Daily Exposure to Mild Intermittent Hypoxia Reduces Blood Pressure in Male Patients with Obstructive Sleep Apnea and Hypertension

Abstract

Rationale

Daily exposure to mild intermittent hypoxia (MIH) may elicit beneficial cardiovascular outcomes.

Objectives

To determine the effect of 15 days of MIH and in-home continuous positive airway pressure treatment on blood pressure in participants with obstructive sleep apnea and hypertension.

Methods

We administered MIH during wakefulness 5 days/week for 3 weeks. The protocol consisted of twelve 2-minute bouts of hypoxia interspersed with 2 minutes of normoxia. End-tidal carbon dioxide was maintained 2 mm Hg above baseline values throughout the protocol. Control participants were exposed to a sham protocol (i.e., compressed air). All participants were treated with continuous positive airway pressure over the 3-week period. Results are mean ± SD.

Measurements and Main Results

Sixteen male participants completed the study (experimental n = 10; control n = 6). Systolic blood pressure at rest during wakefulness over 24 hours was reduced after 15 days of MIH (142.9 ± 8.6 vs. 132.0 ± 10.7 mm Hg; P < 0.001), but not following the sham protocol (149.9 ± 8.6 vs. 149.7 ± 10.8 mm Hg; P = 0.915). Thus, the reduction in blood pressure from baseline was greater in the experimental group compared with control (−10.91 ± 4.1 vs. −0.17 ± 3.6 mm Hg; P = 0.003). Modifications in blood pressure were accompanied by increased parasympathetic and reduced sympathetic activity in the experimental group, as estimated by blood pressure and heart rate variability analysis. No detrimental neurocognitive and metabolic outcomes were evident following MIH.

Conclusions

MIH elicits beneficial cardiovascular and autonomic outcomes in males with OSA and concurrent hypertension.

Clinical trial registered with www.clinicaltrials.gov (NCT03736382).

At a Glance Commentary

Scientific Knowledge on the Subject

Severe intermittent hypoxia induced by sleep apnea has been linked to negative autonomic and cardiovascular outcomes in humans. However, therapeutic administration of mild doses of intermittent hypoxia may elicit beneficial autonomic and cardiovascular outcomes in patient populations.

What This Study Adds to the Field

In the present study, we show that repeated daily exposure to mild intermittent hypoxia for 15 days, in conjunction with in-home continuous positive airway treatment, reduces blood pressure in participants with obstructive sleep apnea.

Obstructive sleep apnea (OSA) is characterized by intermittent hypoxia and hypercapnia, substantial swings in intrathoracic pressure, and arousal from sleep. These perturbations, particularly severe levels of intermittent hypoxia, have been linked to autonomic, cardiovascular, cognitive, and metabolic dysfunction (1–3). The gold standard treatment for OSA is continuous positive airway pressure (CPAP). However, treatment adherence is low (4–6), and its efficacy in mitigating detrimental outcomes is limited in some cases (6). Thus, an additional treatment that directly targets comorbidities linked to OSA while potentially improving adherence to CPAP could contribute to significant improvements in outcome measures.

One possible adjunctive treatment is repeated daily exposure to mild intermittent hypoxia (MIH). MIH administered during wakefulness has been shown to significantly reduce blood pressure in untreated patients with hypertension without documented sleep apnea (7, 8) and in older men with or without coronary artery disease (9, 10). Although appropriate controls were not always employed in these studies, the reductions in blood pressure were significantly lower than reductions often reported in patients with OSA after treatment with CPAP (6). Thus, in contrast to the potential detrimental effects linked to naturally occurring severe intermittent hypoxia, mild forms administered therapeutically during wakefulness might directly improve autonomic and cardiovascular function. Moreover, exposure to MIH may mitigate cognitive (11) and metabolic (12) dysfunction in patients with hypertension.

In addition to its direct effects, MIH might indirectly improve detrimental outcome measures in patients with OSA by increasing adherence to CPAP. Exposure to MIH initiates long-term facilitation of upper airway muscles that is coupled to reductions in airway collapsibility and resistance during sleep in humans (13). Moreover, acute exposure to MIH during sleep is coupled to reductions in therapeutic CPAP without modification in airflow (13), and reductions in mean therapeutic pressure have been associated with improved adherence to CPAP (14).

