INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Deterioration of lung function and nutritional status is frequently observed in patients with cystic fibrosis. Clinical deterioration in these patients is associated with decreased exercise tolerance. Nutritional status and exercise tolerance in CF patients are associated with prognosis and survival (1, 2), and preservation of exercise tolerance and nutritional status is therefore important (3, 4). Exercise programs have been shown to improve exercise tolerance in adults and children with CF (5).

Fat-free mass and body weight are strong predictors of cycling maximal workload (Wmax) and maximal oxygen consumption (O₂max) in healthy children (9). In clinically deteriorated children with CF, significantly reduced Wmax and O₂max have been demonstrated. After correction for decreased body weight, Wmax but not O₂max was reduced as compared with the reference population (10). In these patients, reduced maximal work capacity during cycling is also associated with reduced maximal work capacity in standardized six-min walking tests (11). In patients with CF, oxidative work performance by skeletal muscle is reduced, possibly as a result of diminished nutritional status or decreased oxygen delivery (12). These findings indicate that the upper limits of physical activity patterns in CF patients are reduced, but the mechanisms involved in this reduction have not yet been satisfactorily elucidated.

For the present study, we hypothesized that children with CF have peripheral muscle weakness in concomitance with reduced oxidative work efficiency and reduced maximal exercise performance. This relationship was studied in patients with mild symptoms and in healthy controls. Also, in order to distinguish effects mediated by nutritional status and pulmonary function, we compared patients with more severe clinical symptoms with the control subjects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects were enrolled from a roster of patients with CF who were regularly attending the outpatients' clinic of our hospital, with a total of 63 patients in the desired age range between 10 and 18 yr. Patients with an unstable clinical condition, those unable to perform a reproducible exercise test, and those with additional morbidity, such as recurrent pneumothorax, hemoptysis, and arthritis were excluded. Of 40 eligible patients, 12 refused to cooperate for personal or practical reasons. Thus, 28 patients with CF were included in the study. All but two patients were affected by pancreatic insufficiency, and received supplemental pancreatic enzymes and vitamins A, D, and E. One group was composed of 15 CF patients with moderately severe disease, as shown by diminished pulmonary function (FEV₁ < 80% predicted [13]) and/or diminished nutritional status (weight for age < 1 SD as compared with a Dutch reference) (14). Several of these patients had shown diminished weight gain during the 2 yr preceding the study. The other group was composed of 13 patients with mild symptoms. These patients had normal pulmonary function (FEV₁ > 80% of predicted), and their body weight was not more than 1 SD below the median of the Dutch reference. None of these patients had experienced diminished growth velocity for height and/or weight at any time. Thirteen healthy controls were also selected. They were enrolled from a secondary school (total population approximately 1,000 pupils), where weight and height were measured in 430 children aged 10 to 18 yr. From this sample, children with a history of respiratory complaints or asthma were excluded. The remaining children were matched for gender, age, weight, and height with individual CF patients with mild symptoms, and the child that most closely matched the index CF patient was eligible for the study. Consent was refused for personal reasons by two children, and these were replaced by the second-best matching child. Thus, 13 children with mild CF were matched with an equal number of healthy peers. Characteristics of the three groups are presented in Table 1. The study was approved by the medical ethical committee of our hospital, and informed consent was obtained in all selected cases.

Anthropometry

Fat-free mass (FFM) was calculated from skinfold thickness measured at four standard sites (biceps, triceps, subscapular, and suprailiac), as described by Durnin and Rahaman (15). Mid-upper-arm circumference was measured with a tape, and mid-upper-arm muscle area was calculated from arm circumference and biceps and triceps skinfold thicknesses, as described by Gerver and De Bruin (14).

Pulmonary Function Tests

FVC and FEV₁ were obtained from maximal expiratory flow-volume curves (Masterscreen; Jaeger, Breda, The Netherlands). RV and TLC were measured with a volume-constant body plethysmograph (Masterlab; Jaeger). Values are expressed as the percent of predicted values (13).

Habitual Physical Exercise

Structured questions were asked, and answers were recorded in the procotol notes for each subject as follows: (1) normal school attendance, including full participation in supervised gymnastic activity (complete/incomplete/no); and (2) participation in supervised group sport activities of at least 1 h/wk (yes/no). Unsupervised individual physical activity was recorded, but is not reported in this study.

