INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Weight loss is a common feature in patients with chronic obstructive pulmonary disease (COPD) (1). The clinical importance of weight loss, particularly loss of fat-free mass (FFM), has been demonstrated in its adverse effects on physical performance (2) and quality of life (3). Moreover, weight loss and a low body weight are unfavorable prognostic factors in survival, independent of lung function (4).

Weight loss is generally considered as the result of an imbalance between energy intake and expenditure. Although the dietary intake of patients with COPD does not differ from the recommended daily allowances (5), it is insufficient to match energy requirements in weight-losing patients. An increased resting energy expenditure (REE) has been reported in 26% of the patients with COPD (6). Furthermore, an elevated total daily energy expenditure (TDEE), independent of REE, was observed in patients with COPD compared with healthy age-matched subjects (7).

Recent data have shown that a systemic inflammatory response is present in patients with COPD, based on elevated concentrations of acute-phase proteins, tumor necrosis factor (TNF)- receptors, and soluble adhesion molecules in peripheral blood (8, 9). In addition, clear evidence for a relationship between weight loss and plasma TNF- has been shown in COPD (10, 11). Besides the fact that inflammatory cytokines may induce anorexia such as observed in experimental studies (12), enhanced levels of acute-phase proteins have been related to an increased REE in COPD (8).

In several studies an attempt has been made to reverse the negative energy balance in depleted patients with COPD with nutritional repletion therapy (13). A substantial number of patients failed to respond to the nutritional support, i.e., they were not able to gain weight. The underlying causes of nonresponse to nutritional therapy are unknown. Nonresponse may be due to factors such as noncompliance, an inadequate energy intake relative to energy requirements, the inability of the patient to ingest the extra calories, or underlying disease-specific problems leading to inadequate metabolic handling. The clinical relevance of nonresponse to nutritional therapy was emphasized by a recent nutritional intervention study that revealed weight gain as a significant, independent predictor of the mortality rate in patients with COPD (4).

Because of the huge impact of nonresponse on functional status and even on survival in COPD, we set out to unravel the underlying mechanisms, hypothesizing that nonresponse is related to biologic characteristics. This hypothesis is, firstly, based on the results of the previous nutritional intervention study of our group in which patients received 420 kcal/d (50% combined with anabolic steroid treatment) for 8 wk; in this study there was a substantial number of nonresponders, even though the study was performed in a controlled, inpatient setting so that factors such as noncompliance to the nutritional supplementation therapy could largely be excluded. Secondly, since weight loss is related to systemic inflammation in a part of the COPD population, we hypothesized that an elevated systemic inflammatory response might play a role in nonresponse to nutritional intervention therapy.

Therefore the present study aimed to investigate prospectively the nonresponse to 8 wk of oral nutritional therapy with respect to lung function impairment, body composition, energy balance, and systemic inflammatory profile in depleted patients with COPD. The orally administered supplementation therapy consisted of 500 to 750 kcal/d and was implemented as part of an inpatient pulmonary rehabilitation program.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

The study group consisted of 24 depleted patients (16 men; mean age [SD], 63 [8] yr) with severe COPD (FEV₁, 33 [12]% of predicted) consecutively admitted to a pulmonary rehabilitation center (Asthma Center Hornerheide, Horn, The Netherlands). Depletion was defined as a body mass index (BMI; body weight/height²)  23 kg/m² and/or a FFM index (FFMI; FFM/height²)  15 (women)/16 (men) kg/m². Patients were included if they fulfilled the criteria for COPD according to the ATS guidelines (17). Furthermore, the FEV₁ had to be less than 70% of the reference value and the increase in FEV₁, after inhalation of a ₂-agonist, less than 10% of the reference value. Patients with concomitant confounding diseases such as malignant disorders, gastrointestinal abnormalities, recent surgery, or severe endocrine disorders were excluded. Only patients in a clinically stable condition (not suffering from a recent respiratory tract infection) and without clinical signs of edema were included in the study. The study was approved by the medical ethical committee of the University Hospital Maastricht, and all subjects gave their informed consent.

Body Composition

Body height was determined to the nearest 0.5 cm (WM 715; Lameris, Breukelen, The Netherlands) with subjects standing barefoot. Body weight was assessed with a beam scale to the nearest 0.1 kg (SECA, Hamburg, Germany), with subjects standing barefoot and in light clothing. The patients were asked to report the amount of weight loss during the last 3 mo before admission to the rehabilitation center.

