Oxygen desaturation rate as a novel intermittent hypoxemia parameter in severe obstructive sleep apnea is strongly associated with hypertension

Abstract

Study Objectives:

To investigate the effects of different intermittent hypoxemia properties on blood pressure (BP) and short-term blood pressure variability (BPV) in severe obstructive sleep apnea (OSA) patients.

Methods:

Nocturnal BP was continuously monitored by measuring pulse transmit time. Apnea-related systolic BP elevation values were used to reflect BPV. Beat-to-beat R-R interval data were incorporated in polysomnography for heart rate variability analysis. The low-frequency/high-frequency band ratio was used to reflect sympathovagal balance. The rate of pulse oxyhemoglobin saturation (SpO₂) decrease was counted as the change in the percentage of SpO₂ per second after obstructive apnea and expressed as the oxygen desaturation rate (ODR). Patients with severe OSA (n = 102) were divided into 2 groups according to the median ODR: faster ODR (FODR group: ODR > 0.37, n = 50) and slower ODR (ODR ≤ 0.37, n = 52).

Results:

Comparisons between the 2 groups showed significantly higher systolic BP (SBP) values in the FODR group than in the slower ODR group (awake SBP 149.9 ± 18.3 vs 131.8 ± 15.6 mm Hg; asleep SBP: 149.6 ± 19.9 vs 128.7 ± 15.6 mm Hg; both P < .001), as well as short-term BPV (15.0 ± 4.8 vs 11.6 ± 3.6 mm Hg; P < .001), and the prevalence of hypertension (74.0% vs 26.9%; P < .001). Multiple linear regression analyses revealed that after adjusting for body mass index, functional residual capacity, expiratory reserve volume, and baseline SpO_2, ODR, as assessed by ΔSpO₂/Δt, had the strongest association with both BP and short-term BPV. Correlation analysis showed that ODR was positively correlated with the low-frequency/high-frequency band ratio (r = .288, P = .003).

Conclusions:

ODR, as a novel hypoxemia profile, was more closely associated with the elevation of BP and BPV in patients with severe OSA. FODR might be associated with enhanced sympathetic activity.

Clinical Trial Registration:

Registry: ClinicalTrials.gov; Name: Characteristics of Obstructive Sleep Apnea Syndrome Related Hypertension and the Effect of Continuous Positive Airway Pressure Treatment on Blood Pressure; URL: https://clinicaltrials.gov/ct2/show/NCT03246022; Identifier: NCT03246022.

Citation:

Weng N, Meng Z, Ding N, et al. Oxygen desaturation rate as a novel intermittent hypoxemia parameter in severe obstructive sleep apnea is strongly associated with hypertension. J Clin Sleep Med. 2020;16(7):1055–1062.

BRIEF SUMMARY

Current Knowledge/Study Rationale: The oxygen desaturation rate in patients with severe obstructive sleep apnea should be closely monitored, as extremely rapid exposure to hypoxemia after obstructive apneas could be a marker for hypertension and enhanced cardiovascular risk.

Study Impact: Oxygen desaturation rate can serve as a novel hypoxemia metric more closely associated with elevated blood pressure and blood pressure variability in patients with severe obstructive sleep apnea compared with traditional intermittent hypoxemia parameters.

INTRODUCTION

Obstructive sleep apnea (OSA) is characterized by recurrent collapse of the upper airway during sleep with consequent chronic intermittent hypoxemia (IH). IH is recognized as a major potential contributing factor to the pathogenesis of OSA-related hypertension.^1,2 The oxygen saturation fluctuations in the clinical disorder of OSA are cyclic drops in saturation rather than a sustained drop in oxygen saturation during sleep. High-frequency IH characterized by cycles of hypoxemia with reoxygenation differs from sustained low-frequency hypoxemia and induces excessive production of reactive oxygen species, a key inducer of endothelial dysfunction that can further contribute to hypertension development.^3,4 As the role of IH is critical to cardiovascular disease, the effects of different IH profiles (including the intensity, duration, and depth) on developing hypertension have not been fully investigated.

