Energy Cost of Walking and Functional Aerobic Capacity During Moderate Intensity Exercise in Adults with Obstructive Sleep Apnea: A Cross-Sectional Study

Introduction

Walking is the most common form of human physical activity that requires metabolic energy. It is the outcome of an integration of multiple physiological systems working together to sustain walking pace for the required metabolic demand. Any impairment in the body systems might have a catastrophic effect on the walking economy and functional aerobic capacity (FAC). Meanwhile, patients with obstructive sleep apnea (OSA), a repetitive upper airway obstruction leads to abnormal nocturnal gas exchange (Guilleminault and Abad, 2004), arterial oxygen desaturation, and chronically elevated sympatho-adrenal activity (Carlson et al., 1993; Somers et al., 1995). Further, sleep disorders have been shown to limit physical capacity in adults (Puri et al., 2017) and is commonly associated with significant impairment in daytime physical functioning, fatigue, and daytime sleepiness that may limit daily life activities (Aguillard et al., 1998). By comparison, maximal oxygen consumption in persons is reduced (Mendelson et al., 2018), this indicates that persons with OSA would expend more energy and use a greater percentage of maximal metabolic capacity when completing everyday activities of daily living. Despite extensive and ongoing scientific attention, understanding of (Aguillard et al., 1998)walking energy cost and mechanical efficiency for a given submaximal speed of walking in individuals with OSA have scarcely been studied.

Previous studies on OSA have reported that the presence of OSA negatively influences the aerobic capacity as measured by the maximal oxygen uptake (VO₂ max) (Berger et al., 2019; Lin et al., 2006; Nanas et al., 2010; Ucok et al., 2009; Vanhecke et al., 2008; Vanuxem et al., 1997a). Likewise, OSA severity is associated with an attenuated rate of oxygen consumption, when calculated using a nomogram based on age, sex, and baseline activity level, and high post-exercise blood pressure (Mansukhani et al., 2013). However, it remains unknown whether the disease severity as measured by apnea hypopnea index (AHI) is an important predictor of walking economy.

To our knowledge, walking performance and net oxygen consumption during submaximal exercise in patients with OSA is not clearly understood. Therefore, the aims of the present study (Fig. 1) are (1) To characterize the functional determinants of walking performance and functional aerobic capacity in individuals with OSA compared with healthy individuals, and (2) To examine the relationship between OSA and walking performance outcomes in people with OSA. Quantification of physiological and functional profile of level walking and during submaximal exercise may add new and clinically meaningful perspective to the assessment of treatment outcomes following exercise interventions.

Materials And Methods

Study design and participants

In this cross-sectional study, a convenience sample of 26 volunteers was recruited from a local sleep medicine clinic and the local community; 13 patients with OSA and 13 healthy adults participated in this study. Patients who met the inclusion criteria were enrolled in the study and invited to the functional performance laboratory at George Mason University. An overnight diagnostic polysomnography (PSG) assessment was completed within five years of participants enrolling in the study and was used to determine the presence of OSA and sleep parameters. The severity of OSA was determined using the apnea-hypopnea index (AHI); AHI determines the number of apneas and hypopneas (10 seconds or longer per event) that occur per hour of sleep.

The participants in this study were adults between the ages of 25 and 60 with a body mass index (BMI) between 25 and 39 kg/m² who had undergone a recent third party sleep study with an AHI of 5 or greater and who were capable of walking on an exercise treadmill, were non-smoking, and were free of cardiovascular, metabolic, and lung disease. Participants who reported that they had been or were presently being treated by a physician for any form of cardiovascular, pulmonary, or metabolic disease and/or had any blood clotting history or vascular disease were excluded from this study. Participants were also required to answer no to all questions on the Physical Activity Readiness Questionnaire (PARQ⁺)(American College of Sports Medicine, 2018) and not be at high risk for a cardiac event as determined by the health history questionnaire and the American College of Sports Medicine (ACSM) risk assessment model (American College of Sports Medicine, 2018).

Each study participant underwent a standard assessment of height, weight, and body mass index as well as neck and abdominal circumference measurement and resting heart rate and blood pressure measurement. The body composition examination was performed using Bioimpedance Analysis (Tanita Corp., Tokyo, Japan).

Ethical consideration

The research protocol was approved by the relevant Institutional Review Boards of George Mason University prior to initiation. The written informed consent was obtained from each participant as a condition of enrollment in the study in accordance to the Declaration of Helsinki (General Assembly of the World Medical, 2014).

