Effect of Continuous Positive Airway Pressure on Blood Pressure Variability in Patients With Obstructive Sleep Apnea

Abstract

Obstructive sleep apnea (OSA) is a common risk factor for cardiovascular disease. Continuous positive airway pressure (CPAP) improves OSA symptoms and blood pressure (BP) control. The effect of CPAP on BP variability (BPV) in patients with and without hypertension treated with autotitrating CPAP (APAP) for 2 weeks was studied. A total of 78 participants (76.9% men, 49% hypertensive, mean body mass index 36.2 [6.9] kg/m², age 49.0 [12.9] years) underwent 2 weeks of APAP therapy. Office BP, BPV (standard deviation of three BP measurements), and pulse rate were measured before and after treatment. Systolic BPV (5.3±4.9 vs 4.2±3.4 mm Hg, P=.047) and pulse rate (78.0±14.5 vs 75.5±15.8 beats per minute, P=.032) decreased after treatment, particularly in hypertensive participants. Mask leak was independently associated with reduced changes in systolic BPV (r=−0.237, P=.048). Short‐term APAP treatment reduced BPV and pulse rate, particularly in hypertensive patients with OSA.

Obstructive sleep apnea (OSA) is a well‐recognized risk factor for cardiovascular (CV) disease.1 Its prevalence is rising as a result of the current obesity epidemic.2 OSA is caused by recurrent obstructions of the upper airway when asleep. These obstructions result in apneas and hypopneas, which lead to repeated oxygen desaturations and arousals from sleep, and hence activation of the sympathetic nervous system.3 Increased sympathetic tone impacts on blood pressure (BP), heart rate,4 and importantly on the risk of CV morbidity.

Continuous positive airway pressure (CPAP) maintains upper airway patency and abolishes or reduces obstructive events in patients with OSA,5 also reducing BP.6 Recent studies have revealed that optimal CPAP control reduces BP, with more marked effects in patients with resistant hypertension,7 implying a role of the autonom ous nervous system in relation to changes in BP.

BP variability (BPV) is a marker of autonomic nervous system output and an independent predictor of CV morbidity and mortality.8 It is defined as the fluctuation of BP between different measurements over a defined time interval.9 Visit‐to‐visit BPV comparison is related to an increased risk of CV events.10, 11

Patients with OSA are known to have an increased sympathetic tone, which causes increased BPV12, 13 and leads to raised levels of absolute BP and increased risk of CV morbidity and mortality. Patients with hypertension exhibit higher BPV when compared with normotensives as expression of an enhanced sympathetic tone.9

We aimed to investigate how CPAP treatment modifies the risk of increased sympathetic activation, through its impact on BPV, in patients with obstructive sleep apnea.

Methods

Patients with a diagnosis of OSA who had been referred to a tertiary sleep center (Sleep Disorder Centre, Guy's & St Thomas' Hospitals, London, UK) were enrolled in this study between June 2013 and December 2013.

OSA was diagnosed according to the National Institute of Health and Care Excellence (NICE) guidelines14 with a 4% oxygen desaturation index (ODI) >5 per hour combined with excessive daytime sleepiness (Epworth Sleepiness Scale [ESS] >10). The study was approved by the local institution's review board (2014/3081) and all patients gave written informed consent.

Hypertension was defined according to current international guidelines.15 In order to investigate any differential impact of CPAP, patients with OSA were divided into two subgroups: patients with a known diagnosis of hypertension or BP >140/90 mm Hg on three occasions comprised the hypertensive subgroup, and patients with BP <140/90 mm Hg on three occasions comprised the normotensive subgroup.

Following baseline recording of demographic data and BP measurements, patients underwent nocturnal pulse oximetry and were provided with an autotitrating CPAP (APAP) device for home use. At 2‐week follow‐up, compliance data were obtained and BP was measured again and BPV was calculated.

Baseline Assessments

At baseline, participant's age, sex, height, weight, body mass index (BMI), and ESS, as a measure of daytime sleepiness,16 were recorded.