Follow-up studies have revealed that hypoxemia duration and depth are closely correlated with blood pressure (BP) elevation.^5,6 However, more attention should be paid to another associated IH profile, the oxygen desaturation rate (ODR), as extremely rapid desaturation in apnea can cause profound ventilation-perfusion mismatch, possibly resulting in organ damage.⁷ It has not been investigated previously whether ODR could serve as a novel parameter associated with the incidence of hypertension. Therefore, the main objective of this study was to investigate the effects of different IH properties, especially ODR, during obstructive apnea events on blood pressure (BP) changes and blood pressure variability (BPV).

In this study, pulse oxyhemoglobin saturation (SpO₂), as assessed by finger oximeter, was used as an easy means to measure blood oxygen stores. The rate of SpO₂ decrease (desaturation rate) has been used to reflect the oxygen consumption potential in a closed system (apnea event).^8,9 To better assess the IH and BP elevation as well as BPV, different aspects of IH profiles were investigated among patients with severe OSA, including the intensity, duration, depth, and ODR.

MATERIALS AND METHODS

Study population

Between March 2018 and April 2019, 565 consecutive patients with suspected OSA who underwent full-night polysomnography were recruited in this study. The inclusion criteria were (1) severe OSA (apnea-hypopnea index [AHI] > 30 events/h), (2) BP < 180/110 mm Hg, and (3) agreeing to cease oral antihypertensive drugs. The exclusion criteria were (1) use of sedatives or muscle relaxants, (2) those with atrial fibrillation or autonomic nervous system disease that might influence BP measurements, (3) patients who had been hospitalized for cardiac or respiratory exacerbations < 6 weeks prior to recruitment, and (4) those unwilling to participate in the study. In total, the data of 102 patients were included and analyzed.

Patients taking antihypertensive medication (n = 15) were asked to discontinue treatment prior to study enrollment over a supervised washout period. The exact timeframes for ceasing antihypertensive medications were determined by the last drug intake time and the drug half-life. To reduce the long-term effects of ceasing antihypertension drugs, we only recruited those whose medication stop time was shorter than 3 days. During the withdrawal period, patients whose BP was consistently ≥ 180/110 mm Hg or who complained of clinical symptoms (dizziness, headache, or palpitations) were excluded from the study and immediately given clinical intervention. The 102 recruited patients (69 men and 33 women) were between 23 and 70 years of age (50.3 ± 12.2 years) with a body mass index (BMI) of 29.5 ± 3.7 kg/m²; the median ODR was 0.37 (interquartile range: 0.29–0.47). Patients were divided into 2 groups based on the median ODR: the faster ODR (FODR) group (ODR > 0.37, n = 50) and the slower ODR (SODR) group (ODR ≤ 0.37, n = 52). All participants provided written informed consent before study participation. This study protocol was approved by the Scientific Research and Technology Ethics Committee of Huai'an First People's Hospital and registered in ClinicalTrials.gov (Clinical Trials ID: NCT 03246022).

Polysomnography monitoring

Patients underwent polysomnography (SOMNOmedics GmbH, Randersacker, Germany), and we recorded the following data: electroencephalogram, left and right electro-oscillograms, bilateral chin electromyography, nasal airflow using a nasal cannula and pressure transducer, naso-oral airflow using a thermistor, and respiratory effort using chest and abdominal strains. SpO₂ was monitored using a pulse oximeter. All tracings were scored manually according to the American Academy of Sleep Medicine guidelines.¹⁰ Apnea was defined as a decrease in airflow of ≥ 90% from baseline for a period of ≥ 10 seconds. Hypopnea was defined as a > 50% decrease in naso-oral airflow that was associated with a ≥ 3% oxygen desaturation. Arousal was defined as an abrupt shift in electroencephalogram lasting ≥ 3 seconds. The following sleep variables were calculated during the sleep period: preapneic SpO₂, the point value of the start of desaturation, baseline SpO_2, mean value of SpO₂ during10-minute quiet wakefulness period before sleep onset, respiratory event-related arousals (RERAs), oxygen desaturation index (ODI), the percentage of sleep time with oxygen saturation < 90% (T90), the mean and minimum value of SpO₂ during sleep stage (MeanSpO₂, MinSpO₂), while the number of apnea and hypopnea events per hour of sleep was determined to obtain AHI.

