INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Dyspnea, particularly during physical activity, is one of the predominant complaints of patients with chronic obstructive pulmonary disease (COPD). The severity of dyspnea in COPD patients and its correlation with various pulmonary physiologic tests have been studied: in some reports no correlation (1, 2) or a weak correlation (3) was observed; in others, significant correlation was reported (6, 7). It has also been shown that various clinical techniques for rating dyspnea, including the modified Medical Research Council (MRC) Scale, the Baseline Dyspnea Index (BDI), the Oxygen Cost Diagram (OCD), and the Health-related Quality of Life (HRQL) Questionnaires are all significantly interrelated and virtually identical in evaluating dyspnea in patients with COPD (4, 8). The severity of dyspnea measured by these clinical techniques correlates significantly with FVC and FEV₁ (4, 8).

A substantial number of patients with COPD lose weight and become malnourished (9). Since malnutrition alone, in the absence of COPD, is associated with significant impairment of respiratory muscle strength and endurance (13), its presence in COPD may aggravate the already existing respiratory muscle dysfunction that is caused by chronic air-flow limitation and hyperinflation. The severity of the sensation of dyspnea in patients with COPD, therefore, may be affected by weight loss.

In most studies dealing with the severity of dyspnea in patients with COPD, the state of nutrition and body weight, as possible factors in the genesis of dyspnea, were not taken into account. Efthimiou and associates (14) showed that mean breathlessness and oxygen cost scores at baseline were similar in underweight and normal weight patients with COPD. With 3 mo of dietary supplementation and weight gain, the severity of dyspnea improved significantly in the underweight group. In addition, maximum inspiratory pressure was significantly lower in the poorly nourished patients than in the well-nourished subjects. Both inspiratory and expiratory pressures increased with nutritional support. Gray-Donald and associates (15) assessed the severity of dyspnea in a selective group of patients with severe COPD (mean predicted FEV₁ of 30 ± 11%) and different percentage ideal body weights. Despite small differences in respiratory muscle function and maximal exercise performance, the underweight patients did not fare well in terms of the sensation of dyspnea.

The goals of the present study were (1) to assess the severity of dyspnea in well defined groups of stable, normal weight, and underweight patients with COPD with various degrees of airway obstruction; and (2) to determine whether pulmonary function tests, gas exchange parameters, and respiratory muscle strength differed in the two groups and correlated with the severity of dyspnea.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design and Population

During a 12-mo period, all consecutive male patients referred to the Pulmonary Function Test Laboratory at the Cincinnati Veterans Affairs Medical Center with a clinical diagnosis of COPD who met the following inclusion criteria were studied: (1) an FEV₁ of less than 75% pred, (2) an FEV₁/FVC ratio of less than 75%, and (3) a body mass index (BMI) of less than 28 kg/m². The latter criterion was used to exclude obese subjects from the study. Patients with mixed obstructive and restrictive ventilatory defects, acute COPD exacerbation for any reason, other illnesses known to affect nutritional status or body weight (e.g., malignancies, malabsorption, endocrine disorders, or chronic renal failure), significant cardiac diseases, neuromuscular disorders, and pre- or postoperative states were also excluded. We, therefore, selected for this study a group of patients with COPD who, to the best of our knowledge, were free of coexisting conditions that could have influenced their sensation of dyspnea and their respiratory muscle function. History of oral prednisone treatment during the 2 mo prior to this study was obtained from patient records.

On the basis of BMI measurements, patients with COPD were divided into two groups: normal weight (BMI of 21-28 kg/m²) and underweight (BMI < 21 kg/m²). In all patients, the severity of dyspnea was assessed according to the modified MRC Scale (7) shown in Table 1. The study was approved by our Institutional Review Board.

Measurements

Body weight and height were measured with subjects wearing indoor clothing and shoes with 2.5-cm heels, and BMI was calculated as weight divided by height squared (kg/m²).