Based on these findings, MIH may improve cardiovascular, metabolic, and neurocognitive measures both directly and indirectly via improved adherence to CPAP. Thus, we hypothesized that repeated daily exposure to MIH, coupled with in-home CPAP treatment, would reduce blood pressure in untreated patients with hypertension and OSA compared with patients treated with CPAP alone.

Methods

The institutional review board of Wayne State University and John D. Dingell Veterans Affairs Medical Center approved the protocol (#030617M1FV), which is registered in a clinical database (ClinicalTrials.gov #NCT03736382). Participants provided written consent prior to enrolling in the study.

Sixteen male participants completed the protocol (experimental group n = 10, control group n = 6) (Figure 1). A full description of the study design is in the online supplement (Figure E1). On Day 1, participants were screened to confirm the presence of untreated hypertension (i.e., ⩾130/80 mm Hg) without accompanying comorbidities besides OSA. We verified that participants were not treated with any medication or CPAP. Additional details related to the inclusion/exclusion criteria are in the online supplement. On Day 2, an overnight polysomnogram was completed to confirm the presence of OSA (see Polysomnography in the online supplement for details). Following completion of the polysomnogram, participants were randomly assigned to an experimental or control group using simple randomization. Participants were blinded to the planned intervention. On Day 3, CPAP was titrated to determine the therapeutic pressure. Moreover, measures of the active critical closing pressure were made to obtain measures of upper airway collapsibility (see the online supplement for details).

Figure 1.

Consolidated Standards of Reporting Trials flow diagram.

During completion of the MIH or control protocol (see below), participants were treated with therapeutic CPAP (Dreamstation; Respironics) at home on a nightly basis, except on the nights that corresponded to the middle and final days of the protocol. On these days, participants slept in the laboratory, and CPAP was titrated to determine the therapeutic pressure. If changes in therapeutic pressure were measured during the middle of the protocol, then the therapeutic CPAP was adjusted and used while sleeping at home.

Twenty-Four-Hour Blood Pressure Measurements

On Day 4, blood pressure was measured at home over a 24-hour period via a Holter monitor (ABPM-05; Meditech). When awake, participants were instructed to remain seated and relaxed in a quiet environment whenever possible. In addition, participants were not treated with CPAP during sleep. Blood pressure was measured every 20 minutes beginning Sunday at 6:00 a.m. and ending Monday at 6:00 a.m. Twenty-four-hour blood pressure measurements were repeated on the Sunday following completion of the MIH or sham protocol. Participants were provided a copy of their activity diary from the initial 24-hour blood pressure recording and were asked to complete similar activities.

Metabolic Parameters, Neurocognitive Tests, and Mild Intermittent Hypoxia/Sham Protocol

After the initial 24-hour blood pressure measures were completed, participants visited the laboratory on Monday morning, which was Day 5 of the study. Blood was drawn (22 ml) to obtain measures of blood biomarkers. Thereafter, neurocognitive tests were completed to assess sleepiness, vigilance, and memory. Additional information regarding the blood biomarkers and neurocognitive tests can be found in the online supplement. Subsequently, participants were exposed to the MIH or sham protocol. The experimental group was exposed to MIH between 7:00 and 9:00 a.m., 5 days a week (i.e., Monday–Friday), over 3 consecutive weeks. Similar considerations were employed when the control group was exposed to the sham protocol. Additional details of the MIH and sham protocol can be found in the online supplement (Figure E2). Blood draw and neurocognitive tests were also completed on the final day (i.e., Day 19) of exposure to MIH or the sham protocol.

