Peripheral muscle dysfunction has been consistently related to local tissue depletion in patients with chronic obstructive pulmonary disease (COPD). Several studies have found that the preferential loss of muscle mass in the lower limbs has profound effects in patients' daily functioning(Reference Bernard, LeBlanc, Whittom, Carrier, Jobin, Belleau and Maltais1–Reference Franssen, Broekhuizen, Janssen, Wouters and Schols5).
A number of methods have been used to assess segmental body composition in this patient population(Reference Bernard, LeBlanc, Whittom, Carrier, Jobin, Belleau and Maltais1, Reference Steiner, Barton, Singh and Morgan6–Reference Kilduff, Fuld, Neder, Pitsiladis, Carter, Stevenson and Ward8). Although some of them are highly accurate and precise, e.g. dual-energy X-ray absorptiometry (DEXA) and computed tomography, they are not widely available and require sophisticated equipment in a purpose-designed setting. For a widespread use in the clinical context, however, the ‘ideal’ method needs to be practical, portable and inexpensive. Moreover, it should be sensitive to detect tissue depletion and present with a high-degree of external validity, i.e. its estimates need to correlate well with independent measures of functional capacity. In this context, the anthropometrically based method of Jones & Pearson(Reference Jones and Pearson9) to estimate leg lean volume (LLV) could be valuable(Reference Degens, Sanchez-Horneros, Heijdra, Dekhuijzen and Hopman10). Despite this method being extensively used in healthy subjects(Reference Doré, Bedu, França and Van Praagh11–Reference Tothill and Stewart13), including the elderly(Reference Marsh, Paterson, Govindasamy and Cunningham14, Reference Pearson, Cabbold, Orrell and Harridge15), no previous study has evaluated its clinical usefulness in patients with COPD presenting with variable degrees of tissue depletion.
The present objective, therefore, was to determine the feasibility, accuracy and external validity of LLV estimates as established by the Jones & Pearson approach(Reference Jones and Pearson9) in depleted and non-depleted patients with stable COPD.
Methods
Study population
Forty-eight (ten females) patients, aged 50 or older, with stable, moderate-to-severe COPD according to the Global Initiative for Chronic Obstructive Lung disease criteria(Reference Pauwels, Buist, Calverley, Jenkins and Hurd16), were studied. They were referred from the COPD outpatients clinic of the Federal University of São Paulo (São Paulo, Brazil). All patients had forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) ratio < 0·70 and post-bronchodilator FEV1 < 70 % predicted; in addition, they were clinically stable with no disease exacerbation in the past 4 weeks and they had quit smoking for at least 6 months. Patients presenting with physical disabilities, other wasting chronic diseases (e.g. chronic renal failure, diabetes), under long-term oxygen therapy and BMI higher than 30 kg/m2 were excluded. No patient had taken part in a pulmonary rehabilitation programme in the preceding 6 months. All patients were optimized in terms of standard medical therapy: maintenance medication included long- and short-acting β2-agonists, theophylline and inhaled steroids. No patient was receiving oral steroids or diuretics. The Local Ethics Committee approved the study and written informed consent was obtained from all participants.
Measurements
Pulmonary function tests
Spirometric tests were performed by using the CPF System™ (Medical Graphics Corporation, St Paul, MN, USA) with airflow being measured by a calibrated pneumotachograph. The subjects completed at least three acceptable maximal forced expiratory manoeuvres before and after 400 μg inhaled salbutamol; in the present paper, only post-bronchodilator data are reported. FVC (litre) and FEV1 (litre) were recorded and expressed as percentage of the predicted value(Reference Pereira, Barreto, Simões, Pereira, Gerstler and Nakatani17). A constant-volume and differential-pressure body plethysmograph (1085 Profiler & Elite, Series D; Medical Graphics Corporation) was used to measure static lung volumes (total lung capacity (litre) and residual volume (litre), respectively). Carbon monoxide diffusing capacity (DLCO) was measured by the modified Krogh technique (single breath); the subjects performed three acceptable and reproducible tests, with the results being within 10 % or 3 ml CO/min per mmHg. Predicted values for lung volumes and DLCO were those proposed by Neder et al. (Reference Neder, Andreoni, Castelo-Filho and Nery18, Reference Neder, Andreoni, Peres and Nery19) for the adult Brazilian population. Arterial blood was taken from the radial artery according to standard anaerobic conditions for pH, arterial oxygen tension (mmHg) and arterial carbon dioxide tension (mmHg) measurements.
