ADVERTISEMENT

Issue Navigator

Volume 14 No. 12
Earn CME
Accepted Papers





Scientific Investigations

Objectively Measured Disrupted Sleep Is Independently and Directly Associated With Low Exercise Capacity in Males: A Structural Equation Model

Ren-Jing Huang, PhD1; Shin-Da Lee, PhD2,3,4; Ching-Hsiang Lai, PhD5; Shen-Wen Chang, RN6; Ai-Hui Chung, RN6; Chiung-Wei Chen, MD7; I-Ning Huang, MD7; Hua Ting, MD6,7,8
1Department of Medical Imaging and Radiological Sciences, Chung Shan Medical University, Taichung, Taiwan; 2Department of Physical Therapy, Graduate Institute of Rehabilitation Science, China Medical University, Taichung, Taiwan; 3Department of Occupational Therapy, Asia University, Taichung, Taiwan; 4School of Rehabilitation Medicine, Shanghai Seventh People's Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai, China; 5Department of Medical Informatics, Chung Shan Medical University, Taichung, Taiwan; 6Sleep Medicine Center, Chung Shan Medical University Hospital, Taichung, Taiwan; 7Department of Physical Medicine and Rehabilitation, Chung Shan Medical University Hospital, Taichung, Taiwan; 8Institute of Medicine, Chung Shan Medical University, Taichung, Taiwan

ABSTRACT

Study Objectives:

We investigated the interaction between objective sleep disturbance and obesity, sedentary lifestyle, and lung dysfunction and whether it is negatively associated with cardiorespiratory fitness.

Methods:

In this community cohort study of 521 men (age 46.6 ± 7.5 years), measures of anthropometry, pulmonary function, overnight sleep polysomnography, and cardiopulmonary exercise testing were processed stepwise using structural equation modeling (SEM).

Results:

A univariate correlation analysis was used to group the corresponding variables (in parentheses) into the following eligible latent variables for lower exercise capacity: obesity (body mass index, waist-to-hip ratio), irregular exercise, impaired lung function (predicted values of forced expiratory volume in the first second, forced vital capacity, maximal ventilatory volume, and lung diffusion capacity for carbon monoxide), disrupted sleep (total sleep time, percentage of slow-wave sleep, sleep efficiency), and sleep-disordered breathing (apnea-hypopnea index, lowest oxygen saturation, percentage of total period of oxygen saturation < 90%). Advanced SEM analyses produced a well-fitted final confirmatory model that obesity (direct strength βd = .366, P < .001), irregular exercise (βd = .274, P < .001), and impaired lung function (βd = .152, P < .001), with their mutual interactions, as well as disrupted sleep (βd = .135, P = .001) were independently and directly associated with low exercise capacity. By contrast, sleep-disordered breathing (βd = 0, P = .215) was related to low exercise capacity indirectly through obesity into the mutual interaction cycle of obesity, irregular exercise, and impaired lung function. Sleep-disordered breathing was robustly and mutually correlated with obesity (mutual relationship index = .534, P < .001).

Conclusions:

Objectively measured disrupted sleep is directly and independently associated with low exercise capacity; however, sleep-disordered breathing is indirectly mediated by obesity and mutual interactions among obesity, lung dysfunction, and sedentary lifestyle and is linked to low exercise capacity. Our findings indicate that individuals with limited exercise capacity without definite causes should undertake a sleep study, particularly in those describing symptoms of sleep-disordered breathing or insomnia.

Citation:

Huang RJ, Lee SD, Lai CH, Chang SW, Chung AH, Chen CW, Huang IN, Ting H. Objectively measured disrupted sleep is independently and directly associated with low exercise capacity in males: a structural equation model. J Clin Sleep Med. 2018;14(12):1995–2004.


BRIEF SUMMARY

Current Knowledge/Study Rationale: Although life expectancy can be predicted by absolute values of maximal oxygen uptake, the underlying factors and their interactions remain elusive. We therefore hypothesized that obesity, irregular exercise, superficial sleep, sleep-disordered breathing, and impaired lung function were mutually interactive with each other, in addition to each factor being associated individually with low exercise capacity.

Study Impact: In our study of male workers, structural equation modeling is a promising method for unraveling the complex potential factors linked to low cardiorespiratory fitness. Our results suggest that objectively measured disrupted sleep is associated not only with several medical problems, as identified by previous studies, but also with low cardiorespiratory fitness. In addition, individuals with limited exercise capacity without definite causes should undertake a sleep study, particularly in those describing symptoms of sleep-disordered breathing or insomnia.

INTRODUCTION

A low level of physical fitness, which is in part gene dependent,1 has been indicated as an independent risk factor associated with incidental cardiometabolic morbidity and all-cause premature mortality, independent of all effects of obesity and metabolic syndrome2 in the general population.3,4 A longitudinal study of physical fitness, measured by cardiopulmonary exercise testing, was used to predict cardiovascular and all-cause death in asymptomatic females5 and showed that the age-adjusted hazard ratio was 1.20 for each single metabolic equivalent (MET) decrement in their peak exercise capacity. Another longitudinal cohort study of males demonstrated that the absolute value of maximal oxygen uptake (VO2max), (more strongly than its aged-adjusted percentage) was the most powerful predictor of incident mortality regardless of comorbid cardiovascular disease.6 Accordingly, we posit that finding and minimizing adjustable associates with low exercise capacity could possibly attenuate cardiometabolic sequelae and even prevent premature death.