BP measurements

After participants had been seated for at least 15 minutes, 3 readings of systolic and diastolic BP (cuff BP) were obtained at 5-minute intervals by conventional mercury sphygmomanometry according to the recommendations of American Society of Hypertension.¹¹ Beat-to-beat BP was monitored continuously and synchronized with polysomnography via the pulse transit time (PTT)-based method. Several studies have shown that PTT has an inverse line correlation with BP.^12,13 PTT calibration was implemented when measurement by cuff realized 3 consecutive stable results after a supine resting period of 10 minutes. After BP calibration, all systolic and diastolic BP values, as well as pulse oximetry values were synchronized with polysomnography and analyzed using DOMINO software (version 2.7.0, GmbH, Randersacker, Germany). BP parameters were calculated as follows: (1) awake SBP: the average of systolic BP (SPB) values taken during a supine resting period of 10 minutes in the awake state, (2) asleep SBP: the average of SBP values during sleep, and (3) event-related BP elevation (ΔBP): the gap between the peak value of postapneic SBP and the lowest SBP during an obstructive apnea event, which was used to reflect short-term BPV.⁵

Heart rate variability analysis

Beat-to-beat RR interval data was incorporated in polysomnography measurements. Heart rate variability was assessed in the frequency domains. For the frequency domain, a fast Föurier transform spectral analysis was used. The results were in specific frequency bands, which correspond with specific physiological mechanisms. High-frequency (HF) bands reflect parasympathetic activity, while low-frequency (LF) bands reflect both parasympathetic and sympathetic activity. As a result, the LF/HF band ratio is a marker of sympathetic activity.^14,15

Oxygen desaturation rate

Hypopnea events were not included for analysis in this study since previous research has shown that the value of the ODR could be calculated only in a closed system.¹⁶ When hypopnea occurs, the upper airway is not completely obstructed and, as such, there is still some airflow. As hypopnea was not included in a closed system, our analysis was performed only on obstructive apneas. The length of desaturation was measured to the nearest 0.5 second (Δt). The decrease in SpO₂ during apnea was calculated as the gap from the start of the desaturation to the nadir of desaturation (ΔSpO₂). SpO₂ measurements were made to the nearest whole percentage. ΔSpO₂/Δt was expressed as the change in the percentage of SpO₂ per second during an obstructive apnea event and was used to reflect ODR.

Statistical analysis

Continuous data are presented as mean ± standard deviation. Normal distribution was statistically compared between 2 groups using the t test. Nonnormal distributions were transformed prior to analysis. Chi-square (χ²) tests were used to assess differences in categorical data. The Student–Newman–Keuls method was used for post hoc multiple comparisons. The relationship among variables was first examined using Pearson's correlation, and then after adjustment for BMI, functional residual capacity (FRC), expiratory reserve volume (ERV) and baseline SpO₂ by multiple linear regression analysis. A P value < .05 was considered statistically significant. All statistical analyses were performed with the SPSS statistical software package, version 16.0 (SPSS Inc., Chicago, IL).

RESULTS

Baseline characteristics

Comparisons of anthropometric, spirometry, and lung volume data between the 2 groups are shown in Table 1. No differences were observed between the 2 groups with regard to age, sex distribution, or forced expiratory volume in 1 second/forced vital capacity (all P > .05); however, patients in the FODR group had higher BMI (30.9 ± 3.4 vs 28.1 ± 4.4 kg/m²; P < .001), Epworth Sleepiness Scale scores (14.2 ± 4.4 vs 9.7 ± 4.4; P < .001), lower FRC (3.10 ± 0.54 vs 3.31 ± 0.47 L; P = .042) and ERV (1.07 ± 0.20 vs 1.16 ± 0.21 L; P = .024) compared with the SODR group.

