A randomized, double-blind comparison of antiepileptic drug treatment in the elderly with new-onset focal epilepsy.

To compare the effectiveness of controlled-released carbamazepine (CR-CBZ) to levetiracetam (LEV) and to lamotrigine (LTG) in elderly patients with newly diagnosed focal epilepsy. Randomized, double-blind, parallel-group trial conducted between January 2007 and August 2011, in 47 ambulatory or hospital sites in Germany, Austria, or Switzerland. Eligible participants were aged ≥ 60, had new-onset epilepsy, had no acute illness as the cause of their seizures, and had no contraindication to the drugs in the trial. Patients were randomized 1:1:1 to CR-CBZ, LTG, or LEV. Doses were up-titrated for 6 weeks and could be maintained or adjusted depending on seizure relapse or tolerability over an additional period of 52 weeks. Primary outcome was the retention to treatment at week 58; secondary measures related to seizure and adverse event frequency. Of 361 randomized patients, 359 were included (CR-CBZ n = 121, LTG n = 117, LEV n = 122) in the modified intent-to-treat population (mean age [range] 71.4 [60-95] years). At week 58, the retention rate for LEV was significantly higher than for CR-CBZ (61.5% vs. 45.8%, p = 0.02), and similar to LTG (55.6%). Seizure freedom rates at weeks 30 and 58 were not different across the groups. Twice as many patients receiving CR-CBZ discontinued due to adverse events or death compared to those in the LEV group (32.2% vs. 17.2%; odds ratio 2.28, 95% confidence interval [CI] 1.25-4.19, p = 0.007), whereas discontinuation was intermediate for LTG (26.3%). Median daily doses of completers (n = 195) were CR-CBZ 380.0 mg/day (333.0-384.0), LTG 95 mg/day (94.0-97.0), and LEV 950 mg/day (940.0-985.0). In the initial monotherapy of focal epilepsy in the elderly, 1-year retention to LEV was higher compared to CR-CBZ due to better tolerability. Retention of LTG was intermediate and close to LEV, but did not differ significantly from either comparators. NCT00438451, www.clinicaltrials.gov.