In a multicentre, double-blind trial 150 elderly patients (mean age 77 years) with newly diagnosed epilepsy were randomised in a 2:1 ratio to treatment with lamotrigine (LTG) or carbamazepine (CBZ). Following a short titration period, the dosage was individualised for each patient while maintaining the blind over the next 24 weeks. The main difference between the groups was the rate of drop-out due to adverse events (LTG 18% versus CBZ 42%). This was in part a consequence of the lower rash rate with LTG (LTG 3%, CBZ 19%; 95% CI 7-25%). LTG-treated patients also complained less frequently of somnolence (LTG 12%, CBZ 29%; 95% CI 4-30%). Although there was no difference between the drugs in time to first seizure, a greater percentage of LTG-treated patients remained seizure-free during the last 16 weeks of treatment (LTG 39%, CBZ 21%; P = 0.027). Overall, more patients continued on treatment with LTG than CBZ (LTG 71%, CBZ 42%; P < 0.001) for the duration of the study. The hazard ratio for withdrawal was 2.4 (95% CI 1.4-4.0) indicating that a patient treated with CBZ was more than twice as likely to come off medication than one taking LTG. In conclusion, LTG can be regarded as an acceptable choice as initial treatment for elderly patients with newly diagnosed epilepsy.