In patients with COPD, the diaphragm is chronically overloaded, which is more serious with increasing severity of the disorder. As a result, several adaptations occur, including a shift from fast, glycolytic, type II fibers to slow, oxidative, type I fibers (^{Levine et al 1997}; ^{Mercadier et al 1998}). Furthermore, metabolic changes occur, resembling those that emerge after endurance training, such as an increase in oxidative capacity (^{Noble and Ianuzzo 1985}; ^{Green et al 1995}; ^{Proctor et al 1995}). Mitochondrial function is enhanced in the diaphragm of patients with COPD (^{Ribera et al 2002}), as indicated by increases in the maximum rate of mitochondrial respiration and efficiency of the mitochondrial electron transport chain. This is considered to be beneficial, because these changes cause an increase in endurance capacity.

Despite this shift towards a more fatigue-resistant muscle, the majority of patients with severe COPD die from respiratory failure (^{Braghiroli et al 1997}). This respiratory failure may have multiple causes, including insufficient force production or insufficient endurance capacity of the respiratory muscles, or a combination of these factors. For example, ^{Levine et al (2003)} observed reduced force-generating capacity in the diaphragm of COPD patients.

Anabolic steroids may be helpful in preventing respiratory failure. Clinical studies have mainly investigated the effect of treatment with anabolic steroids on muscle force-generating capacity. For example, it has been shown that treatment with nandrolone decanoate (ND) improved respiratory muscle function and exercise capacity after a pulmonary rehabilitation program in COPD patients using oral glucocorticosteroids (^{Creutzberg et al 2003}). Treatment with testosterone resulted in increases in quadriceps muscle strength and endurance with and without a resistance training program in COPD patients (^{Casaburi et al 2004}). In this study, testosterone did not change respiratory muscle strength.

In animal models, the effects of anabolic steroid treatment on oxidative capacity in several skeletal muscles have been investigated. Controversy exists in the literature about the effect of anabolic steroids on mitochondrial function and associated endurance capacity. The effect of anabolic steroids on mitochondrial function seems to be dependent on fiber type, with fast-twitch fibers (^{Saborido et al 1991}) or aerobic muscle (^{Egginton 1987}) being most sensitive to an increase in oxidative capacity in rats.

For preventing respiratory failure, it is important to know if treatment with anabolic steroids results in increased respiratory muscle oxidative capacity and mitochondrial function, especially in emphysema. Therefore, we investigated whether administration of ND increased the activity of mitochondrial respiratory chain complexes in the diaphragm of emphysematous hamsters. For this purpose, we treated healthy and emphysematous hamsters with ND or placebo and determined the activity of several parts of the mitochondrial energy-generating system.

This is the first study to investigate the effects of treatment with an anabolic steroid on the activity of mitochondrial respiratory chain complexes in the diaphragm of emphysematous hamsters. It shows that the activity of mitochondrial respiratory chain complexes in the diaphragm was not different between normal and emphysematous hamsters. Treatment with ND did not change the activity of mitochondrial respiratory chain complexes in either normal or emphysematous hamsters. However, in emphysematous animals ND treatment decreased the activity of SCC compared with ND treatment in normal hamsters. Mitochondrial content was not changed after ND treatment. Furthermore, bodyweight of hamsters treated with ND was decreased after treatment compared with initial values, and serum testosterone levels were significantly lower in animals treated with ND than in controls.

