Obstructive sleep apnea seems to have an important influence on the autonomic nervous system. In this study, we assessed the relations of sleep apnea–related parameters with 24-hour heart rate variability (HRV) in a large population of young and healthy adults. Participants aged 25 to 41 years with a body mass index <35 kg/m 2 and without known obstructive sleep apnea were included in a prospective population-based cohort study. HRV was assessed using 24-hour electrocardiographic monitoring. The SD of all normal RR intervals (SDNN) was used as the main HRV variable. Apnea-Hypopnea Index (AHI) and oxygen desaturation index (ODI) were obtained from nighttime pulse oximetry with nasal airflow measurements. We defined sleep-related breathing disorders as an AHI ≥5 or an ODI ≥5. Multivariable regression models were constructed to assess the relation of HRV with either AHI or ODI. Median age of the 1,255 participants was 37 years, 47% were men, and 9.6% had an AHI ≥5. Linear inverse associations of SDNN across AHI and ODI groups were found (p for trend = 0.006 and 0.0004, respectively). The β coefficients (95% CI) for the relation between SDNN and elevated AHI were −0.20 (−0.40 to −0.11), p = 0.04 and −0.29 (−0.47 to −0.11), p = 0.002 for elevated ODI. After adjustment for 24-hour heart rate, the same β coefficients (95% CI) were −0.06 (−0.22 to 0.11), p = 0.51 and −0.14 (−0.30 to 0.01), p = 0.07, respectively. In conclusion, even early stages of sleep-related breathing disorders are inversely associated with HRV in young and healthy adults, suggesting that they are tightly linked with autonomic dysfunction. However, HRV and 24-hour heart rate seem to have common information.
Sleep-related breathing disorders, such as obstructive sleep apnea (OSA), are highly prevalent and remain often undiagnosed. OSA is independently associated with an increased risk of hypertension, coronary artery disease, and sudden death. The autonomic nervous system is involved in numerous physiologic processes of the cardiovascular system. Heart rate variability (HRV) is a validated and easily obtainable measurement of the influence of autonomic function on the heart and by itself associated with cardiovascular risk factors and mortality. An experimental study showed markedly greater sympathetic nerve activity in patients with OSA without treatment than that in treated patients. Another small study has found a worse HRV profile and an increased heart rate (HR) directly after an apnea. An unbalanced autonomic function in patients with manifest OSA could therefore be one potential reason for their increased cardiovascular risk. Population-based studies assessing the relation of the autonomic nervous function with OSA are scarce, and the shape of the relation is currently unknown. In addition, it is unclear whether autonomic function is altered or influenced by sleep-related breathing disorders among otherwise asymptomatic subjects with subclinical sleep-related breathing disorders. Finally, HR by itself is an independent risk factor for cardiovascular events, seems to be modified in patients with OSA, and is tightly linked with HRV. In this context, it is unknown whether HRV contains any additional information beyond HR. To address some of these issues, we evaluated the association of HRV with several sleep apnea–related parameters in a large cohort of young and healthy adults from the general population.
Methods
From 2010 to 2013, inhabitants of the Principality of Liechtenstein aged from 25 to 41 years were invited to participate in the “Genetic and Phenotypic Determinants of Blood Pressure and Other Cardiovascular Risk Factors” (GAPP) study, a prospective population-based cohort study. Study design and method have been published previously. Study exclusion criteria were known OSA, renal failure, current intake of antidiabetic drugs, a body mass index (BMI) >35 kg/m 2 , established cardiovascular disease, or other severe illnesses. Of the 2,170 participants enrolled, 1,410 participants performed a nighttime pulse oximetry and nasal flow measurement. Of these, participants were excluded if the sleep study duration was <180 minutes (n = 114) or if they had missing or incomplete 24-hour electrocardiographic (ECG) recordings (n = 19) or other missing covariates (n = 22), leaving 1,255 participants for the current analysis. Written informed consent was obtained from every participant, and the study protocol was approved by the local ethics committee.
Every participant underwent a 24-hour 3-channel Holter ECG recording with a validated device (AR12plus; Schiller AG, Baar, Switzerland). When the ECG monitoring duration was <80% of the target time (i.e., <19.2 hours), the study was repeated whenever possible. ECG devices were started in the morning after the study examination. All Holter studies were postprocessed using a dedicated software (MediLog Darwin V2; Schiller AG) to remove artifacts and redefine premature ventricular and atrial beats. Time- and frequency-domain HRV and mean HR were automatically calculated by the software. Spectral analysis was performed using an autoregressive method. The SD of all normal RR intervals (SDNN) was predefined to be the main HRV variable. In addition, low frequency (LF), high frequency (HF), and total power (TP) were used for this analysis. LF and HF were normalized, by calculating LF/(TP−VLF) × 100 and HF/(TP−VLF) × 100, respectively.
