A randomized, controlled trial of ZYN002 cannabidiol transdermal gel in children and adolescents with fragile X syndrome (CONNECT-FX)

Patients

A total of 212 patients were randomized at 21 clinical sites in the USA, Australia, and New Zealand. The first patient was randomized on August 8, 2018, and the final patient visit occurred on May 15, 2020. Patient disposition is shown in Fig. 2. Of the 212 patients randomized (ZYN002, 110; placebo, 102), one did not receive study treatment. Thus, 211 patients were included in the safety analysis set. One treated patient did not have a post-baseline efficacy measure, resulting in 210 patients in the full analysis set (FAS).

Fig. 2figure 2

Flow diagram of participants in the CONNECT-FX trial. a indicates the following: failure to meet randomization criteria. b indicates the following: received at least one dose of double-blind treatment. c indicates the following: all patients in the safety analysis set with both a baseline and at least one post-baseline efficacy measurement

Baseline demographics and disease characteristics

The patient populations were well matched in terms of demographic and disease characteristics (Table 1). Most patients were white (78.3%) and male (75%).

Table 1 Baseline demographics

As described in the “Methods” section, ad hoc analyses were conducted to evaluate the effect of ZYN002 vs placebo in patients with ≥ 90% methylation of the promoter region of the FMR1 gene and in patients with 100% methylation of the FMR1 gene. The ≥ 90% methylation group represented 79.7% of the total study population. The ≥ 90% methylation patients were similar to the overall intention-to-treat (ITT) population with respect to baseline demographic and disease characteristics (Table 1). The 100% methylation group (64.8% of total study population) also had similar baseline characteristics to the ITT population and the cohort of patients with ≥ 90% methylation.

Consistent with the published literature [14, 48], patients with ≥ 90% methylation generally presented with a more severe phenotype than those who were < 90% methylated (Table 1). Baseline demographics and disease characteristics were well matched between placebo and ZYN002 in the ≥ 90% methylation group of patients (Table 1).

Results for the primary analyses (FAS)ABC-CFXS subscale scores

The primary efficacy end point was the change from baseline to week 12 in social avoidance (SA) as measured by the Aberrant Behavior Checklist-Community FXS specific (ABC-CFXS) SA subscale score [45, 49, 50]. Week 12 changes from baseline in ABC-CFXS subscale scores measuring primary and secondary end points in the FAS population are shown in Fig. 3A. Although improvements in SA, irritability (Irr), and social unresponsiveness/lethargy (SU/L) (indicated by decreases in score) were greater in the ZYN002 group than in the placebo group, the differences were not statistically significant.

Fig. 3figure 3

Changes from baseline in ABC-CFXS subscale scores. Least square mean (LSM) changes from baseline in ABC-CFXS subscale scores are shown for each treatment group along with the LSM treatment differences. A Results from the full analysis set. B Results from the patients who had ≥ 90% methylation of the promoter region of the FMR1 gene. FAS, full analysis set

Meaningful change threshold

Mean change scores on the ABC-CFXS subscales by CaGI-S baseline to week 12 change category are shown in Supplemental Table S1. For the change categories indicating improvement (i.e., 1 or 2 category change), mean change scores on the ABC-CFXS SA, Irr, and SU/L subscales ranged from − 3.0 (2.96) to − 5.6 (3.06), − 8.9 (9.43) to − 13.8 (11.73), and − 3.9 (5.69) to − 7.2 (5.48) respectively for the CaGI-S domain specific and − 3.3 (4.85) to overall behavior items, respectively. A one category improvement (e.g., severe to moderate problem) on the CaGI-S were reported as a meaningful improvement in SA, Irr, and SU/L and overall behavior in the qualitative interviews [47].

Mean change scores on the ABC-CFXS subscales by CaGI-C value at week 12 are presented in Supplemental Table S2. For the values indicating improvement (i.e., a little better, moderately better, and much better), mean change scores on the ABC-CFXS SA, Irr, and SU/L subscales ranged from − 2.6 (3.47) to − 4.7 (3.44), − 8.6 (9.71) to − 17.4 (13.12), and − 5.1 (6.00) to − 7.8 (5.60). Any positive change on the CaGI-C was reported as meaningful or important in the caregiver cognitive interviews. As with the CaGI-S, meaningful change on the CaGI-C was considered by caregivers to imply improvement [51].

The eCDFs show clear separation between the CaGI-S change categories and the CaGI-C values representing improvement and deterioration (Supplemental Figs. 1 and 2).

Triangulating the results from the anchor-based analyses, the visual plots, and the levels of meaningful change reported by caregivers in the qualitative study, MCTs for changes from baseline to week 12 were determined to be 3 or more points on the ABC-CFXS SA subscale, 9 or more points on the ABC-CFXS Irr subscale, and 5 or more points on the ABC-CFXS SU/L.