Data Analysis

Twenty-four-hour blood pressure measurements

Blood pressure was initially separated into three categories based on data collected from an actigraph watch matched to self-reported activities documented in a journal. The journals were reviewed with the participant when the blood pressure monitor was returned. The three categories were wake-rest, wake-active, and sleep. Data were averaged each hour based on these three categories, and an overall average was subsequently calculated. Overall, the wake-rest and sleep categories accounted for 81.3% of the blood pressure measurements. The wake-rest and sleep categories were initially separated because factors that might influence blood pressure, and hence mask the effect of MIH or the sham protocol, were not controlled during sleep and may have varied before and after completion of the protocol. These factors could include potential differences in sleep architecture and variations in severity of breathing instability within an individual, since blood pressure measures were obtained without CPAP treatment before and after exposure to MIH or the sham protocol. Nonetheless, measures obtained during wake-rest alone and wake combined with sleep measures revealed similar findings (see Results). Wake-active measures were not reported in part because they were fewer in number. Likewise, the activities for a given participant were often dissimilar before compared with after the protocol (e.g., walking vs. driving), even though participants were instructed to follow a similar schedule and similar types of activity before and after the protocol.

Blood pressure and heart rate variability

On the initial (Day 5) and final day (Day 19) of the MIH or sham protocol, baseline beat-to-beat blood pressure (Portapres; Finapres Medical Systems) was obtained, along with an ECG, from the final 5 minutes of the first baseline period (see Figure E1). Additional blood pressure measures from the remaining portions of the protocol are not included in the present manuscript. Thereafter, blood pressure and heart rate variability and baroreflex sensitivity was determined. Details related to sampling, processing of the signals, and interpretation of the data can be found in the online supplement. In short, a low-frequency (0.04–0.15 Hz) and high-frequency range (0.15–0.40 Hz) was calculated. Spectral components for heart rate and blood pressure variability were expressed as normalized units, which were calculated using the following equation: [absolute power of the components ÷ (total power – very low-frequency power)] × 100. When normalized for total power, the low-frequency range is an estimated index of sympathetic activity to the heart and microcirculation, and the high-frequency range estimates parasympathetic activity (15, 16). Vt and breathing frequency were monitored to determine their impact, if any, on measures of heart rate and blood pressure variability.

Outcome Variables

Our primary outcome variable was blood pressure (systolic, diastolic, and mean arterial pressure) measured during wake-rest over the 24-hour period. Secondary outcomes included blood pressure measured during sleep over 24 hours, along with the measurement of beat-to-beat blood pressure, autonomic nervous system activity, blood biomarkers, and neurocognitive tests. Additional secondary outcomes included measures of upper airway collapsibility, therapeutic pressure, and adherence.

Statistical Analysis

A sample size calculation is included in the online supplement. Differences in baseline participant characteristics were compared using an unpaired t test. A 2 × 2 repeated measures ANOVA in conjunction with a Student-Newman-Keuls post hoc analyses was used to compare differences in the primary or secondary outcome measures. The factors in the analyses were group (experimental vs. control group) and time (before vs. after protocol or initial vs. final day). Unpaired t tests were used to compare the difference in the change from baseline in the experimental and control group. A Pearson correlation coefficient was used to determine if 24-hour blood pressure recordings during wake-rest were correlated to beat-to-beat blood pressure measures obtained on the initial and final day of the protocol. Likewise, a Pearson correlation coefficient was used to determine if the change in baseline breathing frequency or Vt were correlated to the change in heart rate and blood pressure variability from the beginning to the end of a given protocol. Lastly, Pearson correlations were completed to explore if increases in CPAP adherence were correlated to reductions in blood pressure in the experimental group. The P value was adjusted based on the three primary outcome measures. Consequently, P ⩽ 0.017 was considered statistically significant. Results are presented as mean ± standard deviation together with 95% confidence intervals in the figures and tables. In addition, individual participant results are presented in all figures.