In fact, exercise capacity was associated with self-reported habitual physical activity and could be improved by regular moderate to vigorous exercise.7 Furthermore, individuals with obesity, prone to a sedentary lifestyle,8 were found to frequently have myocardial insufficiency,9 with diastolic and/ or systolic dysfunction in the left ventricle10,11 and pulmonary dysfunction12; this scenario collectively lowered exercise capacity. Furthermore, the severity of insomnia symptoms13 was inversely associated with VO2max. Susceptible to pulmonary hypertension and bilateral ventricular failure,14 patients with sleep-disordered breathing (SDB) supposedly have a lower exercise capacity than their unaffected counterparts. However, this hypothesis has been supported in only a subset of studies15,16 or among patients with comorbid obesity17 or congestive heart failure.18 Interestingly, the links between low exercise capacity and disturbed sleep architecture, self-reported sleep duration/bedtime,19 and insomnia symptoms13 have been reported frequently in young women,20 the elderly,19,21 and in otherwise healthy13 individuals. Nevertheless, the effect of SDB has not been taken into account in these prior studies, and the role of objective sleep duration has not been examined.

To date, underlying and related factors of low values of VO2max and their comprehensive interactions are not completely clear. We therefore hypothesized that obesity, irregular exercise habits, disrupted sleep, SDB, and impaired lung function are mutually interactive, in addition to each being individually associated with low exercise capacity. Accordingly, we have collected several measures relevant to constrained physical fitness from male workers to examine potential latent factors of interest and the interactions among them that are related to low exercise capacity.

METHODS

Participants

A total of 612 otherwise healthy male workers received their regular annual physical checkups at the Chung-Shan Medical University Hospital between July 2007 and March 2010. All those willing to join this cohort study were included, except those with renal or cardiac insufficiency, psychiatric conditions, acute infections, or physical limitations. Participants prescribed beta blockers were instructed to abstain on the morning prior to exercise testing. This study was approved by our Medical Research Ethics Committee (CMUH-IRB No. DMR98-IRB-183). Data were not included in the analysis if the total sleep time (TST) was less than 3 hours, any measure of anthropometric and testing variables was not completed, or if written consent forms were unavailable. Finally, 521 eligible participants (see Table 1) with their complete measures of sleep study, pulmonary function test, and exercise testing participated in the study.

Measures of anthropometric characteristics, pulmonary function, and polysomnographic variables and physical fitness (n = 521).

jcsm.14.12.1995.t01.jpg

table icon
Table 1

Measures of anthropometric characteristics, pulmonary function, and polysomnographic variables and physical fitness (n = 521).

(more ...)

Protocol

Participants were asked to abstain from caffeinated food or drink after lunchtime on the day of testing. They arrived at the sleep center at 9:00 pm to complete anthropometric assessments, questionnaires, and a physical examination. Sleep was evaluated using polysomnography (PSG) between 10:30 pm and 6:00 am, whereas pulmonary function tests and incremental cardiorespiratory exercise testing were performed in the exercise laboratory in the afternoon.

Lifestyle Questionnaire

Our questionnaire included items regarding tobacco-smoking habits (current, former, or never), alcohol consumption, medical history, and recreational exercise habits (none, mild, moderate, and vigorous physical activity were defined as < 1, 1–2, 2–5, and > 5 h/wk, respectively). Responses of none or mild recreational exercise habits were categorized as an irregular exercise lifestyle. Among all participants, 276 (53%) reported a lifestyle consistent with irregular exercise.

Sleep PSG

As reported in our previous studies,22 the procedures, electroencephalography sleep staging, and scoring standards of the sleep parameters such as sleep efficiency, TST, the apneahypopnea index (AHI), arousal index, lowest oxygen saturation (LOS), percentage of the total period for which oxygen saturation was less than 90% (T%dSat), rapid eye movement (REM%), non-REM stage 1 or 2 (N12%), and non-REM slow wave sleep (N3%) as a proportion of TST were all defined using guidelines set forth by The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications (2007).

Pulmonary Functional Testing

In brief, the standard procedures of spirometer testing for maximal voluntary ventilation (MVV), total lung capacity, forced expiratory volume in the first second (FEV1), forced vital capacity (FVC), and lung diffusion capacity for carbon monoxide (DLCO) were performed following the recommendations of the American Thoracic Society/European Respiratory Society.23,24 The prediction equations were as follows:

  • FVC = 0.0844 × height (cm) – 0.0298 × age (years) – 8.7818;

  • FEV1 = 0.0668 × height (cm) – 0.0292 × age (years) – 6.5147;

  • MVV = 3.39 × height (in) – 1.26 × age (years) – 21.40;

  • DLCO = 15.5 × bodily surface area (m2) – 0.238 × age (years) + 6.8.

Accordingly, the percentage-predicted values of the aforementioned parameters were calculated.