In mitochondria, energy is produced by the mitochondrial oxidative phosphorylation system, which is embedded in the mitochondrial inner membrane and comprises five enzyme complexes (I–V) and two electron carriers (coenzyme Q and cytochrome c) (^{Hatefi 1985}; ^{Saraste 1999}). The main function of the system is the coordinated transport of electrons and protons through the different complexes, which leads to the production of adenosine triphosphate (ATP) (^{Smeitink et al 2001}). This ATP can be used for diverse cell functions, including contractions in skeletal muscle. In our study, we found that mitochondrial energy-generating capacity of the diaphragm was not changed after treatment with anabolic steroids, either in normal or emphysematous hamsters. This is in agreement with the results of ^{Saborido et al (1991)}, who found that in rat extensor digitorum longus (EDL), but not in soleus muscle, mitochondrial function was enhanced after treatment with fluoxymesterone and meth-androstanolone, suggesting a higher sensitivity for anabolic steroid in fast-twitch (EDL) muscle. In the same study, the number of mitochondria was not changed after administration of anabolic steroids. In contrast, ^{Egginton (1987)} suggested that aerobic muscles such as soleus and diaphragm are more susceptible to an increase of mitochondrial capacity after nandrolone phenylpropionate treatment. ^{Beck and James (1978)} showed an increase in mitochondrial volume in rat diaphragm after treatment with 19-nortestosterone. This is in agreement with the findings of ^{Satoh et al (2000)}, who found an increase in cross-sectional areas of mitochondria after treatment with nandrolone phenylpropionate in mouse diaphragm.

Comparison with previous results is difficult, because different studies use different dose, species, duration of treatment, age, mode of administration, and activity level. These factors have been shown to influence the effects of anabolic treatment (^{Bresloff et al 1974}; ^{Beck and James 1978}; ^{Kopera 1985}; ^{Prezant et al 1993}). For example, treatment of male rats with 19-nortestosterone in a low dose resulted in an increase in mitochondrial volume proportions in type I and intermediate fibers, whereas treatment with a high dose resulted in a less marked increase in mitochondrial volume proportions, particularly in intermediate fibers, suggesting that the response on anabolic steroids was lower in the high-dose group (^{Beck and James 1978}). The variation in the abovementioned factors may explain the fact that the literature is very inconsistent about the effect of treatment with anabolic steroids on mitochondrial function.

Another finding in our study is that after treatment with ND, SCC activity per mitochondrion is decreased in the diaphragm of emphysematous compared with normal hamsters. SCC activity measures complexes II and III and the electron carrier coenzyme Q, which are located in the mitochondrial inner membrane (^{Molano et al 1999}). Molano and colleagues reported decreased activity of enzymes in the mitochondrial inner membrane of rat liver after treatment with stanozolol and fluoxymesterone, but no change in CS activity, which is located in the matrix space. The authors suggested that anabolic-androgenic steroids could affect the mitochondrial membrane owing to their hydrophobic nature. In the same study, no change in C-IV activity was found after treatment with stanozolol and fluoxymesterone, suggesting that electron transport is disturbed in the electron transport chain before C-IV by these anabolic steroids (^{Molano et al 1999}). However, in our study we did not find a decrease in C-III activity after treatment with ND. This discrepancy in findings may be due to the difference in tissue studied, namely hamster diaphragm in our study versus rat liver in the study of Molano and colleagues. Furthermore, in the latter study the anabolic steroids were administered orally, which produces a first-pass effect in the liver, where the enzyme activities were measured. A possible explanation for our finding that SCC activity is decreased in the diaphragm of emphysematous hamsters treated with ND is that perhaps this specific mitochondrial part is more susceptible to the effects of ND administration in more active muscles (as is the diaphragm in emphysema).

It has recently been shown that the maximum rate of mitochondrial respiration and the efficiency of the electron transport chain for ATP production are increased in the diaphragm of COPD patients (^{Ribera et al 2002}). The maximum respiratory rate of the mitochondria in the diaphragm of COPD patients was twice as high as in the diaphragm of control subjects. Endurance training elicits an increase in maximum respiratory rate of the mitochondria in human vastus lateralis muscle (^{Walsh et al 2001}). However, this increase is much less than that observed in the diaphragm of COPD patients (^{Ribera et al 2002}). The respiratory rate is dependent on the muscle investigated in rats varying from 9.6 μmol O₂/min/g dry weight in soleus muscle to 32μmol O₂/min/g dry weight in cardiac muscle (^{De Sousa et al 2001}). This rate also seems to be dependent on species, because the maximum mitochondrial respiration rate reported by ^{Ribera et al (2002)} in human diaphragm was 5.28 μmolO₂/min/g dry weight, whereas the maximum respiration rate in rat diaphragm averaged 12μmol O₂/min/g dry weight (^{De Sousa et al 2001}). Data on interventions eliciting changes in human diaphragmatic mitochondrial function are not available. Therefore, it can not be predicted if the increase in mitochondrial respiration as found by Ribera and colleagues could be augmented by an intervention.

