Body Composition, Physical Function and Exercise Capacity in Chronic Obstructive Pulmonary Disease

Abstract Current literature yields unequivocal results regarding the effect of body composition on physical function in patients with chronic obstructive pulmonary disease and disproportionately includes a majority of males. The purpose of this study was to determine whether specific body composition measures are significantly associated with physical function and exercise capacity in patients with chronic obstructive pulmonary disease with equal representation of males and females. Independent variables included sex, total body mass, total body lean and fat mass, appendicular total mass, and appendicular lean and fat mass. Dependent variables included peak oxygen consumption, 6-minute walk distance and self-reported physical function. Patients (n = 170) with dual-energy X-ray absorptiometry data, 6-minute walk distance, and self-reported physical function were used for these analyses. A sub-set of 145 of these patients with peak oxygen consumption data were also analysed. Hierarchical multiple regression analysis was used to determine if sex and body composition measures correlated with physical function and exercise capacity and if they explained additional variance after controlling for disease severity. After controlling for disease severity, appendicular lean mass, total body fat mass, and sex explained an additional 16.5% of the variance in peak oxygen consumption (p < 0.001). Appendicular lean mass explained an additional 8.9% of the variance in 6-minute walk distance (p < 0.001) and an additional 2.5% of the variance in self-reported physical function (p = 0.057). Body composition measures may further predict exercise capacity, 6-minute walk distance, and self-reported physical function in patients with chronic obstructive pulmonary disease.


Introduction
Chronic obstructive pulmonary disease (COPD) is a slowly progressive disease characterized by persistent respiratory symptoms and airflow obstruction [1].In 2013, 6.4% of American adults reported physician-diagnosed COPD [2].Due to its prevalence, COPD is the fourth leading cause of death in the United States [3] and creates a tremendous financial and medical burden [4,5].Individuals with COPD experience impairments in physical functioning, exercise capacity, and health-related quality of life [6].
Skeletal muscle dysfunction is often experienced in patients with COPD and negatively impacts functional outcomes, hospitalization rates, and mortality [1,7].Complications associated with COPD, such as poor exercise tolerance and dyspnoea on exertion, discourage regular physical activity, which may exacerbate skeletal muscle dysfunction.
Similar to healthy older adults, patients with COPD undergo alterations in body composition, particularly a reduction in skeletal muscle mass and an increase in visceral fat [8,9].Concurrent obesity and low skeletal muscle mass may result in even worse health outcomes than either of the two conditions on their own [10].Previous investigations which have used body mass index (BMI) as a surrogate for obesity have shown a higher BMI to be associated with lower levels of physical activity, impaired physical function, and reduced exercise capacity in patients with COPD [11,12].There lacks a consensus, however, as several studies have shown similar levels of exercise capacity and physical function when comparing overweight and obese patients with COPD to normal weight patients [13,14].Few of these investigations examined specific components of body mass, such as fat mass or fat-free mass, which have opposing effects on physical function, nor did they include a representative sample of females.
Recently, Wan et al. found that dual energy X-ray absorptiometry (DXA) assessed fat mass was inversely related to physical function; however, there was no relationship between lean mass and physical function [12].These results contrast other studies which have shown physical function and exercise capacity to be positively correlated with fat-free or lean mass [14][15][16][17].Similar to Wan et al., these studies included very few females in the sample.The prevalence of COPD among women has risen in recent years such that prevalence rates are similar in men and women [18].Additionally, sex can influence the clinical expression and symptoms of COPD and patients' response to pulmonary rehabilitation [19,20].Because of conflicting reports regarding the effect of various components of body composition on physical function and exercise capacity and the small numbers of females in these previous studies, the purpose of this investigation was to determine whether specific body composition measures are significantly associated with physical function and exercise capacity in a cohort of patients with COPD with equal representation of males and females.