Nighttime pulse oximetry with nasal flow measurement was performed in every participant using a validated device (ApneaLink; ResMed, San Diego, California). Participants were instructed to place the nasal flow cannula and the finger pulse oximetry probe, to start the device before falling asleep, and to stop the measurement when waking up in the morning. Recording length had to be at least 180 minutes for both nasal airflow measurement and pulse oximetry, otherwise participants were asked to repeat the measurement. Apnea-Hypopnea Index (AHI) was defined as the average number of apnea and hypopnea episodes per hour of sleep. An apnea was defined as a nasal airflow reduction of at least 80% during ≥10 seconds. A hypopnea was defined as a nasal airflow reduction of ≥30% with a concomitant decrease in oxygen saturation of ≥4%. oxygen desaturation index (ODI) was defined as the mean number of oxygen desaturations of ≥4% per hour of recording. Sleep-related breathing disorders were defined as an AHI ≥5 or an ODI ≥5.
Questionnaires were used to obtain information about personal, medical, lifestyle, and nutritional factors. Smoking status was self-assessed and classified as never, past, or current smoker. Physical activity was assessed using the International Physical Activity Questionnaire, and regular physical activity was defined as ≥150 minutes of moderate activity or ≥75 minutes of vigorous activity per week. Information about fruit and vegetable consumption was dichotomized into ≥5 versus <5 servings per day. Highest educational status achieved was self-reported and classified into the categories high school, college and university degree. Daytime sleepiness was assessed using the Epworth Sleepiness Scale (ESS) Questionnaire. Height and weight were measured in a standardized way by trained study nurses. BMI was calculated as body weight in kilograms divided by height in meters squared. Office blood pressure was measured in a sitting position after 5 minutes of rest. Three blood pressure measurements were performed. For the current analysis, the mean of the second and third measurement was used. Low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, glycated hemoglobin A1c, creatinine, and copeptin were measured from a fasting venous blood sample immediately after the blood draw according to the standard method. Estimated glomerular filtration rate was calculated using the creatinine-based Chronic Kidney Disease Epidemiology Collaboration formula.
Baseline characteristics were stratified according to the presence or absence of an AHI ≥5. Distribution of continuous variables was checked using skewness, kurtosis, and visual inspection of the histogram. Continuous variables are presented as medians (interquartile ranges) and categorical variables as numbers (%). Group comparisons were done using Wilcoxon rank-sum tests or chi-square tests, as appropriate.
To evaluate the relation of AHI or ODI with HRV, separate multivariable linear regression models were constructed using different HRV variables as the dependent variable. To assess the linearity of the relations with HRV, AHI, and ODI were categorized into 0 (reference category), 1, 2, 3 to 4, and ≥5 episodes per hour. LF, HF, and TP were log transformed. All HRV variables were converted into z-scores to improve comparability. All models were adjusted for a predefined set of covariates, including gender, age, BMI, current smoking, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, copeptin, prediabetes, systolic blood pressure, estimated glomerular filtration rate, regular physical activity, fruit and vegetable consumption, and educational status. We also evaluated the association of HR with sleep-related breathing disorders in similar models described previously. To evaluate the incremental information of HRV beyond HR, all HRV-based models were additionally adjusted for HR at rest and 24-hour HR in a separate step. To assess potential effect modification of the relation of HRV with sleep-related breathing disorders, prespecified subgroup analyses for gender, smoking, BMI, age, and physical activity were performed. Multiplicative interaction terms were entered in the nonstratified models to evaluate significant subgroup effects. A p value <0.05 was prespecified to indicate statistical significance. SAS, version 9.4 (SAS Institute Inc., Cary, North Carolina), was used for all analyses.
Results
Baseline characteristics stratified by the presence or absence of an AHI ≥5 are presented in Table 1 . Overall, 120 participants (9.6%) had an AHI ≥5, 83% of them were men. Compared with participants with a normal AHI, participants with an AHI ≥5 had a greater BMI (27 vs 24 kg/m 2 , p <0.0001), were more often current smokers (29% vs 21%, p = 0.04), and had a greater blood pressure (128 vs 120 mm Hg for systolic and 85 vs 78 mm Hg for diastolic blood pressure, p <0.0001 for both comparisons). The ESS was not significantly different between both groups (median score 7 vs 7, p = 0.38). SDNN and HF were lower in participants with an AHI ≥5 (147 vs 151 ms, p = 0.04 and 13 vs 16 ms 2 , p = 0.002, respectively). Normalized LF and TP did not differ between the 2 groups.