BP and BP Variability

At each visit, baseline, and at 2‐week follow‐up, three BP measurements were taken with 1‐minute intervals, using an automatic sphygmomanometer (Mindray VS‐800; Mindray Medical International Limited, Shenzhen, China). Patients rested for >5 minutes and were seated in an upright position.17

The average of pulse rate and systolic and diastolic BP was calculated from these readings. Systolic and diastolic range was calculated as the difference between the maximum and minimum BP values.

Systolic and diastolic BPV was calculated as the standard deviation (SD) of the three BP measurements.9

All data were compared between baseline and 2‐week follow‐up (following APAP treatment), which included the change (delta) in systolic and diastolic BP range, means/average BP and SD, and the delta in pulse rate.

APAP Treatment

Following baseline measurements, patients were issued an APAP (S8/S9; ResMed Ltd, Sydney, Australia) for 2 weeks of home use.

The following compliance and treatment indices were downloaded and calculated at the end of the 2‐week period: average daily APAP usage (hours), days of APAP usage >4 h/d, days of APAP usage <4 h/d, percentage of days of APAP usage >4 h/d, air leak (L/min), and the 95th percentile of APAP to control respiratory events (cmH₂O).

The primary outcome parameter was change in BPV between baseline measurements and after 2 weeks of treatment. Secondary outcomes included change in pulse rate, change in absolute BP values, and comparison of these parameters in patients with and without hypertension. Lastly, we compared patients who were compliant with CPAP treatment, defined as usage for more than 4 hours per night for at least 70% of the total days of usage, with patients who had a suboptimal CPAP compliance to determine whether BPV changes were dependent on the treatment received.

Sample Size Analysis

To understand the significance of the difference found in BPV before and after treatment we performed a power calculation stating a significance level (adjusted for sidedness) of 0.025, the total number of patients (78), and the obtained difference in means of 1.2 mm Hg. The calculated power was 99% that the study would detect a difference, if the true difference between treatments was 1.2 times the SD.

Statistical Analysis

Data were analyzed using SPSS version 21 (IBM, New York, NY) and tested for normal distribution using the Shapiro‐Wilk normality test. Normally distributed data are presented as mean (SD) and were analyzed with paired and unpaired t tests. Nonnormally distributed data are presented as median (interquartile range) and were analyzed with the Wilcoxon rank sum test when paired and Mann‐Whitney U test when unpaired.

A bivariate analysis using Pearson correlation coefficient was performed when variables were normally distributed and the Spearman correlation coefficient was used when variables were nonnormally distributed.

Where bivariate analyses revealed significant correlations with either BPV or pulse rate parameters, linear regression analyses were used to identify independent correlations. Age, sex, BMI, and APAP compliance were considered as independent variables. A subanalysis compared BPV in patients with optimal and suboptimal APAP compliance. Nonnormally distributed data were log‐transformed prior to regression analysis and comparison. In order to facilitate their inclusion in regression analyses, nonnormally distributed data were log‐transformed to adjust for the distribution.

Results

Baseline Data

We recruited a total of 78 participants with OSA: 76.9% men, BMI 36.2±6.9 kg/m², age 49.0±12.9 years. At baseline, the systolic and diastolic BP values were 130.9±15.5 mm Hg and 82.7±10.4 mm Hg, respectively. Nocturnal pulse oximetry data confirmed severe sleep‐disordered breathing with a mean 4% ODI of 27.4±19.9 × hour⁻¹ and a mean 3% ODI of 33.4±20.3 × hour⁻¹, while average oxygen saturation was 93.9±3.2% and baseline ESS was 11.3±7.6 points (Table 1).

Table 1.