To measure total body water (TBW), each patient received in the late evening around 10:00 P.M. a weighted (1 g/L estimated TBW) oral dose of deuterium-labeled water (99.84 atom percent excess) mixed into 70 ml water. For the estimation of extracellular water (ECW) 60 mg sodium bromide/L predicted TBW were added to the deuterium dose. Just before and approximately 10 h later, after complete emptying of the bladder, urine and venous blood samples were obtained. Urine was analyzed for deuterium with an isotope ratio mass spectrometer according to the Maastricht protocol (18). Deuterium dilution space was calculated from the quantity of administered deuterium and the urine deuterium concentrations before and after complete distribution. TBW was calculated from deuterium dilution space by applying a conversion factor of 1.04, which corrects for the exchange of labile hydrogen that occurs in humans during the equilibration period. FFM was calculated assuming a hydration factor of 0.73. FM was calculated by subtracting FFM from body weight. The ECW compartment was analyzed using bromide detection. ECW was estimated by the corrected bromide space, which was calculated from the quantity of administered bromide and the serum bromide concentrations before and after complete distribution. Bromide concentration in serum ultrafiltrate was determined by HPLC according to the anion exchange chromatographic method (19). Intracellular water (ICW) was calculated by subtracting ECW from TBW.

Body weight and composition were assessed before the nutritional intervention and after 8 wk of treatment.

Resting Energy Expenditure

REE was measured in the early morning (8:30 A.M.) by an open-circuit indirect calorimetry system using a ventilated hood (Oxycon Beta; Mijnhardt, Bunnik, The Netherlands) before and after 8 wk of nutritional therapy. The system was calibrated daily at the start of the experiment. The accuracy of the system was regularly assessed using a methanol combustion test. Patients were in a fasting state for at least 10 h and had a period of at least 30 min bed rest prior to the measurements. Patients received their maintenance medication 2 h before measurements started. REE was measured when subjects were comfortably lying on a bed in the supine position. REE was calculated from oxygen consumption and carbon dioxide production using the abbreviated Weir formula (20).

Dietary Intake

Before and in the last week of the nutritional therapy, patients were asked to register their food intake for 4 consecutive days to assess dietary intake. The obtained information was coded for computer nutrient analysis. The nutrient data base was derived from the Dutch food composition tables (21). Caloric intake was calculated by taking the average intake over 4 d. Dietary records were used since this method is more valid than the dietary history method; it gives a good representation of the actual consumption of calories because the food intake is averaged over 4 consecutive days so day-to-day variations are largely discarded.

Lung Function

FEV₁ and inspiratory vital capacity (IVC) were calculated from the flow-volume curve using a spirometer (Masterlab; Jaeger, Würzburg, Germany). Diffusing capacity for carbon monoxide (DL_CO) was determined using the single-breath method (Masterlab; Jaeger). Lung function parameters were expressed as percentages of reference values (22). Blood was drawn from the brachial artery while the patient was breathing room air. Pa_O2 and Pa_CO2 were analyzed on a blood gas analyzer (ABL 330; Radiometer, Copenhagen, Denmark). In addition the number of patients receiving a continuous oxygen supply was assessed.

Lung function and blood gases were measured before and after the nutritional therapy.