Study design
The data presented here are from the Reconditioning Exercise and Chronic Obstructive Pulmonary Disease Trial II, a single-center, single-blinded, randomized control trial sponsored by the National Institute of Health.All participants involved in the parent study signed informed consent that was approved by the Wake Forest University Institutional Review Board.The primary purpose of this trial was to compare the effectiveness of a 12-month group-mediated cognitive behavioural intervention against a traditional three-month exercise therapy session in the promotion of long-term physical activity.The complete design, aims, rationale, and results of this study are described elsewhere [21,22].
In total, 170 patients with COPD were included from the original trial.Descriptive statistics for the patient group are presented in Table 1.Of the 170 patients enrolled in this investigation, 145 had peak oxygen consumption (V ̇O2peak ) data that were used in the analyses.Data for this subset are also presented in Table 1.Peak oxygen consumption data was not available on 25 of the 170 patients because of technical difficulties with the gas analysis system.To be eligible for inclusion in the study, participants had to: have an expiratory airflow limitation such that forced expiratory volume in one second (FEV 1 ) to forced vital capacity ratio (FEV 1 /FVC) was ≤ 70% and FEV 1 was ≥ 20% of predicted, reported shortness of breath when completing activities of daily living, be free of severe cardiovascular disease, uncontrolled hypertension, or diabetes, not undergoing active cancer treatment, and not have participated in a structured rehabilitation or exercise program within the past 3 months.A medical history, physical exam, and graded exercise test were used for the determination of a medical diagnosis that would preclude participation in the study.Inclusion of patients into the study was based on the Global Initiative for Chronic Obstructive Lung Disease criteria [23].All participants were evaluated at baseline, 3, 6, and 12 months.For these secondary analyses, only baseline patient data was examined.

Body composition
Dual energy X-ray absorptiometry (DXA) was used to determine body composition.Total body DXA scans using a whole-body fan-beam Delphi ATM, Hologic, Inc. (Waltham, MA) scanner were performed by an experienced DXA technologist and reviewed by a board-certified radiologist.Scans were conducted in accordance with the manufacturer's recommendations for patient positioning, scan protocols, and scan analysis and with the patient lying in a supine position.From each scan, total body lean and fat mass and appendicular lean and fat mass were obtained.Appendicular lean mass and fat mass were calculated, respectively, as the summation of the total lean mass of the right and left arms and legs and the summation of the total fat mass of the right and left arms and legs.

Pulmonary function testing
Tests of pulmonary function were conducted using a Medical Graphics Corporation 1085D plethysmograph (Medical Graphics Corporation, St. Paul, MN, USA).Guidelines from the American Thoracic Society were used to perform the spirometry and lung volume measurements [24].Patients were requested not to use bronchodilator medications three to four hours prior to testing.FEV 1 (% predicted), FVC, and FEV 1 /FVC were assessed, with FEV 1 (% predicted) serving as a measure of disease severity.Prediction equations used to calculate the percent of predicted values were those from the NHANES III study [25].1.9 ± 0.6 2.0 ± 0.6 abbreviations: sD, standard deviation; bmi, body mass index; fev 1 , forced expiratory volume in one second (expressed as a percent of predicted); fvC, forced vital capacity; fev 1 /fvC, ratio of the fev 1 to the fvC; v ̇O2peak , peak oxygen consumption.results of the physical function analyses, both objectively measured (6-minute walk distance) and self-report (fitness arthritis and seniors trial 23 item inventory) were derived from all 170 patients.results of the exercise capacity analysis (v ̇O2peak ) were derived from the 145 patient subset.

Exercise capacity
Exercise capacity was defined as peak oxygen consumption (V ̇O2peak ) measured during a graded exercise test on a motor driven treadmill (Quinton 55XT, Quinton Cardiology, Deerfield, WI, USA).The test was administered using a modified Naughton protocol.Patients were given standardized instructions prior to the start of the test and were instructed to continue exercising for as long as possible.
Patients received verbal encouragement during the test.Indications for test termination were in accordance with those set by the American College of Sports Medicine.
Oxygen consumption (ml/kg/min) was measured throughout the test using a CPX-D gas exchange system (Medical Graphics, St. Paul, MN, USA).Prior to the start of the test, calibrations of the pneumotachograph and gas analyzers were performed according to the manufacturer's specifications.Gases used during calibration were all certified standard gases that had been verified via Haldane chemical analysis.Gas exchange and minute ventilation were calculated on a breath-by-breath basis and were reported in minute intervals.Patients breathed through a rubber mouthpiece that was connected to a disposable Pitot tube flow meter.Expiratory airflow was measured using the Pitot tube flow meter and rapid responding analyzers were used to determine relative fractions of O 2 and CO 2 .Oxygen consumption was calculated using O 2 and CO 2 concentrations and the volume of expired air.While all study patients completed the graded exercise test to rule out severe cardiovascular disease, V ̇O2peak data were not available on 25 of these patients due to technical issues with the gas exchange system.