Results

Participant Characteristics and 24-Hour Blood Pressure

Participant characteristics are shown in Table 1. Anthropometric and baseline physiological measures were similar in the experimental and control groups (Table 1). Likewise, systolic, diastolic, and mean arterial pressure blood pressure measured during wake-rest, over a 24-hour period, was similar in the experimental and control group before the start of the protocol (P ⩾ 0.212 for all blood pressure comparisons) (Figures 2A and 2B). Compared with baseline, systolic, diastolic, and mean arterial blood pressure were reduced after exposure to MIH (P < 0.001 for each blood pressure comparison) (Figure 2A), but not after exposure to the sham protocol (systolic blood pressure: P = 0.915; diastolic blood pressure: P = 0.443; mean arterial pressure: P = 0.394) (Figure 2B). Consequently, the change in systolic, diastolic, and mean arterial pressure from baseline was greater in the experimental group compared with control (Table 2). When blood pressure measures during wake-rest and sleep were combined, the reduction in blood pressure from baseline remained greater in the experimental group compared with the control group (Table 2). The results were upheld when an intent-to-treat analysis was completed (see the online supplement).

Table 1.

Participant Characteristics

Variable	Experimental (n = 10)	Control (n = 6)
Age, yr	40.7 ± 9.8	46.2 ± 10.3
Sex	M	M
Height, cm	174.2 ± 9.9	180.7 ± 6.6
Weight, kg	98.7 ± 22.8	105.2 ± 17.8
BMI, m2/kg	32.4 ± 4.4	32.3 ± 4.7
AHI, events/h	45.4 ± 17.2	44.7 ± 29.1
Apnea index	23.8 ± 17.2	26.0 ± 28.1
Hypopnea index	21.6 ± 11.0	18.7 ± 14.0
Apnea duration, s	20.4 ± 4.3	18.5 ± 2.4
Hypopnea duration, s	22.1 ± 4.5	18.8 ± 2.4
O2 desaturation, %	91.2 ± 2.3	91.8 ± 2.7
Baseline therapeutic CPAP, cm H2O	11.9 ± 2.1	11.5 ± 1.6
Mallampati score	2.4 ± 0.4	2.8 ± 0.8
Systolic blood pressure, mm Hg	147.2 ± 6.4	152.7 ± 4.8
Diastolic blood pressure, mm Hg	88.5 ± 9.3	92.9 ± 7.6
Mean arterial blood pressure, mm Hg	108 ± 7.8	112.0 ± 6.0
Ethnicity	African American (4), Asian (1), White (4), Hawaiian (1)	African American (6)

Definition of abbreviations: AHI = apnea–hypopnea index; BMI = body mass index; CPAP = continuous positive airway pressure.

Figure 2.

(A and B) Scatterplots that show average values (white circles) ± SD with 95% confidence intervals (orange circles overlaying the SD bars) for systolic, diastolic, and mean arterial blood pressure calculated from measures obtained during wake-rest over a 24-hour period before and after completion of the intermittent hypoxia (A) or sham protocol (B). Individual data for each participant are also shown (white circles connected by solid lines). (A and B) Note that blood pressure during wake-rest over the 24-hour period was reduced after repeated daily exposure to intermittent hypoxia (A) but not after exposure to the sham protocol (B). Experimental group n = 8 (blood pressure was not available in 2 participants because of technical difficulties); control group n = 6. A 2 × 2 repeated measures ANOVA in conjunction with a Student-Newman-Keuls post hoc analysis was used to compare differences within and between groups. *Significantly different from baseline. A = after protocol; B = before protocol.

Table 2.

Change from Baseline following Exposure to the Mild Intermittent Hypoxia or Sham Protocol

