Animals and study design
All animals were healthy, pure or mixed-breed dogs of both sexes. The five studies were conducted under a controlled and blinded design, with dogs randomly allocated to two groups of eight dogs each (control and treated). Prior to treatment, all dogs underwent a physical examination conducted by a veterinarian to ensure that they were healthy. To detect the presence or absence of any treatment-related or unrelated health abnormality or adverse event, health observations were conducted daily from the start to end of all studies as well as at hourly intervals for 4 h immediately after treatment. All animals were managed similarly, with due regard for their well-being and in compliance with the Merial Ethics Committee and Merial Institutional Animal Care and Use Committee requirements.
The studies were designed in accordance with the World Association for the Advancement of Veterinary Parasitology (WAAVP) guidelines for evaluating the efficacy of parasiticides for the treatment, prevention and control of flea and tick infestation on dogs and cats  and were conducted in accordance with Good Clinical Practices as described in International Cooperation on Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products .
Studies 1, 3, and 4 were conducted in the United States and studies 2 and 5 were conducted in South Africa.
Infestations were induced with three different laboratory-maintained C. felis strains not known to be resistant to any ectoparasiticide. Fleas used in study 1 were obtained from the Bertek, Inc. colony (Greenbrier, Arkansas), which was started in 2004 using US sourced fleas. Wild caught fleas were added to this colony in 2011. US PLRS strain fleas were used in studies 2 and 5. This strain originated from the US and had been maintained in colony in South Africa for approximately 20 months prior to being used in these studies. Fleas from Elward II labs (Soquel, California) were used in studies 3 and 4. This strain has been maintained under laboratory conditions for approximately 33 years with the addition of wild caught fleas to the colony approximately every three to six months.
Dogs assigned to the five control groups were not treated. On study Day 0, each dog in the five treated groups received a topical application of 6.76%w/v fipronil + 50.48% w/v permethrin, (Frontline Tri- Act®/Frontect®) at either the minimum recommended dose of 0.1 mL/kg bodyweight corresponding to a minimum dose of 6.76 mg/kg fipronil and 50.48 mg/kg permethrin (study 1) or the commercial dose of the product based on bodyweight (pipette dose, studies 2, 3, 4 and 5). For studies 2 through 5, the applied fipronil dose ranged from 6.90 mg/kg to 13.39 mg/kg and the permethrin dose ranged between 51.51 mg/kg and 99.96 mg/kg. Treatments were applied once on Day 0 directly on the skin, after parting the hair, at two spots on the midline of the neck, between the base of the skull and the shoulder blades.
Flea infestations and adult flea counts
Each dog was infested with 100 (±5) unfed adult fleas on Days −1, 7, 14, 21 and 28. All live fleas remaining on the dogs were removed and counted via thorough combing of all body areas with a fine-tooth flea comb either on Day 1 at 24 h after treatment and at 24 h after each of the subsequent weekly flea infestations for study 1, or at 48 h after treatment or subsequent weekly infestation for studies 2, 3, 4 and 5.
No water exposure or shampooing was performed in study 1. In studies 2 and 3, dogs were submitted to water immersion carried out by thoroughly wetting dogs (including the head) with spray from a bathing wand for at least 1 min. The dogs were dried with a blow dryer before returning them to their cages (study 2) or allowed to air dry (study 3). Water immersion was conducted on Days 10 and 24 in study 2 and on Days 10, 17 and 24 in study 3. In studies 4 and 5, dogs were shampooed following a similar protocol for both studies on Day 17 with Bio Guard® shampoo (Farnam Companies, Inc., USA) as follows: the coat of the dog was wetted thoroughly with warm water. The shampoo was spread across the animal to form a foamy lather. It was massaged into the wet coat of the entire animal and left on for 5 min. The animal was then rinsed with clean water. The dogs were dried with a blow dryer before returning them to their cages (study 5) or were allowed to air dry (study 4).
For the evaluation of efficacy against adult fleas, the flea counts were transformed to the natural logarithm of (count + 1) for calculation of geometric means for each treatment group as previously described and in accordance with WAAVP guidelines . The percent efficacy was calculated as 100×[(C-T)/C], where C is the geometric mean of the untreated dogs and T is the geometric mean of the treated dogs. The log counts of the treated groups were compared to the log counts of the untreated groups using an F-test adjusted for the allocation blocks used to randomize the dogs to treatment group. The mixed procedure in SAS® version 9.1.3 was used for the analysis, with the treatment groups listed as a fixed effect and the allocation blocks listed as a random effect. All testing was two-sided at the significance level of p <0.05.