Body composition assessment
Body anthropometry
Body height was determined to the nearest 0·1 cm with subjects standing barefoot. Body weight was assessed with the beam scale to the nearest 0·1 kg, with subjects standing barefoot and in light clothing. BMI was calculated as weight/height2 (kg/m2).
Leg lean volume anthropometry
The lean (fat-free) volume of the right leg (litre) was calculated by the Jones & Pearson method(Reference Jones and Pearson9), which is based on the summation of truncated cones (Fig. 1). Briefly, with the patient standing erect and the feet slightly apart seven circumferences were taken with a metric tape at predetermined sites: the gluteal furrow (C1), one-third of the subischial height up from the tibial–femoral joint space (C2), the minimum circumference above the knee (C3), the maximum circumference around the knee (C4), the minimum circumference below the knee (C5), the maximum calf circumference (C6) and the minimum ankle circumference (C7). The heights (h) above the floor level for each circumference were obtained by using a stadiometer. In addition, anterior and posterior skinfold thicknesses were measured at C2 (thigh) and C6 (calf) using a Harpenden caliper following standard techniques(Reference Harrison, Buskirk, Carter, Johnston, Lohman, Pollock, Roche, Wilmore, Lohman, Roche and Martorell20). All measurements were performed by a nutritionist or dietitian.
Fig. 1 Schematic illustration of the sites from which the anthropometric measurements should be taken in order to obtain the six truncated cones (1–6) for leg lean volume estimation according to the Jones & Pearson technique(Reference Jones and Pearson9). b, diameter of ankle; h, height above floor level. Reproduced with the permission of Wiley-Blackwell Publishing Ltd from Jones PR & Pearson J (1969) Anthropometric determination of leg fat and muscle plus bone volumes in young male and female adults, Journal of Physiology 204, 63–66.
In order to obtain the volume of a truncated right circular cone (Vc), the following equation was used:
where h is the distance between the circumferences, r is the radius of the upper plane and R is the radius of the lower plane. The volume of the foot (Vf) was also calculated by assuming the foot to be wedge-shaped:
where l is the foot's length and b is the diameter calculated from C7 (Fig. 1). LLV (litre) was then calculated by summing up the volumes of the six lean cones (i.e. by subtracting the skinfold readings from their respective diameters) plus Vf. In addition, LLV was expressed as a function of leg height (m) to further correct for inter-individual differences in leg dimensions. All equations have been previously entered into a widely available data sheet software for faster results (Microsoft™ Office Excel, 2003).
Dual-energy X-ray absorptiometry
Total body and right leg fat-free masses (FFM; kg) were also measured by DEXA fan beam technology (QDR-4500A; Hologic Inc., Bedford, MA, USA). In this method, the subjects are scanned with photons produced by an X-ray source at two different energy levels. Bone ash (calcium hydroxyapatite) tissue and soft tissue are separated based on the degree of photon attenuation. The differential absorption with soft tissue is also measured and the ratio of absorbency of the two energy level photons has been shown to be linearly related to the percentage of fat in these tissues. In order to obtain lean (bone plus muscle) leg volumes by DEXA, we divided the mass of each compartment (fat, muscle and bone) by their mean respective densities(Reference Cheng, Li, Lu, Keyak and Lang21–Reference Wang, Bachrach, Van Loan, Hudes, Flegal and Crawford23). In the present study, nutritional depletion was established if BMI ≤ 21 kg/m2 and/or the FFM index (FFM/height2) from whole-body DEXA was ≤ 15 kg/m2 for females and ≤ 16 kg/m2 for males(Reference Vermeeren, Creutzberg, Schols, Postma, Pieters, Roldaan, Wouters and on behalf of the COSMIC study group24).
Exercise and peripheral muscle performance
Six-minute walking distance
Functional exercise capacity was measured by the 6 min walking distance test (m) in a 50 m in-hospital corridor. Technical procedures were those recommended by the American Thoracic Society(25). The tests were performed in duplicate after familiarization and the highest value was recorded.
