We have previously demonstrated an altered pattern of tumor necrosis factor (TNF) secretion in patients with familial Mediterranean fever (FMF). To examine whether TNF determination could assist in diagnosing FMF, we stimulated heparinized blood of 51 asymptomatic FMF patients with lipopolysaccharide (LPS) and then measured TNF production in response to inducers, compared to unstimulated blood cells and to cells from a control group of 12 matched healthy subjects. Following LPS pretreatment, which induced TNF release, FMF patients produced significantly less TNF than controls, whether production was ‘spontaneous’ or induced by either LPS or phytohaemagglutinin (p≤0.003). Such ‘exhaustion’ did not occur in untreated cells. We then used these results to classify a further group of 29 FMF patients and 10 matched healthy controls (‘validation’ group) who underwent the same studies. The test correctly identified 25/29 patients as having FMF and 7/10 controls as not having FMF; a sensitivity of 86% and a specificity of 70% (likelihood ratios 2.9 (positive test) and 0.2 (negative)). Identification of a blunted TNF response following previous stimulation by a simple assay, may help the diagnosis of FMF in asymptomatic patients, provided it is interpreted in conjunction with supportive clinical data.