Incremental Cardiopulmonary Exercise Testing

The calibration procedures, testing protocol, and the breath-by-breath respiratory analyzer (Vmax Spectra 29 Sensor Medics Inc., Yorba Linda, California, United States) used for cardiopulmonary exercise testing were the same as in a previous study.25 All participants had exercised on a cycle ergometer while measurements of gas exchange were made at 3-minute rest, 3-minute unloaded exercise, 6- to 10-minute incremental exercise followed by a 4-minute recovery period. The incremental size was selected based on the individual's history of particularly daily activity, physical examination, and pulmonary function evaluation. At large, the participant was encouraged to continue as long as safely possible. The value of VO2max was defined as the mean of the three highest values of the 20-second average oxygen consumption. These values, normalized by body weight and divided by 3.5, were transformed into METs.

Statistical Analyses

There are several drawbacks to processing data with path analysis. It seems impractical to use a single indicator to fully employ the intricacies of such a construct required in path analysis.5 Furthermore, path modeling does not allow the likelihood of a degree of interrelationship among the residuals associated with variables. Therefore, it appears unrealistic that there can be an influence only from one variable to another. For example, not only might obesity conceivably affect habits of regular exercise but sedentary lifestyle might also cause obesity.

To comprehensively investigate how the factors could be associated with low exercise capacity, we used a structural equation modeling (SEM) approach for the following reasons. Our hypothesis required a confirmatory factor analysis to verify the existence of underlying latent factors. Fulfilling multiple statistical demands (such as causal modeling, confirmatory factor analysis, regression models, and correlation structure models), as well as handling nonnormally distributed variables and not being biased by measurement error, SEM is a powerful multivariate statistic for the simultaneous analysis of covariance matrices. Using factor analysis and multiple equation modeling based on theory, experimental evidence, and logic, SEM serves as a method of model fitness testing of the hypothesis, which is ultimately confirmed, disproved, or modified.

We set latent variables including obesity, poor sleep quality, impaired lung function, and self-reported physical activity as potential factors relating to low exercise capacity (eg, low MET values). Furthermore, poor sleep quality was categorized into disrupted sleep and sleep-disordered breathing based on PSG-recorded sleep architecture and breathing events, respectively. The primary objective of the analysis was to evaluate the loading of these potential latent (unobservable) variables on observable variables of anthropometric characteristics, including pulmonary function testing and overnight sleep PSG, in addition to loading on MET values. Observable variables, among which the covariation was examined, converge to a smaller number of latent variables. Although observed variables are traditionally depicted as a square or rectangle graphically, unobserved variables, also termed latent factors, are depicted with circles or ovals.

The overall hypotheses are shown as an exploratory structural equation model (Figure 1). After all sole factors, latent factors, and their corresponding manifest factors were set up, a secondary constituent part of the SEM analysis, a kind of multivariant analysis, was performed to investigate whether associative relationships between any two factors existed and whether each factor was directly or indirectly (ie, mediated by other factor[s]) related to low exercise capacity.

The hypothetical (exploratory) structural equation model (model 1).

The latent variables (illustrated in ovals) including impaired lung function, obesity, sleep-disordered breathing, and disrupted sleep cannot be assessed straightforwardly but “load” on manifest variables (illustrated in squares). By univariant SEM path analysis, on MET (low exercise capacity, single arrowhead lines), impaired lung function, obesity, sleep-disordered breathing, and superficial sleep were found loaded on (single arrowhead lines) FEV1%, FVC%, MVV%, DLCO%; WHR, BMI; LOS, T%dSat, AHI; and N3%, sleep efficiency, TST, the manifest variables, respectively. Thus, this hypothesized model would be set up by simultaneously considering the potential associations and modulations (double arrowhead lines) among irregular exercise and these four latent factors of interest. Age, smoking habit, and alcohol consumption were treated as confounding factors, and their effects on impaired lung function were adjusted. AHI = apnea-hypopnea index, BMI = body mass index, DLCO% = the predicted percentage of lung diffusion for carbon monoxide, FEV1% = the predicted percentage of forced expiratory volume in the first second, FVC% = the predicted percentage of forced ventilatory capacity, LOS = lowest oxygen saturation, MET = metabolic equivalent units (maximal oxygen consumption normalized by kilogram body weight/3.5), MVV% = the predicted percentage of maximal ventilatory volume, N3% = percentage of total period in slow wave sleep against total sleep time, SE = sleep efficiency, T%dSat = percentage of total period when oxygen saturation less than 90% against total sleep time, TST = total sleep time, WHR = waist-to-hip ratio.

jcsm.14.12.1995a.jpg

jcsm.14.12.1995a.jpg
Figure 1

The hypothetical (exploratory) structural equation model (model 1).

(more ...)

All data are presented as mean ± standard deviation. Statistical analysis was performed using SPSS 17.0. (SPSS Inc., Chicago, Illinois, United States) The SEM path analysis and interaction processing were conducted using Amos 18.0 (IBM, Armonk, New York, United States) to estimate all parameters simultaneously. The level of significance was set to 5%.

