Background: In cancer, reduced exercise capacity is frequently seen, its origin is unclear. Metabolic abnormalities within the skeletal muscle, independent of left ventricular (LV) function may contribute to exercise limitation. The objective of the study was to investigate potential mechanisms for exercise limitation in patients with pancreatic cancer (PCA). Methods: We examined 25 patients with PCA (age 61+/-11 years, mean +/- SD, Karnofsky index 85+/-10%, time since diagnosis 281+/-222 days). Exercise capacity (treadmill, RQ 1.1+/-0.1 vs. 1.2+/-0.1), LV systolic and diastolic function (echo), body composition (DEXA-scan) and limb muscle blood flow were assessed. 12 healthy subjects served as controls (age 58+/-7 years). Results: LV ejection fraction (LVEF) was not different between controls (62+/-5%) and PCA (60+/-8%, p=0.4) as was LVEDD (46+/-6 vs. 49+/-7mm, p=0.3). Diastolic function showed no differences between controls and PCA (9+/-4 vs. 8+/-4, p=0,6), nor did e' alone (0,1+/-0,03 m/s vs. 0,1+/-0,3 m/s, p=0,3) or e'/a' (0.9+/-0.2 vs. 1.1+/-0.7, p=0.5). BMI (25+/-5 vs. 22+/-3kg/m2, p=0.02) was reduced in PCA. Mean weight loss in PCA was 14±5kg. Patients with PCA had similar total lean tissue (48+/-11 vs. 51+/-10kg, p=0.5) but lower fat tissue mass (13+/-5 vs. 22+/-10kg, p=0.0006) and fat tissue content (20+/-7 vs. 31+/-10%, p=0.0004). Exercise capacity (peak VO2 20+/-6 vs. 26+/-6 mL/min/kg, p=0.008) and AT (13+/-2 vs. 15+/-4 mL/min/kg, p=0.03) were reduced in PCA, although the VE/VCO2-slope was similar (35+/-9 vs. 29+/-6, p=0.1). In controls, total peak VO2 related to total lean tissue (r=0.9, pConclusion: Exercise capacity is impaired in PCA, independent of lean tissue mass, systolic and diastolic function, ventilatory efficiency and muscle blood flow. We hypothesise according to the muscle hypothesis in CHF that symptom generation and exercise intolerance in cancer patients may be due to intrinsic metabolic changes within the skeletal muscle.