Flow rates, lung volumes, and single-breath carbon monoxide diffusing capacity (DCO) were determined with automated equipment (model GS/plus; Warren E. Collins, Braintree, MA). Forced inspiratory and expiratory maneuvers were performed three times and the best values obtained from the maximum inspiratory and expiratory flow-volume curves were used for comparison. FRC was measured by nitrogen washout technique and residual volume (RV) was obtained as FRC minus expiratory reserve volume (ERV). Total lung capacity (TLC) was calculated as RV plus vital capacity (VC). DCO was performed in duplicate and the higher value was reported. Recommendations for standardized procedures for various lung function test measurements were followed (16, 17).

Maximum voluntary ventilation (MVV) was determined after instructing the subject to breathe as fast and as deep as possible for a period of 12 s. If the frequency of breathing was less than 60 min¹ and tidal volume was less than 40% of VC on the first manuever, the patient was instructed to breathe faster and deeper on the next maneuver; the greatest of the two efforts was used for comparison (18). Reference predicted values from the following reports were used to assess various parameters: Knudson and associates (19) for spirometry, Goldman and Becklake (20) for lung volumes, Baldwin and associates (21) for MVV, and Gaensler and Smith (22) for DCO. Maximum static inspiratory (PI_max) and expiratory (PE_max) mouth pressures were measured at RV and TLC, respectively, using the method of Black and Hyatt (23). Three to five measurements were recorded; the goal was to obtain two high scores within 10% of each other. An index of respiratory muscle strength (RMS) was calculated from (PI_max + PE_max)/ 2 according to Aldrich and coworkers (24).

Arterial blood samples were drawn from the radial artery with the patient in a sitting position while breathing room air. Arterial blood gas analysis was performed with an ABG-520 (Radiometer American, Westlake, OH).

Statistical Analysis

For each parameter measured or calculated, values for individual patients were averaged per group, and the SEM was calculated. Differences between groups were tested by an unpaired t test; a p value of less than 0.05 was considered significant. Pearson product-moment correlation coefficients of various pulmonary function tests on the dyspnea scale, as the dependent variable, were compared. We also applied stepwise multiple regression analysis, using the dyspnea scale as the dependent variable and all other variables as independent candidate predictors.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ninety stable patients with COPD met the criteria for inclusion in this study. In 57 (63.3%) patients the BMI was within the normal range (normal weight group) and in 33 (36.7%) the BMI was low (underweight group). As shown in Table 2, age and height were similar in the two groups. Only five (8.7%) normal weight and two (6.1%) underweight patients were taking small doses of prednisone during the 2 mo before this study. The severity of dyspnea was significantly greater in the underweight (3.1 ± 0.9) compared with the normal weight (2.5 ± 1.2) patients with COPD (p = 0.0358). The frequency distribution histogram of the MRC dyspnea scale in both groups is represented in Figure 1. None of the underweight patients was free of dyspnea, while 7.0% of the normal weight subjects were not troubled by breathlessness except with strenuous exercise (scale 0). A larger percentage of normal weight patients (43.8 versus 27.2%) complained of slight to moderate dyspnea (scale 1-2), whereas a larger percentage of underweight subjects (63.6 versus 40.3%) had moderately severe to severe dyspnea (scale 3-4). The distribution of very severe dyspnea (scale 5) was similar in the two groups.

View larger version (12K): [in this window] [in a new window]   Figure 1.   Frequency distribution histogram of the severity of dyspnea in normal weight and underweight patients with chronic obstructive pulmonary disease, determined according to the modified Medical Research Council (MRC) Dyspnea Scale.

Mean values of pulmonary function test parameters were similar in the normal weight and underweight groups (Table 3).

The mean values of Pa_O2, Pa_CO2, and pH were similar in the underweight and normal weight patients with COPD. Diffusing capacity was reduced in both groups, but significantly more so in the underweight patients (Table 4).

In Table 5, mean values of maximum respiratory pressures are listed. In both groups of patients with COPD, PI_max, PE_max, and RMS were lower than the predicted normal range (23). Furthermore, PI_max, both in absolute values and as percent predicted, was significantly more reduced in the underweight compared with normal weight patients. PE_max was similar in the two groups. The calculated mean RMS was also significantly lower in the underweight group, but the differences in percent predicted values between the two groups reached only borderline significance.