Physical activity questionnaire

Weekly physical activity status was determined by using the seven-day Physical Activity Recall Questionnaire (PAR) (Hayden - Wade et al., 2003; Sallis et al., 1997). Physical activity (PA) status and total duration of PA were determined for all participants.

Cardiopulmonary exercise testing

The participants visited the laboratory on two separate occasions. During the first visit, the participants underwent a maximal integrated cardiopulmonary exercise test (ICPET) on the treadmill using an individualized ramp protocol. The workload of the treadmill protocol was pre-determined on the basis of each participant’s body weight and height and their self-selected comfortable walking speed (Myers et al., 1992). The ICPET were performed on a Medgraphics® indirect calorimetry gas analysis system (Cardio₂, St. Paul, MN) for gas exchange measurements. CPET test preparation followed the guidelines of the American College of Sports Medicine (ACSM) (American College of Sports Medicine, 2018). The primary data obtained from the motorized treadmill exercise test included peak oxygen uptake (VO_2peak) and VO₂-at-anaerobic threshold (AT) from pulmonary gas exchange analysis. Heart rhythm and rate were assessed using continuous electrocardiogram (ECG) monitoring throughout the exercise testing.

Constant work rate test (CWRT)

The second visit was completed within 3–5 days of the first and included three bouts of constant work rate (CWR) treadmill exercise at 85% of AT, followed by a 10-minute over-ground walking test (10MWT). AT was determined using the following criteria: 1) V-slope method, defined as inflection point from linearity of carbon dioxide output (VCO₂) against oxygen uptake (VO₂) panel; 2) point of increase in end-tidal oxygen tension (PETO₂); and 3) point of increase in the ventilatory equivalent of oxygen (VE/VO₂) and a concomitant reduction of end tidal of carbon dioxide (PETCO₂) (Beaver et al., 1986; Marcus et al., 1971; Wasserman et al., 1973).

Oxygen uptake (VO₂) and walking speed at AT were determined. The net VO₂ (VO_2net, mL/kg. min) was calculated by subtracting the resting value of VO₂ from the VO₂ at the steady state work rate during treadmill exercise that corresponded to 85% of the AT (Ludlow and Weyand, 2017). Steady state VO₂ was calculated by averaging the last 2-minute VO₂ values of the three consecutive 6-minute bouts. The energy cost of walking during submaximal exercise corresponded to 85% of the AT.

Figure 2. Schematic illustration of visit two and constant work rate (CWR) test at 85% of Anaerobic Threshold (AT).

Over-ground 10-minute walking test (10MWT)

Participants were instructed to walk as far as possible around a 25-meter indoor walking track for 10 minutes. The distance covered was recorded every 2.5 minutes and after 10 minutes. The target for this test was the completion of 10 minutes of walking. The participants were encouraged to complete the test without stopping. Walking performance during the 10MWT was expressed as the total distance covered in 10 minutes, walking speed, and the energy cost of walking. The energy cost of over-ground walking is known as the amount of oxygen a person consumes per kilogram of body weight per distance traveled. For the present study, it was derived from the 10MWT and was calculated as the amount of VO₂ over the speed, where the speed was defined as the distance walked in 10 minutes (i.e., mL. kg⁻¹. m⁻¹) (Waters and Mulroy, 1999).

Statistical analysis

For descriptive purposes, the data are presented as mean and standard deviation. After testing for normality, statistical comparisons were made between the OSA group and the non-OSA group (comparison group). The indices used to determine physiological aerobic capacity were VO_2peak and VO₂-at-AT, whereas the net VO₂ and energy expenditure were obtained from the CWR protocol. Functional walking outcome variables were derived from the 10MWT, including gait speed over 10 minutes (meters/m), over-ground walking energy expenditure (mL.kg^− 1.m^− 1), and distance (meters).

The baseline characteristics of the two groups were compared using the independent t-test. Analysis of covariance (ANCOVA) was used to determine whether significant difference in FAC, the energy cost of walking, and net VO₂ emerged between the two groups, using weight, BMI, and age as covariates in the model.Multiple linear regression corrected for weight, age, and BMI was conducted to examine the associations between VO_2peak (mL.kg^− 1.min^− 1), 10MWT distance (m), energy cost of walking (mL.kg^− 1.min^− 1), treadmill energy cost of walking, and AHI. Statistical significance was set at p ≤ 0.05. All data analysis was conducted using the STATA version 16 statistical program.