Collection and Analysis of Laboratory and Inflammatory Parameters

Before and at 8 wk after nutritional intervention blood was obtained from the patients in the fasting state by venipuncture at 9:00 A.M. and collected in evacuated blood collection tubes (Sherwood Medical, St. Louis, MO) containing EDTA. Plasma was separated from blood cells by centrifuging at 1,000 g for 10 min at 4° C within 2 h after collection. Separated plasma was again centrifuged at 1,000 g for 10 min at 4° C. Plasma samples were stored at 80° C until analysis. Most inflammatory markers were measured using sandwich ELISA (enzyme-linked immunosorbent assay) as described previously (23). Soluble tumor necrosis factor receptor 55 (sTNF-R55) and sTNF-R75 were measured using the monoclonal antibodies MR1-1 and MR2-2 for coating on immunoassay plates (Nunc-Immuno Plate Maxisorp; Nunc, Roskilde, Denmark). The standards used were recombinant human sTNF-R55 and sTNF-R75. Specific biotin-labeled polyclonal rabbit antihuman sTNF-R IgG was used as detector reagens followed by streptavidin peroxidase conjugate (Dako, Glostrup, Denmark). Photospectrometry at 450 nm was performed using a micro-ELISA autoreader. The detection limit of both assays was 100 pg/ml. Polyclonal rabbit anti-rh lipopolysaccharide (LPS) binding protein (LBP) IgG was used as coating for the LBP ELISA and biotin-labeled polyclonal rabbit anti-rh LBP IgG was used for detection of LBP. The standard used was rh LBP. Washing and dilution was performed in buffer containing 40 mM MgCl to prevent disturbance by LPS of LBP recovery in the ELISA. The detection limit of the assay was 200 pg/ml. For detection of the adhesion molecules soluble E-selectin and soluble intercellular adhesion molecule (sICAM), the monoclonal mouse antibodies ENA-1 and ENA-2 were used, respectively. Thereafter, detection of the adhesion molecules took place by incubation with rabbit antimouse biotin.

In addition, serum concentrations of glucose, albumin, creatinine, and C-reactive protein (CRP) were determined by spectrophotometric analysis (Cobas Mira; Hoffmann-La Roche, Basel, Switzerland).

Nutritional Intervention

After baseline screening, the patients were treated for 8 wk with nutritional therapy consisting of three supplements per day containing a total of 500 to 750 kcal. Patients were allowed to choose from an assortment of products with variable consistency and flavor (Nutridrink, Fortimel, Ensini, Fortipudding). The supplementation consisted of 61 energy percent of carbohydrates, 19 energy percent of fat, and 20 energy percent of protein. The total protein content of the supplements together with that of the consumed regular meals was more than enough to ensure the amount of protein needed for the optimal protein synthesis that is recommended for repletion of malnourished subjects (1.5 to 1.7 g protein/kg body weight/d) (24). The supplements were labeled with the name of each individual patient and handed out thrice daily in order to have a stringent control over their intake. Patients were, in addition to taking the supplements, encouraged to continue the consumption of their regular meal portions as well.

The nutritional therapy was embedded in an 8-wk, standardized, inpatient rehabilitation program consisting of general physical training, with particular attention to exercise in relation to daily activities, cycle ergometry, treadmill walking, swimming, sports, and games. In addition, an educational program on the disease and medication use was given. When appropriate, psychologic or ergotherapeutic treatment was implemented. The metabolic cost of the exercise program was kept at the same level for all patients by adjusting the exercise activities to the physical performance of each individual patient. For instance, the bicycle training was set at cycling twice daily for 20 min at 50 to 70% of peak work load reached in a symptom-limited incremental bicycle ergometry test, during which work load was increased by 10 W every minute. The test was performed before and after 4 wk and 8 wk, so the training work load had in general to be adjusted halfway though the rehabilitation period. Because the exercise program was offered in an inpatient setting, there was complete control over the compliance of the patients to the different aspects of the program. Furthermore, since the patients in general trained together, they exerted an additional control on the compliance of each other.

Statistics

A computer simulation model on body weight and energy balance, taking into account the patient's age, sex, body composition, and REE, was used for the estimation of the weight response after nutritional therapy (25). On the basis of this model, a mean gain in body weight of  5% was expected. Patients were divided into the following subgroups: (1) no weight gain, defined as a weight gain less than 2% of baseline body weight after 8 wk of intervention (Group 1; n = 5, 4 men); (2) expected weight gain, defined as a weight gain of 5% or more (Group 3; n = 10, 4 men); and (3) medium weight gain defined as a weight gain between 2 and 5% (Group 2; n = 9, 6 men).

All parameters (baseline values and changes after the nutritional intervention) were, firstly, checked on normality of distribution. When variables were normally distributed and equal variances could be assumed, Student's t test for independent samples (in this case one-way analysis of variance [ANOVA] because three groups had to be compared) was applied to compare the differences between the three groups. The ANOVA test was followed by post-hoc analysis to test the least significant differences between the groups by two-sample t tests. When variables were not normally distributed, nonparametric analysis (Mann-Whitney U test), was performed. Baseline differences between the three groups were given, when appropriate, after correcting for sex and age by analysis of covariance. Differences in dichotomous variables such as sex were statistically compared using the chi-square test. Differences within the three groups before and after nutritional therapy were analyzed using Student's paired t test. After calculation of Pearson's product moment correlation coefficients, a linear model was fitted to the data to enable the variables that contributed to the variation in weight change after nutritional therapy using multiple regression analysis. Significance was determined at the level of 5%. Data are expressed as mean (SD). Data were analyzed according to the guidelines of Altman and colleagues (26), using SPSS/PC+ (Statistical Package for the Social Sciences, Version 6.0 for Windows; SPSS Inc., Chicago, IL).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline Characteristics

