The Effects of Ambulatory Blood Pressure Monitoring on Sleep Quality in Men and Women With Hypertension: Dipper vs. Nondipper and Race Differences

Abstract

BACKGROUND

The nondipping circadian blood pressure (BP) profile is associated with both poor sleep quality and increased cardiovascular risk. The present study aimed to clarify the potential confounding effects of 24-hour ambulatory blood pressure monitoring (ABPM) used to characterize the circadian BP profile by assessing its impact on sleep quality.

METHODS

Participants were 121 middle-aged men and women with untreated hypertension (age = 46 ± 8 years; 43% women; 45% African-American). Subjective sleep quality was assessed using the Pittsburgh Sleep Quality Index. Wrist actigraphy was used to measure sleep quality objectively as sleep efficiency (SE) and total sleep time (TST) on 7 consecutive non-ABPM days (baseline) and 3 subsequent 24-hour ABPM days.

RESULTS

Average ambulatory BP was 137.2 ± 10.8/84.3 ± 8.5 mm Hg during the day and 119.6 ± 12.4/69.5 ± 9.8 mm Hg at night. Using the criterion of <10% dip in systolic BP (SBP) to define nondippers, there were 40 nondippers (SBP dip = 7.3 ± 2.6%) and 81 dippers (SBP dip = 15.5 ± 3.4%). There was no effect of time on SE or TST over non-ABPM and ABPM days, suggesting that ABPM does not adversely affect sleep quality. Sleep quality was generally poorer (lower SE) in nondippers compared with dippers (P = 0.033), but differences were independent of whether or not participants were undergoing 24-hour ABPM. African-American race (P = 0.002) was also associated with lower SE.

CONCLUSION

Sleep quality generally appears to be poor in men and women with untreated hypertension and especially among African-Americans. Importantly, for both dippers and nondippers, we found no evidence that ABPM had an adverse effect on sleep quality.

Ambulatory blood pressure monitoring (ABPM) studies using noninvasive monitoring devices have demonstrated the presence of a distinct circadian rhythm, whereby blood pressure (BP) falls during the nighttime sleep period. A normal healthy circadian BP profile includes a nighttime BP dip of 10% or more. In contrast, a nondipping circadian BP profile, typically defined as <10% fall in average BP from daytime to nighttime, is a strong indicator of increased risk of cardiovascular morbidity and mortality for both hypertensive and nonhypertensive individuals.^1–6 A recent meta-analysis found that BP dippers were characterized by better subjective sleep quality than nondippers.⁷ Objective measures of sleep quality derived from polysomnography and actigraphy also have shown that poor sleep quality is associated with a nondipping circadian BP profile.^8–12

Although the prognostic significance of blunted nighttime BP dipping supports the inference that both poor sleep quality and BP nondipping are related individual difference characteristics, the possibility that cuff inflations during the ABPM assessment procedure may disrupt sleep and confound the measurement of the circadian BP profile has been a cause for concern. A sleep laboratory study of 6 healthy participants showed that ABPM measurements could cause arousal from sleep and may result in notable transient increases in recorded BP.¹³ In contrast, a study of 10 hospitalized patients with 24-hour intra-arterial BP recording reported that noninvasive ambulatory measurement of BP every 15 minutes during the daytime and every 30 minutes during the sleep period did not affect circadian BP profiles.¹⁴ Another small study of 44 healthy men found that ABPM induced only modest sleep disturbances that did not distort the circadian BP profile.¹⁵ However, it has been suggested that the potential confounding effects of cuff inflation on sleep/nighttime BP may be more problematic in patients with hypertension because of the need for higher peak cuff inflation pressures.¹⁶

The present study sought to help resolve the conflicting evidence of whether ABPM confounds the assessment of nighttime BP dipping by disturbing sleep. In a study sample of men and women with untreated hypertension, we used wrist actigraphy to monitor sleep quality in the absence of ABPM for 7 continuous days. Participants subsequently underwent 3 assessments of 24-hour ABPM, each separated by 1 week. Our primary measure of sleep quality was sleep efficiency (SE), measured by wrist actigraphy, with total sleep duration also of interest. The study objectives were to determine (i) whether sleep quality was disrupted by ABPM in men and women with untreated hypertension, (ii) whether ABPM-related disruption of sleep was more evident in nighttime BP nondippers than dippers, and (iii) if repeat ABPM assessments would result in less sleep disruption than the first exposure to ABPM. We also explored whether individual difference characteristics, including age, sex, race, and subjective sleep quality, were related to sleep disruption associated with ABPM.