Maximal Work Capacity

A maximal exercise test was performed on a cycle ergometer (Lode Examiner, Groningen, The Netherlands) as previously described (10). Briefly, tests at stepwise-increments of 15 W/min (in seven patients, with FEV₁ < 70% predicted and/or height < 1.70 m, increments of 10 W/min were used) were performed until exhaustion. Continuous measurements of ventilation, oxygen consumption, and carbon dioxide production were made with a valveless mouthpiece (Oxycon Champion; Jaeger). Internal gas and volume calibrations were made before each measurement. Gas volumes are expressed under standard conditions (0° C and 10⁵ N/m²). Recordings of gas volumes and workload were stored automatically in a computer. Heart rate and transcutaneous oxygen saturation were continuously monitored. The highest workload sustained for 1 min was taken as the maximum. Three minutes after a subject reached Wmax, capillary blood samples were taken from the middle finger for immediate measurement of the lactate concentration. All performed tests were considered to be maximal, according to published criteria (10). Breathing reserve at peak exercise was calculated as (MMV  Emax)/MVV · 100%, where MVV is the maximal voluntary ventilation calculated from the maximal expiratory flow-volume curves as 40 · FEV₁ and Emax is the ventilation at peak exercise. From recordings stored during the exercise tests, the oxygen cost of exercise was computed as the slope of the oxygen consumption and workload changes between 0 and 60% O₂max, and the same was done for the slope of ventilation and carbon dioxide production and ventilation and workload change.

Peripheral Muscle Strength

Isometric muscle force was measured bilaterally with subjects in standard positions, using a hand-held myometer (Penny and Giles, Cristchurch, UK), as described by Bäckman (16). Maximal voluntary force measurements were made for six skeletal muscle groups: shoulder abductors, elbow flexors, wrist extensors, hip and knee extensors, and ankle dorsal flexors. Each muscle group was tested three times. The highest value was taken.

Statistics

Normality of distribution was present for all variables. Comparisons between the moderate CF patient group and controls were made with t tests for unrelated samples. Comparisons between the patients with mild symptoms and controls were made with paired t tests. Results for peripheral muscle strength are presented as total maximal muscle force (i.e., summed bilateral maximal force in six muscle groups, since factor analyses showed a single-factor solution [Eigenvalue = 10.0, 77% of total variance]). Correlation analyses (Pearson's r) were performed for anthropometric data, Wmax, O₂max, lung function parameters, and total muscle force. Linear regression analyses were made for FFM, total muscle force, and Wmax, and also for the oxygen cost of exercise and the slope of ventilation versus carbon dioxide production. Comparisons were made for between-group differences in slopes and the vertical differences between regression lines with the method described by Altman and Gardner (17), with appropriate analyses of covariance (ANCOVAs) for independent and paired data. Multiple regression analyses were performed for Wmax and O₂max with FFM and total muscle force, and with additional independent variables (FEV₁, FVC, gender, and group effects), as well as with interaction effects. Differences were considered significant if p < 0.05 (two-tailed), except for correlation analyses, in which p < 0.01 was chosen.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Anthropometric Data, Lung Function, and Habitual Physical Exercise

The CF patients with mild symptoms and matched controls had similar FFM and upper arm muscle area (Table 1). Lung function parameters in these CF patients were 100% of predicted values; in the control subjects FEV₁ was higher than predicted (p < 0.05). All but one of the 13 patients with mild symptoms, and all controls, attended school gymnastics. In these two groups, three and two children, respectively, did not participate in group sports. Patients with moderate CF had significantly lower clinical (Shwachman [18]) scores and deteriorated lung parameters than did the patients with mild CF. In the group with moderately severe symptoms, three of 16 patients were not engaged in school gymnastics, and seven patients did not participate in group sport activities.

