A total of 32 RCTs were included (n=5,212 participants); five were paediatric trials and the remaining 27 were adult trials. Significant heterogeneity was not reported for any of the following analyses.
Primary endpoint - 50% responder rate:
The relative risk ratio of the 50% responder rate between adults and children was 0.67 (95% CI 0.51 to 0.89). The meta-regression confirmed these findings showing a significant relation between age of patients and treatment effect, with a larger effect seen in adults (p=0.02). The 50% responder rate for patients receiving placebo was significantly higher in children (19%) than adults (9.9%) based on logistic regression (p<0.001). Restricting the analyses to comparable doses did not alter the results.
Secondary endpoints:
The relative risk of seizure-free rate (data available only for gabapentin, levetiracetam, oxcarbazepine and topiramate) was not significantly different between paediatric and adult patients.
Total withdrawal rates were significantly lower in paediatric trials (RR ratio 0.65, 95% CI 0.43 to 0.98); this finding was confirmed by the meta-regression (p=0.004). Withdrawal rates for adverse events were also significantly lower in children (RR ratio 0.51, 95% CI 0.26 to 0.98), although this was not confirmed by the meta-regression (p=0.08). There were no significant differences between adults and children for seizure aggravation withdrawal rates, although logistic regression showed that children receiving placebo dropped out more frequently than adults for seizure aggravation (p=0.001).