METHODS

Participants

Participants were 121 men (N = 69) and women (N = 52) between the ages of 30–60 years, including 55 African-Americans and 66 White Americans. BP inclusion criteria were clinic systolic BP (SBP) of 130–159 mm Hg and/or diastolic BP of 85–99 mm Hg, which includes stage 1 and stage 2 hypertension according to the 2017 American College of Cardiology/American Heart Association Guideline for High Blood Pressure in Adults.¹⁷ Exclusion criteria were body mass index (BMI) >35 kg/m²; age <40 years or >60 years; current use of BP or cardiovascular medications; diabetes mellitus; previously diagnosed obstructive sleep apnea; pacemaker; atrial fibrillation; myocardial infarction, percutaneous coronary intervention, or coronary artery bypass graft surgery within 6 months of enrollment; heart failure; severe uncorrected primary valvular disease; uncorrected thyroid heart disease; oral contraceptive use; pregnancy; hormone replacement therapy; alcohol or drug abuse within 12 months; renal or hepatic dysfunction; dementia; inability to comply with the assessment procedures; or inability to provide informed consent. Participants were recruited by advertisements in the Piedmont region of North Carolina, from the general population of over 1 million that reside within a 30-mile radius of Duke University Medical Center. The study sample was selected from a total of 598 people who responded, were screened by phone, and met all eligibility criteria. The study protocol was approved by the Institutional Review Board at Duke University Medical Center. All eligible individuals provided written informed consent prior to participation in the study.

Clinic BP screening

Clinic BP was determined on 3 separate visits, each approximately 1 week apart. After 5 minutes seated in a quiet, temperature-controlled room, 4 seated BP readings, each 2 minutes apart, were taken using a mercury sphygmomanometer and stethoscope. SBP and diastolic BP for each visit were calculated as the means of the last 3 readings, and clinic BP eligibility was based upon whether these average office BP values met the study’s inclusion criteria. Data were collected on age, gender, race, height, and weight; BMI was calculated as weight (kg)/height (m).²

Ambulatory blood pressure monitoring

Twenty-four-hour ambulatory BP (ABP) was assessed on 3 separate occasions for each participant during normal weekdays (Monday–Friday), with an interval of 1 week between monitoring sessions. ABP was measured with the Oscar 2 ABP monitor (Suntech Medical, Raleigh, NC), which has been validated previously.^18,19 The monitor was programmed to take BP measurements every 20 minutes throughout the waking hours and every 30 minutes during the nighttime sleep period.

Subjective sleep quality

The Pittsburgh Sleep Quality Index (PSQI) was used to assess subjective sleep quality. The PSQI is a 19-item self-rated questionnaire for evaluating subjective sleep quality over the previous month.²⁰ The PSQI provides a global subjective sleep quality score ranging from 0 to 21, with higher scores indicating worse sleep quality. A cutoff score of 5 on the PSQI has a sensitivity of 89.6% and specificity of 86.5% for identifying cases with sleep disorder.²¹ All participants completed the PSQI once only, at baseline.