Variable	Experimental (Δ from Baseline)	Control (Δ from Baseline)	Difference (95% CI)	P Value
SBPw24, mm Hg	−10.91 ± 4.12	−0.17 ± 3.55	−10.74 (–15.32 to –6.16)	0.0003
DBPw24, mm Hg	−8.38 ± 4.72	−1.39 ± 3.77	−7.00 (–12.12 to –1.88)	0.012
MAPw24, mm Hg	−8.92 ± 4.17	−1.38 ± 3.29	−7.54 (–12.04 to –3.03)	0.004
SBPw+s24, mm Hg	−10.86 ± 4.33	−0.89 ± 5.36	−9.97 (–15.60 to –4.34)	0.002
DBPw+s24, mm Hg	−6.67 ± 3.41	−2.01 ± 2.20	−4.67 (–8.15 to –1.18)	0.013
MAPw+s24, mm Hg	−7.79 ± 2.98	−1.94 ± 2.89	−5.84 (–9.30 to –2.38)	0.003
SBPBB, mm Hg	−14.28 ± 4.89	−1.71 ± 3.43	−12.57 (–17.46 to –7.67)	0.0001
DBPBB, mm Hg	−8.00 ± 6.93	−0.94 ± 4.16	−7.06 (–13.81 to – 0.32)	0.041
MAPBB, mm Hg	−10.07 ± 4.87	−1.20 ± 3.48	−8.88 (–13.78 to –3.98)	0.002
BPV LF, n.u.	−11.43 ± 7.83	4.84 ± 12.09	−16.27 (–26.87 to –5.67)	0.005
BPV HF, n.u.	11.30 ± 8.33	−4.67 ± 11.62	15.97 (5.30 to 26.64)	0.006
HRV LF, n.u.	−11.44 ± 9.92	5.91 ± 15.37	−17.34 (–30.80 to –3.88)	0.015
HRV HF, n.u.	11.84 ± 12.76	−4.11 ± 10.27	15.95 (2.74 to 29.17)	0.021
Pcrit, cm H2O	−3.07 ± 1.89	0.28 ± 0.58	−3.36 (–5.08 to –1.63)	0.001
TP, cm H2O	−2.60 ± 0.84	0.00 ± 0.63	−2.60 (–3.46 to –1.74)	0.00001
CPAP adherence, h/night	1.64 ± 1.95	−0.34 ± 0.59	1.98 (0.21 to 3.76)	0.006

Definition of abbreviations: 24 = blood pressure measured over the 24-hour period; BB = beat-to-beat blood pressure; BPV = blood pressure variability; CI = confidence interval; CPAP = continuous positive airway pressure; DBP = diastolic blood pressure; HF = high-frequency; HRV = heart rate variability; LF = low-frequency; MAP = mean arterial pressure; n.u. = normalized units; Pcrit = critical closing pressure; SBP = systolic blood pressure; TP = therapeutic pressure; w = blood pressure recorded during wake; w + s = blood pressure recorded during wake and sleep.

Beat-to-Beat Blood Pressure and Autonomic Nervous System Activity

Beat-to-beat systolic, diastolic, and mean arterial pressure were similar in the experimental and control group during baseline on the initial day (P ⩾ 0.495 for each blood pressure comparison) (Figures 3A and 3B). Significant reductions in beat-to-beat systolic, diastolic, and mean arterial blood pressure were evident (P < 0.001 for each blood pressure comparison) on the final day of the MIH protocol (Figure 3A) but not the sham protocol (P ⩾ 0.359 for each blood pressure comparison) (Figure 3B). Thus, the reduction in systolic and mean arterial pressure from baseline was greater in the experimental group compared with control (Table 2), and the decrease in diastolic pressure approached significance. The beat-to-beat measures were correlated to the systolic and diastolic blood pressure measured during wake-rest over a 24-hour period, before and after completion of the MIH or sham protocol (before protocol: systolic R = 0.73, P = 0.004; diastolic R = 0.70, P = 0.005; after protocol: systolic R = 0.86, P < 0.001, diastolic R = 0.80, P < 0.001).

Figure 3.

(A and B) Scatterplots that show average values (white circles) ± SD with 95% confidence intervals (orange circles overlaying the SD bars) calculated from beat-to-beat systolic, diastolic, and mean arterial blood pressure recorded during baseline on the initial and final day of the intermittent hypoxia (A) or sham (B) protocol. (C and D) Likewise, blood pressure variability in the low- and high-frequency regions of the power spectrum determined from beat-to-beat blood pressure measurements obtained during baseline on the initial and final day of the intermittent hypoxia (C) or sham protocol (D). (E and F) Lastly, scatterplots that show heart rate variability in the low- and high-frequency regions of the power spectrum determined from an electrocardiogram obtained during baseline on the initial and final day of the intermittent hypoxia (E) or sham protocol (F). Individual data for each participant are also shown (white circles connected by solid lines). Note that beat-to-beat blood pressure measured during baseline on the final day, before exposure to the intermittent hypoxia or sham protocol, was reduced in the experimental group compared with the control group (A). (C and E) In addition, note that in response to repeated daily exposure to mild intermittent hypoxia, blood pressure and heart rate variability markers of sympathetic nervous system activity (i.e., low-frequency) and parasympathetic nervous system activity (i.e., high-frequency) were reduced (C) and increased (E), respectively. (D and F) In contrast, these changes were not evident after repeated daily exposure to the sham protocol . Experimental group n = 10; control group n = 6. A 2 × 2 repeated measures ANOVA in conjunction with a Student-Newman-Keuls post hoc analyses was used to compare differences within and between groups. *Significantly different from baseline. F = final day; HF = high-frequency; I = initial day; LF = low-frequency.