Table 1.

Anthropometric, spirometric, lung volumes data in FODR and SODR groups.

Variables	FODR Group (n = 50)	SODR Group (n = 52)	P
Age, years	49.4.±12.8	51.1 ± 11.7	.482
Neck circumferences, cm	41.6 ± 3.3	39.7 ± 3.5	.005*
Waist circumferences, cm	109.3 ± 8.6	104.7 ± 8.7	.01*
BMI, kg/m2	30.9 ± 3.4	28.1 ± 4.4	< .001*
Sex(male/female)	37/15	32/18	.392
ESS score	14.2 ± 4.4	9.7 ± 4.4	< .001*
FEV1, L	3.05 ± .46	3.16 ± .46	.239
FVC, L	3.36 ± .48	3.47 ± .49	.265
TLC, L	5.58 ± .50	5.53 ± .49	.603
FRC, L	3.10 ± .54	3.31 ± .47	.042*
ERV, L	1.07 ± .20	1.16 ± .21	.024*

Values are means ± standard error. *P < .05, FODR versus SODR. Sex was tested by χ² test; other values were tested by unpaired Student’s t test. There was no significant difference in age, sex distribution, FEV1, FVC, or TLC. BMI, FRC, ERV, neck and waist circumferences were higher in the FODR group compared with the SODR group. BMI = body mass index, ERV = expiratory reserve volume, ESS = Epworth Sleepiness Scale, FEV1 = forced expiratory volume in 1 second, FODR= fast oxygen desaturation rate, FRC = functional residual capacity, FVC = forced vital capacity, SODR = slow oxygen desaturation rate, TLC = total lung capacity.

Comparison of BP Values and IH Profiles

As shown in Table 2, compared with the SODR group, patients in the FODR group displayed significantly higher BP levels (awake SBP: 149.9 ± 18.3 vs 131.8 ± 15.6 mm Hg; asleep SBP: 149.6 ± 19.9 vs 128.7 ± 15.6 mm Hg; all P < .001), LF/HF band ratio (2.89 ± 0.95% vs 2.40 ± 0.85%; P = .007), BPV (15.0 ± 4.8 vs 11.6 ± 3.6 mm Hg; P < .001) as well as the prevalence of hypertension (74.0% vs 26.9%; P < .001). Additionally, patients in the FODR group were more likely to experience severe sleep breath disturbance, as assessed by AHI (70.8 ± 16.1 vs 55.3 ± 16.9 events/h; P < .001), apnea index (58.2 ± 14.2 vs 39.7 ± 12.6 events/h; P < .001), and RERA occurrences (42.9 ± 18.0 vs 27.9 ± 11.3 events/h; P < .001) than those in the SODR group. When we compared the IH parameters between the 2 groups, IH exposure conditions in the FODR group were more severe than those in the SODR group, including intensity (ODI: 66.1 ± 16.3 vs 51.1 ± 15.4 events/h; P < .001), duration (desaturation length: 195.6 ± 65.7 vs 164.7 ± 60.3 min; P = .015), depth (MeanSpO₂: 89.9 ± 3.2% vs 93.2 ± 2.0%; MinSpO₂: 65.1 ± 7.5% vs 73.4 ± 8.5%; T90: 37.6 ± 18.2% vs 17.0 ± 14.6%; all P < .001), and especially ODR when apnea events occurred (ΔSpO₂/Δt: 0.50 ± 0.11% vs 0.28 ± 0.06%; P < .001). It is worth noting that patients in the FODR group had longer overnight desaturation durations but shorter single IH lengths (29.9 ± 5.6 vs 33.1 ± 8.8 seconds, P = .033) than those in the SODR group. In all patients, ODR was correlated with almost all measures of nocturnal hypoxemia (ODI: r = .578, P < .001; MeanSpO₂: r = −.695, P < .001; MinSpO₂: r = −.592, P < .001; and T90: r = .682, P < .001). Correlations between ODR with pulmonary function as well as with the LF/HF ratio are shown in Figure 1. Both FRC and ERV were negatively correlated with ODR, while ODR was positively correlated with BMI and the LF/HF band ratio.

