Steinhoff 2005

Methods	Randomised, open‐label, parallel‐group trial conducted in 24 centres across Germany. 4 treatment arms: LTG (two arms), CBZ and VPS Participants with partial and generalised epilepsy randomised separately to LTG or CBZ and LTG or VPS respectively
Participants	Adults and children > 12 years with newly diagnosed epilepsy; at least 1 seizure and EEG imaging suggesting epilepsy Number randomised not stated, number included in analysis: LTG = 88, CBZ = 88 (partial); LTG = 33, VPS = 30 (generalised) 106 male participants (64%) in partial epilepsy group, 27 male participants (43%) in the generalised epilepsy group 166 out of 239 total included in analysis have partial epilepsy (69%) Not stated how many participants had received previous AED treatment Mean age (years): LTG (partial) = 46.6, CBZ = 43.1, LTG (generalised) = 22.3, VPS = 23.3 Range not stated
Interventions	Monotherapy with LTG, CBZ or VPS 4‐week escalation phase leading to LTG = 100 mg/d‐200 mg/d, CBZ = 600 mg/d‐1200 mg/d in adults and 600 mg/d‐1000 mg/d in children aged 11‐15, VPS = 600 mg/d‐1200 mg/d for children aged 6‐14, 600 mg/d‐1500 mg/d for adolescents over 14 years and 1200 mg/d‐2100 mg/d for adults Trial duration: 26 weeks, range of follow‐up: not stated
Outcomes	Number of seizure‐free patients during trial weeks 17‐24 "Leaving the study" (retention rates) Adverse event rates
Notes	IPD requested from trial sponsor Glaxo Smith Kline but data could not be provided due to restrictions over the de‐identification of datasets from trials conducted in Germany. Aggregate data extracted from graphs in the publication. Data from participants with partial epilepsy is the randomised comparison of LTG and CBZ and data from participants with generalised epilepsy is the randomised comparison of LTG and VPS (see Data extraction and management)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as randomised, no other information provided
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label trial
Blinding of outcome assessment (detection bias) All outcomes	High risk	Open‐label trial
Incomplete outcome data (attrition bias) All outcomes	High risk	Number of participants randomised to each group not reported (254 randomised and 239 analysed in the four arms of the trial). Reasons for exclusion stated but not which drug these participants were randomised to
Selective reporting (reporting bias)	Low risk	No protocol available but clinical trial summary provided by the sponsor. Seizure outcomes and adverse events well reported
Other bias	Low risk	None identified