Measures of sleep quality during ABPM and non-ABPM days using acitigraphy

Mini-Mitter Actiwatch actigraphs (Mini-Mitter, Sunriver, OR) were used to derive objective estimates of sleep parameters. Previous research has shown that actigraphy provides estimates of sleep parameters that are similar to polysomnography.²² Sleep quality assessed by actigraphy has previously been related to nighttime BP dipping.^11,23 The Actiwatch contains a calibrated accelerometer and a memory storage apparatus housed in a casing that, in size and shape, resembles a wristwatch. The accelerometer samples’ movement/activity at a rate of 32 times per second. In accordance with the Society for Behavioral Sleep Medicine (SBSM) actigraphy monitoring guidelines,²⁴ the sleep period was defined by self-report in the actigraphy log recorded on each assessment day as the time from when study participants switched off the lights and planned to initiate sleep until the time when they got out of bed the following morning. Data from the actigraph were used to confirm the actigraphy log times and to derive SE (the percent of time asleep during the sleep period) and total sleep time (TST; the total hours asleep during the sleep period). A SE of <85% was used to define poor sleep quality.

Data reduction and statistical analysis

Waking and nighttime sleep periods, based on participant self-report, were used to calculate mean waking and mean nighttime sleep values for SBP and diastolic BP for each participant’s 3 ABPM sessions. Values for the 3 sessions were averaged to derive a robust measure of waking and nighttime ABP for each participant. Dipping was assessed by subtracting the mean nighttime sleep SBP from the mean waking SBP. Percent dip was derived by dividing the mean dipping value by the mean wake value and multiplying by 100. BP dipper status was defined as a dip in mean SBP ≥10% (dipper) or ˂10% (nondipper). Twenty-four-hour actigraphy data were recorded on 7 consecutive non-ABPM days prior to the first 24-hour ABPM assessment, as well as on each of the 3 ABPM days. Sleep parameters, including SE and TST, were calculated for each of the 7 non-ABPM days and averaged to provide baseline sleep quality characteristic of each participant. The same sleep quality parameters were computed, separately, for each of day on which ABP monitoring occurred (i.e., ABPM-1, ABPM-2, and ABPM-3), as well as averaged to obtain an overall mean for SE and TST across 3 ABPM days. Demographic variables, BPs, and measures of sleep quality were calculated and expressed as mean ± SD or n (%) for the cohort as a whole and also for dippers and nondippers. Repeated measures analyses of covariance tests were used to examine the associations of demographic and anthropomorphic variables and PSQI scores with sleep quality and to evaluate the relative change in sleep quality over time (i.e., non-ABPM baseline, ABPM-1, ABPM-2, and ABPM-3), controlling for age, gender, race, BMI, and PSQI. All statistical analyses were conducted using the SAS system (SAS 9.2, SAS Institute, Cary, NC) with significance set at P = 0.05.

RESULTS

Study sample characteristics

Characteristics of the study sample according to dipper vs. nondipper classification are shown in Table 1. The 121 participants included 30 African-American women, 25 African-American men, 22 White American women, and 44 White American men, with a mean age of 45.8 ± 8.3 years and screening BP of 140 ± 7/90 ± 5 mm Hg. The average ABP for the study cohort was 137.2 ± 10.8/84.3 ± 8.5 mm Hg during the day and 119.6 ± 12.4/69.5 ± 9.8 mm Hg at night. Using the criterion of ˂10% dip in SBP to define nondippers, there were 40 nondippers (SBP dip = 7.3 ± 2.6%) and 81 dippers (SBP dip = 15.5 ± 3.4%). As shown in Table 1, compared with dippers, nondippers were more likely to be African-American (P = 0.001) and have a higher BMI (P = 0.043). The average PSQI was 6.1 ± 3.8; 59% of participants had a PSQI ≥5.

Table 1.

Sample characteristics by dipper status

	All (n = 121)	Nondipper (n = 40)	Dipper (n = 81)	P
Age (years)	45.8 + 8.3	45.5 + 8.2	45.9 + 8.3	0.785
Sex (n, % male)	69 (57%)	24 (60%)	45 (56%)	0.646
Race (n, % African-American)	55 (45%)	27 (68%)	28 (35%)	0.001
BMI (kg/m2)	28.1 + 3.6	29.0 + 3.0	27.6 + 3.8	0.043
Clinic SBP (mm Hg)	140.2 + 7.1	141.0 + 7.2	139.8 + 7	0.380
Clinic DBP (mm Hg)	89.8 + 5.2	92.3 + 4.2	88.6 + 5.2	<0.0001
Awake SBP (mm Hg)	137.2 + 10.8	138.7 + 10.8	136.5 + 10.8	0.297
Awake DBP (mm Hg)	84.3 + 8.5	85.8 + 8.3	83.6 + 8.5	0.182
Sleep SBP (mm Hg)	119.7 + 12.4	128.6 + 10.7	115.2 + 10.8	<0.0001
Sleep DBP (mm Hg)	69.5 + 9.8	75.5 + 9.0	66.3 + 8.7	<0.0001
PSQI	6.1 + 3.8	6.9 + 3.8	5.7 + 3.7	0.112

Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; PSQI, Pittsburgh Sleep Quality Index; SBP, systolic blood pressure.

Sleep quality measures at baseline and during 24-hour ABPM days

Sleep efficiency.

The mean and SE at baseline was 77.9 ± 9.5; 75% of subjects had a SE <85%. In our analysis of associations of subject characteristics with SE, there were no significant main effects for age, gender, or BMI. There were significant main effects, however, for race (P = 0.020), subjective sleep quality (P ≤ 0.0001), and dipper status (P = 0.033). Means ± SD for SE by dipper status, sex, race, and subjective sleep quality are presented in Table 2. SE was lower in African-Americans than Whites, and a significant Time × Race interaction (P = 0.006) indicated that this race difference in SE was more evident on non-ABPM days. SE was also lower in those with poor subjective sleep quality (PSQI ≥5) than in those with better subjective sleep quality (PSQI <5). Importantly, there were no overall effects of time on SE over non-ABPM and ABPM days, suggesting that ABPM was not associated with any impairment in SE. As illustrated in Figure 1, nondippers evidenced lower SE compared with dippers across both non-ABPM and ABPM days.

Table 2.

Sleep efficiency and total sleep time (mean ± SD) by dipper status, sex, race, and subjective sleep quality (PSQI) for baseline (non-ABPM days) and 24-hour ABPM days

Sleep efficiency (%)
	N	Baseline	ABPM-1	ABPM-2	ABPM-3
Dipper	81	80 ± 8**	82 ± 9*	83 ± 10*	83 ± 10*
Nondipper	40	73 ± 11	77 ± 14	78 ± 11	77 ± 13
Men	69	77 ± 10	81 ± 10	82 ± 11	80 ± 12
Women	52	79 ± 9	79 ± 12	80 ± 12	81 ± 11
White	66	82 ± 6**	82 ± 10	84 ± 7**	83 ± 9**
African-American	55	72 ± 10	79 ± 12	77 ± 13	78 ± 13
PSQI <5	50	81 ± 9**	85 ± 7**	83 ± 9	83 ± 10
PSQI ≥5	71	76 ± 9	77 ± 13	79 ± 12	79 ± 12
Total sleep time (hours)
	N	Baseline	ABPM-1	ABPM-2	ABPM-3
Dipper	81	5.7 ± 1.1**	6.3 ± 1.1**	6.1 ± 1.1	6.1 ± 1.1
Nondipper	40	4.9 ± 1.4	5.6 ± 1.6	5.7 ± 1.4	5.7 ± 1.5
Men	69	5.4 ± 1.4	5.9 ± 1.3	5.9 ± 1.2	5.8 ± 1.3
Women	52	5.6 ± 1.1	6.1 ± 1.4	6.0 ± 1.2	6.2 ± 1.2
White	66	6.1 ± .98**	6.3 ± 1.2*	6.3 ± 1.0**	6.3 ± 1.0**
African-American	55	4.7 ± 1.2	5.7 ± 1.5	5.5 ± 1.3	5.5 ± 1.4
PSQI <5	50	5.8 ± 1.3*	6.3 ± 1.1	6.1 ± 1.1	6.1 ± 1.3
PSQI ≥5	71	5.2 ± 1.2	5.9 ± 1.5	5.9 ± 1.3	5.8 ± 1.3

Between-group difference is significant **P ≤ 0.001, *P ≤ 0.05. Please interpret these unadjusted multiple comparisons of group differences at baseline, ABPM-1, ABPM-2, and ABPM-3 with caution, as in the absence of Group × Time interactions shown in Results section; these comparisons, although of possible interest, are not statistically justified. Abbreviations: ABPM, ambulatory blood pressure monitoring; PSQI, Pittsburgh Sleep Quality Index.