Maximal Work Capacity and Nutritional Status

Results for physical performance are shown in Table 2. Maximal workload and Wmax/kg FFM were significantly lower in the two groups of CF patients than in the controls. O₂max and O₂max/kg FFM were significantly decreased in the patients with moderate CF as compared with controls. Oxygen saturation at maximal cycling and postexercise plasma lactate concentrations were not significantly different in the two groups. In the patients with mild symptoms of CF, Wmax/kg FFM was decreased in comparison with that of the matched controls (3.9 W/kg versus 4.6 W/kg, respectively; difference = 0.7 W/kg [95% CI = 1.2 to 0.3 W/kg]). However, O₂max in these patients was not significantly different from that in matched controls (for O₂max/kg FFM = 57 O₂ · kg¹ · min¹ versus 61 ml O₂ · kg¹ · min¹, respectively; difference = 4 O₂ · kg ¹ · min¹ [95% CI = 10 to 3 O₂ · kg¹ · min¹). FFM was associated with Wmax and O₂max in all groups (Table 3). Body weight was associated with Wmax and O₂max in controls and in CF patients with moderate symptoms; this association was not found in the patients with mild symptoms, however, because of a wider range of fat mass in this group. Single regression analyses of Wmax and FFM are shown in Figure 1. Comparisons between groups showed similar slopes, but significant differences were found in fitted Wmax between controls and CF patients with mild and those with moderate symptoms (difference in vertical distance = 35 W [95% CI = 14 to 56 W] and 47 W [95% CI = 28 to 66 W], respectively). These differences reflect the lower Wmax for FFM in subjects with CF as compared with the Wmax predicted from FFM for healthy children, and are in agreement with values for Wmax/FFM (Table 2). Regression analyses for O₂max and FFM showed no difference in slope or vertical distance.

View larger version (12K): [in this window] [in a new window]   Figure 1.   Fat-free mass and maximal workload in CF patients with mild (closed squares; dotted line) and moderate symptoms (closed triangles; dashed line), and controls (X; solid line). Regression equations (X = FFM [kg]; Y = maximal workload [W]): mild CF: Y = 22 + 4.43 · X (r = 0.75, p = 0.003); moderate CF: Y = 40 + 4.49 · X (r = 0.82, p < 0.001); controls: Y = 7 + 4.46 · X (r = 0.93, p < 0.001). Differences were analyzed for independent samples (moderate CF versus controls) according to Altman and Gardner (17), and for matched pairs (mild CF versus controls) with linear regression analyses. No significant covariance was found. Results are described in the text.

Peripheral Muscle Strength and Maximal Work Capacity

Muscle force was significantly decreased in both groups of CF patients as compared with controls (Table 2). Means for total muscle force in patients with mild CF versus controls were 2.7 kN versus 3.1 kN, and the mean (95% CI) difference was 0.4 kN (0.8 to 0.1 kN); similar trends were seen in separate analyses for upper- and lower-extremity muscle force (summed bilateral force in three leg muscle groups = 1.6 ± 0.3 kN versus 1.9 ± 0.4 kN, p = 0.037). In correlation analyses, total muscle force was significantly associated with Wmax and O₂max in all groups. Results for Wmax and total muscle force are shown in Figure 2. Regression analyses of Wmax (Y, given in W units) as a function of total muscle force (X, given in units of kN) showed very similar functions for both groups of CF patients (regression equation for moderate CF: Y = 63 + 0.090 · X, for mild CF: Y = 69 + 0.087 · X), with proportionally lower values in the patients with moderately severe symptoms (and lower mean FFM) as compared with their peers with mild symptoms. Four of 13 CF patients with mild symptoms and one of 15 patients with moderate CF had Wmax values higher than predicted values for total muscle force fitted from control data (Figure 2). Total muscle force was associated with a disproportionate decrease in Wmax, as shown by a significant difference in slope from that of controls (regression equation: Y = 67 + 0.043 · X). The difference (95% CI) in slopes between patients with moderate CF and controls was 0.047 W/kN (0.004 to 0.090 W/kN), and that between patients with mild CF symptoms and controls was 0.044 W/kN (0.001 to 0.087 W/kN). For O₂max versus total muscle force, no significant differences in slopes were found.

