Bimatoprost for Eyelash Growth in Japanese Subjects: Two Multicenter Controlled Studies

The reported multicenter, double-masked, randomized, parallel-group, vehicle-controlled studies had a similar design and a duration of 5 months. Study 1 was conducted at eight centers in Japan with subjects who had idiopathic hypotrichosis of the eyelashes, and study 2 was conducted at nine centers in Japan with subjects who had chemotherapy-induced hypotrichosis of the eyelashes. Three centers participated in both studies.

The trial period of each study ran from July 2011 to May 2012. Each study was performed in compliance with Japanese Good Clinical Practice, Good Clinical Practice, and the International Conference on Harmonisation guidelines. Institutional review board approval was obtained at each site before initiation of the study, and all the subjects gave written informed consent.

In both studies, subjects 20 years of age or older were eligible if they had a Global Eyelash Assessment (GEA) score of 1 or 2 using the GEA with the associated Photonumeric Guide for Japan (GEA-J), a best-corrected visual acuity score of 20/100 or better for each eye, and an intraocular pressure (IOP) of 20 mmHg or lower in each eye. In study 2, the eligible subjects also were required to report chemotherapy-induced eyelash hypotrichosis after completion of standard adjuvant chemotherapy for an early-stage solid tumor. The chemotherapy had to be completed 4–24 weeks before baseline assessments, and all side effects related to the chemotherapy had to be grade 1 or less on the Common Terminology Criteria for Adverse Events (CTCAE, version 4.03). Subjects were excluded from both studies if they had any uncontrolled systemic disease, any known ocular disease or abnormality, or a history of ocular surgery within the previous 3 months.

The subjects were screened for eligibility 3–14 days before randomization. The eligible subjects were randomly assigned in a 1:1 ratio to bimatoprost 0.03 % or vehicle by an interactive voice or by a Web response system on day 1 (baseline). The study medication was dosed once nightly for 4 months. One drop of study medication was applied to a sterile, single-use applicator and brushed along the upper eyelid margins. Separate applicators were used for each eye. The subjects were advised to remove makeup and contact lenses before application of the study medication. Contact lenses could be reinserted after 30 min. The subjects returned to the study center for evaluation at the end of week 1, then at months 1, 2, 4 (end of treatment), and 5 (posttreatment follow-up assessment).

The primary efficacy measurement was overall eyelash prominence as determined by the GEA score using the GEA-J. The GEA-J scale is identical to the original GEA scale except that the associated photoguide contains photographs of Japanese subjects exclusively. Both the GEA and the GEA-J scales have been validated in previous clinical studies [15]. A GEA score of 1 indicates none to minimal eyelash prominence; a score of 2 indicates moderate eyelash prominence; a score of 3 indicates marked eyelash prominence; and a score of 4 indicates very marked eyelash prominence.

The investigators assigned GEA scores based on live, in-person assessments of overall eyelash prominence across both eyes. The secondary efficacy end points included measurements of upper eyelash length (mm), eyelash thickness (mm²), and eyelash darkness (intensity units) using a validated digital image analysis (DIA) of superior-view digital eyelash photographs taken with standardized equipment and subject positioning at each study visit (except for week 1). Patient-reported satisfaction with various attributes of eyelashes also was assessed at each visit except for week 1 using the 9-item Eyelash Satisfaction Questionnaire (ESQ-9).

The ESQ-9 consists of nine questions and three domains. Each question is measured on a 5-point Likert-type scale (i.e., very satisfied, satisfied, neutral, unsatisfied, very unsatisfied), and each domain consists of three questions. The three domains are Length, Fullness, and Overall Satisfaction (LFOS: domain 1, assessing subject satisfaction with physical attributes of eyelashes); Confidence, Attractiveness, and Professionalism (CAP: domain 2, assessing subject satisfaction with subjective attributes of eyelash health); and Daily Routine (DR: domain 3, assessing subjects’ perceptions of the degree of bother associated with making eyelashes presentable). Positive changes in ESQ-9 scores and domains indicate improvement. The ESQ-9 has been translated, culturally adapted, and psychometrically validated with the Japanese population.

Ophthalmic examinations were performed at screening and at months 1 and 4. The exams included biomicroscopy and dilated ophthalmoscopy as well as measurement of best-corrected visual acuity and IOP. Adverse events (AEs) were recorded at each visit. Physical examinations and clinical laboratory testing were performed at screening and at month 4, and vital signs were obtained at each visit except for week 1.