Table 2.

Comparison of polysomnographic and blood pressure parameters between two groups.

Variables	FODR Group (n = 50)	SODR Group (n = 52)	P
awake SBP, mm Hg	149.9 ± 18.3	131.8 ± 15.6	< .001*
asleep SBP, mm Hg	149.6 ± 19.9	128.7 ± 15.6	< .001*
Hypertension, n (%)	37 (74.0)	14 (26.9)	< .001*
TST, hours	6.2 ± 1.0	6.1 ± .9	.783
AI, events/h	58.2 ± 14.2	39.7 ± 12.6	< .001*
AHI, events/h	70.8 ± 16.1	55.3 ± 16.9	< .001*
ODI, events/h	66.1 ± 16.3	51.1 ± 15.4	< .001*
T90, %	37.6 ± 18.2	17.0 ± 14.6	< .001*
BaselineSpO2, %	96.1 ± 1.30	96.7 ± 1.16	.01
MinSpO2, %	65.1 ± 7.5	73.4 ± 8.5	< .001*
MeanSpO2, %	89.9 ± 3.2	93.2 ± 2.0	< .001*
RERAs, events/h	42.9 ± 18.0	27.9 ± 11.3	< .001*
Single IH length, s	29.9 ± 5.6	33.1 ± 8.8	.033*
LF/HF, %	2.89 ± .95	2.40 ± .85	.007*
ΔSpO2/Δt, %/s	0.50 ± .11	0.28 ± .06	< .001*
ΔBP, mm Hg	15.0 ± 4.8	11.6 ± 3.6	< .001*
Desaturation duration, min	195.6 ± 65.7	164.7 ± 60.3	.015*

Values are means ± SE. *P < .05, FODR versus SODR. Hypertension was tested by χ² test; other values were tested by unpaired Student’s t test. ∆SpO₂ = the difference between the predesaturation value and the nadir of SpO₂ after the apnea, ∆SpO₂/∆t = used to reflect the ODR during apnea event, ∆t = the average length of desaturation event, AHI = apnea-hypopnea index, AI = apnea index, asleep SBP = the average value of the systolic blood pressure during sleep period, awake SBP = mean systolic blood pressure taken by PTT during10 minute quiet wakefulness period before sleep onset, Baseline SpO_{2 =} mean value of SpO₂ during10-minute quiet wakefulness period before sleep onset, Desaturation duration = the total time of nocturnal desaturation, FODR = faster oxygen desaturation rate, HF = high-frequency, LF = low-frequency, MeanSpO₂ = the mean SpO₂ during sleep, MinSpO₂ = the minimal SpO₂ during sleep, ODI = oxygen desaturation per hour of sleep, Single IH length = the average length of single desaturation event during sleep period, SODR = slower oxygen desaturation rate, T90 = percentage of sleep time with oxygen saturation < 90%, TST = total sleep time, ΔBP = the gap between the peak value of postapneic SBP and the lowest SBP during an obstructive apnea event, which was used to reflect short-term BPV.

Figure 1. Correlation of ODR with BP and short-term BPV.

Correlation with BP (A,B) and short-term BPV (C). BP = blood pressure, BPV = blood pressure variability, ODR = oxygen desaturation rate, SpO₂ = pulse oxyhemoglobin saturation.