RESULTS

Structural Equation Models

Low exercise capacity was equivalent to negative MET values (Table 2). Using a univariate SEM path confirmatory analysis, we identified latent variables including impaired lung function, obesity, sleep-disordered breathing, and disrupted sleep loaded on low exercise capacity. Herein, FEV1%, FVC%, MVV%, and DLCO% (negatively), waist-to-hip ratio and body mass index (BMI) (both positively), LOS (negatively), T%dSat (positively), AHI (positively), N3% (negatively), sleep efficiency (negatively), and TST (negatively) were eligibly loaded sequentially into the aforementioned latent variables of this SEM model (Table 2, values of P < .001 were noted in all potential latent factors forward linking to low exercise capacity). That is to say, for example, that the latent variable “disrupted sleep,” a characteristic combination of a lower percentage of slow wave sleep, low sleep efficiency, and shorter TST, was found to be associated with “low exercise capacity” by primary univariate path analysis. Notably, age, alcohol consumption, and smoking habit were adjusted for as confounding factors at the loading of impaired lung function. Irregular exercise was also significantly and positively forward linked to low exercise capacity (P < .001; see Table 2). Furthermore, we created a hypothetical structural equation model that integrated the aforementioned four latent variables and irregular exercise and then built common paths forward to low exercise capacity (Table 3, model 1 and Figure 1) while assuming the simultaneous existence of all mutual relationships (double arrowhead lines) between each of the two latent factors. After repetitive trials of SEM regression and path analyses, a pairwise manner remained between each of impaired lung function, irregular exercise, and obesity and also between obesity and SDB. However, the forward path from SDB to low exercise capacity (Table 3, model 2 and Figure 2) eventually vanished. Notably, the arousal index and periodic limb movement index were not eligible at loading into the latent construct “disrupted sleep.”

Path estimates and significances of potential latent variables and irregular exercise to low exercise capacity and their manifest variables by univariant SEM path analysis.

jcsm.14.12.1995.t02.jpg

table icon
Table 2

Path estimates and significances of potential latent variables and irregular exercise to low exercise capacity and their manifest variables by univariant SEM path analysis.

(more ...)

Estimates and significance of associations among the parameters in the structural equation models: the associations among irregular exercise and four latent variables causing low exercise capacity.

jcsm.14.12.1995.t03.jpg

table icon
Table 3

Estimates and significance of associations among the parameters in the structural equation models: the associations among irregular exercise and four latent variables causing low exercise capacity.

(more ...)

Confirmatory structural equational model (model 2) showing all significant paths with standardized coefficients.

For normalized data, the structural equation low exercise capacity(N) = .135 × disrupted sleep(N) + .366 × obesity(N) + .274 × irregular exercise(N) + .152 × impaired lung function(N) + error of low exercise capacity(N) can be built to estimate low exercise capacity. Therefore, with original data, each 1 standard deviation increased in the value of disrupted sleep, obesity, irregular exercise, and impaired lung function was associated with an increase of .135 standard deviation, .366 standard deviation, .274 standard deviation, and .152 standard deviation, respectively, in the value low exercise capacity, represented by reversed MET values. In addition, the indices of the mutual interaction between sleep-disordered breathing and obesity, obesity and irregular exercise, irregular exercise and impaired lung function, and impaired lung function and obesity were .534, .173, .090, and .126, respectively. Asterisks indicate statistical significance: ** = P < .01, *** = P < .001. AHI = apnea-hypopnea index, BMI = body mass index, DLCO% = the predicted percentage of lung diffusion for carbon monoxide, FEV1% = the predicted percentage of forced expiratory volume in the first second, FVC% = the predicted percentage of forced ventilatory capacity, LOS = lowest oxygen saturation, MET = metabolic equivalent units (maximal oxygen consumption normalized by kilogram body weight/3.5), MVV% = the predicted percentage of maximal ventilatory volume, N3% = percentage of total period in slow wave sleep against total sleep time, SE = sleep efficiency, T%dSat = percentage of total period when oxygen saturation less than 90% against total sleep time, TST = total sleep time, WHR = waist-to-hip ratio.

jcsm.14.12.1995b.jpg

jcsm.14.12.1995b.jpg
Figure 2

Confirmatory structural equational model (model 2) showing all significant paths with standardized coefficients.

(more ...)

The indices of fit for the confirmatory SEM model of low exercise capacity are shown in Table 4. This final model was acceptable in terms of a good model fit with normed fit index, incremental fit index, nonnormed fit index, comparative fit index, goodness-of-fit index and adjusted goodness-of-fit index > 0.90, root mean square error of approximation, and standardized root mean square residual < 0.08, minimum discrepancy divided by its degrees of freedom (χ2/df) < 5.0, and critical N for a significance level of 0.05.26,27 Obviously, the final confirmatory model had a better fit than the hypothetical one.

Fitting indices for the hypothetical model and final model.26,27

jcsm.14.12.1995.t04.jpg

table icon
Table 4

Fitting indices for the hypothetical model and final model.26,27

(more ...)

