INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The frequent occurrence of weight loss and subsequent tissue depletion in patients with chronic obstructive pulmonary disease (COPD) is associated with an increased morbidity and even mortality (1). In addition, several reports have provided evidence that weight loss negatively affects the prevalence and outcome of acute disease exacerbations of COPD. The risk of being hospitalized for an acute exacerbation was increased in patients with a low body mass index (BMI; body weight/ height²) (2). A low BMI and weight loss were reported as an unfavorable index of outcome during an exacerbation of COPD, i.e., predicting the need for mechanical ventilation (3). Furthermore, the survival time following a disease exacerbation was found to be independently related to the BMI (4).

In a subgroup of patients, weight loss is suggested to follow a stepwise pattern related to acute disease exacerbations. Previously we reported an impaired energy balance in patients with COPD during the first days of an acute exacerbation of their disease, predominantly due to a severely depressed dietary intake and, to a lesser extent, to an increased resting energy expenditure (REE) (5). Limited data are, however, available on the underlying disturbances in the (dis)regulation of the energy balance during an acute exacerbation of COPD.

Previous experimental and clinical research indicates the involvement of leptin in body weight homeostasis. Leptin is a hormone produced by the adipose tissue and its circulating concentrations are proportional to amount of fat mass (FM). Leptin regulates the energy balance in a feedback mechanism in which the hypothalamus is involved (6). In animals, administration of leptin results in a reduction in food intake (7) as well as in an increase in energy expenditure (8). In stable depleted patients with COPD, dietary intake as well as weight gain after nutritional therapy were inversely related to plasma leptin (9).

The normal leptin feedback mechanism can be disturbed by several factors. In animals, administration of endotoxin, tumor necrosis factor- (TNF-), or interleukin-1 (IL-1), inflammatory cytokines known for their anorectic effects, resulted dose dependently in an up-regulation of leptin mRNA in fat cells and in an increase in circulating leptin concentrations (10, 11). In stable patients with emphysema, leptin was found to be positively related to plasma soluble TNF receptor 55 (9). During an acute exacerbation of COPD, the involvement of the systemic inflammatory response may even be more pronounced than in stable patients (12).

Glucocorticosteroids are also reported to stimulate leptin production. They can act on leptin metabolism in a direct manner (13) or via the induction of insulin resistance, as glucose and insulin are also able to induce leptin expression (14, 15). The influence of glucocorticosteroids on leptin metabolism may be of special interest during acute exacerbations of COPD. Under these conditions, intravenous treatment with high-dose prednisolone is still part of the medical therapy of the patients, despite the moderate effects on clinical outcome (16).

For these reasons we hypothesize that disturbances in leptin metabolism underlie the impaired energy balance during an acute exacerbation of COPD. These disturbances might be induced by the systemic inflammatory response as well as by the glucocorticosteroid treatment. The aim of this study was to investigate the course of the energy balance in relation to leptin and glucose metabolism and the systemic inflammatory response in patients with COPD hospitalized for an acute exacerbation of their lung failure.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

The patient group consisted of subjects consecutively admitted to the hospital suffering from an acute exacerbation of COPD. An exacerbation was defined as a recent increase in dyspnea, cough, and sputum production of sufficient severity to warrant admission to the hospital. The presence of an exacerbation was determined by an independent chest physician. COPD was defined according to the criteria of the American Thoracic Society (ATS) (17). Patients were excluded when they were suffering from concomitant diseases and when their stay in the hospital was less than 7 d. The final study group consisted of 17 patients (11 men). Duration time of admission to the hospital was mean (SD) 13 (6) days. Of patients 14 of 17 were current or ex-smokers. The patients were treated with a standard protocol of medication consisting of nebulized short-acting ₂-sympathicomimetics (salbutamol >20 mg/24 h), inhaled anticholinergics (ipratropium bromide), and intravenously administered theophylline and prednisolone. Prednisolone was given in decreasing doses throughout the exacerbation. The patients received, depending on body weight, 50 or 75 mg/24 h during the first 4 d of the exacerbation. From Day 4 to Day 7, the initial dose was halved. The plasma theophylline concentrations of the patients fell within the therapeutic range (mean [SD] 10.7 [4.5] mg/ml). In case of a bacterial infection which was confirmed by a sputum culture (12 of 17 patients), specific antibiotic therapy was given. Only one patient had fever on the first day after admission to the hospital. During hospitalization, 11 of the 17 patients were on additional oxygen therapy.