Correlation between IH profiles and BP parameters

Pearson's correlations were performed between ODR and BP value. In all patients, ODR was positively correlated with awake SBP, asleep SBP, and BPV elevation (all P < .001) (Figure 2). To determine whether ODR was the strongest factor associated with hypertension and BPV elevation, multiple linear regression analyses were performed to examine the association between various measures of sleep-disordered breathing and hypoxemia during sleep (ie, ODR, AHI, ODI, RERAs, T90, MinSpO₂, MeanSpO₂, desaturation duration, predesaturation SpO₂, single IH length) with BP parameters (BPV and SBP at awake and asleep) after adjusting for BMI, FRC, ERV, and baseline SpO₂.The results of these analyses, which showed that ODR played a dominant role in both the elevation of BP and BPV, are summarized in Table 3, Table 4, and Table 5.

Figure 2. Correlation analysis for relationship of ODR with LF/HF band ratio, BMI, and lung volume.

Correlation with LF/HF band ratio (A), BMI (B), and lung volume (C,D). BMI, body mass index, LF/HF = low-frequency/high-frequency, ODR = oxygen desaturation rate, SpO₂ = pulse oxyhemoglobin saturation.

Table 3.

Analysis with awake SBP as a dependent variable (model 1).

Independent variables	Pearson's Correlation		Multiple Regression
	r	P	β	Lower 95%	Upper 95%	P
ODR	.542	< .001	.605	40.255	120.736	< .001
ODI	.332	.001	.420	−1.005	.084	.096
AHI	.386	< .001	.388	−.063	.880	.089
T90	.411	< .001	.311	−.199	−.165	.781
MinSpO2	−.260	.008	.177	−.141	.895	.151
MeanSpO2	−.382	< .001	.132	−1.794	3.425	.536
RERAs	.304	.002	−.113	−.444	.186	.417
Desaturation duration	.225	.023	.136	−.049	.130	.374

In multiple regression model, adjusted variables include BMI, lung function variables, and baseline SpO₂. AHI = apnea-hypopnea index, Desaturation duration = the total time of nocturnal oxygen desaturation, MeanSpO₂ = the mean SpO₂ during sleep, MinSpO₂ = the minimal SpO₂ during sleep, ODI = oxygen desaturation index during sleep, ODR (ΔSpO₂/Δt) = the rate fall in SpO₂ during desaturation event, RERAs = respiratory event-related arousals, T90 = percentage of sleep time with oxygen saturation < 90%.

Table 4.

Analysis with asleep SBP as a dependent variable (model 2).

Independent Variables	Pearson's Correlation		Multiple Regression
	r	P	β	Lower 95%	Upper 95%	P
ODR	.582	< .001	.574	40.898	123.574	<.001
ODI	.377	< .001	−.340	−.960	.158	.158
AHI	.421	< .001	.347	−.091	.878	.110
T90	.477	< .001	.385	−.076	.896	.097
MinSpO2	−.337	.001	.111	−.276	.788	.342
MeanSpO2	−.438	< .001	.143	−.727	3.634	.482
RERAs	.323	.001	−.134	−.488	.158	.314
Desaturation duration	.256	.010	.103	−.059	.125	.481

In multiple regression model, adjusted variables include BMI, lung function variables, and baseline SpO₂. AHI = apnea-hypopnea index, Desaturation duration = the total time of nocturnal oxygen desaturation, MeanSpO₂ = the mean SpO₂ during sleep, MinSpO₂ = the minimal SpO₂ during sleep, ODI = oxygen desaturation index during sleep, ODR (ΔSpO₂/Δt) = the rate fall in SpO₂ during desaturation event, RERAs=respiratory event-related arousals, T90 = percentage of sleep time with oxygen saturation < 90%.

Table 5.

Analysis with short-term BPV as a dependent variable (model 3).