Baseline measures of low- and high-frequency blood pressure variability were similar in the experimental and control group during baseline on the initial day of the protocol (P ⩾ 0.557 for both comparisons) (Figures 3C and 3D). Significant increases in high-frequency (P = 0.002) and decreases in low-frequency measures (P = 0.002) were evident on the final day of the MIH protocol (Figure 3C) but not the sham protocol (P ⩾ 0.236 for both comparisons) (Figure 3D). Therefore, the increase in high-frequency measures and the decrease in low-frequency measures from baseline was greater in the experimental group compared with the control group (Table 2).

Measures of heart rate variability revealed similar findings to those outlined for blood pressure variability. No differences in baseline measures were evident between groups (P ⩾ 0.148) (Figures 3E and 3F). Significant increases in high-frequency (P = 0.007) and decreases in low-frequency measures (P = 0.010) were evident on the final day of the MIH protocol (Figure 3E) but not the sham protocol (P ⩾ 0.254 for both comparisons) (Figure 3F). Accordingly, the increase in high-frequency measures from baseline was greater in the experimental group compared with control (Table 2). The decrease in low-frequency measures from baseline was also greater in the experimental group compared with control, but the difference did not achieve statistical significance (Table 2). Baroreceptor sensitivity was similar between the groups (P = 0.718) and was similar within each group at the start and end of the protocol (P = 0.524).

No significant correlations between the change in breathing frequency or Vt and the change in low-frequency or high-frequency measures of blood pressure or heart rate variability from the initial to final day of the protocol were evident.

Critical Closing Pressure, Therapeutic Pressure, and CPAP Adherence

The active critical closing pressure was less positive after completion of the protocol compared with baseline (P < 0.001) in the experimental group (Figure 4) but not the control group (P = 0.662) (Figure 4). Therefore, the reduction in the active critical closing pressure from baseline was greater in the experimental group compared with the control group (Table 2).

Figure 4.

Scatterplots that show the average values (white circles) ± SD, with 95% confidence intervals (orange circles overlaying the SD bars), of the active critical closing pressure measured during sleep before and after exposure to the mild intermittent hypoxia (n = 10) or sham protocol (n = 6). Note that the active critical closing pressure was reduced following exposure to mild intermittent hypoxia but not after exposure to the sham protocol. Individual data for each participant is also shown (white circles connected by solid lines). A 2 × 2 repeated measures ANOVA in conjunction with a Student-Newman-Keuls post hoc analyses was used to compare differences within and between groups. *Significantly different from baseline. A = after protocol; B = before protocol.

Similarly, the therapeutic CPAP was reduced in the experimental group at the middle (not shown) and postprotocol time points compared with baseline (P < 0.001) (Figure 5A). The therapeutic pressure was similar at each time point in the control group (Figure 5A) (P = 1.000). As a result, the reduction in therapeutic CPAP from baseline was greater in the experimental group compared with the control group (Table 2). The change in the active critical closing pressure was correlated to the change in the therapeutic pressure (R = 0.72; P = 0.002).

Figure 5.

