Impaired skeletal muscle oxidative phosphorylation in patients with severe mitochondrial respiratory chain defects results in disabling exercise intolerance that is associated with a markedly blunted capacity of muscle to increase oxygen utilization in relation to circulatory and ventilatory responses that increase oxygen delivery to muscle during exercise. The range of oxidative limitation and the relationship between the severity of oxidative defects and physiological responses to exercise among a broader spectrum of mitochondrial respiratory chain defects has not been defined. We evaluated oxidative capacity and circulatory and ventilatory responses to maximal cycle exercise in 40 patients with biochemically and/or molecularly defined mitochondrial myopathy (MM) associated with varying levels of exercise tolerance, and compared responses with those in healthy sedentary individuals. In the MM patients, mean peak work capacity (0.88 ± 0.6 W/kg) and oxygen uptake (VO2, 16 ± 8 ml/kg/min) were significantly lower (P < 0.01) than in controls (mean work capacity = 2.2 ± 0.7 W/kg; VO2 = 32 ± 7 ml/kg/min), but the patient range was broad (0.17–3.2 W/kg; 6–47 ml/kg/min). Oxidative capacity in patients was limited by the ability of muscle to extract available oxygen from blood [mean peak systemic arteriovenous O2 difference (a–vO2); patients = 7.7 ± 3.5, range 2.7–17.6 ml/dl, controls = 15.2 ± 2.1 ml/dl], as indicated by a linear correlation between peak VO2 and peak systemic a–vO2 difference (r2 = 0.69). In the patients, the increase in cardiac output relative to VO2 (mean ΔQ/ΔVO2 = 15.0 ± 13.6; range 3.3–73) and ventilation (mean peak VE/VO2 = 65 ± 24; range 21–104) were exaggerated compared with controls (mean ΔQ/ΔVO2 = 5.1 ± 0.7; VE/VO2 = 41.2 ± 7.4, P < 0.01). There was a negative exponential relationship between ΔQ/ΔVO2 and peak systemic a–vO2 difference (r2 = 0.92) and between peak VE/VO2 and systemic a–vO2 difference (r2 = 0.53). In patients with heteroplasmic mtDNA mutations, we found an inverse relationship between the proportion of skeletal muscle mutant mtDNA and peak extraction of available oxygen during exercise (r2 = 0.70). We conclude that the degree of exercise intolerance in MM correlates directly with the severity of impaired muscle oxidative phosphorylation as indicated by the peak capacity for muscle oxygen extraction. Exaggerated circulatory and ventilatory responses to exercise are direct consequences of the level of impaired muscle oxidative phosphorylation and increase exponentially in relation to an increasing severity of oxidative impairment. In patients with mtDNA mutations, muscle mutation load governs mitochondrial capacity for oxidative phosphorylation and determines exercise capacity.