Objective physical function
Physical function was defined as the distance walked in six minutes.The 6-minute walk was conducted in a dedicated gymnasium according to the guidelines from the American Thoracic Society [26].Patients were instructed to walk as far as possible in the six minutes; however, they were allowed to stop and rest if needed.No verbal feedback was given to the patient during the 6-minute walk test.

Self-reported physical function
Self-reported physical function was assessed using a functional performance questionnaire originally developed for the Fitness Arthritis and Seniors Trial [27].This inventory, comprised of 23 questions, prompted participants to indicate how much difficulty they experienced performing select activities of daily living over the past month.Responses were rated along a Likert scale from one (usually done with no difficulty) to five (unable to do).A composite disability score was developed by averaging each of the 23 scores.This composite index has an α reliability of 0.79.

Statistical analysis
Pearson product moment correlations were used to examine the relationships between the various lean and fat mass measures, objective and subjective physical function measures, and exercise capacity.Lean and fat mass measures most highly correlated with physical function and exercise capacity were then evaluated using hierarchical multiple regression analysis.More specifically, after controlling for lung function (FEV 1 % of predicted), a stepwise entry method was employed to determine if sex and specific measures of lean and fat mass helped explain additional variance in the measures of physical function and exercise capacity.Regression analyses were initially conducted with the data stratified by sex.Differences in the regression coefficients between men and women were then tested.The only regression coefficients found to differ between men and women was FEV 1 for the 6-minute walk analysis (p = 0.045).Because no other differences were found in the regression coefficients between men and women, data from both sexes were combined and the results of those analyses are presented.All analyses were performed using Statistical Package for the Social Sciences (SPSS) software version 26.0 for Windows (Chicago, IL, USA).Significance was set at the 0.05 level for all analyses.

Exercise capacity
Weak relationships were found between exercise capacity (V ̇O2peak ) and FEV 1 % predicted (r = 0.250, p = 0.001), total body lean mass (r = 0.251, p = 0.001), appendicular lean mass (r = 0.277, p < 0.001), appendicular fat mass (r = −0.153,p = 0.033), and sex (r = −0.179,p = 0.016).Total body fat mass was not found to be significantly correlated with V ̇O2peak (r = −0.087,p = 0.150).Results of the hierarchical regression analysis with V ̇O2peak as the dependent variable are shown in Table 2.After controlling for FEV 1 % predicted, appendicular lean mass, total body fat mass, and sex entered into the model.FEV 1 % predicted accounted for 6.3 percent of the variance in V ̇O2peak .Appendicular lean mass, total body fat mass, and sex accounted for an additional 16.5 percent of the variance.

Objective physical function
Weak relationships were found between 6-minute walk distance and FEV 1 % predicted (r = 0.203, p = 0.004), appendicular lean mass (r = 0.327, p < 0.001), total body lean mass (r = 0.287, p < 0.001) and sex (r = −0.210,p = 0.003).Neither total body fat mass nor appendicular fat mass were found to be significantly correlated with 6-minute walk distance (r = −0.016,p < 0.418 and r = −0.032,p = 0.340, respectively).Results of the hierarchical regression analysis with 6-minute walk distance as the dependent variable are shown in Table 3.After controlling for FEV 1 % predicted, both appendicular lean mass and total body lean mass entered into the model.However, because of problems with multicollinearity, total body lean mass was removed from the model.The FEV 1 % predicted was found to account for 4.1 percent of the variance in 6-minute walk distance.Appendicular lean mass accounted for an additional 8.9 percent of the variance.Sex was not a significant contributor to the model.

