This study was designed to compare the efficacy of two ectoparasiticides against adult fleas on dogs: a topical (DPP, dinotefuran-permethrin-pyriproxyfen) and a systemic (S, spinosad). Dogs (n = 48; 10.21-22.86 kg BW) were allocated to six groups of eight dogs each (C1, C4, DPP1, DPP4, S1, S4). Dogs in the treated groups were administered a topical (3.6 mL of DPP) or a tablet (665 or 1040 mg of S) on day 0. Infestations with 100 unfed fleas (Ctenocephalides felis) occurred on days -6, -1, 2, 7, 14, 21 and 28. An additional untreated group (QC, n = 6) was involved to evaluate the flea-anti-feeding efficacy. These dogs were infested once with 150 fleas prior to combing of at least 50 live fleas from each dog 5 or 10 min after infestation. In the treated group, dislodged dead and moribund fleas were collected from dogs 5, 10, 15 and 60 min (DPP1, S1) or 5, 10, 30 and 240 min (DPP4, S4) post-treatment and subsequent flea infestations on pans placed underneath the cages. Fleas were counted and removed from dogs by combing 1 (C1, DPP1, S1) or 4 h (C4, DPP4, S4) post-treatment and subsequent infestations. Quantitative PCR analysis of the canine cytochrome b gene was conducted on dislodged fleas collected from treated and control (QC) dogs 5 and 10 min after post-treatment infestations. The number of gene copies was used as a marker of blood volume ingested by fleas. Dislodgeability and insecticidal efficacy were calculated using arithmetic means. A rapid onset of killing was observed for DPP with 12.7 % of dead and moribund fleas being dislodged in average from dogs as soon as 5 min after infestation. DPP exhibited a significantly higher and sustained speed of kill than S. The average insecticidal efficacy was 86 ± 8.8 and 95.3 ± 2.1 % with DPP, whereas it was only 33.7 ± 19.9 and 57.6 ± 18.6 % with S at respectively 1 and 4 h after weekly reinfestations. The DPP combination significantly inhibited the feeding of fleas (89 % reduction) up to onset of flea mortality for 1-month post-treatment.