Results

Of the twenty-six participants, 13 participants with OSA had an average AHI of 45 ± 28, and a mean age, weight, and BMI of 47.53 ± 9.6 years, 88 ± 1Kg, and 29 ± 5 kg/m², respectively. The non-OSA group also consisted of 13 individuals with a mean age, weight, and BMI of 40 ± 9 years, 67 ± 11 kg, and 23 ± 3 kg/m², respectively. The majority of the OSA participants (69%) had severe AHI, 15% had moderate AHI, and 15% had mild AHI. Participants in both groups underwent a treadmill CPET at both maximal and sub-maximal intensity of 85% of AT. Six of the 13 OSA participants were CPAP users. Participants’ baseline physical and demographic characteristics are provided in Table 1. There were no significant differences in baseline values between the two groups other than their weight. The physical activity level obtained through 7-PAR did not detect differences between the two groups (p = 0.64) (Table 1).

Table 1

Demographic and physical characteristics of the study population
Variables	OSA n = 13	Non-OSA n = 13	p-value
Variables	Mean ± SD	Mean ± SD	p-value
Weight (kg)	87.62 ± 19.4	68.93 ± 11.27	0.006*
Height (cm)	172.15 ± 6.8	169.46 ± 8.9	NS
BMI (cm²/kg)	29 ± 5	24 ± 3	0.002*
Age (years)	47.53 ± 9.6	40.46 ± 8.8	NS
Resting HR (b/min)	74.22 ± 7.3	77.11 ± 14.68	NS
Resting VO₂(mL/kg/min)	65.38 ± 4.8	67.54 ± 8.05	NS
Gender (M: F)	9:4	10:3	NS
RER at VO_{2 peak}	1.23 ± 0.09	1.24 ± 0.09	NS
Resting VO₂(mL/kg/min)	3.5 ± 0.78	3.85 ± 1.35	NS
Sleep duration (Hours/Week)	49 ± 6	51 ± 5	NS
7-day Physical activity recall (METs)	7.5 ± 3.4	8 ± 3.1	NS
^{Values are expressed as Mean (SD) or frequency} ^{Abbreviation: OSA, Obstructive Sleep Apnea; VO2, oxygen uptake; RER, respiratory exchange ratio; NS, not significant} ^{* Denotes statistically significant p≤0.05}

Data on maximal cardiopulmonary capacity are presented in Table 2. There were significant differences in VO_2peak (OSA 29.69 ± 5.6 mL/kg/min vs. non-OSA 37.49 ± 6.5 mL/kg/min, p = 0.03). Compared to the non-OSA group, the OSA participants had significantly (P < 0.001) lower walking performance, and the OSA patients tended to have a higher energy cost of walking compared to the non-OSA group as well as a higher energy expenditure (Table 3). The VO_{2 net} of the participants with OSA during CWR was significantly lower compared to that of the non-OSA participants after controlling for age, weight, BMI, and speed m/min (13 ± 6 vs 19 ± 5, p = 0.02). In Table 3 the distance covered during the 10MWT was significantly lower for the OSA participants compared to the non-OSA participants (p < 0.001). During submaximal exercise, the oxygen pulse VO₂/HR (mL O₂/beat) of the OSA group was not statistically significantly different from that of the non-OSA group (11.70 ± 3.31 vs. 19 ± 5, 10.56 ± 2.85, p = 0.1).

Table 2

Comparison of cardiopulmonary variables at peak exercise of the study population
Variables	OSA (n = 13)	Non-OSA (n = 13)	p-value
Variables	Mean ± SD	Mean ± SD	p-value
Time to AT (min)	7.0 ± 1.41	8.11 ± 2.44	0.1
VO₂ at AT (mL/kg/min)	19 ± 5	25 ± 6	0.05
Test Duration (min)	10.2 ± 1.5	10.8 ± 2.7	0.08
VCO₂ (mL/min)	2967 ± 763	2983 ± 717	0.3
PETO₂	115 ± 3	115 ± 6	0.2
PETCO₂	32 ± 2	31 ± 4	0.5
HR_Peak (b/min)	160 ± 13	167 ± 21	0.9
VE/VO_2Peak ratio (no unite)	31 ± 4.2	37 ± 6.3	0.6
VE/VCO_2Peak ratio (no unite)	25 ± 6.6	29 ± 3.6	0.03
VO₂/HR_Peak (mL O₂/beat)	17.92 ± 5.4	17.46 ± 5.6	0.2
^{Value are means± SD} ^{Abbreviation: AT, anaerobic threshold; CWR, constant work rate; HR, heart rate} ^{ANCVOA testing were used corrected for the age, weight, BMI}