Independent Variables	Pearson's Correlation		Multiple Regression
	r	P	β	Lower 95%	Upper 95%	P
ODR	.474	< .001	0.349	1.114	20.964	.030
ODI	.416	< .001	-0.108	−.150	.094	.646
AHI	.438	< .001	0.214	−.063	.171	.364
T90	.461	< .001	0.143	−.086	.153	.577
PredesaturationSpO2	−.100	.320	0.036	−.659	.912	.750
MeanSpO2	−.410	< .001	0.082	−.564	.807	.726
RERAs	.454	< .001	0.134	−.040	.112	.343
Single IH length	.057	.567	0.091	−.074	.184	.398

In multiple regression model, adjusted variables include BMI, lung function variables, and baseline SpO₂. AHI = apnea-hypopnea index, MeanSpO₂ = the mean SpO₂ during sleep, ODI = oxygen desaturation index during sleep, ODR (ΔSpO₂/Δt) = the rate fall in SpO₂ during desaturation event, Predesaturation SpO₂ = the point value of the start of desaturation, RERAs = respiratory event-related arousals, Single IH length = the average length of single desaturation event during sleep period, T90 = percentage of sleep time with oxygen saturation < 90%.

DISCUSSION

It has been increasingly recognized that IH elicits chemoreflex stimulation with consequent sympathetic activation and vasoconstriction, thus contributing to hypertension. An analysis of data from the study by Punjabi et al¹⁷ demonstrated that only those hypopneas associated with a desaturation of 4% or more were independently associated with cardiovascular diseases, while there was no association with the hypopnea event accompanied by lower degree of desaturation. Additionally, we previously demonstrated that T90 had a stronger association with elevated awake BP and asleep BP levels compared with other sleep breathing disturbance parameters (eg, AHI and ODI).⁵ These studies suggest that the adverse cardiovascular effects of IH are primarily dependent upon the severity of intermittent hypoxemia. In contrast with the previous studies, we provide a novel insight into the hemodynamic response to different IH pattern. In this study, we investigated the response of BP to the speed of desaturation when an obstructive apnea occurred. A major finding of our study is that instead of traditional IH parameters (T90 and ODI), ODR played had a stronger association with the elevation of both awake and sleep BP levels and short-term BPV, which are all in turn associated with altered autonomic function, as shown by increased sympathetic activity on heart rate variability analysis.

During an obstructive apnea event, ODR can be altered by changing oxygen stores or by increasing the rate of metabolic oxygen consumption.^16,18 Of these 2 factors, oxygen stores are influenced by lung volume. A reduction in FRC and ERV are proportional to obesity, which could affect lung oxygen stores.¹⁹ Once an apnea event has occurred, the body becomes a closed system, meaning that ODR is heavily influenced by the levels of lung O₂ store. Therefore, the lower the thoracic gas volume, the lower the alveolar O₂ will be and oxygen stores will be depleted more rapidly and ODR will be faster. Consistent with previous findings, our study observed that patients in the FODR group had higher BMI and shorter desaturation event duration than those in the SODR group. Moreover, obesity can also lead to a restrictive ventilatory defect. It has been suggested that airflow resistance is increased in the extrathoracic and peripheral airways in OSA, especially in patients with obesity and OSA.²⁰ Patients with FODR had a significantly higher BMI and AHI compared with patients with SODR. Thus, to reopen the more frequently collapsed pharyngeal airway and overcome the excess mechanical load, the diaphragm load must be higher, and even accessory respiratory muscles may need to be recruited. The augmented respiratory effort could increase the metabolic rate, and increased metabolism could further augment tissue O₂ utilization during apnea, which has been suggested to be another important factor for FODR.

