SummaryBackground LEO 43204 is a novel ingenol derivative in development for the treatment of actinic keratosis. Objectives To compare the safety and preliminary efficacy of three doses of LEO 43204 with ingenol mebutate in actinic keratoses (AKs). Methods Patients with at least three visible, discrete, nonkeratotic AKs on four separate selected treatment areas on the forearms received LEO 43204 gel (0Á025%, 0Á05% and 0Á075%) and ingenol mebutate 0Á05% gel, by investigator-blinded, randomized allocation, for 2 consecutive days. Patients were assessed at 8 weeks. Primary outcomes included maximum composite local skin response (LSR) score and adverse events (AEs). Secondary outcomes included a reduction in the number of visible AKs. Results Forty patients completed the trial. For all treatments, mean LSR scores peaked at week 1, and were below baseline by week 8. Mean maximum composite LSR scores for LEO 43204 0Á025%, 0Á05% and 0Á075% were 9Á2 (Dunnett adjusted P = 0Á02), 10Á1 (Dunnett adjusted P = 0Á90) and 11Á2 (Dunnett adjusted P < 0Á01), respectively, vs. ingenol mebutate 0Á05% gel (10Á0). The most frequent AEs across all treatments were application site pruritus, burning sensation and tenderness. Mean reduction in the number of AKs was comparable for ingenol mebutate and the two lowest doses of LEO 43204 (71Á9-73Á1%), but LEO 43204 0Á075% gave a significantly larger reduction (81Á8%; Dunnett adjusted P = 0Á04). Conclusions LEO 43204 had a similar safety profile to ingenol mebutate and a dose-response relationship for LSRs was demonstrated. The highest LEO 43204 dose (0Á075%) significantly reduced the AK count when compared with ingenol mebutate.