Responder analysis—ABC-CFXS

Responder analyses using the MCT estimates did not find a statistically significant difference between placebo and ZYN002 for SA (nominal P = 0.254) and SU/L (nominal P = 0.190) subscale scores; however, there was a trend toward statistical significance in favor of ZYN002 for Irr (nominal P = 0.057). The model-based estimates of percent improved at week 12 were higher for the ZYN002 group than the placebo group for all 3 subscales: 54% vs 46% for SA, 37% vs 24% for Irr, and 41% vs 32% for SU/L.

Clinical Global Impression, Improvement (CGI-I)

For the key secondary end point of CGI-I, the percentage of patients with improvement at week 12 was higher in the ZYN002 group compared with the placebo group, but the difference was not statistically significant. Ratings of “Very much improved,” “Much improved,” or “Minimally improved” were reported for 40.5% of the placebo group vs 46.8% of the ZYN002 group (P = 0.376).

Caregiver Global Impression-Change (CaGI-C)

The percentage of patients whose parents/caregivers indicated that their child was “a little better,” “moderately better,” or “much better” on the CaGI-C was higher for patients receiving ZYN002 compared with those receiving placebo for SA/isolation (least squares [LS] means 57.1% vs 47.6%, nominal P = 0.184), social interactions (61.5% vs 44.9%, nominal P = 0.021), irritable/disruptive behavior (48.6% vs 38.7%, nominal P = 0.185), and overall behavior (55.9% vs 44.9%, nominal P = 0.145).

Ad hoc analyses: results in the ≥ 90% methylation groupABC-CFXS subscale scores

In this study, there was clear evidence of a threshold effect at 90% methylation based on a statistically significant treatment-by-subgroup interaction in change from baseline to week 12 ABC-CFXS SA, (nominal P = 0.002); the 2 subgroups were qualitatively different, i.e., the treatment effect was reversed in the < 90% methylation group. There was no statistical evidence of a similar or greater treatment difference between subgroups below 90% methylation. The treatment effect observed in patients with ≥ 90% methylation was more than double the treatment effect observed in the overall sample.

Analysis of week 12 changes from baseline in ABC-CFXS subscale scores in the ≥ 90% methylation group demonstrated statistically significant improvement in the ZYN002 patients vs placebo patients for the primary end point of SA as measured by the ABC-CFXS SA subscale (treatment difference of − 1.00, nominal P = 0.020) (Fig. 3B). The median percent of improvement from baseline was 40.0% in the ZYN002 group vs 21.1% in the placebo group. Although statistical significance was not observed for the ABC-CFXS subscales of Irr (nominal P = 0.091) and SU/L (nominal P = 0.135) in the ≥ 90% methylation group, the mean changes in score were greater for the ZYN002 group than for the placebo group for both subscales (Fig. 3B).

The distribution of responses to ZYN002 (change from baseline in SA subscale scores) in the ≥ 90% methylation group illustrated the response to treatment, as shown in Fig. 4. In the ≥ 90% methylation group, the mode for change from baseline among patients receiving ZYN002 was − 4 vs 0 for patients receiving placebo. In the < 90% methylation group, the mode was 0 for both placebo and ZYN002; a similar number of patients in the ZYN002 treatment group improved (5/17) or worsened (6/17).

Fig. 4figure 4

Individual patient changes from baseline to week 12 in ABC-CFXS Social Avoidance subscale scores: ≥ 90% methylation group. Distribution of changes in ABC-CFXS SA subscale scores at 12 weeks in the patients who had ≥ 90% methylation of the promoter region of the FMR1 gene. Each circle represents one patient in the study. The mode for the distribution of each group is shown in the figure

Responder analysis—ABC-CFXS

In the responder analysis for clinically meaningful within-patient change for the ≥ 90% methylation group, significant differences between placebo and ZYN002 were observed for the percent of patients improved at week 12 in SA (odds ratio 2.04, nominal P = 0.031) and Irr (odds ratio 2.17, nominal P = 0.036) (Fig. 5). The model-based estimate of percent improved in the ZYN002 group was 58.2% for SA and 40.3% for Irr. The improvement at Week 12 in SA corresponded to a number needed to treat (NNT) of 5.7.

Fig. 5figure 5

Meaningful change in ABC-CFXS subscales in the ≥ 90% methylation group. Percentage of patients who experienced meaningful changes from baseline in ABC-CFXS subscale scores, defined as a change of ≥ 3 for SA, ≥ 9 for Irr, or ≥ 5 for SU/L. The data represent results from the patients who had ≥ 90% methylation of the promoter region of the FMR1 gene

Clinical Global Impression, Improvement (CGI-I)

In the ≥ 90% methylation group, ratings of “Very much improved,” “Much improved,” or “Minimally improved” were reported in 37.7% of the placebo population vs 51.1% in the ZYN002 group (nominal P = 0.056).

Caregiver Global Impression-Change (CaGI-C)

In the ≥ 90% methylation group, the percentage of patients whose parents/caregivers indicated that their child was “a little better,” “moderately better,” or “much better” was statistically significantly higher for patients receiving ZYN002 compared with those receiving placebo for all 3 items in CaGI-C, SA/isolation, social interactions, and irritable/disruptive behavior (all nominal P < 0.05) and neared significance for overall behavior (nominal P = 0.052) (Fig. 6).