BP Parameters Before and After 2 Weeks of APAP Treatment in All Patients^a

BP Parameter	Before APAP	After APAP	P Value	Mean Δ	95% CI
Systolic BP, mean (SD), mm Hg	130.9 (15.5)	129.9 (14.9)	.486	−0.98	−3.74 to 1.77
Diastolic BP, mean (SD), mm Hg	82.7 (10.4)	82.2 (10.5)	.615	−0.50	−2.46 to 1.45
Systolic SD, mm Hg	5.3 (4.9)	4.2 (3.4)	.049	−1.10	−2.16 to −0.03
Diastolic SD, mm Hg	2.0 (2.1)	2.3 (2.0)	.410	0.38	−0.25 to 1.01
Systolic range, mm Hg	10.0 (8.8)	8.0 (6.8)	.04	−2.32	−4.46 to −0.19
Diastolic range, mm Hg	4.0 (4.0)	4.0 (4.0)	.483	0.68	−3.74 to 1.77
Pulse rate, beats per min	78.0 (14.5)	75.5 (15.8)	.033	−2.04	−7.60 to 3.52
Nocturnal pulse rate, mean (SD), beats per min	70.0 (8.8)	69.7 (8.7)	.416	−0.68	−2.32 to 0.95
Pulse rise index (h−1), mean (SD)	39.0 (20.4)	41.3 (19.6)	.703	0.61	.54 to .67
Nocturnal oxygen saturation, %	93.9 (3.2)	94.3 (3.1)	.993	0.05	−0.08 to 0.18
3% ODI	26.8 (28.4)	35.6 (29.7)	.841	0.98	−7.30 to 9.27
4% ODI	20.6 (27.0)	28.3 (26.3)	.675	−5.22	−9.98 to −0.47
Epworth Sleepiness Scale	11.0 (10.8)	7.5 (8.0)	<.001	−2.58	−5.01 to −0.14

Abbreviations: APAP, autotitrating continuous positive airway pressure; BP, blood pressure; CI, confidence interval; ODI, oxygen desaturation index; SD, standard deviation. Data are presented as median (interquartile range) unless otherwise indicated. ^aTotal of 78 hypertensive and normotensive participants with obstructive sleep apnea (OSA).

Thirty‐eight participants had hypertension, of these 17 (45%) were untreated, 13 (34%) were treated with angiotensin receptor blockers, three (8%) with β‐blockers, 12 (32%) with calcium channel blockers, five (13%) with diuretics, and one (3%) with α‐adrenergic blockers.

The hypertensive group (n=38) was older (51 [16.5] vs 43 [13] years, P=.008) and had a higher BMI (37.3 [8.0] kg/m² vs 32.8 [8.9] kg/m², P=.019) than the normotensive OSA group (n=40). No other baseline differences existed between the two groups (Table 2).

Table 2.

Baseline Characteristics and OSA Severity of Patients With and Without HTN

Parameter	HTN (n=38)	No NTN (n=40)	P Value
Male sex, %	71.0%	82.5%	.367
Age, y	51.0 (16.5)	43.0 (13.0)	.008
BMI, kg/m2	37.3 (8.0)	32.9 (8.9)	.019
Epworth Sleepiness Scale	12.0 (10.0)	10.0 (11.25)	.960
3% ODI, events per h	33.2 (31.9)	22.0 (19.2)	.121
4% ODI, events per h	25.1 (29.6)	17.9 (16.5)	.114
Mean oxygen saturation, %	93.5 (3.6)	94.2 (2.9)	.162

Abbreviations: BMI, body mass index; HTN, hypertension; ODI, oxygen desaturation index; OSA, obstructive sleep apnea. Data are expressed as median and interquartile range, and group differences were analyzed with Mann‐Whitney U test, unless otherwise indicated.

Two‐Week Follow‐Up

All 78 patients were seen at 2 weeks for follow‐up. No dropouts were recorded. There was a significant reduction in symptoms for all patients, as measured by the ESS (11.0±10.8 points vs 7.5±8.0 points, P<.001). The 95th percentile APAP was 13.3±5.0 cmH₂O and air leaks were acceptably low (at 95th percentile 0.30±0.60 L/s). APAP was used for 14.0±0.0 days and daily APAP usage was 3.0±2.3 hours per night. A total of 41.0% of patients used APAP for more than 4 hours per night and the total hours of APAP usage over 14 days was 43.3±31.4 hours.

There were no differences in compliance data when comparing the hypertensive and normotensive groups (Supporting Information, Table S1).