The measurements were performed in the early morning (8:30 A.M.), when patients were in the fasting state for at least 10 h. The measurements were done at Days 1, 3, 5, and 7 after hospitalization. The patients received their bronchodilating medication 2 to 3 h before the measurements started. The study was approved by the medical ethics committee of the University Hospital Maastricht and all subjects gave their informed consent.

For reference values of the laboratory parameters, fasting blood was collected from 23 (16 men) healthy, elderly volunteers. They were not exhibiting any acute or chronic disease nor taking medication that could influence energy or glucose metabolism.

Lung Function

Forced expiratory volume in 1 s (FEV₁), inspiratory vital capacity (IVC), and peak expiratory flow (PEF) were calculated from the flow-volume curve using a portable spirometer (Jaeger, Würzburg, Germany). Lung function was expressed as percentage of reference value (18). Blood was drawn from the brachial artery and arterial oxygen tension (Pa_O2) and carbon dioxide tension (Pa_CO2) were analyzed on a blood gas analyzer (Radiometer, ABL 330, Copenhagen, Denmark).

Body Composition

Body height was determined to the nearest 0.5 cm (Lameris, WM 715, Breukelen, The Netherlands) with subjects standing barefoot. Body weight was assessed to the nearest 0.1 kg using a digital weighing chair while subjects wore light clothing and no shoes. Body composition was estimated using single-frequency (50 kHz) bioelectrical impedance analysis (BIA) (Xitron Technologies Inc., San Diego, CA) while subjects were in a supine position. Fat-free mass (FFM) was calculated using the disease-specific equation of Schols and coworkers (19). In the healthy volunteers, FFM was calculated from total body water assessment using the deuterium dilution technique described in the Maastricht protocol (20). Previously it was established that in healthy subjects as well as in patients with COPD, FFM assessed with BIA did not significantly differ from FFM assessed with the deuterium dilution technique (21). FM was calculated by subtracting FFM from body weight.

Energy Balance

Habitual dietary intake was evaluated using the diet history method with cross-check. All interviews were performed by the same trained dietician. On the day of admission to the hospital (Day 0) and during hospitalization, dietary intake was registered by use of dietary records and an automated food distribution system. The patients had the opportunity to choose their menus. The daily 500-ml infusion solution contained 5% glucose, which is equivalent to 100 kcal/d. This amount of energy was incorporated in the calculations of the daily dietary intake in the hospital. The food composition data were coded for computer energy and nutrient analysis. The database was derived from the Dutch food composition tables (22).

REE was measured by an open circuit indirect calorimetry system using a ventilated hood (Oxycon ; Mijnhardt, Bunnik, The Netherlands). The system was calibrated daily at the start of the experiment, while the accuracy was regularly assessed using a methanol combustion test. Patients had a period of at least 30 min bed rest prior to the measurements started. When patients were on additional oxygen during the hospitalization, the oxygen was temporarily withdrawn 30 min before and during the measuring of REE. The patients were comfortably lying on a bed in a supine position allowing them to watch television. REE was calculated from oxygen consumption (O₂) and carbon dioxide production (CO₂) using the abbreviated Weir formula (23). REE was compared with the reference values (24).

Collection and Analysis of Laboratory and Inflammatory Parameters

For the measurement of glucose, blood was collected in an evacuated blood collection tube containing heparin and centrifuged at 4000 rpm for 5 min at 4° C. After separation from the blood cells, trichloride acetic acid was added to the plasma. The suspension was vortexed and stored at 70° C until analysis by enzymatic, fluorimetric spectrophotometry (Cobas Mira; Hoffmann-La Roche, Basel, Switzerland). To assess insulin, blood was collected in an evacuated coagulation tube and centrifuged for 15 min at 3,000 rpm. Serum was collected and stored until assessment using a radioimmunoassay.