(A) Scatterplots that show the average values (white circles) ± SD, with 95% confidence intervals (orange circles overlaying the SD bars), showing the therapeutic continuous positive airway pressure measured during sleep at baseline and following exposure to mild intermittent hypoxia (MIH) or the sham protocol. Individual data for each participant are also shown (white circles connected by solid lines). Note that the therapeutic pressure was reduced following exposure to MIH but not the sham protocol. (B) Scatterplots that show the average values (white circles) ± SD, with 95% confidence intervals (orange circles overlaying the SD bars), showing adherence to in-home treatment with continuous positive airway pressure during the initial and final half of the MIH or sham protocol. Note that adherence improved in the final half of the protocol in those participants exposed to MIH but not the sham protocol. Individual data for each participant are also shown (white circles connected by solid lines). A 2 × 2 repeated measures ANOVA in conjunction with a Student-Newman-Keuls post hoc analyses was used to compare differences within and between groups. *Significantly different compared with baseline. A = after protocol; B = before protocol; F = final half of the protocol; I = initial half of the protocol.

The reduction in the therapeutic pressure in the experimental group was coupled to an increase in adherence to CPAP during the final half compared with the initial half of the protocol in the experimental group (P = 0.006) (Figure 5B). No increase in adherence was evident in the control group (P = 0.613). Consequently, the increase in adherence from baseline was greater in the experimental group compared with the control group (Table 2). The change in therapeutic pressure was correlated to the change in adherence to CPAP (R = −0.62; P = 0.011). In the experimental group, the therapeutic CPAP during the initial half of the protocol and final half, which was accompanied by a reduction in pressure, was effective as indicated by an apnea–hypopnea index that was below 5 events per hour (P = 0.698) (Figure E7). As expected, this was also the case for the control group during the initial and final half of the protocol. There was no correlation between the magnitude of change in adherence and the magnitude of change in 24-hour systolic (P = 0.655), diastolic (P = 0.356), or mean arterial pressure (P = 0.430) in the experimental group.

Neurocognitive and Metabolic Outcomes

Improvements in some neurocognitive measures were evident following MIH. In addition, no detrimental metabolic outcomes were evident. All neurocognitive and metabolic data are presented and discussed in the online supplement (Tables E2 and E3).

Discussion

The results from our study showed that repeated daily exposure to MIH, coupled with CPAP treatment, significantly reduced blood pressure during wake-rest in hypertensive participants with OSA. The findings were robust given that measures were obtained over a 24-hour period. Moreover, secondary measures of beat-to-beat blood pressure immediately before exposure to MIH on the initial and final day of the protocol revealed similar findings, which were correlated with the 24-hour blood pressure measures. Our findings also showed improved neurocognitive function and no detrimental metabolic adaptations following exposure to MIH (see the online supplement for discussion). Furthermore, MIH improved upper airway function during sleep, and these improvements were related to increased in-home CPAP adherence. In contrast to the findings in the experimental group, our results showed that blood pressure, upper airway function, and CPAP adherence were not altered in a group of patients with hypertension and OSA who were treated with CPAP alone.

A number of mechanisms could be responsible for the reductions in blood pressure that we observed. As a start to addressing the issue, we obtained indirect measures of parasympathetic and sympathetic nervous system activity via measures of blood pressure and heart rate variability (17–20). Our results indicated that sympathetic nervous system activity decreased and parasympathetic nervous system activity increased following repeated daily exposure to MIH. The modifications in blood pressure and heart rate variability that we observed were independent of breathing frequency and Vt, which are known to influence variability measures (17, 18). More specifically, breathing frequency and Vt during baseline were not altered after exposure to MIH. Likewise, no correlation was evident when changes in breathing frequency and Vt, after compared with before MIH, were correlated with changes in blood pressure and heart rate variability. Thus, our indirect evidence suggests that modifications in autonomic nervous system activity might have a role in the reduction in blood pressure that we observed. Direct measures of muscle sympathetic nervous system activity before and after repeated daily exposure to MIH will provide further support for these findings.

The modification in autonomic activity could be owing to modifications in centrally located nuclei, as baroreflex sensitivity remained unchanged following exposure to mild intermittent hypoxia, unlike the response to chronic continuous hypoxia (21) and chronic severe intermittent hypoxia (22). Other potential physiological mechanisms may also be responsible for the observed change in blood pressure. The prevailing thought is that MIH activates transcription factors that ultimately lead to angiogenesis and improved microvascular function (23). These vascular effects are thought to be mediated by upregulation of oxygen-sensitive genes that promote the release of hypoxia-inducible factor 1α (24), vascular endothelial growth factor (25), and alterations to nitric oxide production (8) that could result in reduced total peripheral resistance and subsequently lower blood pressure. Future studies that explore modifications in microvascular function following exposure to MIH will contribute to our understanding of the mechanisms that contribute to the reduction in blood pressure.