The sample size determination in study 1 indicated that a total enrollment of 126 subjects would provide 90 % statistical power for detecting a 30 % difference in the response rate between the treatment groups, assuming a 30 % response in the vehicle control group and a two-sided type 1 error of 0.05. Study 2 was primarily a safety study not powered for efficacy and enrolled a much smaller cohort.

All efficacy analyses were conducted on the intent-to-treat (ITT) population, which comprised all of the randomized subjects whether they had received study treatment or not. The primary efficacy end point was prospectively identified as the clinical response and defined as at least a one-grade increase in GEA score from baseline at month 4. Between-group comparisons of the response rate were made using Pearson’s Chi square test or Fisher’s exact test. The same method was used to compare the response rates between the treatment groups at months 1, 2, and 5.

The secondary efficacy end points of change from baseline to month 4 (the primary time point) in eyelash length, progressive eyelash thickness, and eyelash darkness on DIA were calculated as the average of right and left upper eyelash measurements. Between-group comparisons were performed using the Wilcoxon rank-sum test. The ESQ-9 scores for each question and domain (three domains consisting of three questions each) were analyzed. The change from baseline on each question was analyzed by descriptive statistics.

Correlation analyses of the change from baseline to month 4 in GEA versus ESQ-9 question 3 (“Overall, how satisfied are you with your eyelashes?”), domain 1, and domain 2 were performed to evaluate the direction and strength of the correlation using the Spearman rank correlation coefficient (ρ). Its significance was assessed by the t test.

Safety parameters were evaluated in the safety population, which consisted of all the subjects who received at least one dose of study medication. The AEs were coded using the Medical Dictionary for Regulatory Activities, version 15.0. Comparisons of safety parameters were made using appropriate nonparametric statistical methods (Fisher’s exact test, Pearson’s Chi square test, or Wilcoxon rank sum test) or parametric tests (analysis of variance). No imputation of missing safety data was made.

For study 1, a total of 173 subjects were randomized: 88 subjects to bimatoprost and 85 subjects to vehicle. The two treatment arms were comparable in terms of baseline characteristics. The overall mean age was 40.8 years, and 89.6 % of the subjects were women. A majority of the patients (69.9 %) had a baseline GEA score of 2 (Table 1). One subject with a baseline GEA score of 3 was randomized to the bimatoprost group but did not receive treatment. This subject was included in the efficacy analyses as part of the ITT population.

Study 2 randomized 36 subjects to treatment, with 18 subjects in each group. The groups were comparable except for age. All of the subjects were women, and the majority (72.2 %) had a GEA score of 1 at baseline (Table 1).

The proportion of subjects who responded to treatment with at least a one-grade improvement in GEA score from baseline to month 4 was significantly greater with bimatoprost than with vehicle in both study 1 (77.3 vs 17.6 %; P < 0.001) and study 2 (88.9 vs 27.8 %; P < 0.001). In study 1, significantly higher response rates with bimatoprost, reflecting improved eyelash prominence, were evident by month 1 and maintained at the month 5 posttreatment assessment (Fig. 1). In addition, a significantly greater proportion of patients in the bimatoprost group than in the vehicle group experienced at least a two-grade improvement in GEA score from baseline, which started at month 2 and continued through the posttreatment assessment at month 5 (38.1 vs 1.2 % at month 5; P < 0.001).

Although study 2 was not powered for efficacy, the statistically significant between-group difference in eyelash prominence achieved at the month 4 primary time point was maintained at the month 5 posttreatment assessment (83.3 vs 27.8 %; P < 0.001). An example of a subject with a two-grade improvement at month 4 in study 1 and a one-grade improvement at month 4 in study 2 is shown in Fig. 2.

In both studies, the bimatoprost-treated subjects experienced significantly greater improvements than the vehicle group from baseline to month 4 in upper eyelash length, thickness, and darkness. In study 1, from baseline to month 4, the bimatoprost-treated subjects experienced mean improvements of 1.62 mm (24 %) in upper eyelash length, 0.35 mm² (44.8 %) in upper eyelash thickness, and −12.02 intensity units (negative value represents darkening; 8 %) in upper eyelash darkness (Fig. 3). In contrast, these eyelash parameters were essentially unchanged in the vehicle group.