The response of the sympathetic nervous system to hypoxemia may also be correlated with the hypoxemia pattern. A previous study in a murine model demonstrated that sympathetic activation and the systemic hypertensive response occurred when the animals were exposed to chronic intermittent hypoxia, exposure but not to sustained hypoxic exposure.²¹ A primary stimulus for OSA-related hypertension is nocturnal exposure to chronic intermittent hypoxemia. Chronic intermittent hypoxemia profiles comprise a wide range of parameters, including the length and intensity of hypoxemia exposure. Comparison of IH characteristics between the 2 groups showed that patients who experienced more short durations of single IH and higher desaturation rates per night were more likely to have serious effects on BP elevation, which were associated with enhanced sympathetic tone. Our data supported that the LF/HF ratio, an index for cardiac sympathetic modulation, was closely correlated with the slope of ODR. The HF band reflects parasympathetic activity, while the LF band reflects both parasympathetic and sympathetic activity. Consequently, LF/HF is a ratio of parasympathetic to sympathetic activity, and an elevation of this ratio is a sign of higher sympathetic activity.^14,15 Bivariate correlation was performed between desaturation rate and the LF/HF ratio in our population, which showed that FODR was associated with a high LF/HF band ratio. This suggests that high ODR associated with OSAS could lead to severe autonomic imbalance, favoring sympathetic activity, which implies an elevated BP. ODR resulted in rapidly ascending and descending oxygen saturation, which could lead to maladaptive responses by differential modulation of hypoxia-inducible factor 1 and 2, which are key factors in endothelia dysfunction.²² However, endothelial dysfunction has important implications for vasomotor function and plays a crucial role in hypertension.²³ In contrast, slow oxygen consumption could result in longer desaturation, lasting nearly a minute. This could even be modeled to sustain the hypoxemia profile, which can be seen as adaptive responses leading to an increase in erythropoiesis; in this hypoxemia exposure condition, no BP elevation was observed.²⁴ Moreover, the carotid body, the main O₂ chemoreceptor, plays a key role in the generation of hypertension after chronic intermittent hypoxemia.^25,26 However, not all IH conditions can fully alter carotid body sensitivity to hypoxemia. A previous report by Peng and Prabhakarl²⁷ demonstrated that arterial BP was significantly elevated in short durations of IH but not in longer durations of IH, which was correlated with the enhanced hypoxic sensitivity of the carotid body. The differences between the effects of short durations of IH and longer durations of IH on the carotid body suggest that frequent brief IH events, which were closely associated with a rapid desaturation rate, have more serious effects on chemosensory activity than less frequent, but longer IH events. Thus, we suggest that the effects of ODR on the pathogenesis of hypertension are multifactorial, including increased sympathetic activation, endothelial dysfunction, and enhanced hypoxic sensitivity of the carotid body.

Although this study revealed important discoveries, the following limitations of our study should be addressed. First, as the study population was confined to severe OSA, our findings cannot be extrapolated to patients with mild or moderate OSA. Second, we only inquired about the years of snoring; however, the duration of OSA in the 2 groups was unknown. Therefore, the total duration of exposure to hypoxemia could not be compared between the 2 groups because BP elevation may be also correlated with hypoxemia exposure time. The LF/HF ratio analysis was only performed using nighttime observations, so these results might not be applicable for autonomic nerve function during the day. Lastly, the cross-sectional nature of the study does not allow us to establish causality.

CONCLUSIONS

In summary, our study shows that ODR is a clinically significant variable in patients with severe OSA, for which those with FODR might have higher BP and BPV than those with SODR.

DISCLOSURE STATEMENT

All authors have seen and approved the manuscript. This work was funded by the National Natural Science Foundation of China (81900084).

ABREVIATIONS

AHI

apnea-hypopnea index

BMI

body mass index

BP

blood pressure

BPV

blood pressure variability

ERV

expiratory reserve volume

FODR

faster oxygen desaturation rate

FRC

functional residual capacity

HF

high-frequency

IH

intermittent hypoxemia

LF

low-frequency

MeanSpO2

mean SpO₂

MinSpO2

minimum SpO₂

ODI

oxygen desaturation index

ODR

oxygen desaturation rate

OSA

obstructive sleep apnea

PTT

pulse transit time

RERA

respiratory event-related arousal

SBP

systolic blood pressure

SODR

slower oxygen desaturation rate

SpO2

pulse oxyhemoglobin saturation

T90

percentage of sleep time with oxygen saturation < 90%