Table 3

Outcome of walking economy during submaximal exercise and 10-minute walk test (10 MWT) among the study groups
Variables	OSA (n = 13)	Non-OSA (n = 13)	p-value
Variables	Mean ± SD	Mean ± SD	p-value
10MWT VO₂ (mL/kg/min)	11.27 ± 0.48	14.45 ± 0 .80	< 0.001*
10 MWT Speed (m/min)	178 ± 5	110 ± 8	< 0.001*
10 MWT Distance (m)	777.48 ± 48	1095.64 ± 83	< 0.001*
10 MWT Energy cost of walking (mL/kg/m)	0.14 ± 0.003	0.13 ± 0.002	< 0.001*
Energy cost of CWR (mL/kg/m)	0.019 ± 0.003	0.015 ± 0.004	0.0007
CWR VO₂/HR (mL O₂/beat)	11.70 ± 3.31	10.56 ± 2.85	0.1
^{Value are means± SD} ^{*denote statistically significant p≤0.05} ^{Abbreviation: AT, anaerobic threshold; CWR, constant work rate; HR, heart rate} ^{ANCVOA testing were used corrected for the age, weight, BMI}

A multiple regression model, after adjusting for confounders including age, weight, and BMI, was run to assess the association between the energy cost of walking, 10MWT speed, distance, VO_2peak, and AHI severity. The findings showed that AHI significantly predicted the energy cost of walking (F = 27.6, p < .001, R₂ = 0.87) and that AHI explains 87% of the variance in the energy cost of walking during 10MWT. There is lack of significant association between VO_2peak and AHI severity (P = 0.4). The significant associations between AHI and walking speed, distance, and VO₂ at anaerobic threshold (p < 0.001) are reported in Table 4.

Table 4

The Multiple linear regression analyses for the association between VO2 and walking performance variables during 10 MWT and AHI of the study populationa
Variables	AHI	p-value
Variables	R²	p-value
VO_2peak (mL/kg/min)	0.45	0.4
VO2-at-AT	0.92	< 0.001*
10MWT distance (m)	0.85	< 0.001*
10MWT speed (m/min)	0.85	< 0.001*
Energy cost of walking (mL/kg/min)	0.87	< 0.001*
Treadmill energy cost of walking	0.59	0.04*
^{Value are means± SD} ^{Abbreviation: AHI: apnea hypopnea index; VO2−at−AT, oxygen consumption at anaerobic threshold} ^*_{Denotes p−value ≤0.05 for the correlation coefficients} ^a. _{corrected for age, weight and BMI}

Discussion

The major finding of this study is that the OSA participants had impaired walking performance and aerobic walking capacity when compared to the non-OSA adults. Our results indicate that the metabolic cost of walking was high in the OSA group compared to the control group, and by comparison, their maximal oxygen consumption was reduced. The severity of OSA, as reflected by AHI, was found to be associated with the rate of oxygen consumption at the AT, which is a robust predictor of aerobic performance exercise. There was a significant relationship between OSA severity (AHI) and the energy cost level of walking, walking distance, and net VO₂ during the submaximal treadmill test. Collectively, these findings indicate that individuals with more severe OSA are more likely to expend more energy when completing activities of daily living, which may increase perceptions of fatigue. In this study, the lack of significant differences between OSA and non-OSA participants in ventilatory efficiency (VE/VO₂ and VE/CO₂ slope) and peak and submaximal oxygen pulse (VO₂/HR is used as proxy for ventilatory and cardiac performance) suggests that the role of the cardiopulmonary response to exercise was not a limiting factor in the participants’ walking performance.