n=1255 | Apnea-Hypopnea Index | p-value | |
|---|---|---|---|
| < 5 n=1135 | ≥ 5 n=120 | ||
| Age (years) | 35 (30; 39) | 38 (33; 40) | 0.001 |
| Men | 492 (43.4%) | 100 (83.3%) | <0.0001 |
| Smoker | 0.08 | ||
| Current | 238 (21.0%) | 35 (29.2%) | |
| Past | 260 (22.9%) | 29 (24.1%) | |
| Never | 637 (56.1%) | 56 (46.7%) | |
| Body mass index (kg/m 2 ) | 23.9 (23.9; 26.5) | 27.4 (24.5; 30.9) | <0.0001 |
| Systolic blood pressure (mmHg) | 120 (111; 127) | 128 (120; 134) | <0.0001 |
| Diastolic blood pressure (mmHg) | 78 (73; 84) | 85 (79; 89) | <0.0001 |
| Hypertension | 142 (12.5%) | 33 (27.5%) | <0.0001 |
| Low-density lipoprotein (mmol/l) | 2.85 (2.33; 3.44) | 3.34 (2.75; 3.92) | <0.0001 |
| Low-density lipoprotein (mg/dl) | 110.2 (90.1; 133.2) | 129.2 (106.2; 151.7) | <0.0001 |
| High-density lipoprotein (mmol/l) | 1.53 (1.27; 1.79) | 1.22 (1.02; 1.48) | <0.0001 |
| High-density lipoprotein (mg/dl) | 59.1 (49.1; 69.1) | 47.1 (40.0; 57.1) | <0.0001 |
| Prediabetes | 244 (21.5%) | 38 (31.7%) | <0.0001 |
| High-sensitivity C-reactive protein (mmol/l) | 1.0 (0.5; 2.0) | 1.5 (0.6; 3.2) | 0.002 |
| Education | 0.01 | ||
| High school | 81 (7.1%) | 18 (15.0%) | |
| College | 629 (55.4%) | 60 (50.0%) | |
| University degree | 425 (37.4%) | 42 (35.0%) | |
| Apnea-Hypopnea Index | 1 (0; 2) | 8 (6; 12) | <0.0001 |
| Oxygen Desaturation Index | 1 (0; 2) | 9 (6; 13) | <0.0001 |
| Oxygen Desaturation Index ≥5 | 36 (3.2) | 105 (87.5) | <0.0001 |
| Epworth Sleepiness Scale | 7 (5;10) | 7 (5; 11) | 0.38 |
| Epworth Sleepiness Scale ≥11 | 221 (19.5) | 30 (25.0) | 0.15 |
| Resting heart rate (bpm) | 61 (56; 67) | 63 (58; 68) | 0.04 |
| Ambulatory heart rate (bpm) | 74 (69; 80) | 76 (71; 82) | 0.09 |
| Standard deviation of all normal RR intervals (ms) | 151.3 (128.3; 177.7) | 147.1 (118.2; 170.4) | 0.04 |
| High frequency nu (ms/s 2 ) | 15.5 (11.4; 19.9) | 12.7 (9.1; 17.4) | 0.002 |
| Low frequency nu (ms/s 2 ) | 53.6 (47.5; 59.0) | 56.8 (51.2; 61.9) | 0.55 |
| Total power (ms/s 2 ) | 3625 (2502; 5058) | 3459 (2119; 4887) | 0.12 |
The relations between HRV categories and AHI are presented in Table 2 . After multivariable adjustment, SDNN significantly decreased across increasing AHI categories (p for trend = 0.006). There was no significant relation of AHI with normalized LF, normalized HF, and TP. Additional adjustment for HR at rest and 24-hour HR attenuated these relations. Similar findings were obtained when AHI categories were replaced by ODI categories, as listed in Table 3 . After multivariable adjustment, we found significant inverse relations for SDNN and TP (all p for trend <0.05) and a positive association with normalized LF (p for trend = 0.02). After additional adjustment for either HR at rest or 24-hour HR, only SDNN remained linearly associated with ODI (p for trend = 0.006 and 0.03, respectively; Table 3 ). Similar findings were also obtained, when the AHI or ODI were entered into the models as dichotomous variable (≥5 vs <5), as listed in Table 4 .