Absolute BP, BPV, and Pulse Rate at 2 Weeks

BPV, as expressed by systolic BP SD (5.3±4.9 vs 4.2±3.4, P=.049), pulse rate (78.0±14.5 vs 75.5±15.8, P=.033), and systolic BP range (10.0±8.8 vs 8.0±6.8, P=.040) decreased following 2 weeks of treatment compared with baseline. There was no difference in absolute BP (systolic and diastolic), diastolic SD, or diastolic range (Table 1).

CV Parameter Changes in Hypertensive vs Normotensive Participants

A significant change in pulse rate was observed in the hypertensive OSA cohort (P=.026) but not within the normotensive cohort (P=.471, Figure S1, Table S2, S3). A subanalysis of patients did not show any impact of compliance on changes in BP or BPV in the two groups but revealed a greater change in the pulse rate (0.02±8.92 beats per minute vs −4.44±6.72 beats per minute, P<.05) in the group with higher APAP compliance (Table S5).

APAP Compliance and BP and BPV

We found that delta mean systolic BP was associated with increased APAP pressure (r=0.293, P=.009), and delta BPV (systolic SD) was inversely associated with an increased air leak at an APAP 95th percentile (r=−0.237, P=.048). A decrease in the pulse rate (delta pulse) correlated with an increased number of days when APAP was used for >4 hours (r=−0.408, P<.001), a lower number of days when APAP was not used (r=0.334, P=.003), an increase in the total hours of usage (r=−0.355, P=.002), and an increased average daily usage (r=−0.352, P=.002) (Table 3). Regression analyses did not reveal any compliance measures to be independent predictors of a change in pulse rate (APAP days used >4 hours (P=.276), APAP days not used (P=.695), APAP total hours used (P=.847), and APAP average daily usage (P=.985) (Table 4).

Table 3.

Bivariate Analysis of APAP Compliance Variables With Delta Pulse

Delta Pulse	APAP Compliance
	Days>4 hours	Days Not Used	Total Hours	Average Daily Usage
r value	−0.408	0.334	−0.355	−0.351
P value	0.0001	0.003	0.002	0.002

Abbreviation: APAP, autotitrating continuous positive airway pressure.

Table 4.

Multiple Regression of APAP Compliance Data and Delta Pulse

APAP Compliance	Coefficients (Dependent Variable=Delta Pulse)
	Unstandardized Coefficients		Standardized Coefficients
Model	β	SE	β	t	Significance
Constant	.185	3.587		.052	.959
Days>4 hours	−.780	.711	−.464	−1.098	.276
Days not used	.134	.340	.074	.393	.695
Total hours	.037	.189	.140	.194	.847
Average daily usage	.043	2.302	.012	.018	.985

Abbreviations: APAP, autotitrating continuous positive airway pressure; SE, standard error.

When comparing patients with optimal and suboptimal CPAP compliance, there was no difference in the change in BP (Table S5 and S7), but a reduction in pulse rate in patients with optimal compliance (−5.50±[8.25] vs −0.50 [10.50] beats per minute; P<.05).

Discussion

Two weeks of APAP treatment showed a beneficial effect on BPV in patients with OSA. This effect seems to occur through modulation of BPV in hypertensive patients, most likely as a result of an improved sympathovagal balance.

These results indicate that short‐term treatment of OSA with CPAP has a favorable impact on BPV. BPV has a prognostic value on potential future CV events and this should be considered when assessing patients with OSA and hypertension.18 Although most correlations were modest, this was not unexpected, as hypertension is a multifactorial condition.

There is evidence showing that assessing BPV is important when determining the overall CV risk of a patient with hypertension. Rothwell and colleagues8 have demonstrated that systolic BPV and maximal systolic BP are strong predictors of stroke, independent of the mean systolic BP. An increased residual variability of systolic BP in patients with treated hypertension is associated with a high risk of vascular events. In comparison, patients with OSA are subjected to repeated surges of sympathetic nervous system activity,19 which explains the higher BPV in these patients.20

The observed decrease in BPV in patients with OSA treated with APAP is accompanied by a decrease in pulse rate, suggesting a beneficial effect of CPAP on overall sympathetic activation.21 Heart rate, represented by its surrogate marker pulse rate, can also predict long‐term BP changes in patients with OSA treated with CPAP22 and could be evaluated in prospective studies of patients with sleep‐disordered breathing who are treated with CPAP in clinical sleep services.