Existing literature evaluating the cardiopulmonary response to exercise in OSA patients using a motor-driven treadmill reports conflicting results. In several studies, no difference was shown in the maximal exercise capacity between healthy adults and OSA patients (Daiana Mortari et al., 2014; Flore et al., 2006). This contradictory finding may be attributed to the selection criteria for study participants (who were newly diagnosed with OSA), the presence of co-morbidities, and varying levels of OSA severity that may limit the generalizability of the results (Flore et al., 2006). In line with our findings, other studies investigating the response to exercise testing of OSA patients have shown a decrease in exercise capacity (Lin et al., 2006; Nanas et al., 2010; Ucok et al., 2009; Vanhecke et al., 2008; Vanuxem et al., 1997a). As expected, the value of the peak oxygen consumption reported in this study was within the reference range (29.6 ± 6 mL/kg/min) for the OSA population reported previously (Daiana Mortari et al., 2014). Systematic reviews showed that the reduction in VO_2peak was found to be larger in non-obese patients (body mass index < 30 kg/m²) such as our participant group (Berger et al., 2019; Mendelson et al., 2018).

There are a number of etiology factors by which OSA may appear to reduce functional aerobic capacity and influence the energy cost of walking in persons with severe obstructive sleep apnea. Sleep fragmentation and daytime somnolence are known to influence aerobic capacity, which may partially contribute to the reported decline in exercise tolerance and the increase in the energy cost of walking of persons with OSA (Hong and Dimsdale, 2003; Martin, 1981; Mougin et al., 1991). A potential physiological mechanism, that were not being analyzed due methodological limitations, is decreased maximal lactate concentration and delayed lactate elimination. This has been reported previously in OSA compared to age-matched controls, which may suggest impaired glycolytic and oxidative metabolism (Vanuxem et al., 1997a). The findings of previous studies may explain the reported impairment of OSA patients’ functional aerobic capacity.

The findings of exercise intolerance in our study demonstrated by a high energy cost of walking and decreased speed and distance in OSA patients may contribute to excessive daytime somnolence. A number of observational studies have revealed diminished exercise performance among OSA patients under the influence of sleep deprivation (Aguillard et al., 1998; Martin, 1981; Van Helder and Radomski, 1989). Extreme daytime somnolence in healthy men has been shown to reduce tolerance to tasks and increase the rating of perceived exertion during exercise testing (Temesi et al., 2013). Therefore, we cannot exclude the possibility of the influence of sleep deprivation on the performance of the OSA participants in our study. Among several factors that must be considered are delayed clearance of post-exercise lactate in OSA patients and increased concentration of catecholamines. It has been reported that in OSA, the lactate threshold was attained at lower workload levels compared to age- and weight-matched controls, and an abnormal build-up of lactate was directly related to exercise capacity(Bonanni et al., 2004). These findings could be explained, at least in part, by the lower oxidative capacity at the muscle that promoting an increased reliance on anaerobic metabolism in patients with OSA. Although local muscle metabolic impairment is not clearly understood, it would support both an early onset of metabolic acidosis and a subsequent increase in perceived fatigue (Keyser, 2010). Vanuuxem et al. (Vanuxem et al., 1997a) reported a slow rate of lactate elimination, which could indicate a primary defect in oxidative metabolism secondary to the repeated events of nocturnal hypoxemia. Unfortunately, lactate concentration was not assessed in our study and would require further evaluation. Furthermore, repetitive cycles of hypoxemia and reoxygenation induce autonomic system instability, which diminishes aerobic performance through changes in the muscle bioenergetic response (Bonanni et al., 2004; Vanuxem et al., 1997b). Several studies demonstrated structural changes in the skeletal muscle fibers and bioenergetic system of the active muscle parallel to the modulation that is observed in the condition of chronic hypoxia (Beitler et al., 2014)^; (Sauleda et al., 2003). Abnormalities of the skeletal muscles, such as structural and bioenergetics changes in skeletal muscle fibers (Sauleda et al., 2003), which have been identified in OSA patients in a previous study using muscle biopsy, may contribute to their lower maximal exercise capacity. Lower net VO₂ during submaximal exercise may in part be explained by the recruitment of glycogenolytic fibers occurring at lower levels of muscle contraction(Chwalbinska-Moneta et al., 1989). The reduced resistance to fatigue in these patients could be related to an increase in the energy cost of walking and a decrease in the workloads even after controlling for confounders such as weight, BMI, and speed in the analysis.

