The euglycaemic glucose clamp technique has been used to assess insulin resistance in patients with colorectal adenocarcinoma. Ten cancer patients were studied and compared with control subjects matched for age, sex and nutritional status. Forty-one euglycaemic clamps were performed at one of five different insulin infusion rates (20, 30, 40, 100 or 200 milliunits min-1 m-2). Glucose disposal was significantly decreased in the cancer group at all insulin infusion rates, whilst attained insulin levels and metabolic clearance rates of insulin were comparable in the control and cancer groups. Analysis of dose-response data allowed assessment of sensitivity (insulin concentration of half maximal glucose disposal) and responsiveness (maximal glucose disposal). Responsiveness was significantly reduced in the cancer group (40.3 versus 71.5 mumol kg-1 min-1; P less than 0.001), whilst sensitivity was similar (93.7 milliunits l-1 in controls versus 90.8 milliunits l-1 in cancer patients), suggesting a postreceptor defect in insulin action in the cancer group.