Figure 1.

Sleep efficiency (mean ± SE) by dipper status for baseline (average of 7 non-ABPM days) and each 24-hour ABPM day.

Total sleep time.

Analyses were repeated for TST, with means ± SD also shown in Table 2. The mean TST for the entire cohort as baseline was 5.7 ± 1.1 hours. There were no significant main effects for age, BMI, subjective sleep quality, or dipper status on TST. There was a significant main effect for gender (P = 0.044); as shown in Table 2, women slept slightly longer than men. There also was a significant Time × Race interaction (P = 0.001), which was due to lower TST in African-Americans than Whites, that was particularly evident during the non-ABPM days. There were no effects of time over non-ABPM and ABPM days, indicating that ABPM was not associated with adverse effects (shortening) of TST.

DISCUSSION

The present findings are consistent with a number of previous studies that have found a blunted pattern of circadian BP variation, and the nondipping nighttime BP profile, to be associated with poor sleep quality.^7–12 Importantly, our observations for sleep quality, assessed objectively by wrist actigraphy in 121 men and women with untreated hypertension, show that both SE and TST during ABPM assessment were no different than sleep quality defined according to these same parameters recorded during a typical week involving no ABPM assessments. It is nonetheless noteworthy that our study sample of men and women with untreated hypertension had relatively poor sleep quality, whether measured subjectively (average PSQI >5) or objectively (SE <85%) on both ABPM and non-ABPM days. The observation that hypertension is associated with poor sleep quality and insomnia has been documented previously.²⁵ Our findings allay concerns that ABPM may have a confounding effect on the extent to which BP falls during the nighttime sleep period by disrupting normal sleep patterns.^26,27 The current observations indicate that poorer sleep quality observed in nondippers compared with dippers does not appear to be a consequence of the ABPM assessment, as it is also clearly evident during non-ABPM assessment days. Other individual difference characteristics that were examined in our study, including sex, race, and subjective sleep quality, also were related to differences in actigraphy-based measures of sleep quality, but there was no evidence that any of these characteristics identified a susceptibility to sleep disruption associated with ABPM.

Our study participants slept at least as well, on average, during ABPM as they did on non-ABPM days. We also found that repeating ABPM on 3 separate occasions, each 1 week apart, resulted in no habituation to monitoring; there was no improvement in sleep quality over the 3 days of ABPM assessments. This finding contrasts with evidence that disruptive effects of ABPM were slightly reduced during the second 24-hour phase of a 48-hour ABPM in 823 White men and women with mild to moderate hypertension.²⁸ Although our findings do not suggest that ABPM results in sleep disturbance for our study sample as a whole, our results do not imply that multiple ABPM monitoring sessions are not of value, as 24-hour ABP profiles do show variability due to a range of factors, including individual variations in daytime physical activity, behavioral stress, and in sleep quality.^29,30 Thus, as is the case for the measurement of any phenomenon that shows variation within individuals, averaging over multiple assessments may yield a more representative ABP profile for a given individual than any single assessment alone.

