999.
- Revision received March 2, 1999.
- Accepted June 9, 1999.
O’Brien E, Sheridan J, O’Malley K. Dippers and non-dippers. Lancet. 1988;13:397.
Verdecchia P, Porcellati C, Schillaci G, Borgioni C, Ciucci A, Battistell M, Guerrieri M, Gatteschi C, Zampi I, Santucci A, Santucci C, Reboldi G. Ambulatory blood pressure: an independent predictor of prognosis in essential hypertension. Hypertension. 1994;24:793–801.
Kario K, Matsuo T, Kobayashi H, Imiya M, Matsuo M, Shimada K. Relation between nocturnal fall of blood pressure and silent cerebrovascular damage in elderly hypertensives: advanced silent cerebrovascular damage in extreme dippers. Hypertension. 1996;27:130–135.
Pickering TG. The clinical significance of diurnal blood pressure variations: dippers and nondippers. Circulation. 1990;81:700–702.
Imai Y, Abe K, Sasaki S, Minami N, Nihei M, Munakata M, Murakami O, Matsue K, Sekino H, Miura Y, Yoshioka K. Altered circadian blood pressure rhythm in patients with Cushing’s syndrome. Hypertension. 1988;12:11–19.
Uzu T, Kazembe FS, Ishikawa K, Nakamura S, Ikenaga T, Kimura G. High sodium sensitivity implicates nocturnal hypertension in essential hypertension. Hypertension. 1996;28:139–142.
Kohara K, Nishida W, Maguchi M, Hiwada K. Autonomic nervous function in non-dipper essential hypertensive subjects: evaluation by power spectral analysis of heart rate variability. Hypertension. 1995;26:808–814.
Kario K, Motai K, Mitsuhashi T, Suzuki T, Nakagawa Y, Ikeda U, Matsuo T, Nakayama T, Shimada K. Autonomic nervous dysfunction in elderly hypertensive patients with abnormal diurnal blood pressure variation: relation to silent cerebrovascular disease. Hypertension.. 1997;30:1504–1510.
Kario K, Eguchi K, Nakagawa Y, Motai K, Shimada K. Relationship between extreme dippers and orthostatic hypertension in elderly hypertensive patients. Hypertension. 1998;31:77–82.
Schwartz JE, Grossman S, Pickering TG. How much is the awake-sleep difference in ambulatory blood pressure (dipping) due to sleeping vs. being supine? J Hypertens. 1998;16(suppl 2):S200. Abstract.
Schwartz JE, Warren K, Pickering TG. Mood, location and physical position as predictors of ambulatory blood pressure and heart rate: application of a multi-level random effects model. Ann Behav Med. 1994;16:210–220.
Schnall PL, Schwartz JE, Landsbergis PA, Warren K, Pickering TG. Relation between job strain, alcohol, and ambulatory blood pressure. Hypertension. 1992;19:488–494.
Devereux RB, Roman MJ. Ultrasonic techniques for the evaluation of hypertension. J Hypertens.. 1994;3:644–651.
Portaluppi F, Cortelli P, Avoni P, Vergnani L, Contin M, Maltoni P, Pavani A, Sforza E, degli Uberti EC, Gambetti P, Lugaresi E. Diurnal blood pressure variation and hormonal correlates in fatal familial insomnia. Hypertension. 1994;23:569–576.