An evacuated tube containing EDTA (Sherwood Medical, St. Louis, MO) was used to collect blood for the measurements of the inflammatory parameters and leptin. Plasma was separated from blood cells by centrifugation at 1,000 × g for 10 min within 2 h after collection. Separated plasma was again centrifuged at 1,000 × g for 10 min. Plasma samples were stored at 70° C until analysis. C-reactive protein (CRP), sTNF-R55, sTNF-R75, and interleukin-6 (IL-6) were measured using sandwich enzyme-linked immunosorbent assay (ELISA) as described previously (25). Plasma IL-6 concentrations were measured using the murine anti-human IL-6 antibody 5E1 (developed in our laboratory). Diluted plasma samples (1/1) and the standard titration curve with recombinant IL-6 were added to the immunoassay plates (Nunc-Immuno Plate Maxisorp, Roskilde, Denmark). The amount of IL-6 bound to the wells was quantified by sequential incubation with a polyclonal rabbit anti-IL-6 antibody, followed by adding peroxidase-conjugated goat anti-rabbit IgG and substrate. The lower detection limit of IL-6 was 0.01 ng/ml. IL-6 was not measured in the plasma of the healthy volunteers. Plasma leptin was measured with a double antibody sandwich ELISA (BioVendor Laboratory Medicine Inc., Brno, Czech Republic). The wells were precoated with anti-human leptin antibody. After a thorough wash, anti-human leptin antibody labeled with horseradish peroxidase (HRP) was added to the wells and incubated with the immobilized antibody-leptin complex. Following washing, the remaining HRP-conjugated antibody was allowed to react with the substrate tetramethylbenzidine. The reaction was stopped by adding an acidic solution. Absorbency was measured spectrophotometrically at 450 nm using a micro-ELISA autoreader. The lower and upper limits of detection were respectively 0.2 and 50 ng/ml. The intra- and interassay variation was 6% and 9%, respectively.

Statistics

For comparisons within the individual among Days 1, 3, 5, and 7 of the exacerbation, the paired t test was used. The Bonferroni correction was applied to correct for the multiple comparisons that were performed. Because per time point three comparisons were done, significance was determined at a p value of 0.017 (= 0.050/3). For dietary intake, five comparisons per time point were done, so significance was determined at a p value of 0.010 (= 0.050/5). Differences between healthy volunteers and patients with COPD on Days 1 and 7 of the exacerbation respectively were compared using a Student's t test for independent samples, if the data were normally distributed and equal variances could be assumed. Otherwise, nonparametric analysis (Mann- Whitney U test) was performed. Significance was determined at a p value of 0.05. Plasma leptin was logarithmically transformed by calculating its natural logarithm (LN leptin) for use in correlation and regression analysis. Pearson's product-moment correlation coefficients on the metabolic and inflammatory parameters were calculated on Days 1 and 7, respectively. To adjust for sex, age, and, when appropriate, FM expressed as percentage of body weight (%FM), partial correlation analysis was performed. To investigate which factor(s) could significantly explain the variation in dietary intake/REE ratio on Day 1 and on Day 7 of the exacerbation, linear stepwise regression analysis was done. Data were expressed as mean (SD) in the text and tables and as mean (SEM) in the graphs. Data were analyzed according to the guidelines of Altman and coworkers (26) using SPSS/PC+ (Statistical Package for the Social Sciences, version 7.5 for Windows, SPSS Inc., Chicago, IL).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Group

The characteristics of the patients with COPD on Day 1 after admission to the hospital for an acute exacerbation are shown in Table 1. Although mean BMI was normal, the FFM index (FFMI; FFM/height²) was just above the minimum normal value. Lung function was significantly impaired; on base of the FEV₁, the patients could be considered as having severe COPD. While Pa_O2 was decreased compared with normal, no signs of hypercapnia were seen.

Course of the Parameters during the Exacerbation

Energy balance. Figure 1 shows the habitual dietary intake and the course of dietary intake and REE during the exacerbation. Dietary intake at the day of admission (Day 0) was severely decreased when compared with the habitual intake of the patients, but gradually increased during the course exacerbation. At Day 7 it was even higher than the habitual dietary intake. REE decreased from Day 1 to Day 7, but in absolute terms the drop was not significant. Also when expressed as percentage of predicted, the decrease in REE from Day 1 to Day 7 of the exacerbation was not significant and the patients remained hypermetabolic (115 [11] versus 109 [12] %pred, NS). Mean dietary intake/REE ratio increased from 1.28 (0.57) at Day 1 to 1.65 (0.45) at Day 7 (p = 0.005).