Isokinetic dynamometry
Concentric contractions of the quadriceps femoris (knee extension) of the right leg were evaluated by using an isokinetic dynamometer (Con-Trex™; Cybex, Chattanooga, NY, USA). All patients performed a maximum isokinetic strength test with two trials of five sequential contractions at an angular velocity of 60° per s as peak torque (N × m). After a rest period of at least 5 min, patients performed an isometric strength (N × m) test with two trials against a fixed resistance pad at 60° per s. After another resting period, patients performed two knee-extensor tests at 300° per s (thirty repetitions) to record mean total work (J/contraction)(Reference Neder, Nery, Shinzato, Andrade, Peres and Silva26). The highest value was selected for analysis in all tests.
Statistical analysis
Results are presented as means and standard deviations. In order to contrast body composition and functional variables between depleted and non-depleted patients, a non-paired t test was used. Pearson's correlation coefficient was obtained to estimate the level of association between continuous variables. Sensitivity and specificity of LLV by anthropometry in identifying depleted patients were assessed by a ROC curve analysis. The level of statistical significance was set at P < 0·05.
The limits of agreement between LLV estimates by anthropometry and DEXA were investigated by plotting the individual differences against their respective means (Bland–Altman analysis). Heteroscedasticity was examined by plotting the absolute (i.e. ignoring any sign) differences against the individual means and calculating the Spearman's correlation coefficient(Reference Bland and Altman27). If the heteroscedasticity correlation was close to zero, the mean bias and the 95 % limits of agreement were calculated as mean ± 1·96 sd of the between-estimate differences.
Results
Nineteen subjects (39 %) were considered as nutritionally depleted(Reference Vermeeren, Creutzberg, Schols, Postma, Pieters, Roldaan, Wouters and on behalf of the COSMIC study group24). As shown in Table 1, they had more severe airflow obstruction and lower DLCO than non-depleted subjects. In fact, fifteen of the nineteen depleted patients (79 %) were classified as stages III–IV according to the Global Initiative for Chronic Obstructive Lung disease guidelines(Reference Pauwels, Buist, Calverley, Jenkins and Hurd16); in contrast, twenty-five of the twenty-nine non-depleted subjects (86 %) were on stages II–III. In addition, depleted subjects had significantly lower LLV, exercise capacity and peripheral muscle performance (P < 0·01; Table 1).
Table 1 Demographic, body composition and functional characteristics of depleted and non-depleted patients with chronic obstructive pulmonary disease
DLCO, carbon monoxide diffusing capacity; FEV1, forced expiratory volume in 1 s; FFM, fat-free mass; FFMI, fat-free mass index; FVC, forced vital capacity; PaCO2, arterial carbon dioxide tension; PaO2, arterial oxygen tension; RV, residual volume; TLC, total lung capacity.
Mean values were significantly different from those of the non-depleted group (non-paired t test): *P < 0·05.
As described in the Methods, LLV by anthropometry was compared with the values derived from DEXA scan of the right leg (Table 2). A Bland–Altman analysis revealed that the mean bias of the LLV differences between anthropometry and DEXA were 0·40 litre (95 % CI − 0·59, 1·39) and 0·50 litre (95 % CI − 1·08, 2·08) for depleted and non-depleted patients, respectively (Fig. 2). There were significant correlations between all of the muscle functional attributes with LLV by anthropometry: the correlation coefficients did not differ substantially from those obtained against leg FFM by DEXA. Interestingly, however, only in non-depleted patients was there a significant correlation between distance walked and leg volume and mass (Table 3).
Table 2 Leg composition by dual-energy X-ray absorptiometry and anthropometry measurements (Jones & Pearson method(Reference Jones and Pearson9)) in depleted and non-depleted patients with chronic obstructive pulmonary disease
C1 to C7, circumferences 1 to 7; FFM, fat-free mass; FM, fat mass; LF, length of the foot; LLC1 to LLC7, leg lean circumferences 1 to 7; LLV, lean leg volume; TLV, total leg volume.
Mean values were significantly different from those of the non-depleted group (non-paired t test): *P < 0·05.
Fig. 2 The limits of agreement (Bland–Altman plot) between anthropometry (ANTHRO) and dual-energy X-ray absorptiometry (DEXA) in estimating the leg lean volume (litre) in depleted (●) and non-depleted (○) patients with chronic obstructive pulmonary disease.
Table 3 Correlation between leg lean mass by dual-energy X-ray absorptiometry and leg lean volume by anthropometry with selected indexes of functional capacity and peripheral muscle strength in depleted and non-depleted patients with chronic obstructive pulmonary disease
Pearson's R correlation coefficients: *P < 0·05.