View larger version (11K): [in this window] [in a new window]   Figure 2.   Total muscle force and maximal workload in CF patients with mild (closed squares; dotted line) and moderate symptoms (closed triangles; dashed line), and controls (X; solid line). Differences were analyzed for independent samples (moderate CF versus controls) according to Altman and Gardner (17). For matched pairs (mild CF versus controls), linear regression analyses were conducted and showed significant covariance (F = 5.12, p = 0.045) between the model (df 2,10) for matched pairs of muscle force as compared with the model (df 1,11) with the paired differences. Results described in the text are given for the appropriate statistical models.

Oxygen Cost of Work, and Breathing during Submaximal Exercise

In the three study groups, O₂max and Wmax showed high correlations (Table 3). Correlations between oxygen consumption and workload, and between ventilation and carbon dioxide production, were also high during the submaximal part of the exercise test (data not shown). Comparisons for the slopes and breathing reserve are shown in Table 2. The mean oxygen cost of exercise was significantly higher in both groups of CF patients than in the controls (mild CF and moderate CF versus controls, respectively: 0.19 ml/J and 0.18 ml/J versus 0.16 ml/J; differences [95% CI]: 0.03 ml/J [0.01 to 0.04 ml/J] and 0.02 ml/J [0.001 to 0.04 ml/J]). Ventilation during submaximal exercise was significantly increased, and breathing reserve was decreased, in proportion to the carbon dioxide production in the CF patients with moderate symptoms as compared with the controls, but no significant differences were found between patients with mild symptoms and controls. Neither the slope of the ventilation versus workload change nor the intercept was significantly different between the groups.

Other Determinants of Work Capacity

In patients with moderate CF, FVC was significantly correlated with Wmax, O₂max, and total muscle force, and FEV₁ was significantly correlated with O₂max (Table 3). In the other two groups, correlations between lung function and physical capacity parameters were not significant. Multiple regression analyses of combined data showed that FFM and muscle force consistently explained large proportions of variance in Wmax and O₂max (multiple R²: 0.84 to 0.91). For patients with moderate CF and controls, multiple regression analyses for O₂max showed significant effects of muscle force (p = 0.01), and significant interactions between FFM and FEV₁ (p < 0.001) and between muscle force and FEV₁ (p = 0.03), with multiple R² = 0.93 and no additional effect for gender. For Wmax, the effects of FEV₁ and gender were not significant. Multiple regression showed no significant effects or interactions for lung function and exercise parameters in CF patients with mild symptoms and controls. Gender had a small additional effect on O₂max (p = 0.03), but not on Wmax (p = 0.051) when FFM and muscle force were accounted for.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We found that as compared with healthy controls: (1) peripheral muscle weakness, diminished work capacity, and an increased oxygen cost of exercise are present in children with CF; (2) muscle weakness in CF patients is associated with a disproportionate decrease in Wmax, even in patients with normal pulmonary volumes and nutritional status; and (3) in patients with deteriorated clinical status, a low FFM further compromised muscle strength, Wmax, and O₂max.

Maximal physical performance in children is strongly related to body mass (muscle mass) and cardiopulmonary capacity: Wmax and O₂max are associated with body weight and, in particular, with FFM (9). FFM increases during childhood and adolescence, and is affected by age, puberty, and gender. Correction of maximal workload for diminished body weight in children with CF underestimates their reduced work capacity, but this is not the case when work capacity parameters are expressed per unit FFM (10). In the present study, significant associations between FFM, Wmax, and O₂max were found within each group. Importantly, in the absence of overt pulmonary symptoms at rest and during exercise, the patients with mild CF closely matched control subjects for relevant confounders, and also for FFM and upper-arm muscle area, but their mean maximal workload and peripheral muscle force were reduced by 17% and 13%, respectively.

Postexercise lactate concentrations and oxygen saturation were not different in CF patients and controls, suggesting similar relative contributions of glycolytic adenosine triphosphate (ATP) synthesis at maximal workload. Cycling at maximal and submaximal levels requires activity of many muscle groups, in order to exert force on the pedals during the effective part of each cycle. All tests were performed on the same ergometer, and with similar cycling rates (of approximately 55 rpm in patients and controls). Thus, a direct relation between dynamic effective leg muscle force (times duration) and workload can be deduced. The disproportional decrease in Wmax in comparison with muscle force in the CF patients may be explained by differences in calcium stress, which may have blunted their maximal isometric muscle force (19, 20), or by differences in sensation of fatigue in the peripheral muscles, as recently reported in subjects with chronic obstructive pulmonary disease (21, 22). No measurements of these mechanisms were conducted in the present study, and these and other explanations thus cannot be excluded.