| AHI 0 n= 388 | AHI 1 n= 397 | AHI 2 n= 204 | AHI 3-4 n= 146 | AHI≥5 n= 120 | p-value | ||
|---|---|---|---|---|---|---|---|
| SDNN | Model 1 | Ref. | -0.06 (-0.19; 0.08) | -0.07 (-0.24; 0.10) | -0.22 (-0.41; -0.02) | -0.29 (-0.50; -0.07) | 0.006 |
| Model 2 | -0.06 (-0.18; 0.06) | -0.06 (-0.21; 0.09) | -0.16 (-0.33; 0.02) | -0.17 (-0.37; 0.02) | 0.04 | ||
| Model 3 | -0.03 (-0.15; 0.09) | 0.00 (-0.15; 0.15) | -0.11 (-0.28; 0.05) | -0.09 (-0.28; 0.10) | 0.23 | ||
| LFnu | Model 1 | Ref. | 0.08 (-0.05; 0.21) | 0.18 (0.01; 0.34) | 0.24 (0.05; 0.43) | 0.14 (-0.07; 0.36) | 0.17 |
| Model 2 | 0.08 (-0.05; 0.21) | 0.17 (0.01; 0.34) | 0.22 (0.03; 0.40) | 0.10 (-0.11; 0.31) | 0.05 | ||
| Model 3 | 0.07 (-0.06; 0.20) | 0.14 (-0.02; 0.30) | 0.19 (-0.00; 0.37) | 0.04 (-0.17; 0.25) | 0.17 | ||
| HFnu | Model 1 | Ref. | -0.03 (-0.17; 0.10) | -0.13 (-0.30; 0.03) | 0.02 (-0.17; 0.21) | -0.05 (-0.27; 0.16) | 0.63 |
| Model 2 | -0.03 (-0.16; 0.10) | -0.13 (-0.29; 0.04) | 0.05 (-0.14; 0.23) | -0.005 (-0.21; 0.20) | 0.88 | ||
| Model 3 | -0.01 (-0.14; 0.11) | -0.09 (-0.24; 0.07) | 0.09 (-0.10; 0.27) | 0.08 (-0.13; 0.28) | 0.44 | ||
| TP | Model 1 | Ref. | 0.04 (-0.09; 0.17) | -0.06 (-0.22; 0.10) | -0.05 (-0.24; 0.13) | -0.19 (-0.40; 0.01) | 0.06 |
| Model 2 | 0.04 (-0.07; 0.15) | -0.05 (-0.18; 0.09) | 0.01 (-0.14; 0.17) | -0.07 (-0.25; 0.10) | 0.31 | ||
| Model 3 | 0.07 (-0.03; 0.17) | 0.02 (-0.10; 0.14) | 0.07 (-0.07; 0.21) | 0.04 (-0.11; 0.20) | 0.75 | ||
| ODI 0 n= 377 | ODI 1 n= 360 | ODI 2 n= 195 | ODI 3-4 n= 182 | ODI ≥5 n= 141 | p-value | ||
|---|---|---|---|---|---|---|---|
| SDNN | Model 1 | Ref. | -0.003 (-0.14; 0.14) | -0.06 (-0.24; 0.11) | -0.20 (-0.39; -0.01) | -0.37 (-0.58; -0.16) | 0.0004 |
| Model 2 | -0.003 (-0.12; 0.12) | -0.05 (-0.20; 0.11) | -0.15 (-0.31; 0.02) | -0.25 (-0.44; -0.06) | 0.006 | ||
| Model 3 | -0.009 (-0.13; 0.11) | -0.05 (-0.20; 0.11) | -0.12 (-0.28; 0.04) | -0.19 (-0.38; -0.009) | 0.03 | ||
| LFnu | Model 1 | Ref. | 0.11 (-0.02; 0.25) | 0.06 (-0.11; 0.23) | 0.17 (-0.01; 0.36) | 0.26 (0.05; 0.47) | 0.02 |
| Model 2 | 0.11 (-0.02; 0.24) | 0.06 (-0.11; 0.23) | 0.15 (-0.02; 0.34) | 0.22 (0.01; 0.43) | 0.06 | ||
| Model 3 | 0.12 (-0.01; 0.25) | 0.05 (-0.11; 0.22) | 0.13 (-0.05; 0.31) | 0.17 (-0.04; 0.37) | 0.17 | ||
| HFnu | Model 1 | Ref. | -0.06 (-0.19; 0.08) | -0.09 (-0.26; 0.08) | 0.008 (-0.18; 0.19) | -0.17 (-0.38; 0.04) | 0.14 |
| Model 2 | -0.06 (-0.19; 0.08) | -0.08 (-0.25 ; 0.09) | 0.03 (-0.15; 0.21) | -0.12 (-0.33; 0.08) | 0.33 | ||
| Model 3 | -0.06 (-0.19; 0.07) | -0.08 (-0.24; 0.09) | 0.06 (-0.11; 0.24) | -0.06 (-0.26; 0.14) | 0.99 | ||
| TP | Model 1 | Ref. | 0.06 (-0.07; 0.19) | 0.04 (-0.13; 0.20) | 0.01 (-0.17; 0.19) | -0.17 (-0.37; 0.04) | 0.05 |
| Model 2 | 0.06 (-0.05; 0.17) | 0.05 (-0.08; 0.19) | 0.07 (-0.08; 0.22) | -0.04 (-0.21; 0.13) | 0.48 | ||
| Model 3 | 0.05 (-0.05; 0.15) | 0.06 (-0.07; 0.19) | 0.11 (-0.02; 0.25) | 0.05 (-0.11; 0.21) | 0.62 | ||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