The baseline characteristics of the three groups with respect to body composition, energy balance, and lung function are shown in Table 1. The extent of weight loss in the last 3 mo was similar for Groups 1 and 2; however, Group 3 exhibited significantly more weight loss than did Group 2. Although the three groups were not different in body weight, body composition, or REE, baseline daily dietary intake was, also after adjustment for REE, significantly lower in Group 1 than in Group 3. This was predominantly associated with significant differences in total protein and fat intake (data not shown). No differences were seen in baseline lung function or arterial blood gases. The number of patients receiving continuous supplemental oxygen was, however, significantly greater in Group 1 than in Group 3 (2 of 5 versus 0 of 10, p < 0.05). In Group 2, two of nine patients were receiving continuous supplemental oxygen.

The baseline laboratory and inflammatory profile of the patients is shown in Table 2. Fasting glucose was significantly greater in Group 1 than in Groups 2 and 3. The concentration of the acute-phase protein LBP was significantly greater in Group 1 than in Groups 2 and 3. Groups 1 and 2 exhibited significantly greater concentrations of sTNF-R55 and sTNF-R75 than did Group 3. The serum concentrations of creatinine, a relevant renal function parameter, were within the normal range for all subjects, indicating a normal kidney function and thus a normal clearance rate of the TNF-receptors. Therefore, the differences in plasma concentrations of the TNF-receptors between the groups could not be attributed to abnormal clearance rates but to differences in TNF-receptor production.

In order to determine whether the tendency towards a greater ECW/ICW ratio as seen in the nonresponders was related to tissue depletion, we correlated baseline ECW/ICW ratio with baseline FFMI. Indeed, a significant, inverse correlation coefficient was found (r = 0.54, p = 0.006). The proposed link between a systemic inflammatory response and anorexia was investigated by correlation analysis of the inflammatory parameters at baseline with baseline dietary intake. The concentration sICAM significantly, inversely correlated with dietary intake (r = 0.50, p = 0.016). Scatter plots of the respective correlation coefficients are shown in Figures 1 and 2.

View larger version (10K): [in this window] [in a new window]   Figure 1.   Scatter plot of baseline fat-free mass index (fat-free mass/ height2) and extracellular water/intracellular water ratio. Correlation coefficient: 0.54, p = 0.006.

View larger version (9K): [in this window] [in a new window]   Figure 2.   Scatter plot of baseline soluble ICAM and dietary intake. Correlation coefficient: 0.50, p = 0.016.

In order to investigate whether nonresponse was related to the medication that the patients used, the maintenance medication use of the patients was determined: 96% of the patients were receiving ₂-sympathicomimetics by inhalation, 67% oral theophyllines, 83% ipratropium bromide by inhalation, 92% inhalation corticosteroids, and 54% oral corticosteroids. No significant differences in drug use were seen between the three groups. Furthermore, no relationship was found between the daily dose of oral corticosteroidsmean (SD) 7.7 (3.0) mg prednisone/dayand weight change after nutritional therapy.

Response to Nutritional Therapy

The changes in body composition, energy balance, and laboratory and inflammatory mediators after 8 wk of nutritional intervention are shown in Table 3. Body weight increased only in Groups 2 and 3. The 2-wk body weight course, expressed as percentage of baseline body weight for the three groups, is shown in Figure 3. The changes in body weight were reflected in significant increases in FFM in Groups 2 and 3 after 8 wk, but not in Group 1. Although all groups showed adjustment in the intake of their regular meals, Groups 2 and 3 significantly elevated their total dietary intake, whereas the net rise in dietary intake in Group 1 did not reach significance despite a significant rise in carbohydrate intake (48 [21] g/d, p < 0.05). Only in Group 3 did REE increase significantly, but this increase was eliminated after correction for the rise in FFM. In Group 1, a significant rise in serum albumin was seen after 8 wk. The ratio ECW/ICW, serum glucose, and inflammatory parameters remained unchanged in all groups after the nutritional intervention.