The diminished Wmax and peripheral muscle weakness in children with mild CF as compared with their matched controls suggests the presence of intrinsic abnormalities in skeletal muscle in patients with this disorder, consisting either of diminished efficiency of oxidative ATP synthesis (in the mitochondria) or abnormalities in myofibril mechanics (differences in fiber recruitment, or diminished efficiency of force generation by myosin ATP hydroxylase). The increased oxygen cost of work during the progressive exercise test (18% higher in mild CF patients than in controls) support this explanation. In patients with CF, a 19 to 25% diminished efficiency of mitochondrial oxidative ATP synthesis has been previously demonstrated in ³¹P-magnetic resonance studies of forearm muscle during steady-state submaximal voluntary contractions (12). The O₂max in the CF patients with mild symptoms in the present study was only 8% lower than and not significantly different from that of matched controls. Thus, neither ventilatory restrictions nor decreased rates of maximal oxidative phosphorylation in muscle mitochondria appear to explain the lower maximal workload in these patients. A putative defect in the efficiency of mitochondrial oxidative phosphorylation in CF may be primary (i.e., related to expression of mutated CF transmembrane conductance regulator [CFTR], or to a splicing variant thereof, as has been demonstrated in heart muscle [23]), or to functional abnormalities in its adenosine triphosphatase (ATPase) activity (24). However, expression of the CFTR has not yet been reported in human skeletal muscle. Mitochondrial dysfunction in CF muscle could also be secondary (e.g., associated with subclinical pulmonary infections and systemic inflammatory changes). Although these and other possible factors (e.g., differences in intensity of habitual physical exercise) cannot be excluded as causes of peripheral muscle weakness and diminished Wmax in the patients with mild CF, our study provides strong evidence that diminished exercise capacity in patients with CF who are in good clinical condition cannot be attributed to the effects of diminished nutritional status.

In CF patients with moderately severe clinical symptoms, we found that muscle mass was clearly diminished, as evidenced by decreases in upper-arm muscle area and FFM of 22 and 17%, respectively (Table 1). Work capacity is proportionally lower within this patient group owing to their decreased FFM, and Wmax is disproportionately decreased when muscle weakness is present (Figures 1 and 2). Other factors could also have contributed to these patients' diminished exercise tolerance and O₂max. Several patients in this group reported low levels of habitual physical activity. The CF patients with moderately severe symptoms had significant airway obstruction. Moreover, the significantly larger changes of ventilation over carbon dioxide production during submaximal exercise, and the lower breathing reserve, indicate ventilation/perfusion inequality and an impaired ventilatory pump function in this patient group. This may have blunted the increase in oxygen cost of work, and may have contributed to these patients' lower O₂max. In our study, FEV₁ was significantly associated with O₂max in patients with moderate CF, and interacted with the effects of nutritional status and total muscle force on oxygen capacity in this group (but not in patients with mild CF). De Jong and coworkers reported a 76% explained variance between cycling Wmax and FEV₁ as well as subjective dyspnea scores, in patients with CF (25). In contrast, Moorcroft and coworkers found that O₂max in adults with CF was maintained in correlation with increases in body mass, irrespective of changes in FEV₁ (26). Body composition was not reported by De Jong and colleagues, and differences in nutritional status in association with the severity of pulmonary disease may explain the differences between the results of Moorcroft and coworkers and those of our study.

We conclude that in children with CF, peripheral muscle function is diminished and that this is associated with an impaired maximal workload, even in the absence of decreased pulmonary function or nutritional status. Oxygen cost of work was increased in our patients with CF. These findings suggest that diminished physical performance in CF is at least partly due to a pathophysiologic factor in skeletal muscle that cannot readily be attributed to nutritional factors. In patients with CF with more severe clinical symptoms, reduced FFM is associated with additional loss of maximal muscle force and work capacity. In addition to pursuing normal growth and body composition, pursuing normal muscle strength and exercise tolerance becomes an important goal for all children with CF.