The results of the final SEM model are presented in Table 3, Table 5, and Figure 2, supporting that obesity (βd = .366, P < .001), irregular exercise (βd = .274, P < .001), disrupted sleep (βd = .135, P = .001), and impaired lung function (βd = .152, P < .001), but not SDB (βd = 0, P > .05), are positively linked to low exercise capacity. In addition, obesity (Table 3) was mutually correlated with SDB (r = .534, P < .001), impaired lung function (r = .126, P < .014), and irregular exercise (r = .173, P < .001). Impaired lung function was mutually correlated with irregular exercise (r = .090, P < .048).

Association strengths of four latent variables and irregular exercise to low exercise capacity.

jcsm.14.12.1995.t05.jpg

table icon
Table 5

Association strengths of four latent variables and irregular exercise to low exercise capacity.

(more ...)

The total strengths of associations (βt = βd + βid; Table 5) with low exercise capacity in obesity, irregular exercise, super-ficial sleep, SDB, and impaired lung function were .433, .351, .135, .195, and .223, respectively.

DISCUSSION

By using SEM analyses, low exercise capacity was found to be independently linked to superficial sleep as well as to a vicious cycle of obesity, irregular exercise, and impaired lung function. In contrast, SDB demonstrates an indirect association with low exercise capacity, modulated by obesity and its relevant former cycle. To the best of our knowledge, this study is the first to comprehensively identify multiple latent factors by using SEM, mostly objectively measured, relating to exact values of physical fitness, that is, METs. Although limited to a cross-sectional study design, our findings may give insight into the direction of future causal studies for preventing the negative consequences of low exercise capacity in males.

Indeed, the factor of low physical activity was shown to be not only comorbid with a cluster of cardiovascular risk factors28 but also an independent predictor of all-cause early mortality.3,4 A study with a 6-year follow-up period showed that the absolute maximal exercise capacity measured by gas-exchange cardiopulmonary exercise was the most powerful predictor of mortality, stronger than hypertension, smoking habit, diabetes, and many other exercise-related variables in otherwise healthy male controls and male patients with cardiovascular conditions.6 Actually, exercise capacity appeared positively correlated with physical activity29 and could be improved following long-term moderate to high levels of regular exercise.30 By this approach, detrimental cardiovascular consequences31 could be attenuated even in individuals who were overweight. Notably, exercise capacity exhibited a more powerful effect than habitual physical activity in preventing cardiovascular disease,32 possibly because it is a self-reported measure.33 Nevertheless, self-reported irregular exercise poses direct links to low exercise capacity in the current study, suggesting that two exercise periods of 1 hour per week are sufficient to maintain cardiovascular health.

In terms of associations with low physical fitness, pairwise correlations are found among irregular exercise, obesity, and impaired lung function, which appear quite consistent with the results of prior studies: (1) Individuals with low exercise capacity were more likely to live a sedentary life8 and had the propensity to gain weight. (2) Both whole-body obesity and abdominal obesity may mechanically compromise ventilatory efficiency via restrictive and obstructive lung diseases,12 which are in turn detrimental to cardiovascular health. (3) Long-term high levels of regular exercise might protect those with obesity from detrimental cardiovascular consequences for a long period of time.31 We therefore speculatively suggest that these three factors (obesity, sedentary lifestyle, and pulmonary dysfunction) are not only individually linked to but also instigate a cycle that potentially worsens the low level of physical fitness.

SDB predisposes individuals to pulmonary hypertension and bilateral ventricular heart failure.14 However, it is still not certain whether SDB independently affects exercise capacity and whether these cardiovascular consequences could be attenuated by application of continuous positive airway pressure,34,35 which is the standard therapeutic strategy for SDB. In a study by Guillermo et al. among United States Air Force personnel,34 only patients with moderate to severe sleep apnea (AHI > 20 events/h) had lower exercise capacity than individuals with no apnea; however, SDB itself could not predict the deterioration of physical fitness. Lower exercise capacity was found solely in patients with morbid obesity with SDB in comparison with their counterparts matched in age, sex, BMI, and waist and neck circumference.17 On the contrary, Ucok and colleagues16 concluded that the AHI value was an independent predictor of low exercise capacity but was independent of BMI. Obesity is a well-known important risk factor for the development of SDB, and most patients with severe SDB had obesity. In our study, SDB modulated by obesity and its relevant cycle was indirectly associated with low exercise capacity. This finding could further support the speculation of Ucok et al.16 that obesity might attenuate the association of SDB with exercise capacity. We speculated that men with both SDB and low exercise capacity were likely to exhibit behaviors consistent with disrupted sleep or fall prey to the cycle of obesity, sedentary lifestyle, and lung dysfunction. However, overstated assumptions regarding the vicious cycle based on activity levels should be suspended until further research determining causality is completed. Further, slim Taiwanese people with congenital craniofacial abnormalities36 or chronic rhinitis37 are not uncommon among those experiencing SDB. Further studies are needed to evaluate whether any factors for low exercise capacity are race dependent.