Fig. 6figure 6

Caregiver Global Impression-Change, any improvement: ≥ 90% methylation group. Percentage of patients who recorded improvements in Caregiver Global Impression-Change. The data represent results from the patients who had ≥ 90% methylation of the promoter region of the FMR1 gene

Ad hoc analysis: results in the 100% methylation group

An ad hoc analysis of patients with 100% methylation of the promoter region of the FMR1 gene was conducted to further explore the impact of complete methylation on response to ZYN002. This analysis revealed a further increase in treatment effect. The effect sizes were − 0.14 for the primary analysis, − 0.36 for the ≥ 90% methylation group, and − 0.39 for 100% methylation group. In patients with 100% methylation, ZYN002 was associated with 40% median improvement in the ABC-CFXS SA (treatment difference of − 1.08, nominal P = 0.027). Statistically significant effects were also observed for responder analyses for clinically meaningful change in ABC-CFXS SA (≥ 3 points; 56% for ZYN002 vs 37% for placebo [nominal P = 0.030]) and in the CaGI-C for any improvement in social interaction (63% for ZYN002 vs 37% for placebo [nominal P = 0.005]) and irritable/disruptive behaviors (54% for ZYN002 vs 33% for placebo [nominal P = 0.027]).

Safety

Approximately half (54.0%) of the 211 patients included in the safety analysis population experienced at least one treatment-emergent adverse event (TEAE). The frequency of TEAEs was similar for the placebo and ZYN002 treatment groups (50.0% and 57.8%, respectively) (Table 2). In the ≥ 90% methylation group, 55.4% of patients in the safety analysis population (n = 168) experienced at least one TEAE (placebo: 51.9%; ZYN002: 58.2%). In the < 90% methylation group, 47.6% experienced at least one TEAE (placebo: 44.0%; ZYN002: 52.9%). All TEAEs were mild or moderate in severity.

Table 2 Treatment-emergent adverse events occurring in > 1% of patients (safety analysis set)

There were no serious adverse events (SAEs) or severe TEAEs reported during the study. One patient in the placebo group discontinued study treatment due to a TEAE (stereotypy). The most frequently reported TEAEs in either treatment group were upper respiratory tract infections. Psychiatric disorder TEAEs, primarily symptoms of FXS, were reported for 5 (4.9%) patients in the placebo group (anxiety, impulsive behavior, irritability, staring, stereotypy: 1 patient each) and 2 (1.8%) patients in the ZYN002 group (aggression, bruxism: 1 patient each).

Treatment-related TEAEs were reported for 4 (3.9%) patients in the placebo group (total of 6 events) and 11 (10.1%) patients in the ZYN002 group (total of 14 events). In the ≥ 90% methylation group, 8.3% had at least one treatment-related TEAE (placebo: 5.2%; ZYN002: 11.0%). In the < 90% methylation group, 2.4% experienced at least one treatment-related TEAE. The most common treatment-related TEAE was application site pain, reported in 1 (1.0%) placebo-treated patient and 7 (6.4%) ZYN002-treated patients. Application site pain was mild except for 2 events of moderate severity in the ZYN002 group. All other application site TEAEs reported in the ZYN002 group (application site dryness and application site pruritus in 1 patient, application site rash in 1 patient, and application site reaction in 1 patient) were of mild severity. Application site urticaria of moderate severity was reported for 1 patient in the placebo group.

Regarding skin assessments, 97% and 89% of individual daily diary skin irritation scores assessed by caregivers were “0,” no erythema, for placebo- and ZYN002-treated patients, respectively. The percentage of patients with a score of “2,” moderate erythema with sharply defined borders, recorded by caregivers during months 1, 2, and 3 of treatment, respectively, were 2.9%, 2.0%, and 2.0% for the placebo group and 17%, 5.6%, and 1.9% for the ZYN002 group. The only scores of “3,” intense erythema with or without edema, were recorded during month 1 of randomized treatment for 1 patient (1%) in the placebo group and 3 patients (2.8%) in the ZYN002 group. There were no scores of “4,” intense erythema with edema and blistering/erosion, reported by caregivers for any patients.

The highest skin irritation score assigned by investigators was a score of “2” recorded for 2 of 104 (1.9%) patients in the ZYN002 group at week 4, 1 of 98 (1.0%) at week 8, and for 1 of 98 (1.0%) patients in the placebo group at weeks 8 and 12.

Changes from baseline in laboratory values for chemistry and hematology were comparable between the placebo and ZYN002 treatment groups, and there were no clinically relevant abnormalities in either group. There were no clinically significant changes in liver function tests reported in any patient. In both treatment groups, overall changes from baseline in vital signs and ECG parameters were minimal and not clinically significant.

留言 (0)

沒有登入
gif