The fact that in our study we did not find an increase in the activity of mitochondrial respiratory chain complexes after ND treatment, either in normal or emphysematous hamsters, implies that treatment with ND may not increase the endurance capacity of the diaphragm. This suggests that treatment with anabolic steroids may not be able to prevent respiratory failure. However, respiratory failure may also be caused by a decrease in inspiratory muscle force-generating capacity. The role of anabolic steroids in preventing this cause of respiratory failure in patients with COPD is contradictory. For example, ^{Creutzberg et al (2003)} found an increase in maximum inspiratory muscle strength after ND treatment in COPD patients (50 mg ND/2 weeks). The patients in this study performed a pulmonary rehabilitation program consisting of several endurance activities. Only in patients using oral glucocorticosteroids did ND treatment have an additional beneficial effect on respiratory muscle function above the effect of pulmonary rehabilitation. This effect of ND was not observed in patients who did not use oral glucocorticosteroids. This is in accordance with the results of ^{Casaburi et al (2004)}, who found no change in maximum inspiratory muscle strength after administration of testosterone in COPD patients who did not receive long-term oral corticosteroid treatment. The finding that in patients using corticosteroids ND treatment restores respiratory muscle function is in line with animal experimental studies (^{van Balkom et al 1998}, ¹⁹⁹⁹). In a rat model, treatment with ND prevented the loss of diaphragm force induced by long-term, low-dose methylprednisolone administration (^{van Balkom et al 1998}). A subsequent study (^{van Balkom et al 1999}) reported that ND was also able to prevent the loss of diaphragmatic function in emphysematous hamsters treated with long-term, low-dose methylprednisolone. These findings, combined with findings in the abovementioned studies (^{Creutzberg et al 2003}; ^{Casaburi et al 2004}), suggest that treatment with anabolic steroids alone might be useful for preventing respiratory failure only in patients receiving corticosteroids.

In our study, we found that bodyweight was decreased after treatment with ND. This is in agreement with other studies, finding decreased bodyweight in male rats after treatment with ND (^{Bisschop et al 1997}). Circulating endogenic androgens may play a role. Surpassing the physiological level of androgens in males has been reported to result in depression of the natural production of testosterone (^{Ryan 1981}), downregulation of androgen-binding receptors (^{Rance and Max 1984}), decrease in appetite (^{Kochakian and Endahl 1959}), and a metabolic conversion to excess estradiol (^{Hickson and Kurowski 1986}). The natural production of testosterone is indeed decreased in ND-treated hamsters in our study, which is shown by lower serum testosterone levels in the ND groups.

In conclusion, this study shows that the activity of mitochondrial respiratory chain complexes in the diaphragm in both normal and emphysematous hamsters was equal, and that treatment with ND did not change this activity. In emphysematous hamsters, administration of ND decreased the activity of SCC compared with ND treatment in normal hamsters. Furthermore, we have shown that bodyweight of hamsters treated with ND was decreased after treatment compared with initial values, and that treatment with ND resulted in significantly lower serum testosterone levels in both normal and emphysematous hamsters. Taking all results together, our data do not support the use of anabolic steroids in preventing respiratory failure caused by fatigue of the diaphragm.

We thank the technicians of the mitochondrial biochemical diagnostics lab at the Laboratory of Pediatrics and Neurology for performing the mitochondrial measurements. This study was financially supported by Boehringer-Ingelheim, the Netherlands (Grant 4922 BA12).