OBJECTIVE: There are no evidence-based guidelines on the preferred approach to treating early-life epilepsy. We examined initial therapy selection in a contemporary US cohort of children with newly diagnosed, nonsyndromic, early-life epilepsy (onset before age three years). METHODS: Seventeen pediatric epilepsy centers participated in a prospective cohort study of children with newly diagnosed epilepsy with onset under 36 months of age. Details regarding demographics, seizure types, and initial medication selections were obtained from medical records. RESULTS: About half of the 495 enrolled children with new-onset, nonsyndromic epilepsy were less than 12 months old at the time of diagnosis (n = 263, 53%) and about half (n = 260, 52%) had epilepsy with focal features. Of 464 who were treated with monotherapy, 95% received one of five drugs: levetiracetam (n = 291, 63%), oxcarbazepine (n = 67, 14%), phenobarbital (n = 57, 12%), topiramate (n = 16, 3.4%), and zonisamide (n = 13, 2.8%). Phenobarbital was prescribed first for 50 of 163 (31%) infants less than six months old versus seven of 300 (2.3%) of children six months or older (P < 0.0001). Although the first treatment varied across study centers (P < 0.0001), levetiracetam was the most commonly prescribed medication regardless of epilepsy presentation (focal, generalized, mixed/uncertain). Between the first and second treatment choices, 367 (74%) of children received levetiracetam within the first year after diagnosis. CONCLUSIONS: Without any specific effort, the pediatric epilepsy community has developed an unexpectedly consistent approach to initial treatment selection for early-life epilepsy. This suggests that a standard practice is emerging and could be utilized as a widely acceptable basis of comparison in future drug studies.