Indeed, in our study, superficial sleep without mutual interaction or covariate effect with SDB was consistent with the results of a previous study with PSG data38 indicating that sleep apnea and nocturnal myoclonus were equally distributed among patients with insomnia, those in the control group, and patients with other sleep disorders. The manifest variables for this objective disrupted sleep are shorter TST, lower sleep efficiency, and less percentage of non-REM deep sleep, which resembles insomnia or hyperarousal symptoms. In fact, previous studies have reported that elderly women with poor sleep measured by actigraphy21 or elderly persons with both longer sleep duration and longer time in bed19 had worse daily physical function. The researchers speculated that superficial sleep (or disrupted sleep) precluded regular physical activity and in turn lowered physical fitness, although no objective VO2max data were available. In addition, Strand et al.13 observed that VO2max was reduced in a linear fashion by the severity of middle insomnia (sleep maintenance insomnia), and this was independent of conventional cardiovascular risk factors and self-reported physical activity. Because self-reported sleep quality was not assessed in our study, it is still unknown whether our “superficial sleep” variable can completely reflect early or middle insomnia. Although chronic insomnia is diagnosed exclusively on self-reported conditions, objective measures are generally considered ineffective in the diagnosis, severity scaling, or therapeutic response of insomnia.39 However, self-reported insomnia was found not only to have unacceptable diagnostic correctness and imperfect generalizability in clinical practice,40 but it also was not proved to be associated with any medical morbidity.40 In contrast, over the past decade,40 objective short sleep duration and sleep efficiency in insomnia has been found to be related to hypertension (catecholamine excretion and cortisol plasma levels), type 2 diabetes (glucose levels and insulin sensitivity), and even mortality40 in the Penn State Cohort, suggesting objective short sleep duration predicting the biological severity of chronic insomnia, relevant to the activation of the two limbs of the stress system. In addition, an objective short sleep duration (< 6 hours) in the general population independently played a role not only as an enhancer of the metabolic syndrome41 but also as a mediator of hypertension31 on mortality. The current finding that objective disrupted sleep appears directly and independently associated with reduced cardiorespiratory fitness in men apparently further supports that “insomnia with objective short sleep duration is the most biologically severe phenotype of the disorder” speculated by Vgontzas et al.40 We may postulate that the association of objectively measured disrupted sleep with low cardiorespiratory fitness is at least partially relevant to the activation of the stress system,41 affecting hemodynamics and glucose regulation. Further longitudinal neurobiological and cardiometabolic studies along with repetitive complete cardiopulmonary exercise testing may shed light on the mechanisms of exercise limitation. In addition, further clinical trials should examine whether lengthening sleep could improve physical fitness or vice versa.

Our study has some strengths, including using SEM to simultaneously evaluate variables of diverse aspects linked to low cardiorespiratory fitness, measuring the exact values of VO2max as the indicator of exercise capacity, minimizing errors from submaximal exercise data,42 and measuring objective sleep parameters by a standard full-night PSG study.

Limitations

The current study design does not permit definitive conclusions of causality between potential factors and low exercise capacity. Confirmation of the proposed model with findings from prospective, longitudinal, and interventional trials is still required. However, the internal consistency of model fit indicates that the links in the model represent a potential association between these variables. Indeed, low exercise capacity might precede, contribute to, or worsen various latent factors or vice versa in the current study. Furthermore, the reliability of self-reported lifestyles may bias the results. Our sample comprised individuals of low to middle socioeconomic status, which might limit the generalizability of these findings to other groups. However, this sample possessed a substantial degree of homogeneity, particularly in ethnic identity and socioeconomic status, which may affect the observed statistical relationships. Indeed, whether these findings are replicable among women or other ethnic or socioeconomic groups is unknown. Nevertheless, to date, no study has found that sex moderates the interactions among physical activity, maximal exercise capacity, perceived activity levels, and sleep problems. The generalizability of our findings is limited to otherwise healthy male Taiwanese workers. Finally, our model with a good fit still cannot rule out other models based on different hypotheses but with equally good fitness. Future studies may benefit from the inclusion of more direct latent factors on low exercise capacity.

CONCLUSIONS

In our study of male workers, SEM is promising for unraveling the complex potential factors linked to low physical fitness. Although the current study is limited by its cross-sectional design, our results might imply that sleep disruption correlates better with low exercise capacity than indices of severity of SDB. Undoubtedly, further longitudinal studies are warranted. Our findings indicate that individuals with limited exercise capacity without definite causes should undertake a sleep study, particularly in those who describe symptoms of SDB or insomnia.

DISCLOSURE STATEMENT

All authors have seen and approved the manuscript. This study was supported in part by grant MOST 104-2314-B-039 -013 -MY3 from the Ministry of Science and Technology, Taiwan, and Taiwan Ministry of Health and Welfare Clinical Trial and Research Center of Excellence (MOHW105-TDU-B-212-133019). Research was supported by the Program for Professor of Special Appointment (Eastern Scholar, Honorary Chair Professor) at Shanghai Institute of Higher Education (No 2012-47). The authors report no conflicts of interest.

