C Tudur Smith1, A G Marson, P R Williamson. 1. Division of Statistics and Operational Research, Department of Mathematical Sciences, University of Liverpool, Mathematics & Oceanography Building, Peach Street, Liverpool, UK, L69 7ZL. cat1@liverpool.ac.uk
Abstract
BACKGROUND: In developing countries, phenobarbitone is commonly used but its use in Europe and the USA has decreased due to concerns over adverse effects. Carbamazepine is recommended as the drug of choice for partial onset seizures, and there is concern that it may worsen some generalized onset seizure types. We report a review using individual patient data in which carbamazepine and phenobarbitone are compared. OBJECTIVES: To review the effects of carbamazepine compared to phenobarbitone monotherapy for people with partial onset seizures or generalized onset tonic-clonic seizures. SEARCH STRATEGY: The Cochrane Controlled trials register (Cochrane Library Issue 2, 2002); MEDLINE; EMBASE; handsearching; contacting experts and original trial investigators; contacting manufacturers of carbamazepine. SELECTION CRITERIA: Randomized or quasi-randomized, blinded or unblinded controlled trials in children or adults with partial onset seizures or generalized onset tonic-clonic seizures. DATA COLLECTION AND ANALYSIS: Outcome measures were (i) time to withdrawal of allocated treatment, (ii) time to 12 month remission, and (iii) time to first seizure. Data were analysed using a stratified logrank analysis with results expressed as hazard ratios (HR) and 95% confidence intervals (CIs), where a HR>1 indicates an event is more likely on phenobarbitone. A test for interaction between treatment and seizure type (partial versus generalized onset) was also undertaken. MAIN RESULTS: Data are available for 684 participants from four trials, representing 59% of the participants recruited into the nine trials that met our inclusion criteria. The main overall results (HR 95% CI) adjusted for seizure type were, (i) time to withdrawal 1.63(1.23 to 2.15), (ii) time to 12 month remission 0.87(0.65 to 1.17), (iii) time to first seizure 0.85(0.68 to 1.05). The review suggests that time to withdrawal is significantly improved with carbamazepine compared to phenobarbitone. No overall difference between drugs is identified for the outcomes 'time to 12 month remission' and 'time to first seizure'. Statistical heterogeneity was not encountered. An interaction between treatment and seizure type, confirmed statistically, was identified for time to first seizure, where phenobarbitone was favoured for partial onset seizures and carbamazepine for generalized onset tonic-clonic seizures. REVIEWER'S CONCLUSIONS: We found no overall difference between carbamazepine and phenobarbitone for time to 12 month remission or time to first seizure, however, subgroup analyses for time to first seizure suggest an advantage with phenobarbitone for partial onset seizures and a clinical advantage with carbamazepine for generalized onset tonic-clonic seizures. Phenobarbitone is significantly more likely to be withdrawn, indicating that it is less well tolerated than carbamazepine.
BACKGROUND: In developing countries, phenobarbitone is commonly used but its use in Europe and the USA has decreased due to concerns over adverse effects. Carbamazepine is recommended as the drug of choice for partial onset seizures, and there is concern that it may worsen some generalized onset seizure types. We report a review using individual patient data in which carbamazepine and phenobarbitone are compared. OBJECTIVES: To review the effects of carbamazepine compared to phenobarbitone monotherapy for people with partial onset seizures or generalized onset tonic-clonic seizures. SEARCH STRATEGY: The Cochrane Controlled trials register (Cochrane Library Issue 2, 2002); MEDLINE; EMBASE; handsearching; contacting experts and original trial investigators; contacting manufacturers of carbamazepine. SELECTION CRITERIA: Randomized or quasi-randomized, blinded or unblinded controlled trials in children or adults with partial onset seizures or generalized onset tonic-clonic seizures. DATA COLLECTION AND ANALYSIS: Outcome measures were (i) time to withdrawal of allocated treatment, (ii) time to 12 month remission, and (iii) time to first seizure. Data were analysed using a stratified logrank analysis with results expressed as hazard ratios (HR) and 95% confidence intervals (CIs), where a HR>1 indicates an event is more likely on phenobarbitone. A test for interaction between treatment and seizure type (partial versus generalized onset) was also undertaken. MAIN RESULTS: Data are available for 684 participants from four trials, representing 59% of the participants recruited into the nine trials that met our inclusion criteria. The main overall results (HR 95% CI) adjusted for seizure type were, (i) time to withdrawal 1.63(1.23 to 2.15), (ii) time to 12 month remission 0.87(0.65 to 1.17), (iii) time to first seizure 0.85(0.68 to 1.05). The review suggests that time to withdrawal is significantly improved with carbamazepine compared to phenobarbitone. No overall difference between drugs is identified for the outcomes 'time to 12 month remission' and 'time to first seizure'. Statistical heterogeneity was not encountered. An interaction between treatment and seizure type, confirmed statistically, was identified for time to first seizure, where phenobarbitone was favoured for partial onset seizures and carbamazepine for generalized onset tonic-clonic seizures. REVIEWER'S CONCLUSIONS: We found no overall difference between carbamazepine and phenobarbitone for time to 12 month remission or time to first seizure, however, subgroup analyses for time to first seizure suggest an advantage with phenobarbitone for partial onset seizures and a clinical advantage with carbamazepine for generalized onset tonic-clonic seizures. Phenobarbitone is significantly more likely to be withdrawn, indicating that it is less well tolerated than carbamazepine.