Self-Reported physical function
Weak relationships were found between the physical function questionnaire and appendicular lean mass (r = −0.182,p = 0.009), total body lean mass (r = −0.153,p = 0.023) and sex (r = 0.287, p < 0.001).FEV 1 % predicted, total body fat mass and appendicular fat mass were not found to be significantly correlated with physical function questionnaire score (r = −0.009,p = 0.455, r = 0.018, p = 0.409, and r = 0.029, p = 0.355, respectively).Results of the hierarchical regression analysis with physical function questionnaire score as the dependent variable are shown in Table 4.After controlling for FEV 1 % predicted, both appendicular lean mass and total body lean mass entered into the model.Similar to 6-minute walk distance, problems with multicollinearity were present in this model, thus total body lean mass was removed.FEV 1 % predicted accounted for less than one percent of the variance in physical function questionnaire score.Appendicular lean mass accounted for an additional 2.5 percent of the variance.Sex was not found to be a significant contributor to the model.

Discussion
The purpose of this study was to determine if specific body composition measures were associated with exercise capacity and physical function in a sample of patients with COPD featuring equal representation of males and females.Significant correlations were found between appendicular and total body lean mass measures and exercise capacity (V ̇O2peak ), objectively measured physical function (6-minute walk distance), and self-reported physical function (physical function questionnaire).Sex was also significantly correlated with all three dependent variables, but only entered the regression equation for V ̇O2peak .Albeit not significantly correlated with V ̇O2peak , total body fat mass did account for a small percentage of the variability in V ̇O2peak .
After controlling for disease severity, it was found that appendicular lean mass, total body fat mass, and sex explained an additional 16.5% of the variance in V ̇O2peak .Appendicular lean mass explained an additional 8.9% of the variance in 6-minute walk distance and an additional 2.5% of the variance in physical function questionnaire score after controlling for disease severity.While disease severity did account for some of the variance in exercise capacity and physical function, specific measures of body composition accounted for additional variance in each.Previous research has indicated that while COPD adversely affects both the respiratory and limb muscles, a greater level of disability occurs in the limbs [28].As such, it is not surprising that appendicular lean mass was associated with measures of physical function in the current study.
Previous studies examining the role of body composition on physical performance and exercise capacity in COPD patients have used either bioelectrical impedance analysis or DXA and have generally shown estimates of skeletal muscle mass to be positively associated with exercise capacity and physical function.In general, our results support those of previous investigations and show that estimates of skeletal muscle are predictive of exercise capacity and physical function in patients with COPD in both males and females.Studies that have utilized DXA to assess body composition in order to examine the relationship with exercise capacity and physical function are limited.Yoshikawa et al. showed that lean mass is positively correlated with maximal exercise capacity on a cycle ergometer in male patients with COPD, a finding supported by the results of this investigation [29].
In contrast to the findings of Yoshikawa et al. and those of this investigation, Wan et al. did not find lean mass to be correlated with 6-minute walk distance.These investigators, however, found fat mass to be inversely correlated with 6-minute walk distance [12].Differences between the results of Wan et al. and ours may be due to the fact that the subjects in her investigation were almost exclusively male.In a study with approximately equal representation of males and females, Machado et al. found that there was a greater frequency of males with low lean mass in those characterized as overweight or obese; however, the effects of low fat-free mass on exercise capacity were less pronounced in overweight and obese patients.Our patients were primarily overweight as determined from BMI and had equal representation of males and females, whereas the study by Wan et al. included mostly males classified as obese, and this may help explain the lack of consensus.Thus, it appears that in obese patients, fat mass may be a better predictor of physical function and exercise capacity; whereas in non-obese patients, lean mass may be a better predictor.Studies which have stratified patients by BMI are less conclusive about the effect of obesity on physical function with some studies indicating no association between BMI and physical functioning [13] and others indicating poor physical function is associated with a high BMI [11,30].Machado et al. found 6-minute walk distance to be similar among normal weight, overweight and obese patients.Interestingly these investigators showed V ̇O2peak , as a percent of predicted, to be greater in the overweight and obese patients.
The present study provides useful insight into the effect of appendicular lean muscle mass on physical function and exercise capacity in males and females with COPD.Obesity appears to present conflicting effects on exercise capacity in COPD.While obese COPD patients are assumed to have greater muscle mass and strength, resulting in improved physical functioning, increasing fat mass may, conversely, worsen pulmonary mechanics and induce exercise-limiting dyspnoea by reducing end-expiratory lung volume [31].Understanding the quantity and distribution of fat mass and lean mass may help guide exercise and nutrition regimens for COPD patients.Many existing studies utilize BMI as a surrogate for lean mass and fat mass; however, BMI cannot discern body composition values, failing to recognize the different contributions from fat mass versus lean mass in the performance of physical tasks.
This investigation has several strengths.It is one of the first studies to examine the relationships between measures of body composition and exercise capacity and physical function in patients with COPD with equal representation of males and females.Another strength of this study is that we used DXA for the assessment of body composition.DXA has been shown to be an accurate and reliable means of measuring body composition and is the preferred method of assessing body composition in patients with COPD [32].
Limitations to this study include the fact that many of the patients within this sample presented with moderate (FEV 1 50%-79%) or severe (FEV 1 30%-49%) COPD.Physical function and exercise capacity may vary depending on disease severity, calling for a greater representation of those with mild and very severe disease severity.The original purpose of the parent trial, from which this secondary analysis is based, was to improve physical activity in COPD patients.As a secondary analysis, this study cannot determine whether purposefully increasing appendicular and total body lean mass, as the present study suggests, can yield a clinically significant increase in 6-minute walk distance or V ̇O2peak .While self-report measures of physical function have been shown to assess constructs different from those of objectively assessed measures of function, there are limitations with the use of self-report measures.Self-report measures are based on an individual's perception of their level of functioning in relationship to their environment and can be influenced by factors such as depression, anxiety, cognitive impairment, and level of education.As such, individuals may over-or under-estimate their ability to perform certain activities depending on their situation [33].Additionally, Waite et al. reported that there are a number of differences between individuals who volunteer to participate in research studies versus those that do not including levels of education and anxiety [34].As the subjects in this investigation had volunteered to participate in a 3-month behavioral intervention, this could have influenced their self-report physical function scores, and these results should be viewed in the context of these limitations.