We also sought to investigate whether the leg height-corrected LLV estimates would be sensitive enough to differentiate depleted from non-depleted patients. As depicted in Fig. 3, the area under the ROC curve showed high sensitivity and specificity in identifying depleted patients. Therefore, a leg height-corrected LLV ≤ 9·2 litres/m was 95 % sensitive and 80 % specific to indicate nutritional depletion. Moreover, patients with reduced height-corrected LLV had significantly lower exercise capacity and muscle performance than their counterparts who presented with higher values (Table 4; P < 0·05).
Fig. 3 A ROC curve analysis of the diagnostic performance of anthropometric estimates (leg lean volume/leg height, l/m) in identifying patients with nutritional depletion. Note that a threshold value of 9·2 litres/m presented with the best combination of sensitivity and specificity (95 and 80 %, respectively). Depletion was defined according to the following criteria(Reference Vermeeren, Creutzberg, Schols, Postma, Pieters, Roldaan, Wouters and on behalf of the COSMIC study group24): BMI ≤ 21 kg/m2 and/or fat-free mass index ≤ 15 kg/m2 in females and ≤ 16 kg/m2 in males. AUC, area under the curve.
Table 4 Resting and exercise characteristics of chronic obstructive pulmonary disease patients presenting (Group A) or not (Group B) with leg lean volume/leg height≤9·2 litres/m
DLCO, carbon monoxide diffusing capacity; FEV1, forced expiratory volume in 1 s; FFM, fat-free mass; FFMI, fat-free mass index; FVC, forced vital capacity; PaCO2, arterial carbon dioxide tension; PaO2, arterial oxygen tension; RV, residual volume; TLC, total lung capacity.
Mean values were significantly different from those of Group B (non-paired t test): *P < 0·05.
Discussion
The present study has shown that LLV estimates, as determined by the anthropometric technique of Jones & Pearson(Reference Jones and Pearson9), were comparable with those derived from DEXA and they correlated as well as lean (bone plus muscle) DEXA readings with measures of peripheral muscle function in depleted and non-depleted patients with COPD. The present data indicate that this method presented with acceptable accuracy and external validity in this patient population. Moreover, LLV showed high specificity and sensitivity to identify patients with whole-body FFM depletion. This technique, therefore, might prove to be useful in clinical settings where more advanced and expensive techniques for segmental body composition assessment are not accessible.
There is a long-standing interest in developing anthropometry-based methods for segmental body composition evaluation in different clinical populations. More recently, it has been shown that many patients with COPD may develop important functional abnormalities on the peripheral muscle, which seem to be related, at least partially, to local FFM depletion(Reference Serres, Hayot, Prefaut and Mercier2, Reference Franssen, Broekhuizen, Janssen, Wouters and Schols5, Reference Baarends, Schols, Mostert and Wouters28, Reference Malaguti, Nery, Dal Corso, Nápolis, De Fuccio, Lerario, Castro and Neder29). There is also evidence demonstrating that leg FFM depletion is a strong predictor of morbidity and mortality in these patients(Reference Marquis, Debigare, Lacasse, LeBlanc, Jobin, Carrier and Maltais4). However, several investigators have found that appendicular muscle depletion correlates variably with whole-body measurements and direct evaluation of limb tissue depletion is advisable(Reference Degens, Sanchez-Horneros, Heijdra, Dekhuijzen and Hopman10, Reference Couillard, Maltais, Saey, Debigaré, Michaud, Koechlin, LerBlanc and Préfaut30). Unfortunately, most studies involving segmental body composition assessment in patients with COPD used sophisticated and expensive methods (e.g. computed tomography and DEXA) which hampered their practical application in clinical settings.
In the present study, we investigated the feasibility, accuracy and external validity of an alternative, anthropometry-based technique that was proposed almost four decades ago(Reference Jones and Pearson9). The Jones & Pearson approach(Reference Jones and Pearson9) uses standard geometric principles to model the human leg as a series of truncated cones from the gluteal furrow down to the foot (Fig. 1). By obtaining the thickness of anterior and posterior subcutaneous fat deposition on the thigh and on the mid-calf, the inner (lean) leg volume can be estimated. The main advantage of this method relies on its low cost and the lack of a need for technologically advanced equipment. However, it should be recognized that it requires a reasonable amount of manual labour (approximately 20 min/patient) and careful standardization of the places to measure the circumferences and their heights from the floor. As cited, the method also assumes that anterior and posterior skinfold thickness remains invariable across the thigh and leg cones which can introduce a source of error in the estimation of lean volume. In addition, it is advisable that the recorded values be entered into a previously designed computing program for quicker and more precise results.
