The value of positive scans of autologous 99mTc-labelled leucocytes in the detection of intestinal inflammation was assessed in 29 patients with known inflammatory bowel disease and a high probability of intestinal inflammation and 58 with a low probability of intestinal inflammation comprising 37 with intestinal symptoms and 21 with suspected abscess. Autologous leucocytes were labelled with 99mTc using either of the established methods. In patients with inflammatory bowel disease, 3/13 with ileocaecal and 1/11 with colonic accumulation of 99mTc were incorrectly diagnosed in that there was no histological evidence of inflammation in those regions. In the low probability patients, all 26 positive images were false positives: there were 19 in the ileocaecal and 7 in colonic regions and no differences between the two subgroups. False positives were more common in low probability than high probability patients (p = 0.011). The results were independent of the 99mTc cell labelling technique used. Three false positive patients also had normal scans for 111In-labelled granulocytes. The time of first detection of 99mTc in the bowel was earlier in true than in false positive patients (92% v 33% within 1 h, p = 0.008), the false positive rates in the first hour of imaging being 5% in high and 15% in low probability patients. In 99mTc scintigraphy, the appearance of the isotope in the intestine occurs commonly in the absence of inflammation irrespective of the cell labelling technique used. In high probability patients, imaging at 1 h can more reliably discriminate true from false positives and retain clinical usefulness. In low probability patients specificity remains unacceptable and 99mTc-labelled leucocyte scanning in these patients is of little value as a positive predictor of intestinal inflammation.