OBJECTIVE: To determine the feasibility of studying myocardial and skeletal muscle bioenergetics using 31P magnetic resonance spectroscopy (MRS) in babies and young children with congenital heart disease. SUBJECTS: 16 control subjects aged 5 months to 24 years and 18 patients with CHD, aged 7 months to 23 years, of whom 11 had cyanotic CHD, five had cardiac failure, and two had had a Senning procedure. DESIGN: 31P MRS was carried out using a 1.9 Tesla horizontal 65 cm bore whole body magnet to study the myocardium in 10 patients and skeletal muscle (gastrocnemius) in 14 patients, eight of whom were exercised, together with appropriate controls. RESULTS: In hypoxaemic patients, in skeletal muscle at rest intracellular pH (pHi) was abnormally high [7.06 (SEM 0.04) v 7.04 (0.05), P < 0.01] and showed a positive correlation with haemoglobin (P < 0.03). On exercise, hypoxaemic patients fatigued more quickly but end-exercise pHi and phosphocreatine recovery were normal, implying that an equivalent but smaller amount of work had been performed. End-exercise ADP concentration was lower. On recovery, the initial rate of phosphocreatine resynthesis was low. Skeletal muscle bioenergetics were within normal limits in those in heart failure. In the myocardium, the phosphocreatine/ATP ratio was similar in controls and hypoxaemic subjects, but low in those in heart failure. CONCLUSIONS: In heart failure, the myocardial phosphocreatine/ATP ratio was reduced, as in adults, while resting skeletal muscle studies were normal. By contrast, hypoxaemic children had normal myocardial bioenergetics, but showed skeletal muscle alkalinity, and energy reserves were more readily depleted on exercise. On recovery, the initially slow phosphocreatine resynthesis rate reflects a low rate of mitochondrial ATP synthesis, probably due to an inadequate oxygen supply. 31P MRS offers a safe, non-invasive method of studying myocardial and skeletal muscle bioenergetics in children as young as 5 months.