Although blood pressure was the primary outcome measure of our investigation, we also obtained secondary outcome measures of upper airway collapsibility and adherence to CPAP. Our results showed that MIH improved upper airway function and that this improvement was correlated to improved CPAP adherence. This improved adherence could be owing to reductions in airway resistance and collapsibility, along with potential modifications in the arousal threshold (26, 27). Independent of the mechanism, it is possible that MIH might indirectly influence outcome measures by improving treatment adherence, particularly over an extended time period. Indeed, it could be argued that the reduction in blood pressure that we observed was owing primarily to the improved adherence. However, no correlation between these variables was evident. In addition, the reductions in blood pressure observed during wake-rest exceeded reductions typically reported after treatment with CPAP alone (6). Likewise, no reductions in blood pressure were found in the control group despite adequate adherence in some participants. Moreover, profound reductions in blood pressure were observed after only 3 weeks of treatment, even when adherence was inadequate in one of the participants, in the experimental group. Typically, longer periods of treatment with CPAP are required to induce even small reductions in blood pressure (6, 28). To support our contention, further studies are required to determine if MIH alone will induce reductions in blood pressure in patients with hypertension and sleep apnea independent of treatment with CPAP. This group was not used in the present study because the safety and efficacy of MIH in modifying cardiovascular, metabolic, and neurocognitive function was unknown in hypertensive individuals with OSA.

Limitations and Future Considerations

The present small-scale physiological study was designed to examine if MIH reduces blood pressure both directly and indirectly by increasing CPAP adherence. A per-protocol analysis was used to explore the potential mechanisms responsible for the reduced blood pressure and to explore the acute efficacy of MIH when participants are fully compliant with the treatment. Our findings are an initial step to examining the long-term treatment efficacy of MIH and to determining the most effective dose.

Several intermittent hypoxia protocols have been used in the past with different goals in mind. Many protocols used long duration and more severe hypoxic protocols to simulate sleep-disordered breathing to induce the pathophysiology that accompanies this degree of exposure (29–32). However, as we have shown, along with other studies, shorter duration and milder forms of intermittent hypoxia (see reviews for detailed discussion [7, 23, 33, 34]) lead to beneficial outcomes. However, the most effective dose has not been determined. Likewise, whether the treatment can be sustained well beyond the period of administration requires further exploration.

Our findings are also limited to males with OSA and hypertension. Given that sex may impact cardiovascular health (35–40), and the response to MIH (41–43), future investigations are required to explore the efficacy of MIH on blood pressure in females with OSA and hypertension. Similarly, race might impact the response to MIH (33, 39). For example, the response to MIH could be greater in White individuals compared with African Americans. However, our results do not indicate that this is the case. The magnitude of the reduction in blood pressure was independent of race in the experimental group. Likewise, a reduction in blood pressure was observed in African Americans following MIH but not after exposure to the sham protocol. Nonetheless, because of the small n values, further studies are required.

Despite the limitations to our investigation, our data support the completion of a large-scale randomized control trial to investigate the efficacy of MIH on blood pressure in individuals with hypertension.

Conclusions

Repeated daily exposure to MIH coupled with nightly in-home treatment with CPAP may result in decreases in blood pressure that are not observed following treatment with CPAP alone, particularly for short treatment periods. Likewise, exposure to MIH is not accompanied by detrimental neurocognitive or metabolic outcomes. In addition, exposure to MIH improves upper airway collapsibility, which may lead to improved adherence to CPAP. These findings suggest that repeated daily exposure to MIH may reduce blood pressure directly by impacting mechanisms that influence blood pressure (i.e., autonomic nervous system activity and the microvasculature) and possibly by improving treatment adherence to CPAP.