Our objective was to define responder criteria using an anchor-based approach for frequency of cataplexy attacks and excessive daytime sleepiness in patients with narcolepsy undergoing sodium oxybate treatment. We used pooled data from two randomized, placebo-controlled, double-blind, multicentre 4- and 8-week trials of sodium oxybate for narcolepsy with cataplexy and analysed using receiver operator characteristics analysis. The percentage change in frequency of weekly cataplexy attacks and the Epworth Sleepiness Scale outcomes were compared with Clinical Global Impression of Change ratings, used as the anchor to define true response. Participants (n=336) were 39% male, 89% white, with a mean age of 41.5 (15.3)years, reporting a median of 20.5 cataplexy attacks per week and a mean Epworth Sleepiness score of 17.5 at baseline. A majority (51%) were Much Improved or Very Much Improved based on Clinical Global Impression of Change ratings, considered a true response to treatment. Area under the curve values for % reduction in cataplexy attacks (77%) and % change in sleepiness score (78%) supported response definition thresholds of 46% and 12%, respectively. Classification using either response definition agreed with the anchor for approximately 71% of participants. Cataplexy response definition was more sensitive (cataplexy=0.77, Epworth Sleepiness Scale=0.69), while sleepiness was more specific (cataplexy=0.66, Epworth Sleepiness Scale=0.75). Both responder definitions showed a dose-response relationship with sodium oxybate, demonstrating their validity using an external criterion. Weekly cataplexy attacks and Epworth Sleepiness Scale can be used to help document clinical response to narcolepsy treatment using criteria of 46% and 12% reductions, respectively.