Regulatory Decision Summary for Sholyne

The submission for oral Sholyne for the treatment of symptoms of dry mouth in patients with Sjögren’s Syndrome was based on three studies: study SB95US01 (reported by Fife et al.), a randomized, 6-week double-blind dose-finding Phase 2 study, comparing cevimeline 30 milligram (mg), 60 mg and placebo in 75 subjects, two 12-week pivotal Phase 3 studies, SB96US02 (reported by Petrone et al.), and SB96US04, comparing cevimeline 15 mg, 30 mg and placebo in 197 and 212 subjects, respectively.

There was limited data in the toxicology studies to allow for a comparison based on exposure. In non-clinical studies, Sholyne had no effect on fertility in males and a slight decrease in implantation in females, and administration during pregnancy was not associated with any fetal developmental effects.

A thorough cardiovascular assessment, consistent with the International Council on Harmonisation (ICH) S7, should have been provided given that Sholyne acts through receptors that are expressed on cardiac tissue. In the Response to NON (R/NON), the sponsor proposed considerations based on the totality of clinical data, changes to the Canadian product monograph and a commitment to perform an in vitro hERG assay post-approval. Ultimately, the data provided did not support this response.

There was a statistically significant increase in uterine adenocarcinomas in the rat, although the exposure ratio proposed could not be verified as the human exposure the sponsor provided does not support the ratio provided. The sponsor’s approach to confirm exposure in humans in the R/NON is not satisfactory. Thus, a lack of human exposure data obtained using bioanalytical methods and human plasma do not allow the exposure ratios proposed in the Product Monograph to be verified.

Study SB95US01 had two co-primary endpoints. The efficacy of the first co-primary endpoint, the subjective global evaluation of dry mouth at six weeks, with a response of “better” was 35%, 76% and 67% for placebo, cevimeline 30 mg and cevimeline 60 mg group subjects respectively (p = 0.004). Outcome was not significantly better when comparing either cevimeline groups with placebo for the other co-primary endpoint of subjective evaluation of six symptoms of dry mouth on a visual analog scale. Based on these results, only the global evaluation of dry mouth was kept as a primary efficacy endpoint for the Phase 3 studies, and the 60 mg dose, which did not result in better efficacy, was not included in later studies.

Study SB96US02 reached statistical significance of the primary efficacy endpoint with a response of “better” in 37%, 45% and 66% of placebo, cevimeline 15 mg, and cevimeline 30 mg group subjects, respectively, p = 0.0007.

In the largest Phase 3 study SB96US04, none of the results were statistically significant with a response of “better” in 55%, 36%, and 53% of placebo, cevimeline 15 mg, and cevimeline 30 mg group subjects, respectively, p = 0.889). As per the Sponsor, no firm explanation other than random chance can account for failure to show efficacy against placebo for both cevimeline doses, and the higher placebo response is consistent with random variability and uncertainty. Factors that can influence placebo response presented in the R/NON would also apply to the cevimeline arms. As such, these factors could not be responsible for the lack of efficacy observed for two different doses of cevimeline when compared to placebo. The R/NON did not provide confirmation of efficacy. 

Failure to show efficacy of cevimeline versus placebo in the pivotal study (SB96US04) is not only due to high placebo response, as proposed by the Sponsor, but to a lower efficacy observed with cevimeline than in studies SB96US02 and SB95US01. This has not been addressed adequately in the R/NON.

The lack of efficacy is supported by the fact that none of the other subjective endpoints were significant for cevimeline when compared with placebo in any of the studies, including assessment of symptoms of dry mouth rated on a visual analog scale, which was a co-primary endpoint in Phase 2 study SB95US01. Subjects reporting an improvement on global assessments did not report a significantly better outcome on this endpoint when comparing either active group with placebo. The failure to achieve significant results for this endpoint in all studies has not been addressed adequately in the R/NON.