ABBREVIATIONS

AGFI

adjusted goodness-of-fit index

AHI

apnea-hypopnea index

BMI

body mass index

CFI

comparative fit index

CN

critical N for a significance level of 0.05

DLCO

lung diffusion capacity for carbon monoxide

FEV1

forced expiratory volume in the first second

FVC

forced vital capacity

GFI

goodness-of-fit index

IFI

incremental fit index

LOS

lowest oxygen saturation

MET

metabolic equivalent

MVV

maximal voluntary ventilation

MVV%, FEV1%, FVC%, DLCO%

percentage-predicted values of the aforementioned lung function parameters

NFI

normed fit index

NNFI

nonnormed fit index

PCFI

parsimonious comparative fit index

PNFI

parsimonious normed fit index

PSG

polysomnography

REM%, N12% and N3%

rapid eye movement, non-REM stage 1 or 2, and non-REM slow wave sleep as a proportion of TST

RMSEA

root mean square error of approximation

SDB

sleep-disordered breathing

SE

sleep efficiency

SEM

structural equation modeling

SRMR

standardized root mean square residual

T%dSat

percentage of the total period for which oxygen saturation was less than 90%

TLI

Tucker Lewis index

TST

total sleep time

VO2max

maximal oxygen uptake

WHR

waist-to-hip ratio

χ2/df

minimum discrepancy divided by its degrees of freedom

REFERENCES

1 

Ravussin E, Bogardus C. Relationship of genetics, age, and physical fitness to daily energy expenditure and fuel utilization. Am J Clin Nutr. 1989;49 5 Suppl:968–975. [PubMed]

2 

Katzmarzyk PT, Church TS, Janssen I, Ross R, Blair SN. Metabolic syndrome, obesity, and mortality: impact of cardiorespiratory fitness. Diabetes Care. 2005;28(2):391–397. [PubMed]

3 

Lee CD, Blair SN, Jackson AS. Cardiorespiratory fitness, body composition, and all-cause and cardiovascular disease mortality in men. Am J Clin Nutr. 1999;69(3):373–380. [PubMed]

4 

Wei M, Kampert JB, Barlow CE, et al. Relationship between low cardiorespiratory fitness and mortality in normal-weight, overweight, and obese men. JAMA. 1999;282(16):1547–1553. [PubMed]

5 

Schreiber JB, Nora A, Stage FK, Barlow EA, King J. Reporting structural equation modeling and confirmatory factor analysis results: a review. J Educ Res. 2006;99(6):323–337

6 

Myers J, Prakash M, Froelicher V, Do D, Partington S, Atwood JE. Exercise capacity and mortality among men referred for exercise testing. N Engl J Med. 2002;346(11):793–801. [PubMed]

7 

Jackson AS, Kampert JB, Barlow CE, Morrow JR Jr, Church TS, Blair SN. Longitudinal changes in cardiorespiratory fitness: measurement error or true change? Med Sci Sports Exerc. 2004;36(7):1175–1180. [PubMed]

8 

Janssen I, Katzmarzyk PT, Srinivasan SR, et al. Utility of childhood BMI in the prediction of adulthood disease: comparison of national and international references. Obesity Res. 2005;13(6):1106–1115

9 

Zhou YT, Grayburn P, Karim A, et al. Lipotoxic heart disease in obese rats: implications for human obesity. Proc Natl Acad Sci U S A. 2000;97(4):1784–1789. [PubMed Central][PubMed]

10 

Ammar KA, Redfield MM, Mahoney DW, Johnson M, Jacobsen SJ, Rodeheffer RJ. Central obesity: association with left ventricular dysfunction and mortality in the community. Am Heart J. 2008;156(5):975–981. [PubMed Central][PubMed]

11 

Redfield MM, Jacobsen SJ, Burnett JC Jr, Mahoney DW, Bailey KR, Rodeheffer RJ. Burden of systolic and diastolic ventricular dysfunction in the community: appreciating the scope of the heart failure epidemic. JAMA. 2003;289(2):194–202. [PubMed]

12 

Leone N, Courbon D, Thomas F, et al. Lung function impairment and metabolic syndrome: the critical role of abdominal obesity. Am J Respir Crit Care Med. 2009;179(6):509–516. [PubMed]

13 

Strand LB, Laugsand LE, Wisloff U, Nes BM, Vatten L, Janszky I. Insomnia symptoms and cardiorespiratory fitness in healthy individuals: the Nord-Trondelag Health Study (HUNT). Sleep. 2013;36(1):99–108. [PubMed Central][PubMed]

14 

Sajkov D, Cowie RJ, Thornton AT, Espinoza HA, McEvoy RD. Pulmonary hypertension and hypoxemia in obstructive sleep apnea syndrome. Am J Respir Crit Care Med. 1994;149(2 Pt 1):416–422. [PubMed]

15 

Lin CC, Hsieh WY, Chou CS, Liaw SF. Cardiopulmonary exercise testing in obstructive sleep apnea syndrome. Respir Physiol Neurobiol. 2006;150(1):27–34. [PubMed]

16 

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(1):29–36. [PubMed]

17 

Vanhecke TE, Franklin BA, Zalesin KC, et al. Cardiorespiratory fitness and obstructive sleep apnea syndrome in morbidly obese patients. Chest. 2008;134(3):539–545. [PubMed]

18 

Oldenburg O, Lamp B, Faber L, Teschler H, Horstkotte D, Töpfer V. Sleep-disordered breathing in patients with symptomatic heart failure: a contemporary study of prevalence in and characteristics of 700 patients. Eur J Heart Fail. 2007;9(3):251–257. [PubMed]