View larger version (16K): [in this window] [in a new window]   Figure 1.   Habitual dietary intake and the course of dietary intake (solid line) and resting energy expenditure (dashed line) during the first 7 d of an acute exacerbation of COPD. Data are expressed as mean (SEM).

Glucose and insulin. Plasma glucose at Day 1 of the exacerbation was significantly higher than in the healthy subjects (p < 0.001). Thereafter, it gradually decreased in such a way that on Day 7 it was lower than on Day 1 and the difference with the healthy subjects group was eliminated (Figure 2). The pattern of serum insulin during the exacerbation was similar to that of plasma glucose; however, no significances in the drop in insulin could be achieved. On Day 1 as well as on Day 7 serum insulin was elevated compared with the healthy subjects (p < 0.001) (Figure 2).

View larger version (12K): [in this window] [in a new window]   Figure 2.   Course of plasma glucose (solid line) and serum insulin (dashed line) during the first 7 d of an acute exacerbation of COPD. Data are expressed as mean (SEM). Horizontal lines: mean plasma glucose (solid line) and mean serum insulin (dashed line) in healthy, elderly subjects.

Inflammatory parameters and leptin. A temporary increased acute phase protein response (APPR) was seen during the exacerbation. At Day 1 of the exacerbation, plasma CRP was elevated compared with the healthy subjects (33.3 [25.9] µg/ml versus 9.0 [16.0] µg/ml, p < 0.001). Already on Day 3, CRP was dropped to 15.3 (17.6) µg/ml (p = 0.010 versus Day 1) and no longer differed from healthy subjects. CRP on Day 7 amounted up to 11.7 (15.9) µg/ml (p = 0.016 versus Day 1).

No differences were seen in the concentrations of sTNF-R55 and R75 between patients with COPD on Days 1 and 7 compared with the healthy subjects. The TNF receptors did not drop during the exacerbation. Instead, sTNF-R55 showed a temporary rise from Day 1 to Day 3 that already was eliminated at Day 5 (Figure 3). IL-6 concentrations were low in the patient group and no changes were seen throughout the exacerbation (Day 1: 0.28 [0.54] versus Day 7: 0.11 [0.11] ng/ml, NS).

View larger version (14K): [in this window] [in a new window]   Figure 3.   Course of plasma-soluble TNF receptor 55 (solid line) and 75 (dashed line) during the first 7 d of an acute exacerbation of COPD. Data are expressed as mean (SEM). Horizontal lines: mean plasma-soluble TNF receptor 55 (solid line) and 75 (dashed line) in healthy, elderly subjects.

Compared with its concentration in the healthy subjects (12.5 [33.8] ng/ml), plasma leptin was significantly higher in the patients with COPD on Day 1 and still on Day 7 of the exacerbation (76.8 [190.5] ng/ml, p = 0.006 and 62.2 [180.8] ng/ml, p = 0.017, respectively, versus healthy subjects). Leptin concentration decreased gradually throughout the exacerbation. After dividing leptin by %FM, the same pattern was seen (Figure 4). Leptin divided by %FM was significantly elevated on Day 1 and nearly significantly increased on Day 7 compared with the healthy subjects (p = 0.008 and p = 0.018, respectively).

View larger version (11K): [in this window] [in a new window]   Figure 4.   Course of plasma leptin divided by fat mass (as percentage of body weight) during the first 7 d of an acute exacerbation of COPD. Data are expressed as mean (SEM). Horizontal line: mean plasma leptin divided by fat mass (as percentage of body weight) in healthy, elderly subjects.

No differences in the course of the energy balance, glucose and insulin concentrations, inflammatory mediators, or leptin during the exacerbation were observed between patients with and without a bacterial infection confirmed by sputum culture.

Correlation Analyses

To further elucidate the proposed regulation of leptin and the energy balance, correlation analysis was performed on %FM, glucose, insulin, LN leptin, sTNF-R55 and R75, and dietary intake/REE ratio on Days 1 and 7 of the exacerbation.

On Day 1 of the exacerbation, no significant correlations could be revealed between any of the parameters. In contrast, on Day 7 of the exacerbation, LN leptin positively correlated with %FM after correcting for the influences of sex and age (Table 2). In addition, LN leptin was found to be positively correlated with sTNF-R55, but not R75, after correction for sex, age, and %FM (Table 2).