Due to failure of the pivotal Phase 3 study SB96US04 to show efficacy, the Phase 2 study SB95US01 was filed as pivotal. The latter cannot be used as a substitute for the failed pivotal Phase 3 study due to its shorter duration, small number of subjects treated (twenty-five patients on cevimeline 30 mg), and mixed efficacy results on the co-primary outcome measures. Although it was filed as a supportive study, SB96US04 is considered pivotal in the overall assessment of Sholyne.

The secondary endpoint of change from baseline in salivary flow cannot be used as confirmatory evidence of efficacy or as a substitute for the pre-specified primary endpoint which failed to reach significance. The difference in salivary flow between cevimeline 30 mg and placebo in SB95US01 was minimal, with better results at the 60 mg dose. These results are not consistent with the results of the subjective primary endpoint, which was better at the lower dosage. The difference in salivary flow was also smaller between arms in study SB96US02, and similar results were obtained with cevimeline 30 mg and placebo in study SB96US04. As per the Sponsor, salivary flow measurements show high variability and uncertainty, greater at the 30 mg dose. Study SB96US02 cannot be used as primary or confirmatory evidence of efficacy in the context of the failed pivotal study SB96US04.

Moreover, the inconclusive efficacy is supported by a lack of consistent dose response. A response of “better” was reported in more cevimeline 30 mg arm subjects in study SB95US01, with no significant improvement in the 60 mg arm over placebo. The decrease in efficacy at the higher dose is not considered to be indicative of a plateau, as stated in the R/NON, as the efficacy does not plateau at 30 mg but decreases, and is more likely to be indicative of spurious results and lack of efficacy, in line with the lack of efficacy of different doses when compared with placebo. 

Cross-study comparisons presented in the R/NON do not offer robust or confirmatory evidence of efficacy, or compensate for the lack of efficacy shown in SB96US04. A cross-study dose-response analysis presented in the R/NON is not as robust and reliable as comparing doses within the same study, which has shown a lack of dose-response.

Therefore, explanations provided in the R/NON did not address satisfactorily the concerns noted during review and do not confirm efficacy.

Many of the most common adverse events (AEs) reflect the muscarinic activity of the drug. The most common AEs reported in studies of patients with Sjogren’s Syndrome taking cevimeline were excessive sweating, nausea, and diarrhea. There were two cases of myocardial infarction reported in the studies as possibly related to cevimeline and cardiovascular safety was not sufficiently demonstrated. As stated by the Sponsor, association of serious cardiovascular events with use of cevimeline cannot be excluded. In the R/NON, the Sponsor acknowledges the potential risks of QT prolongation and Torsades de Pointes, reported with cevimeline use. Cautionary labelling was proposed to be added in the Product Monograph. However, this risk mitigation is not considered adequate for a medication with no substantial evidence of efficacy for the treatment of dry mouth.

As per the Sponsor, no pharmacokinetic, safety, or efficacy studies were identified on drug interactions or in patients with hepatic or renal impairment, conditions which could result in reduced or delayed clearance and increased plasma concentrations of Sholyne. As such, clinical safety and efficacy have not been adequately demonstrated in these patients. Information provided in the R/NON pertaining to other drugs that exhibit similar clearance mechanisms to assess the changes in exposure due to hepatic or renal impairment, end stage renal disease, and drug-drug interactions only provides a theoretical estimate of risks. The proposed cautionary statements and contraindications for use provided in the R/NON may adress the unknown data, although clinical data in the target population would be more accurate and preferable.

Overall, the lack of evidence to support efficacy of cevimeline as compared to placebo, the inconsistent results on subjective endpoints, and the lack of dose-response suggest an overall lack of efficacy. Explanations provided in the R/NON do not sufficiently address the concerns identified during the review and do not provide confirmation of efficacy. Safety limitations including potential cardiovascular risks, limited data on drug interactions and on use in patients with renal or hepatic impairment have been acknowledged in the R/NON. Proposed labelling does not adequately address these safety issues especially in the context where there is a lack of substantial evidence supporting efficacy for the treatment of dry mouth in patients with Sjögren’s Syndrome.

A negative benefit-harm-uncertainty profile is maintained for the use of Sholyne for the proposed indication.