View larger version (11K): [in this window] [in a new window]   Figure 3.   Body weight course per 2 wk expressed as percentage of baseline body weight. Circles: Group 1 (weight gain < 2%); squares: Group 2 (weight gain 2 to 5%); triangles: Group 3 (weight gain  5%). *p < 0.05, **p < 0.01, ***p < 0.001 versus baseline. Data are given as means (SEM).

To investigate which factors significantly contributed to the variation in body weight change after nutritional therapy, a stepwise multiple linear regression analysis was performed after calculation of Pearson's product moment correlation coefficients. Age and baseline dietary intake/REE ratio, sTNF-R55, and ECW/ICW ratio were selected as independent, significant parameters contributing to a total explained variation of 78% in weight change after nutritional intervention (Table 4). The variables weight loss in the last 3 mo, serum glucose, LBP, and sTNF-R75 were excluded from the model. Although sTNF-R75 also significantly correlated with weight change, regression analysis without sTNF-R55 (because the concentrations of sTNF-R55 and sTNF-R75 were significantly correlated; r = 0.90, p < 0.001) did not, however, reveal sTNF-R75 as a significant parameter.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The etiology of nonresponse to nutritional therapy in patients with COPD is unknown. The present study has provided a unique opportunity to elucidate the biologic factors that might contribute to nonresponse because it was performed in a controlled, inpatient setting. Therefore, possible other explanations such as noncompliance could largely be excluded. The results of this study indicate that nonresponse is related to a variety of independent, but perhaps also partly interrelated, factors. From multiple linear regression analysis, age, baseline dietary intake/REE ratio, sTNF-R55, and ECW/ICW ratio were selected as independent, significant contributors to the variation in body weight change. In addition, nonresponders included a greater number of patients receiving continuous oxygen supply and expressed greater baseline concentrations of glucose, LBP, and sTNF-R75. It must be stressed that the results of this study are preliminary because of the small study groups. Nevertheless, they reveal new possible insights into the etiology of nonresponse to nutritional therapy in patients with COPD.

Age played for a significant part in the variation in weight change after nutritional therapy and was significantly higher in the nonresponders. Ageing is known to be related to decreases in REE and in energy expenditure for activities in healthy subjects, resulting in a lower dietary intake. The lower energy expenditure in the elderly is predominantly related to the decline in FFM caused by reduced physical activity (27), possibly accompanied by an alteration in tissue metabolism and/or an adaptation in REE because of a lower daily caloric intake (28). However, the lower dietary intake in the nonresponders in the present study was independent of age, and no differences in REE or in the amount of metabolically active tissue (FFM) were seen between the groups. In addition, instead of an adaptation in REE caused by weight loss, an increased REE and TDEE have been reported in patients with COPD (6, 7). Therefore, other factors must be involved in the nonresponse to nutritional therapy.

No differences were seen in baseline body weight or body composition. Despite the comparable REE and the same degree of weight loss, baseline dietary intake was significantly lower in the nonresponders than in the responders, also after adjustment for REE. In animal studies exposure to LPS, TNF-, or interleukin (IL)-1-induced anorexia and weight loss (12). In this study we also showed evidence for a relationship between the relative anorexia seen in the nonresponders and an enhanced systemic inflammatory response. In the absence of adequate characterization of systemic inflammation in COPD, a selection of inflammatory parameters was analyzed for the present study. Although one would expect a relationship between dietary intake and the sTNF-receptors, we could not reveal such correlation coefficients. Instead, a significant, inverse correlation between baseline dietary intake and sICAM was found. Soluble adhesion molecules like sICAM and sE- selectin may reflect systemic inflammatory activity and are reported to be elevated in COPD (9). The fact that sICAM correlated with dietary intake and sE-selectin did not, can possibly be explained by the fact that E-selectin is thought to underlie the initial leukocyte rolling on the vessel wall in areas of inflammation, whereas ICAM is thought to be present in tissues throughout the body and may therefore, more than E-selectin, reflect the presence of a systemic inflammatory response (9).