Significant between-group differences were observed for mean change in eyelash length and darkness, starting at month 1, and for mean change in eyelash thickness, starting at month 2. These differences were maintained through month 5 (all P < 0.001 except for eyelash darkness at month 1 [P = 0.012]). In study 2, improvements in these parameters with bimatoprost compared with vehicle were statistically significant at month 4 (mean change from baseline in eyelash length [2.65 vs 0.99 mm; P = 0.003], eyelash thickness [0.76 vs 0.28 mm²; P = 0.007], and eyelash darkness [−21.07 vs −4.60 intensity units; P = 0.040). Examples of improvements in eyelash length, thickness, and darkness for bimatoprost-treated subjects in both studies are shown in Fig. 4.

In both studies, both treatment groups had similar mean baseline scores for the nine questions and three domains of the ESQ-9. In study 1, the bimatoprost group had significantly greater improvements from baseline satisfaction scores in domains 1 (LFOS) and 2 (CAP) (both P ≤ 0.001) at multiple time points and in the individual questions comprising these domains than the vehicle group (P ≤ 0.028). For question 3 in domain 1 (“Overall, how satisfied are you with your eyelashes?”), the bimatoprost group reported significantly greater improvements in satisfaction scores from baseline than the vehicle group at months 2–5 (P < 0.001) (Table 2).

No significant between-group differences were observed for domain 3 (DR) or its individual questions. At month 4, for the bimatoprost group, the improvements from baseline in subject satisfaction with eyelashes, as measured by question 3 (ρ = 0.381), domain 1 (ρ = 0.451), and domain 2 (ρ = 0.409), were positively and significantly correlated with improvements from baseline in eyelash prominence, as measured by GEA score (all P < 0.001). Significant correlations were not seen in the vehicle group.

In study 2, compared with the vehicle group, the bimatoprost group also had significantly greater improvements from baseline in satisfaction scores at multiple time points in domains 1 (P ≤ 0.017) and 2 (P = 0.019) and in the individual questions comprising these domains (P ≤ 0.05). Again, in the bimatoprost group, subject satisfaction with eyelashes, as measured by question 3 (ρ = 0.641; P = 0.004), domain 1 (ρ = 0.516; P = 0.028), and domain 2 (ρ = 0.538; P = 0.021) in ESQ-9, were significantly correlated with improvements in eyelash prominence measured by the GEA score.

Overall, most AEs resolved before the end of the study, and the vast majority were of mild severity. None of the AEs reported as moderate or severe was considered by the investigator to be associated with the study treatment. The serious AEs in the bimatoprost group included one case each of appendicitis and pyelonephritis, which were not considered to be associated with treatment. Three subjects discontinued the study due to AEs, namely, dry eye and facial pain (each considered to be treatment related) in the bimatoprost group and dissociative disorder in the vehicle group.

In study 1, ophthalmic examination showed no statistically or clinically significant differences between the bimatoprost and vehicle groups in terms of change from baseline in IOP (mean change from baseline to month 1, −0.79 vs −0.47 mmHg, P = 0.264; mean change from baseline to month 4, −0.84 vs −0.23 mmHg, P = 0.085). No subject had an IOP of 6 mmHg or lower, and no subject had an AE report of reduced IOP. Visual acuity remained unchanged (between −2 and +2 lines) for more than 90 % of the patients in both groups (P = 0.248) at month 1 and for more than 85 % at month 4 (P = 0.257). Biomicroscopy and ophthalmoscopy findings of a one-grade increase from baseline were observed more frequently in the bimatoprost group and reflected higher rates of conjunctival hyperemia (9.2 vs 0 %; P = 0.007). No two-grade increases were found.

In study 2, the bimatoprost group had a statistically significantly greater reduction in IOP from baseline than the vehicle group at the month 1 time point only (−2.21 vs −0.68 mmHg; P = 0.007). These changes in IOP were not considered clinically meaningful and, at month 4, the changes in IOP did not differ significantly between the treatment groups (−1.50 vs −0.38 mmHg; P = 0.152). As in study 1, no subject had an IOP of 6 mmHg or lower, and no subject had an AE report of reduced IOP. Visual acuity remained unchanged in most of the subjects, except for a decline of more than two lines in one subject and an increase of more than two lines in one subject, both in the bimatoprost group. The biomicroscopy and ophthalmoscopy findings did not differ between the treatment groups. No clinically meaningful changes in vital signs, hematology, blood chemistry, or urinalysis laboratory data were observed during either study.