Previous findings related to the association between OSA severity and physical performance are generally consistent with our results. In a study by Billings et al.(Billings et al., 2016), a low level of recreational activity such as walking was correlated with greater severity of sleep apnea, defined by using the apnea-hypopnea index, especially in male and obese individuals. Our findings suggest that severity of OSA may contribute to some extent to an increase in the energy cost of walking and lower VO₂ at CWR, independent of age, weight, and BMI. This is also supported by a study by Mansukhani et al.(Mansukhani et al., 2013) which offers robust evidence that the severity of sleep-related breathing was associated with the FAC.

Our study has a few limitations. First, the cross-sectional design of our study did not allow us to establish a causal relationship. Second, some of the participants in the OSA group were using continuous positive airway pressure (CPAP), which could confound our results. However, there is conflicting evidence on the effects of long-term use of CPAP on exercise capacity. It has been previously reported that nocturnal use of CPAP improves exercise tolerance and dyspnea in obese patients with OSA (Pendharkar et al., 2011). On the contrary, other studies have reported contradictory findings that CPAP treatment did not change peak oxygen consumption (VO_2peak). To some extent, our findings may have been influenced by the 20% of patients who were using CPAP. Despite that, the differences between OSA and non-OSA participants in walking speed and mechanical efficiency were significant. However, future research should account for the influence of CPAP. Finally, we accounted for visceral obesity and fat distribution in this study by using weight and BMI as a confounder in the statistical analysis, yet there may be unknown confounders associated with obesity that we did not control for.

