Rats were treated for 28 days with increasing doses of dexfenfluramine (0.5, 1, 1.5, 2, 3, 4 and 5 mg/kg bid ip, each dose given for 4 days before being increased) and subsequently studied at intervals between 1 and 60 days following the cessation of treatment. Control rats received vehicle and were allowed food ad libitum or were pair fed with dexfenfluramine-treated animals. Immediately following drug treatment 5-HT immunoreactivity was increased in cortical areas compared to control animals. Subsequently, there was a persistent decrease in fine fibre density and the appearance of coarse truncated fibres. 5-HT levels in cortex were decreased 1 day following dexfenfluramine treatment but recovered to control values by 15 days. GFAP and GAP 43 immunoreactivity was unaffected by dexfenfluramine treatment compared to control animals, indicating a lack of evidence for neuronal degeneration and regeneration. Dexfenfluramine treatment decreased the density of 5-HT uptake sites in the cortex, labelled with [3H]-citalopram, but this partially recovered towards control values at 60 days. These alterations in 5-HT terminal networks conflict with the return of 5-HT levels to normal and the lack of evidence for degenerative changes or neuronal regrowth. On the basis of these results, it cannot be concluded that dexfenfluramine is neurotoxic.