The rise in dietary intake after 8 wk was significant only in the patients exhibiting a weight gain of more than 2% of baseline body weight, although the mean net increase in intake did not differ between the groups. The power of the study might have been insufficient to reveal a significance in the change in dietary intake in group 1. However, on the basis of the substantial rise in dietary intake of 336 kcal/d in Group 1, which is equivalent to two slices of wheat bread with butter and cheese, and because of the fact that the percentage of patients suffering from recent weight loss did not differ between Group 1 and the well-responding Group 3, a greater gain in body weight would be expected in Group 1. It might easily be concluded that the failure of the patients in Group 1 to gain weight was due to the inability to consume enough calories, i.e., to noncompliance to the therapy. However, because of the inpatient setting in which the study was performed, there was a strict control on the consumption of the supplements. Therefore, the lower increase in dietary intake was likely due to an adaptation in the consumption of the regular meals, on which we had less control despite the fact that the patients were encouraged to consume every meal completely. So the gap between the offered 500 to 750 kcal/d and the rise in dietary intake of 336 kcal/d in Group 1, compared with a gap of ± 250 kcal/d in Groups 2 and 3, was predominantly due to less compliance to the consumption of the regular meals and not to noncompliance to the therapy per se, namely, the nutritional supplementation.

Another explanation for the nonresponse to the nutritional intervention could be a variation in the metabolic load of the exercise program. It must be noted, however, that the exercise program was standardized for all patients; the training activities were adjusted to the physical performance state of each individual patient. For instance, the bicycle training load was set at 50 to 70% of the peak work load reached during an incremental bicycle ergometry test. This test was performed before and after 4 and 8 wk, so the training work load had in general to be adjusted halfway into the rehabilitation period. The individual adjustment of the exercise program shows that all patients were trained at the same "metabolic" intensity level of exercise. Because the exercise program was given in an inpatient setting, there was, furthermore, a thorough control of the compliance of the patients to the different aspects of the exercise program. In addition, since the patients in general trained together, they exerted an additional control on the compliance of each other.

Another possible covariate of the observed differences in weight response between the three groups could be an eventual discrepancy regarding the translation from training load intensity to the actual activity-related energy expenditure for each individual patient. In other words, the energy efficiency might have been different between the patients. Indeed, in a previous study of our group, an increased TDEE was observed in patients with COPD compared with age-matched control subjects, independent of REE, which was probably related to an increased energy expenditure for activities (7). The fact that the patients in Group 1 and 3 were in a state of negative energy balance prior to admission to the pulmonary rehabilitation center, despite a mean dietary intake/REE ratio ranging from 1.3 to 1.8, might be indicative for the presence of an elevated wasting of energy. The mechanisms behind this weight loss, however, were probably different for these two groups. The reason for the recent weight loss of the patients in Group 3 was likely an inadequate caloric intake relative to increased energy requirements that could easily be overcome by offering them nutritional supplementation alone. In contrast, the patients in Group 1 seemed to be compromised with additional problems related to their decreased dietary intake and/ or to inadequate metabolic handling.

Because of the stringent control on the compliance to the nutritional therapy and the standardization of the pulmonary rehabilitation program, other (biologic) factors had to be involved in the etiology of nonresponse to nutritional therapy. In other words, other explanations must be sought for the much lower increase in body weight than would have been expected on the basis of the change in dietary intake in Group 1.

The phenomenon of nonresponse might be related to the effects of the enhanced systemic inflammatory profile on metabolism. Baseline concentrations of the acute-phase protein LBP were elevated, together with a trend towards a higher CRP. The concentrations of sTNF-R55 and sTNF-R75 were significantly greater in the patients with a weight gain less than 5% versus the patients with a weight gain of 5% or more. Furthermore, sTNF-R55 significantly contributed to the variation in weight change after 8 wk, independent of baseline dietary intake. Previous studies have shown evidence for a relationship between weight loss and the presence of a systemic inflammatory response in patients with COPD. Increased serum concentrations of the inflammatory mediator TNF- were found in patients with COPD who had involuntarily lost weight in comparison with weight-stable patients and healthy subjects (10). Furthermore, the LPS-induced TNF- production of monocytes was greater in weight-losing patients with COPD than in patients not suffering from weight loss and to healthy subjects (11). A previous study of our group reported increased concentrations of the acute-phase reactant proteins CRP and LBP in hypermetabolic patients with COPD compared with normometabolic patients. In this hypermetabolic subgroup the patients with an elevated CRP had elevated concentrations of LBP, IL-8, sTNF-R55, and sTNF-R75 and expressed a lower FFM (8).