As a dependent variable, the dyspnea scale correlated significantly with various parameters (Table 6); the strongest correlation was with %DCO (r = 0.512, p < 0.001). In the stepwise multiple regression model with dyspnea scale as the outcome, %DCO and %MVV combined were the strongest predictors of the severity of dyspnea (R² = 0.30, p = 0.001).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In 1964, Mitchell and Filley (25) divided patients with COPD into three clinical groups: tussive, dyspneic, and asthmatic. The "dyspneic" patients were characterized by significant weight loss and progressive decrease in ventilatory capacity. In 1968 Filley and associates (26), on the basis of these and other observations, designated the clinical types of pink puffer (emphysematous) and blue bloater (chronic bronchitic). Weight loss was one of the main characteristics of the emphysematous type. Since then, the relationship between emphysema and weight loss, both in experimental animal studies (27) and in humans (9), has been well described. On the basis of clinical studies, including the present article, the physiological profile of underweight patients with COPD consists invariably of more severe reduction in DCO and variable changes in respiratory muscle strength (14, 30); spirometric parameters, lung volumes, and resting arterial blood gases have persistently been shown to be similar in underweight and normal weight patients with COPD. Indeed, among all measures tested in patients with COPD, BMI correlates best with DCO. In a previous study of 123 patients with stable COPD (12), we showed that BMI correlated most strongly with DCO (r = 0.53, p < 0.0001). In the present study of a similar group of patients with COPD, this correlation is again observed (r = 0.51, p < 0.001). As shown in Table 4, mean DCO values are lower than predicted normal values in both COPD groups, but significantly more so in the underweight patients. None of the 33 underweight patients with COPD had a normal DCO, while in 16 of 57 (28%) normal weight patients with COPD the DCO was within the normal range.

Since both malnutrition (13, 33, 34) and chronic air-flow limitation (32, 35, 36) affect respiratory muscle strength and function, it is expected that the combination of the two should lead to a more profound reduction in respiratory muscle strength. Studies of respiratory muscle function in normal weight and underweight patients with COPD, however, have yielded variable results. The present study reveals that respiratory muscle strength, as determined by PI_max, PE_max, and RMS, is lower than predicted normal values in all patients with COPD, PI_max and RMS being significantly lower in the underweight than in the normal weight group (Table 5). Mean PE_max values are similar in the two groups.

In a comprehensive analysis of various pulmonary physiologic factors that affect maximum respiratory pressures, Ringqvist reported that the RV/TLC ratio had the highest negative correlation with both PI_max and PE_max (37). To determine whether the lower PI_max in our underweight group was due to overinflation or intrinsic muscle weakness, we corrected PI_max values for the RV/TLC ratio in all subjects. The corrected PI_max was 117 ± 40% in the normal weight group and 106 ± 40% in the underweight group (p = 0.37), suggesting that the lower PI_max in the underweight subjects with COPD was most likely due to muscle weakness rather than mechanical factors. Absence of more severe overinflation in underweight patients shown in this and other reports and the presence of generalized muscle weakness in undernourished subjects (38) confirm our observations.

The "dyspneic" patients with COPD described by Mitchell and Filley and further characterized as "pink puffer" by Filley and associates are distinguished from the "tussive" or "blue bloater" types by the predominance of breathlessness as the cardinal clinic manifestation. Newer scaling techniques to quantify breathlessness, however, have not been employed in well-defined groups of patients with COPD to test the validity of this clinical dictum. In two previous studies the issue of dyspnea in underweight and normal weight patients with COPD was addressed but, as will become apparent, the appropriateness of the underweight groups chosen for comparison in these studies is questionable.