Furthermore, a significant correlation coefficient was revealed between plasma glucose and %FM, after correction for sex and age, at Day 7 of the exacerbation of COPD (Table 2). Plasma glucose was also correlated with sTNF-R55 on Day 7 after adjusting for the influences of sex, age, and %FM (Table 2).

On Day 7 of the acute exacerbation of COPD, the dietary intake/REE ratio was found to be inversely correlated with the natural logarithm of plasma leptin on the one hand and with the plasma concentration of the sTNF-receptor 55 on the other after correcting for sex, age, and %FM. The respective scatter plots are shown in Figure 5a and 5b. Since both LN leptin and sTNF-R55 were correlated with the dietary intake/ REE ratio, we performed a stepwise regression analysis with sex, age, %FM, glucose, insulin, LN leptin, sTNF-R55, and sTNF-R75. This elucidated that sTNF-R55 significantly explained 66% of the variation in energy balance on Day 7 of the exacerbation (p = 0.009), while the other parameters were excluded from the model.

View larger version (9K): [in this window] [in a new window]   View larger version (11K): [in this window] [in a new window]   Figure 5.   Scatterplot of the dietary intake/resting energy expenditure ratio against the natural logarithm of plasma leptin (A) and against plasma-soluble TNF-receptor 55 (B) at Day 7 of an acute exacerbation of COPD. Respective correlation coefficients after adjusting for the influences of sex, age, and fat mass (as percentage of body weight): 0.74 (p = 0.037) and 0.93 (p = 0.001).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Until now, little is known about the mechanisms underlying the energy imbalance in patients with COPD. The present study was performed to make a first attempt to unravel the factors contributing to the disturbed energy balance in patients with COPD suffering from an exacerbation. Therefore, the course of the energy balance in relation to leptin and glucose metabolism and the systemic inflammatory response were investigated during the first 7 d of hospitalization for an acute exacerbation of COPD.

According to the study by Vermeeren and coworkers (5), the patients exhibited a severely depressed dietary intake compared with habitual on the day of admission to the hospital. Intake gradually restored during the course of the exacerbation. The elevated REE decreased slightly, but remained above normal. Nevertheless, the dietary intake/REE ratio increased throughout the exacerbation from 1.3 on Day 1 to 1.7 on Day 7. The fact that the latter value was equal to the total daily energy expenditure corrected for REE in stable COPD (27) indicates that the energy balance was restored on Day 7.

Leptin divided by %FM was relatively high during the first days of the exacerbation, when intake was low, and decreased thereafter, together with the gradual rise in intake. On Day 7 of the exacerbation, the dietary intake/REE ratio was inversely correlated with LN leptin. This finding was in line with the suggested role of leptin in the regulation of the energy balance (6) and with the involvement of leptin in anorexia in stable COPD (9). The fact that no relation was seen between leptin and intake at Day 1 may indicate that the normal leptin feedback mechanism regulating food intake was completely disturbed during the first days of the exacerbation.

In addition to the inverse relationship between LN leptin and intake, we also revealed a significant, inverse correlation coefficient between dietary intake and sTNF-R55 on Day 7, the latter as a measure of the systemic inflammatory response. This finding provided further evidence for the involvement of inflammation in anorexia and was in line with our recent report in which stable, depleted patients with COPD with a low caloric intake exhibited higher plasma concentrations of the sTNF receptors (28). Animal studies have revealed that administration of inflammatory mediators like endotoxin, TNF-, or IL-1 indeed induced anorexia and cachexia in animals (29). Furthermore, evidence exists for the involvement of the systemic inflammatory response in weight loss and hypermetabolism in COPD (30). Because both LN leptin and sTNF-R55 were inversely correlated with the dietary intake/ REE ratio, a stepwise regression analysis was performed. The finding that sTNF-R55 significantly explained 66% of the variation in energy balance on Day 7 of the exacerbation, while LN leptin was excluded, may indicate that the influence of leptin on energy balance was under control of the systemic inflammatory response.

In accordance with the normal mechanism of leptin regulation (6), LN leptin was found to be correlated with %FM on Day 7 of the exacerbation. The fact that no such relation was seen at Day 1 of the exacerbation probably indicated a temporary dissociation, related to the exacerbation, of the normal feedback regulation of leptin by FM. Mean leptin corrected for %FM significantly decreased throughout the exacerbation. However, its plasma concentration was not only at Day 1, but until Day 7, elevated compared with the healthy, elderly subjects.