The presence of an elevated inflammatory state has consequences for the protein need because of an enhanced protein synthesis rate in the liver. Seven days of exposure to LPS, TNF-, or IL-1 in rats produced anorexia and weight loss and accelerated peripheral protein wasting while preserving liver protein content. In contrast, when the animals were only pair-fed or starved, loss of liver proteins and relative preservation of skeletal muscle protein occurred (12). It is likely that the patients in our study group, especially those with an enhanced inflammatory state, exhibited a protein balance in favor of the liver content. Therefore, it is important that protein intake is adequate in order to increase or preserve peripheral muscle tissue. The finding that the concentrations of the acute-phase proteins, adhesion molecules, and inflammatory cytokines did not change after the nutritional intervention, except for the rise in serum albumin in Group 1, emphasizes that the increased protein need will stay present. In addition, the fact that the concentrations of the inflammatory markers did not change provides evidence against the hypothesis that anorexia and cachexia are causes of an elevated systemic inflammatory response rather than the consequences (29).

The variation in baseline ECW/ICW ratio significantly contributed for 11% to the variation in body weight change after nutritional intervention. In line, a tendency towards a greater ECW/ICW ratio was seen in the patients who failed to respond. This shifting in body water compartments can possibly be extrapolated to a decreased cellular hydration state resulting in cellular shrinkage. Cellular shrinkage can be triggered by a variety of factors such as uremia, inflammatory state, and stress conditions, resulting in hormonal alterations. In the absence of differences in uremic levels, the reported enhanced inflammatory state of the nonresponders could have induced an increase in ECW/ICW ratio. Cellular shrinkage in turn acts as a catabolic signal triggering breakdown of protein and glycogen on the one hand and the expression of gluconeogenic enzymes on the other, together resulting in net protein breakdown (30). This proposed catabolic response can explain the inverse relationship between baseline ECW/ICW ratio and FFMI, the latter as a measure of functional tissue depletion, although one must take into account that the increase in ECW was relative to ICW.

Nonresponders exhibited a significantly greater fasting serum concentration of glucose than did responders, possibly reflecting the presence of insulin resistance. The scarce data on insulin and carbohydrate metabolism in patients with COPD are, however, conflicting. Jacobsson and colleagues (31) reported a lower skeletal muscle glycogen concentration, assessed by a biopsy of the quadriceps femoris, in hypoxemic patients with COPD compared with a control group; no evidence for insulin resistance of the peripheral tissues was, however, found in this COPD group. Hjalmarsen and colleagues (32) in contrast reported an altered glucose metabolism in severely hypoxemic patients with COPD.

The number of patients who were receiving continuous oxygen support was significantly higher in the nonresponding group than in the responding group. Chronic or even intermittent hypoxemia is reported to affect intermediary metabolism (33). Hypoxia is also a stringent depressor of protein synthesis (34). Anorexia, together with a decrease in the growth rate of muscle tissue, was found to be another effect of hypoxia, which could not be attenuated by supplementation of extra protein (35). Hypoxia has, furthermore, a potentiating effect on the in vitro production of inflammatory cytokines (IL-1 and TNF-) after endotoxin stimulation that was postulated to be due to a decreased anti-inflammatory prostaglandin PGE₂ synthesis (36, 37).

Further studies are required in order to investigate whether the enhanced glucose levels and the greater use of additional oxygen represent disturbances in intermediary metabolism and hypoxemia, respectively, as well as the possible relation with nonresponse to nutritional therapy. Furthermore, it must be stressed that the patient population studied in this report was too small to rely on for definitive recommendations.

In conclusion, despite the overall positive effect of high caloric nutritional support incorporated into a pulmonary rehabilitation program in depleted patients with COPD, a substantial number of nonresponders was revealed. Nonresponse in COPD can be considered as a multifactorial problem. Besides the presence of an acute-phase response and an elevated systemic inflammatory response, other factors were shown to play a role: ageing, relative anorexia, and shifting in body water compartments. The findings of this study, although preliminary because of the limited size of the study population, indicate the complexity of metabolic disturbances in patients with COPD characterized by a systemic inflammatory response. The results may have important implications for the treatment of pulmonary cachexia, especially concerning the patients not responding to oral nutritional therapy. Further research is needed to investigate which options will be appropriate in the treatment of depleted patients with COPD not responding to high caloric nutritional support.