Conclusion
The current study indicates that measures of body composition, specifically appendicular and total body lean mass, explained additional variance in V ̇O2 peak , 6-minute walk distance, and physical function questionnaire.Sex was correlated to all three measures but only entered into the regression equation for V ̇O2peak .These findings warrant the need for future research to evaluate whether exercise programs specifically aimed at improving lean mass in patients with COPD can effectively yield significant improvements in exercise capacity and physical function outcomes.

Table 2 .
hierarchical regression analysis with v ̇O2peak as the dependent variable.
note: B are the unstandardized regression coefficients.95% Ci for b are the 95% confidence intervals for the unstandardized regression coefficients.β are the standardized regression coefficients.95% Ci for β are the 95% confidence intervals for the standardized regression coefficients.ΔR 2 is the change in R 2 with the addition of each variable to the model.

Table 3 .
hierarchical regression analysis with 6-minute walk distance as the dependent variable.B are the unstandardized regression coefficients.95% Ci for B are the 95% confidence intervals for the unstandardized regression coefficients.β are the standardized regression coefficients.95% Ci for β are the 95% confidence intervals for the standardized regression coefficients.ΔR 2 is the change in R 2 with the addition of each variable to the model.

Table 4 .
hierarchical regression analysis with physical function questionnaire as the dependent variable.B are the unstandardized regression coefficients.95% Ci for B are the 95% confidence intervals for the unstandardized regression coefficients.β are the standardized regression coefficients.95% Ci for β are the 95% confidence intervals for the standardized regression coefficients.ΔR 2 is the change in R 2 with the addition of each variable to the model.