The elevated leptin concentrations during the exacerbation likely represented an up-regulation of leptin mRNA resulting in an enhanced leptin production, which could have been induced by several factors. Independently of %FM, a positive correlation between LN leptin and the plasma concentration of sTNF-R55 was seen in patients with COPD on Day 7 of the exacerbation. This finding gives further evidence for an inflammation-related disturbance in leptin metabolism in COPD. Previously we also reported in clinically stable patients with emphysema, in a cross-sectional study design, a positive correlation between leptin and sTNF-R55 (9). In patients with solid tumors, infusion of TNF- resulted in a transient increase in serum leptin concentration (33). Administration of recombinant human IL-1 to cancer patients also increased leptin concentration dose dependently, accompanied by a decrease in appetite in the majority of the patients (34).

Controversial results on the relationship between IL-6 and leptin in acute illness are reported in the literature. We revealed very low IL-6 concentrations in the present study, unrelated with leptin and not changing during the exacerbation. Bornstein and coworkers revealed that IL-6 and leptin concentrations were dramatically elevated in critically ill patients suffering from acute sepsis compared with healthy subjects (35). In the study of Torpy and coworkers, elevated leptin and IL-6 concentrations were also seen in critically ill patients suffering from sepsis. However, leptin inversely correlated with IL-6, suggesting that the leptin hypersecretion was not under the influence of IL-6 (36). This was confirmed by a study of Faggioni and coworkers (37), in which turpentine was able to increase leptin in IL-6-nondeficient mice as well as in IL-6-deficient mice. Oppositely, leptin could not be induced in IL-1-deficient mice, but only in IL-1-nondeficient mice. In two other investigations from our laboratory, using the same measurement techniques and plasma handling as in our study, elevated IL-6 concentrations were reported in acute disease (sepsis and hemorrhagic shock syndrome after surgery, respectively) (38, 39). So the low IL-6 concentrations found in the present study could not be attributed to inappropriate measurement techniques, but probably to the disease severity; an acute exacerbation of COPD was likely not able to induce prompt increases in IL-6, opposite to sepsis and shock. Taking the available studies in the literature together with our own finding of very low IL-6 concentrations, unrelated to leptin, in patients suffering from an acute exacerbation of COPD, IL-6 does not seem to exert a great role in the enhancement of leptin secretion, opposite to IL-1.

A second reason for the high leptin concentrations might have been the high dose glucocorticosteroid treatment. The course of plasma leptin seemed to mimic the scheme of prednisolone administration. The first days, during which the patients received the highest dose of prednisolone, leptin was high. Thereafter, lower leptin concentrations were seen, together with a tapering off of the glucocorticosteroid treatment. Glucocorticosteroids in high doses are indeed reported to stimulate leptin production. Two days of oral dexamethasone administration to healthy volunteers increased serum leptin as well as the presence of leptin mRNA in abdominal and gluteal adipose tissues (13).

Glucocorticosteroids in therapeutic doses have a stimulating effect on leptin concentrations via the induction of insulin resistance (40), as glucose and insulin are also able to induce leptin expression. In lean mice, glucose injection resulted in an increase in plasma glucose and in an up-regulation of leptin mRNA, independently of age, nutritional state, and plasma insulin (14). Also injection of insulin in fasted rats induced leptin gene expression (41). At the level of adipose tissue and muscle, glucocorticosteroids antagonize insulin-mediated uptake and utilization of glucose and also exert a permissive effect on lipolysis by promoting the activation of the cAMP-dependent hormone-sensitive lipase, a key enzyme inhibited by insulin (42). Insulin resistance indeed seemed to be present during the exacerbation, predominantly based on the high insulin concentrations and to a lesser extent based on the temporary increased glucose concentrations. It is therefore plausible that the kinetics of glucose and insulin during the exacerbation were (partly) related to the tapering off of the systemic prednisolone treatment.

