Inclusion criteria
Home owners with flea infested cats contacted the “Flea Team” through referrals from the Sunshine Animal Hospital, Tampa, FL, Animal Dermatology South, New Port Richey, FL, and advertisements on Facebook® and CRAIGSLIST®. Members of the team visited each residence, and 40 private residences were selected for inclusion in the study from May 17 through June 14, 2017. Selection criteria included: (i) ≥ five fleas in comb counts on at least one cat at the residence; (ii) ≥ five fleas trapped during a 16–24 h period in two intermittent-light flea traps; (iii) one to 10 healthy cats and dogs living at the private residence; (iv) cats spend ≥ 12 hours/day inside the residence; (v) homeowners agree not to use any other topical, oral or premise flea control products during the study; (vi) cats and dogs cannot be pregnant or nursing; (vii) all cats and dogs must be > 6 months of age and cats need to be at least 1.18 kg and dogs > 2 kg; and (viii) owners sign a consent form and fill out a questionnaire concerning pet habits, flea control history and personal observations concerning potential flea hosts around their residence.
Flea population assessment
Flea infestation in indoor residence areas was assessed using two intermittent-light traps (MyFleaTrap™, Zantey Inc, Tallahassee, FL, USA) [14, 15]. Traps were placed one each in two rooms during 16- to 24-hour collection periods. Rooms were chosen based on where the cat(s) spent most of their time, or where fleas had been observed by owners. At each collection period traps were returned to the same location in the room. Species, number and sex of fleas collected on the adhesive-sheets of the traps were recorded.
Prior to initiating the study all personnel were trained in proper cat handling techniques using American Association of Feline Practitioners and American Animal Hospital Association recommended methods for proper and safe handling of cats. Fleas on cats were counted using a modified combing procedure due to time and safety constraints. Six regions were examined using 10 strokes of a standard flea comb per region: (i) back of the head behind the ears and extending along the dorsal midline to the tail; (ii) left and (iii) right side from axillary region to posterior aspect of the cat’s body, including hair on posterior of each leg; (iv) ventral aspect from chest to inguinal region; (v) ventral neck region from chin to chest; and (vi) top of the head between the ears. As fleas were collected they were placed in a plastic bag with collected hair. Fleas were counted and immediately placed back on the cat.
Fleas on dogs were estimated using a previously described area count methodology [16]. Fleas were counted in five areas on each animal; dorsal midline, tail head, left lateral, right lateral, and inguinal region. Due to the effects of large flea numbers on the accuracy of area counts, flea numbers in each of the five areas was capped at 50; therefore, the maximum total area flea count was 250.
All on-animal and premises flea counts were conducted ± 1 day on days 0, 7, 14, 21, then once between days 28–30, 40–45, 56–60 and 82–86. Personnel conducting pet and premises flea counts were not blinded to treatment groups.
Evaluation of pruritus and dermatologic lesions
Pet owners assessed the severity of pruritus of dogs(s) in each home during each scheduled visit using a previously validated and described non-numeric scale [17, 18]. In each home one owner completed the assessment. There is no equivalent validated technique for pruritus assessment in the cat; therefore, the owner-assessed pruritus visual analogue score (PVAS) used in this study was a modification of a previously published feline pruritus visual analogue score [19]. The severity of pruritus of the qualifying cat(s) in each home was similarly evaluated at each visit by the owner. Owners rated the pruritus level of the qualifying cat(s) using a non-numeric scale on a data capture form with descriptions of increasing severity. The owner pruritus rating was then numerically assessed as previously described for the dog [17, 18].
In both assessments owners did not see, nor were they informed of the numerical score given to their dog(s) or cat(s). In households with more than one owner, the same owner was required to assess the pruritus level of the pet(s) throughout the study.
Blinded clinical dermatologic observations were made on days 0, 30, 60, and 84 of the study (± 3 days) of all qualifying cats resident in the homes. The extent and severity of dermatologic lesions were assessed using the validated scale, Scoring Feline Allergic Dermatitis (SCORFAD) [19]. For this assessment 10 body zones were assessed for excoriations, miliary dermatitis, eosinophilic plaques, and self-induced alopecia using a score from 0–4 for each category and lesion type [19]. The percentage reduction in SCORFAD from baseline was previously determined to be the most valid assessment of clinical response and has been proposed as an assessment tool in feline hypersensitivity dermatitis [18].