Supporting the findings that APAP might improve BP parameters, systolic BP changes were associated with increased APAP.

Air leaks are associated with low CPAP compliance,23 and we found that increased APAP leak reduced the effect of the treatment on BPV. This suggests that air leak influences compliance and that it has an indirect hemodynamic effect with a detrimental impact on BPV.

When we compared subgroups with optimal and suboptimal compliance, the decreases in BPV parameters did not reach statistical significance. This may have been the result of a small sample size, but it can also be related to the absolute baseline BP values. However, optimal CPAP compliance was associated with a greater reduction in pulse rate, suggesting a greater impact on sympathetic nervous system activation.

Limitations

This was a prospective study; however, participants were their own controls (pre‐post analysis) and included a relatively small cohort, which could potentially result in overestimating of the observed effect on BPV. Future investigations into the validity of the effect of APAP on BPV requires further randomized controlled studies.

Participants in this study were not assessed using polysomnography, which is the gold standard for the diagnosis of sleep apnea.24 However, national and international guidelines state that pulse oximetry is an effective diagnostic test for patients with OSA, and can therefore be used to evaluate patients at moderate and high risk of sleep‐disordered breathing.15 Routine clinical practice and sparse public health resources have further contributed to the widespread use of nocturnal home pulse oximetries in sleep services.

Office BP measurements might be influenced by the white‐coat effect,25 a stress‐induced rise in BP that occurs when patients are exposed to doctors in the clinical setting. However, as indicated by the Ohasama Study,26 office BP measurements have an important clinical meaning and predict subsequent risk of CV mortality.

Additional information regarding the number of medications used in treated patients may have been useful in determining efficacy of BP control; further, it is important to point out that this study was not powered to compare parameters in the subgroups of hypertensive and normotensive patients, but for a pretreatment vs posttreatment analysis in the whole cohort. Lastly, while only six of our participants had BPs <120/80 mm Hg on three separate occasions, future work should focus on the specific differences between normotensive and prehypertensive patients.

Conclusions

Two weeks of treatment with APAP in OSA patients is associated with a significant reduction in BPV, which may reflect the impact of treatment on sympathetic tone and could be used as a marker for CV risk assessment. Randomized controlled trials are needed to establish causality and to better understand whether changes in BPV lead to a reduction of CV events in patients with OSA.

Conflict of interest

The authors have no conflicts of interest related to this manuscript.

Supporting information

Table S1. Participant APAP compliance, including the differences between hypertensive OSA and normotensive OSA subgroups compared with paired t/Mann‐Whitney test.

Table S2. BP parameters before and after 2 weeks of APAP treatment in patients with hypertension.

Table S3. BP parameters before and after 2 weeks of APAP treatment in patients without hypertension.

Table S4. Baseline characteristics, OSA severity, and BP parameters of patients with optimal and suboptimal compliance.

Table S5. Changes in BP and BPV in patients with optimal and suboptimal compliance.

Table S6. Baseline characteristics, OSA severity, and BP parameters with extended criteria for optimal and suboptimal compliance (percentage of days and hours/day).

Table S7. Changes in BP parameters with extended criteria for optimal and suboptimal compliance (Δ=difference between before and after APAP treatment).

Figure S1. Line chart illustrating systolic and diastolic standard deviation changes before and after 2 weeks of APAP treatment.

Figure S2. Line chart illustrating pulse rate before and after 2 weeks of APAP treatment.

Figure S3. Line chart illustrating systolic and diastolic range before and after 2 weeks of APAP treatment.

J Clin Hypertens (Greenwich). 2016;18:1180–1184. DOI: 10.1111/jch.12845. ©2016 Wiley Periodicals, Inc.