Other medications that were prescribed to our patients during an acute exacerbation of COPD might also have affected leptin metabolism. Little is reported about the effect of theophyllines on leptin; only one study suggests that leptin release is attenuated by theophylline-like substances (43). More is known about the effects of adrenergic stimulation on leptin metabolism. However, no studies have been performed with specific ₂-adrenergic stimulants such as salbutamol, but only with nonselective -agonists. Intravenous infusion of isoprenaline (a _1,2-agonist) in young, lean, healthy volunteers resulted in a maximal suppression of plasma leptin of 20% of baseline values after 2 h. In the recovery period of 1 h, leptin concentrations rapidly returned to baseline values (44). It must be noted that in the present study, the patients received salbutamol by nebulizer and not intravenously. Therefore probably relatively little active substance might have reached the systemic circulation. With respect to the very high leptin concentrations, it is likely that other factors, such as an inflammatory response and/or the high dose oral glucocorticosteroid treatment, overruled the influence of treatment with ₂-sympathicomimetics and/or theophyllines on leptin metabolism.

Besides the influence of glucocorticosteroids on glucose metabolism, TNF- also is known for its role in insulin resistance (45). The concentration of glucose at Day 1 of the exacerbation was significantly higher than in the healthy subjects and gradually decreased thereafter. In contrast, the elevated insulin concentration did not significantly decrease. On Day 7 of the exacerbation, sTNF-R55 was positively correlated with plasma glucose, independently of the amount of FM. TNF-, which is produced in fat and muscle tissue, is known as a potent regulator of glucose metabolism, decreasing the expression of the glucose transporter Glut4. Expression of TNF- in the m. vastus lateralis was found to be fourfold higher in patients with insulin resistance or non-insulin-dependent diabetes mellitus (NIDDM) than in insulin-sensitive subjects (46). The correlation between glucose and sTNF-R55 in patients with COPD on Day 7 might have been a reflection of disease-specific metabolic disturbances related to the systemic inflammatory response.

During the first day of the exacerbation an increased APPR was seen: CRP, which is commonly used in daily clinical practice as a marker of acute inflammation, was elevated on Day 1 of the exacerbation, but rapidly dropped thereafter. This was confirmed in another recent study in which CRP, irrespective of the presence of a proven bacterial or viral infection, could be identified as a marker of acute exacerbation of COPD, showing a rapid decline after 4-5 d of treatment (47).

In contrast to CRP, plasma sTNF-R55 and R75 were not different between patients and healthy subjects and did not decrease during the exacerbation. The latter phenomenon may be due to a lack in response to the high-dose glucocorticosteroid treatment (48) or to relatively slow kinetics of the TNF receptors (49). The TNF receptors have been suggested to represent a buffer system that prolongs the biological effects of TNF- by forming a "slow-release reservoir." Compared with circulating TNF-, which was not measured in the present study, the sTNF receptor concentrations remain elevated for longer periods of time and are probably of more value for monitoring proinflammatory responses (50).

In the present study, no correlations were revealed between sTNF-R75 and any of the metabolic parameters. This might be explained by the fact that sTNF-R55 represents the major soluble form of the TNF-receptor and, although sTNF-R75 may have a permissive role, it appears that sTNF-R55 has the primary role in controlling the cytotoxic effects of TNF- (50).

Besides its function in weight homeostasis, leptin may also play a protective role in severe stress states. In leptin-defective mice, in which the T-cell-mediated immunity is decreased for unknown reasons, administration of leptin resulted in a complete reversion of the immunosuppressive effects of acute starvation. This might be explained by the fact that leptin, which is low during starvation, can induce proliferation, differentiation, and functional activation of hematopoietic cells and enhances the production and phagocytic activity of macrophages (51). This protective effect of leptin was confirmed in two studies in acute sepsis, in which surviving patients were characterized by higher leptin concentrations than nonsurvivors (35, 36).

Furthermore, a role for leptin in the regulation of respiration has been suggested. The respiratory depression of which leptin-deficient, obese mice were suffering was easily restored by administration of leptin (52). Whether these two additional functions of leptin play a role during an acute exacerbation of COPD is yet unknown, but deserves further attention.

In conclusion, temporary disturbances in the energy balance were seen during an acute exacerbation of COPD, related to increased leptin concentrations as well as to the systemic inflammatory response. Evidence was found that the elevated leptin concentrations were in turn under control of the systemic inflammatory response, and, presumably, the high-dose systemic glucocorticosteroid treatment.