19 

Stenholm S, Kronholm E, Bandinelli S, Guralnik JM, Ferrucci L. Self-reported sleep duration and time in bed as predictors of physical function decline: results from the InCHIANTI study. Sleep. 2011;34(11):1583–1593. [PubMed Central][PubMed]

20 

Lee AJ, Lin WH. Association between sleep quality and physical fitness in female young adults. J Sports Med Phys Fitness. 2007;47(4):462–467. [PubMed]

21 

Goldman SE, Stone KL, Ancoli-Israel S, et al. Poor sleep is associated with poorer physical performance and greater functional limitations in older women. Sleep. 2007;30(10):1317–1324. [PubMed Central][PubMed]

22 

Ting H, Lo HS, Chang SY, et al. Post- to pre-overnight sleep systolic blood pressures are associated with sleep respiratory disturbance, pro-inflammatory state and metabolic situation in patients with sleep-disordered breathing. Sleep Med. 2009;10(7):720–725. [PubMed]

23 

Macintyre N, Crapo RO, Viegi G, et al. Standardisation of the single-breath determination of carbon monoxide uptake in the lung. Eur Respir J. 2005;26:720–735. [PubMed]

24 

Miller MR, Hankinson J, Brusasco V, et al. Standardisation of spirometry. Eur Respir J. 2005;26(4):319–338

25 

Ting H, Lee SD, Chung AH, et al. Effects of bilateral T2-sympathectomy on static and dynamic heart rate responses to exercise in hyperhidrosis patients. Auton Neurosci. 2005;121(1-2):74–80. [PubMed]

26 

Arbuckle JL. Amos [version 7.0]. Chicago, IL: SPSS; 2006.

27 

Chase-Carmichael CA, Ris MD, Weber AM, Schefft BK. Neurologic validity of the Wisconsin card sorting test with a pediatric population. Clin Neuropsychol. 1999;13(4):405–413. [PubMed]

28 

Lee DC, Artero EG, Sui X, Blair SN. Mortality trends in the general population: the importance of cardiorespiratory fitness. J Psychopharmacol. 2010;24 4 Suppl:27–35

29 

Prentice AM, Jebb SA. Obesity in Britain: gluttony or sloth? BMJ. 1995;311(7002):437–439. [PubMed Central][PubMed]

30 

American College of Sports Medicine. ACSM's Guidelines for Exercise Testing and Prescription. 8th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2010.

31 

Fernandez-Mendoza J, He F, Vgontzas AN, Liao D, Bixler EO. Objective short sleep duration modifies the relationship between hypertension and all-cause mortality. J Hypertens. 2017;35(4):830–836. [PubMed]

32 

Blair SN, Jackson AS. Physical fitness and activity as separate heart disease risk factors: a meta-analysis. Med Sci Sports Exerc. 2001;33(5):762–764. [PubMed]

33 

Walsh MC, Hunter GR, Sirikul B, Gower BA. Comparison of self-reported with objectively assessed energy expenditure in black and white women before and after weight loss. Am J Clin Nutr. 2004;79(6):1013–1019. [PubMed]

34 

Guillermo LQ, Gal TJ, Mair EA. Does obstructive sleep apnea affect aerobic fitness? Ann Otol Rhinol Laryngol. 2006;115(10):715–720. [PubMed]

35 

Vgontzas AN, Liao D, Bixler EO, Chrousos GP, Vela-Bueno A. Insomnia with objective short sleep duration is associated with a high risk for hypertension. Sleep. 2009;32(4):491–497. [PubMed Central][PubMed]

36 

Chang ET, Shiao GM. Craniofacial abnormalities in Chinese patients with obstructive and positional sleep apnea. Sleep Med. 2008;9(4):403–410. [PubMed]

37 

Li HY, Wang PC, Hsu CY, Cheng ML, Liou CC, Chen NH. Nasal resistance in patients with obstructive sleep apnea. ORL J Otorhinolaryngol Relat Spec. 2005;67(2):70–74. [PubMed]

38 

Vgontzas AN, Kales A, Bixler EO, Manfredi RL, Vela-Bueno A. Usefulness of polysomnographic studies in the differential diagnosis of insomnia. Int J Neurosci. 1995;82(1-2):47–60. [PubMed]

39 

Littner M, Hirshkowitz M, Kramer M, et al. Practice parameters for using polysomnography to evaluate insomnia: an update. Sleep. 2003;26(6):754–760. [PubMed]

40 

Vgontzas AN, Fernandez-Mendoza J, Liao D, Bixler EO. Insomnia with objective short sleep duration: the most biologically severe phenotype of the disorder. Sleep Med Rev. 2013;17(4):241–254. [PubMed Central][PubMed]

41 

Fernandez-Mendoza J, He F, LaGrotte C, Vgontzas AN, Liao D, Bixler EO. Impact of the metabolic syndrome on mortality is modified by objective short sleep duration. J Am Heart Assoc. 2017;6(5)

42 

Lewicke A, Beck KC. Establishing equivalence of exercise intensities. Open Sports Med J. 2009;3:21–25