OBJECTIVE: In congestive heart failure (CHF) the skeletal muscle of the lower limbs develops a myopathy with atrophy and shift from the slow type to the fast type fibres. The aim was to test the hypothesis that this myopathy is specific and not simply related to detraining, by comparing patients with different degrees of CHF with patients with severe muscle atrophy due to disuse. DESIGN: Case-control study involving 50-150 micrograms needle biopsies of the gastrocnemius muscle. By an electrophoretic micromethod, the three isoforms of myosin heavy chains (MHC) were separated. PATIENTS: Five patients restricted to bed for more than one year because of stroke with disuse atrophy and normal ventricular function, and 19 with CHF were studied. There were seven age matched controls. MAIN OUTCOME MEASURES: The percentage of MHC1 (slow isoform), MHC2a (fast oxidative), and MHC2b (fast glycolytic) was determined by densitometric scan and correlated with indices of severity of cardiac failure. RESULTS: Ejection fraction was 42.5 (SD 15.2)% in CHF, 59.5 (1.0)% in disuse atrophy and 60.3 (1.4)% in controls (P < 0.001 v both). The degree of muscle atrophy as calculated by the body mass index/gastrocnemius cross sectional area, showed a profound degree of atrophy in patients with muscle disuse [0.94 (0.39)]. This was worse than in the controls [4.27 (0.16), P < 0.0005] and the CHF patients [2.60 (1.10), P < 0.005]. Atrophy in CHF patients was also greater than in controls (P < 0.005). MHC1 was lower in CHF than in disuse atrophy [51.83 (15.04) v 84.5 (17.04), P < 0.01] while MHC2b was higher [23.5 (7.4) v 7.25 (7.92), P < 0.001]. There was a similar trend for MHC2a [24.83 (15.01) v 8.25 (9.12), P < 0.05]. Within the CHF group there was a positive correlation between NYHA class and MHC2a (r = 0.47, P < 0.05) and MHC2b (r = 0.55, P < 0.01) and a negative correlation between NYHA class and MHC1 (r = -0.74, P < 0.001). Similarly, significant correlations were found for ejection fraction, diuretic consumption score, exercise test tolerance, and degree of muscle atrophy. CONCLUSIONS: The CHF myopathy appears to be specific and not related to detraining. The magnitude of MCH redistribution correlates with the severity of the disease. The electrophoretic micromethod used is very sensitive and reproducible. Biopsies are so well tolerated that can be repeated frequently, allowing thorough follow up.