Efthimiou and associates (14) found no differences in dyspnea score, OCD, or 6-min walking distance between malnourished and well-nourished patients with COPD; however, DCO was similar in the two groups. Gray-Donald and associates (15) reported no differences in OCD or quality of life index between malnourished and well-nourished patients with COPD. Again, percent predicted DCO was similar in the two groups. As was discussed earlier, more severely reduced DCO is the most distinctive feature of malnourished patients with COPD. In the present study, DCO was 37% lower in the underweight compared with the normal weight patients with COPD. Lack of difference in DCO between malnourished and well-nourished patients with COPD in the two previous studies, therefore, raises the question of appropriateness of groups compared.

O'Donnell and Webb (39) compared physiologic parameters in severely breathless and mildly breathless patients with COPD. Severely breathless patients had significantly lower DCO than the mildly breathless group. Furthermore, baseline dyspnea index was best correlated with DCO in all patients (r = 0.68, p < 0.001); DCO accounted for 46% of the variance in the dyspnea index. Although mean values of body weight were greater than 100% pred ideal body weight in both groups, the severely breathless patients had significantly lower body weights. It is more interesting that, on the basis of calculations of BMI from mean values given in their article, many of the severely breathless patients were clearly underweight, with BMIs less than 20 kg/m². Therefore, although this study was not concerned with the influence of weight loss on the perception of dyspnea in COPD, it clearly showed that those patients with COPD who suffered from weight loss had lower DCO values and were more dyspneic compared with those who had normal weights.

Subjects participating in our study were a well-defined group of stable patients with COPD and without any other known condition that could have affected their sensation of dyspnea or respiratory muscle function. Except for differences in body weight, no other clinical feature differed in the normal weight and underweight groups. The underweight patients were significantly more dyspneic than the normal weight subjects. Other than the BMI, only DCO, PI_max, and RMS were significantly different between the two groups. However, a variety of physiologic parameters correlated significantly with the severity of dyspnea in all patients; the strongest being DCO.

The relationship between the severity of dyspnea and lung function has been investigated, albeit with variable conclusions: some showing no significant correlation (1, 2), some reporting weak correlation (3), and others observing significant correlation (6, 7). Many factors may contribute to the sensations of dyspnea in COPD, including mechanical properties of the lung, hypoxemia, deconditioning, and coexisting cardiac disease. In the present study, spirometry and lung volumes did not discriminate between the two groups. Furthermore, both groups had similar resting CO₂ set-points and arterial oxygen tension; increased chemorespiratory drive at rest, therefore, was not responsible for the differences in the intensity of dyspnea. Differences in DCO, PI_max, and RMS clearly separated the underweight from the normal weight patients with COPD.

The sensation of breathlessness results from a complex interaction of signals arising from receptors in the upper airways, the lung parenchyma, and the chest wall and from within the central nervous system. A complex interplay of mechanical, physiological, metabolic, neurologic, and psychophysical factors activates these signals (40, 41). Although multiple factors are responsible for differences in the intensity of dyspnea in our patient groups, DCO and respiratory muscle strength are most likely among the contributing factors. How lower DCO translates into a more intense sensation of dyspnea is not clearly known, but similar correlations between the severity of dyspnea and DCO have been reported in interstitial lung disease (42) and in a large general population sample (43). Furthermore, DCO correlates significantly with BMI in patients with COPD and is the strongest indicator of the overall pathophysiologic consequences of weight loss in COPD (12).

In addition, MVV is highly dependent on FEV₁ (18) and on PI_max and RMS (24). On the other hand, PI_max is significantly correlated with the RV/TLC ratio (37) and PE_max (32). It is not, therefore, surprising that, by the elimination of interrelated variables in a stepwise multiple regression model, DCO and MVV combined represent the strongest predictors of the severity of dyspnea in underweight patients with COPD.

In summary, results of the present study provide an objective confirmation for the long-held clinical impression that the patients with COPD who are distinguished as pink puffers (emphysematous) and who are characterized by weight loss and lower DCO are indeed more dyspneic than normal weight patients with COPD who have similar degrees of airway obstruction. Although the origin of dyspnea is multifactorial, reduced DCO and respiratory muscle strength are at least in part responsible for the enhanced sensation of dyspnea in the underweight emphysematous patients.