Overall, our observations contrast with several prior studies that reported ABPM to impair sleep quality, albeit typically to a modest degree. Degaute et al.¹⁵ reported a sleep laboratory study of 44 healthy men that showed taking noninvasive BP measurements every 10 minutes resulted in modest sleep disturbance. Van de Borne et al.³¹ observed ABPM to be associated with a small decline in SE in a sample of 15 healthy men in a study conducted in a sleep laboratory environment. Similar observations were made by Davies et al.¹³ in 6 healthy individuals also studied in a sleep laboratory environment. Lenz and Martinez³² also observed that BP measurement in individuals with suspected obstructive sleep apnea could cause arousal and a rise in BP, again when assessed in a sleep laboratory environment, which they concluded might influence dipper vs. nondipper classification. However, other studies contradict these findings. Parati et al.¹⁴ found that ABPM had no effect on nighttime BP recorded in hospitalized patients by intra-arterial catheter. Tropeano et al.³³ studied 70 men and women who were undergoing ABPM for the assessment of hypertension by assessing sleep quality on preceding non-ABPM days and then during the ABPM assessment, yielding results similar to our current findings in that sleep quality assessed objectively by wrist actigraphy was not affected by ABPM. These inconsistent findings across studies are difficult to reconcile. One possible consideration is that ABPM technology has improved quite dramatically over the past 4 decades, with early monitors being noisier and more obtrusive, whereas current day ABPM devices are compact and relatively quiet and thus less likely to disturb sleep. Our finding that sleep quality was undisturbed by ABPM in hypertensive patients also suggests that the more elevated cuff pressures required for measurement of BP during the nighttime in hypertensive patients, and perhaps especially in hypertensive nondippers, does not appear to be problematic.¹⁶ Our results may not apply, however, to patients with markedly elevated BP.

Limitations of our study include the use of wrist actigraphy, rather than the gold standard of polysomnography, to provide an objective approach to assessing sleep quality. However, wrist actigraphy is a validated and accepted measure of sleep quality, is less expensive and burdensome relative to polysomnography, and is a recommended objective measure when staging of sleep architecture is not a fundamental aspect of the research study.³⁴ Although we found no evidence that ABPM had an adverse effect on sleep quality, small effects may have been missed due to the study sample size. The study sample included participants in the relatively narrow age range of 40–60 years, so our findings may not necessarily generalize to younger and older patients with hypertension or to patients who are prescribed medication. The cohort we examined, however, represents an important subset of patients with high office BP in which the US Preventive Services Task Force,³⁵ the The National Institute for Health Care Excellence clinical guideline,³⁶ and the 2017 American College of Cardiology/American Heart Association Guideline¹⁷ recommend ABPM before diagnosing hypertension and initiating medical therapy.

Interestingly, subjective measures of sleep quality based upon self-report appear more likely to show evidence of disruption associated with ABPM than is evidenced by objective measures such as wrist actigraphy.³³ In a study of 2,934 untreated hypertensives, Verdecchia et al.³⁷ found that blunted nighttime BP dipping was associated with increased risk of cardiovascular disease events at 7-year follow-up. However, they also noted that in a subsample of 13.7% of participants who, based on self-report, experienced the ABPM procedure to be disruptive (i.e., deprived them of about 2 hours of nighttime sleep), the prognostic significance of blunted BP dipping was not evident. As acknowledged by the authors, however, analysis of the subset of subjects that experienced sleep disturbance was statistically underpowered for assessment of cardiovascular endpoints. In terms of habituation to potential sleep disruption by ABPM, another study limitation was that 24-hour ABPM sessions were scheduled a week apart rather than over a continuous 72-hour period. However, because we did not find the first ABPM session to result in attenuated sleep quality relative to non-ABPM days, it is unlikely that a more burdensome 72-hour ABPM study would have been any more informative. An important strength of our study is that sleep quality comparisons between ABPM and non-ABPM days were all made in participants’ home environments.

In summary, our study sample of men and women with untreated hypertension was generally characterized by relatively poor sleep, but there was no evidence that ABPM had any adverse impact on sleep quality, which remained similar across ABPM and non-ABPM assessment days. This finding suggests that impaired actigraphy-assessed sleep quality is an individual difference characteristic rather than a confounding effect of the ABPM assessment. Moreover, our results suggest that an attenuated nocturnal fall in BP (nondipping) is a real physiologic characteristic of some individuals with hypertension and not merely an artifact resulting from arousal during nighttime BP measurement. Our observations further support the view that poor sleep quality is associated with a nondipping BP profile. Although the available evidence suggests a clear association between poor sleep quality and the nondipping BP pattern, well-designed studies are still needed to establish causality and also to determine whether interventions that improve sleep quality promote a normal pattern of nighttime BP dipping, thereby also potentially reducing cardiovascular risk.

DISCLOSURES

The author(s) declared no conflict of interest.