Treatment groups
Qualifying homes and all pets in that household were randomly allotted to 1 of 2 treatment groups on day 0. Household entry numbers (1–40) were assigned a random number by Excel (Excel 2016, Microsoft, Redmond, WA) and blocked into groups of 2. The highest random number within each block was assigned to group 1 and the lowest to group 2.
In treatment Group 1 cats were administered a topical fluralaner solution (Bravecto® Topical Solution; Merck Animal Health, Madison, NJ, USA) at the recommended labeled dose (minimum 40 mg/kg). While this study was focused on cats, dogs resident in these households were administered an oral fluralaner chew (Bravecto® Chewable Tablets; Merck Animal Health, Madison, NJ, USA) at the recommended labeled dose (minimum 25 mg/kg).
In treatment Group 2 cats were administered a topical selamectin solution (Revolution®; Zoetis, Whippany, NJ, USA) at the recommended labeled dose (minimum 6 mg/kg). While this study was focused on cats, any dogs resident in these households were also administered an oral sarolaner chew (Simparica®; Zoetis, Whippany, NJ, USA) at the recommended labeled dose (minimum 2 mg/kg).
All animals were weighed on a calibrated scale prior to treatments and products were administered according to product labeling by study personnel. Fluralaner was administered once on day 0. Selamectin and sarolaner were administered three times; once on study day 0; once between days 28–30 and finally between days 56–60. All dogs and cats living at a residence were administered group appropriate treatment and no alternative flea treatments were used during the study on pets or premises. During this study no corticosteroids, antihistamines, antibiotics or medicated shampoos were used to alleviate pruritus or skin lesions. No restrictions were placed on the animals regarding swimming, non-insecticidal baths or movement outdoors.
This study was conducted without a placebo control group because the heavy and constant flea challenge experienced by cats and dogs in subtropical Florida would make inclusion of a non-treated group inhumane. Withholding flea adulticide treatment would be detrimental to the health and welfare of the pets and potentially also to humans in these households.
Data analysis
The animal and trap flea count data were transformed prior to analysis using the Y= loge(x+1) transformation. The log-transformed flea counts on animals were analyzed by a mixed linear model with repeated measures including treatment, day, treatment*day as the fixed effects; and household, and animal as random effects. The log-transformed flea counts in traps were analyzed by a mixed linear model with repeated measures including treatment, day, and treatment*day as the fixed effects and household as random effect.
A Kenward-Rogers adjustment was used to determine the denominator degree of freedom for hypothesis. Akaike’s information criterion (AIC) was used as the criterion to select the covariance structure for repeated measures. The dermatology, pruritus and SCORFAD scores were analyzed by the same mixed linear model with repeated measures as that for the flea counts on dogs. Percentages of animals without fleas were analyzed and compared using Fisher’s exact test. All comparisons were made between treatment groups on each data collection day and also between each collection day and the baseline (day 0) values within each treatment group.
A two-tailed t-test was used for the comparison and significance was declared when P < 0.05; 90% confidence intervals were constructed for the differences between treatment groups for the equivalence declaration. The primary software was SAS version 9.3 (SAS® Language: Reference, Version 9.3, SAS Institute Inc., Cary, NC, USA).
Percent control of flea counts were calculated using geometric means with Abbott’s formula:
$$ \mathrm{Efficacy}\kern0.5em \left(\%\right)\kern0.5em =\kern0.5em 100\kern0.5em \times \kern0.5em \left({\mathrm{M}}_{\mathrm{B}}\kern0.5em -\kern0.5em {\mathrm{M}}_{\mathrm{C}}\right)\kern0.5em /\kern0.5em {\mathrm{M}}_{\mathrm{B}} $$
where MC is the geometric mean number of fleas on flea count day and MB is the geometric mean number of live fleas count on baseline.
Percent reduction of clinical scores were calculated using arithmetic means with Abbott’s formula:
$$ \mathrm{Efficacy}\kern0.5em \left(\%\right)\kern0.5em =\kern0.5em 100\kern0.5em \times \kern0.5em \left({\mathrm{M}}_{\mathrm{B}}\kern0.5em -\kern0.75em {\mathrm{M}}_{\mathrm{C}}\right)\kern0.5em /\kern0.5em {\mathrm{M}}_{\mathrm{B}} $$
where MC is the arithmetic mean of clinical scores on score collection day and MB is the arithmetic mean of clinical scores on baseline.