We found that the limits of agreement between LLV measured by this technique and by DEXA have a small mean bias (within 0·5 litre or 7 %) with a symmetrical distribution of the residuals, either in depleted and non-depleted patients. Therefore, as there was no evidence of heteroscedasticity on data distribution (i.e. error proportional to the mean) and the mean inter-method difference was within 20 % for most subjects, the estimates seem to provide an acceptable accuracy for estimating LLV in these patients (see later). In this regard, however, a note of caution should be made: it is likely that a small error has been introduced in the between-method comparison of the thigh volume, as a diagonal line delimits the upper limit of the thigh by DEXA but the top plane of cone 1 is horizontal (Fig. 1).
We also showed that LLV estimated by anthropometry and FFM readings by DEXA correlate similarly with selected indexes of skeletal muscle functioning, which indicated a comparable degree of external validity. Interestingly, however, neither LLV nor FFM were significantly related to 6 min walking distance in depleted patients. Although the exact reason for this finding is not clear, it should be noted that depleted patients had significantly lower maximal ventilatory capacity than non-depleted subjects (Table 1). Therefore, it is conceivable that pulmonary-ventilatory mechanisms, as opposed to peripheral muscles factors, played a greater role to limit whole-body exercise capacity in depleted patients.
Another interesting finding of the present study was the high specificity and sensitivity of the technique in identifying depleted patients with COPD (Fig. 3). In fact, patients with leg-height corrected LLV ≤ 9·2 litres/m, the best cut-off to indicate nutritional depletion on a ROC curve analysis, had significantly worse lung function and physical performance than patients with higher LLV values (Table 4). To the authors' knowledge, this is the first study to look at the practical value of segmental anthropometry in distinguishing depleted from non-depleted patients with COPD. However, this threshold value should be further validated in larger and more heterogeneous samples (see later).
The present study has some relevant limitations. Firstly, the method inter- and intra-subject variabilities remain to be determined; however, they are expected to be similar to other anthropometric techniques as only a few skinfold measurements are needed and the limb circumferences are not especially difficult for trained subjects to obtain. Secondly, it was not our objective to establish the responsiveness of the technique to nutritional or ergogenic interventions and future studies should address this issue. Thirdly, we did not use more direct techniques for LLV determination in order to compare them with anthropometry: the DEXA estimates, therefore, cannot be assumed as the ‘true’ values. Fourthly, it remains to be determined whether this method compares favourably against other similarly simple, albeit costlier, techniques for the evaluation of segmental body composition, such as appendicular bioimpedance(31,32). Finally, the present results should not be extrapolated to other sub-populations of patients with COPD, including those who are obese (BMI>30 kg/m2), hypoxaemic, clinically unstable and dehydrated.
In conclusion, a simple and inexpensive technique to estimate the LLV by anthropometry(Reference Jones and Pearson9) was shown to present with acceptable accuracy and external validity in depleted and non-depleted patients with stable COPD. This approach seems to be of practical value when more complex and costlier techniques for segmental body composition assessment are not available.
Acknowledgements
The authors would like to thank all colleagues from the Pulmonary Function and Clinical Exercise Physiology Unit (Division of Respiratory Diseases, Department of Medicine, Federal University of São Paulo (UNIFESP), Brazil) for their friendship and support during this study. D. S. V. was supported by a Master Fellowship Grant from CAPES (Coordenadoria de Aperfeiçoamento do Pessoal de Nível Superior, Brazil). J. A. N. is an Established Investigator (level II) of the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Brazil. D. S. V. was in charge of data collection, contributed to data analysis and wrote the first draft of the manuscript. M. C. L. helped to design the study. S. dal C. and L. N. contributed to patient recruitment and participated in data analysis and interpretation. A. L. P. A. participated in data analysis and interpretation. M. L.-C. is the supervisor of the DEXA laboratory of UNIFESP and also contributed to data interpretation. A. S. helped to design the study and contributed to data interpretation. L. E. N. also participated in designing the study and in data interpretation. J. A. N. originally designed the study, supervised the group in data analysis and interpretation, and wrote the advanced versions of the manuscript. All participants have read and approved the final version of the manuscript. There are no conflicts of interest.