References

Aguillard RN, Riedel BW, Lichstein KL, Grieve FG, Johnson CT, Noe SL. Daytime functioning in obstructive sleep apnea patients: exercise tolerance, subjective fatigue, and sleepiness. Appl Psychophysiol Biofeedback. 1998;23:207–17.
American College of. Sports Medicine ACSM’s guidelines for exercise testing and prescription. Philadelphia: Wolters Kluwer; 2018.
Beaver WL, Wasserman K, Whipp BJ. A new method for detecting anaerobic threshold by gas exchange. J Appl Physiol. 1986;60:2020–7.
Beitler JR, Awad KM, Bakker JP, et al. Obstructive sleep apnea is associated with impaired exercise capacity: a cross-sectional study. J Clin Sleep Med. 2014;10:1199–204.
Berger M, Kline CE, Cepeda FX, et al. Does obstructive sleep apnea affect exercise capacity and the hemodynamic response to exercise? An individual patient data and aggregate meta-analysis. Sleep Med Rev. 2019;45:42–53.
Billings ME, Johnson DA, Simonelli G, et al. Neighborhood walking environment and activity level are associated with OSA: the Multi-Ethnic Study of Atherosclerosis. Chest. 2016;150:1042–9.
Bonanni E, Pasquali L, Manca ML, et al. Lactate production and catecholamine profile during aerobic exercise in normotensive OSAS patients. Sleep Med. 2004;5:137–45.
Carlson JT, Hedner J, Elam M, Ejnell H, Sellgren J, Wallin BG. Augmented resting sympathetic activity in awake patients with obstructive sleep apnea. Chest. 1993;103:1763–8.
Chwalbinska-Moneta J, Robergs RA, Costill DL, Fink WJ. Threshold for muscle lactate accumulation during progressive exercise. J Appl Physiol. 1989;66:2710–6.
Daiana, Mortari. Camila Pereira Leguisamo, Simone Chaves Fagondes Exercise capacity in patients with obstructive sleep apnea syndrome. Rev Fac Ciênc Médicas Sorocaba. 2014;16:164–9.
Flore P, Tonini J, Pepin J-L, Levy P, Wuyam B. Obstructive sleep apnoea: hypoapnoea syndrome reversibly depresses cardiac response to exercise. Eur Heart J. 2006;27:1632–3.
General Assembly of the World Medical World Medical Association Declaration. of Helsinki: ethical principles for medical research involving human subjects. J Am Coll Dent. 2014;81:14.
Guilleminault C, Abad VC. Obstructive sleep apnea syndromes. Med Clin North Am. 2004;88:611–30.
Hayden - Wade HA, Coleman KJ, Sallis JF, Armstrong C. Validation of the telephone and in-person interview versions of the 7-day PAR.(physical activity recall). Med Sci Sports Exerc. 2003;35:801–9.
Hong S, Dimsdale JE. Physical activity and perception of energy and fatigue in obstructive sleep apnea. Med Sci Sports Exerc. 2003;35:1088–92.
Keyser RE. Peripheral fatigue: high-energy phosphates and hydrogen ions. PM&R. 2010;2:347–58.
Lin C-C, Hsieh W-Y, Chou C-S, Liaw S-F. Cardiopulmonary exercise testing in obstructive sleep apnea syndrome. Respir Physiol Neurobiol. 2006;150:27–34.
Ludlow LW, Weyand PG. Walking economy is predictably determined by speed, grade, and gravitational load. J Appl Physiol. 2017;123:1288–302.
Mansukhani MP, Allison TG, Lopez-Jimenez F, Somers VK, Caples SM. Functional aerobic capacity in patients with sleep-disordered breathing. Am J Cardiol. 2013;111:1650–4.
Marcus JH, Ingram RH Jr, McLean RL. The threshold of anaerobic metabolism in chronic obstructive pulmonary disease: a promising index of evaluation. Am Rev Respir Dis. 1971;104:490–8.
Martin BJ. Effect of sleep deprivation on tolerance of prolonged exercise. Eur J Appl Physiol. 1981;47:345–54.
Mendelson M, Marillier M, Bailly S, et al. Maximal exercise capacity in patients with obstructive sleep apnoea syndrome: a systematic review and meta-analysis. Eur. Respir. J., 2018, 51.
Mougin F, Simon-Rigaud ML, Davenne D, et al. Effects of sleep disturbances on subsequent physical performance. Eur J Appl Physiol. 1991;63:77–82.
Myers J, Buchanan N, Smith D, et al. Individualized ramp treadmill: observations on a new protocol. Chest. 1992;101:236S–241S.
Nanas S, Sakellariou D, Kapsimalakou S, et al. Heart rate recovery and oxygen kinetics after exercise in obstructive sleep apnea syndrome. Clin Cardiol. 2010;33:46–51.
Pendharkar SR, Tsai WH, Eves ND, Ford GT, Davidson WJ. CPAP increases exercise tolerance in obese subjects with obstructive sleep apnea. Respir Med. 2011;105:1565–71.
Puri S, Herrick JE, Collins JP, Aldhahi M, Baattaiah B. Physical functioning and risk for sleep disorders in US adults: results from the National Health and Nutrition Examination Survey 2005–2014. Public Health. 2017;152:123–8.
Sallis JF, Buono MJ, Roby JJ, Micale FG, Nelson JA. Seven-day physical activity recall. Med Sci Sports Exerc. 1997;29:89–103.
Sauleda J, García-Palmer FJ, Tarraga S, Maimó A, Palou A, Agustí AG. Skeletal muscle changes in patients with obstructive sleep apnoea syndrome. Respir Med. 2003;97:804–10.
Somers VK, Dyken ME, Clary MP, Abboud FM. Sympathetic neural mechanisms in obstructive sleep apnea. J Clin Invest. 1995;96:1897–904.
Temesi J, Arnal PJ, Davranche K, et al. Does central fatigue explain reduced cycling after complete sleep deprivation. Med Sci Sports Exerc. 2013;45:2243–53.
Ucok K, Aycicek A, Sezer M, et al. Aerobic and anaerobic exercise capacities in obstructive sleep apnea and associations with subcutaneous fat distributions. Lung. 2009;187:29–36.
Van Helder T, Radomski MW. Sleep deprivation and the effect on exercise performance. Sports Med. 1989;7:235–47.
Vanhecke TE, Franklin BA, Ajluni SC, Sangal RB, McCullough PA. Cardiorespiratory fitness and sleep-related breathing disorders. Expert Rev Cardiovasc Ther. 2008;6:745–58.
Vanuxem D, Badier M, Guillot C, Delpierre S, Jahjah F, Vanuxem P. Impairment of muscle energy metabolism in patients with sleep apnoea syndrome. Respir Med. 1997a;91:551–7.
Vanuxem D, Badier M, Guillot C, Delpierre S, Jahjah F, Vanuxem P. Impairment of muscle energy metabolism in patients with sleep apnoea syndrome. Respir Med. 1997b;91:551–7.
Wasserman K, Whipp BJ, Koyl SN, Beaver WL. Anaerobic threshold and respiratory gas exchange during exercise. J Appl Physiol. 1973;35:236–43.
Waters RL, Mulroy S. The energy expenditure of normal and pathologic gait. Gait Posture. 1999;9:207–31.