A pharmacogenetic predisposition to thalidomide-induced neuropathy has been investigated. Differences of drug metabolism were examined in 16 patients with severe orogenital ulceration, who were treated with thalidomide (< or = 200 mg/day) for 0.3-5.0 years. Eight had evidence of early peripheral neuropathy according to nerve conduction studies. Rates of C-hydroxylation, N-acetylation, and conjugation reactions with sulphate, glucuronide and glycine, were tested with the probe compounds debrisoquine, sulphadimidine, paracetamol and aspirin, respectively. Urinary drug metabolites were analysed by high pressure liquid chromatography. Results were compared with 16 healthy age- and sex-matched volunteers. Of the patients 6.25% and 13.3% of the controls had a poor Debrisoquine Hydroxylator Ratio (DMR); none of the patients with neuropathy had a poor DMR as compared to 12.5% without neuropathy. Of the patients 40.0% and 35.7% of the controls were slow acetylators; 28.6% with neuropathy were slow acetylators as opposed to 50% without neuropathy. Similarly, there were no significant differences in rates of conjugation between groups. All unaffected patients were active smokers, whereas only two of those with neuropathy smoked. Cumulative dose or duration of therapy were unrelated to risk of neuropathy. In conclusion, changes of nerve conductivity are a frequent and unpredictable adverse effect of thalidomide (< or = 200 mg/day), although smoking may have a protective action against their development. Nerve conduction studies are required before and during treatment, irrespective of the prescribed dose.