Agile's Phase III Data For Twirla Patch Fall Flat, Raise Doubts On Approval

Twirla low-dose birth control patch doesn't work as well in heavier women in Phase III SECURE study, but Agile believes that this is a class effect and an artifact of the stringent, real-world study design requested by FDA.

Agile Therapeutics Inc. is counting on data from the Phase III SECURE study to support a refiling of its Twirla low-dose combined birth control patch, but the results have raised serious questions about the product's approval prospects and spooked investors.

Agile unveiled top-line results from the SECURE study of 2,032 women on Jan. 3, describing "positive" results for the Pearl Index, which measures the rate of unintended pregnancies. Overall, the Pearl Index in the study was 4.80 with the upper bound of the 95% confidence interval at 6.06.

Twirla (AG200-15), which is given once weekly, combines a low dose of ethinyl estradiol with a dose of levonorgestrel, using Agile's proprietary transdermal patch technology. FDA issued a complete response letter on the application and requested a new Phase III study in 2014. (Also see "Agile SEC filing reveals FDA rejection for Twirla contraceptive " - Scrip, 19 Mar, 2014.) The company plans to resubmit an NDA in the first half of 2017.

Agile's release stirred doubts in the market about whether FDA will accept the data – the Pearl Index was 5.76 overall, though the figure was better in compliant patients at 3.1.

In the pivotal study of Teva Pharmaceutical Industries Ltd.'s dose-ascending, extended-regimen oral contraceptive Quartette (levonorgestrel/ethinyl estradiol) the Pearl Index was 2.92 (95% CI, 2.26-3.72). (Also see "Teva’s Quartette data support US filing" - Scrip, 25 Oct, 2012.)

Agile has noted there were differences in trial design, suggesting that direct comparisons are not appropriate, but that point of view is not proving to be persuasive.

"These are disappointing results for Twirla. The intention-to-treat population had a PI of 4.80 (95% CI upper bound of 6.06) which does not compare well to historical data for Quartette (PI of 2.92; 95% CI, 2.26-3.72). Quartette and Twirla were both tested in a wide variety of patients including the overweight and obese," Biomedtracker analyst David Dahan commented.

"Most contraceptives were approved on a PI below 2 and when Quartette was approved it was a signal that the FDA was willing to accept a lower efficacy in return for the health benefits of a low-dose contraceptive pill. It is questionable whether the FDA will accept further reductions in efficacy in return for the increased compliance with a low-dose contraceptive patch," Dahan added.

Agile pointed out during a Jan. 3 investor call that the design of SECURE was stricter than studies of other drugs. For example, cycles in which there was no sexual activity were excluded in SECURE, which was not the case with Quartette. Use of backup contraception was excluded in both development programs.

"Although the FDA will take this into account, we think it unlikely that such variabilities provide a sufficient explanation for the higher PI values reported here," Dahan commented.

Another aspect of the SECURE data that disappointed was the 51.4% discontinuation rate, whereas the dropout rate for Quartette was 40.3%.

Agile's stock price plunged from $5.77 to $1.33 on Jan. 3, rebounding a bit on Jan. 4 to a close of $2.63 and falling further in after-hours trading to $2.37.

On the safety front, Agile reported that the product was well tolerated with deep vein thrombosis and pulmonary embolism rates in line with other hormonal contraceptives.

Execs noted during the call that it was great news that "no woman in our trials so far who is of normal BMI had had a clot event that was considered potentially related to the drug."

Furthermore, patch-related irritation and itching rates were low, the company reported.

Per FDA, "Stringent" Trial Design

In reporting the results, the company noted that the study "was designed in consultation with the FDA, and comprised a number of stringent trial design elements, including exclusion of treatment cycles not only for use of back-up contraception but also for lack of sexual activity."

"SECURE had broad entry criteria, placed no limitations on BMI [body mass index] or other demographic factors during enrollment, and enrolled a large and diverse population from the United States in order to allow for efficacy to be assessed across different groups, as requested by the FDA. These entry criteria resulted in the inclusion of a substantial number of women with high BMI, who have frequently been under-represented in past contraceptive studies," the company said in a statement.

During the investor call, Agile noted that the relationship between higher BMI and lower efficacy has also been recognized for Teva's Quartette and Janssen Pharmaceutical Cos.' Ortho Evra (ethinyl estradiol/norelgestromin) contraceptive patch.

A number of FDA divisions are recognizing that efficacy of drugs may differ for people who are obese – yet the agency does not want to deny access to contraceptives for obese women. Its approach has been to stress compliance for these patients, according to the company. If Twirla is approved, Agile does not expect restrictions based on BMI in labeling, rather it expects that the implications of weight could be addressed in labeling language, enabling a dialogue between women and their providers.

Of the total population in the SECURE study, 35.3% were obese and the data show that efficacy was reduced as weight increased (see table).

Execs said that the company now has a robust dataset that is much more representative of the US population, "so that when doctors and women sit down and have those very personal conservations about the absence of safety and efficacy and lifestyles they can have those discussions armed with data as opposed to a lack of data."

Dahan noted that with SECURE, Agile achieved FDA objectives when it came to the diversity of the patient population and that the study is the first to contraceptive trial to report results for BMI subgroups.

"A meta-analysis completed by the FDA in 2015 showed that obesity was associated with reduced effectiveness of hormone contraceptives and reported PI values of 2.53 (95% CI, 1.88-3.41) and 3.14 (95% CI, 2.33-4.22) for non-obese (BMI <30) and obese (BMI >30) women, respectively. While SECURE reproduces this effect of obesity on contraceptive efficacy, the PI values are higher than expected for both the non-obese and obese subgroups," he said.

SECURE Efficacy Results, By Body Weight
BMI Category	% Trial population	Pearl Index	Upper bound of 95% confidence interval
Normal	39%	3.03	4.62
Overweight	25%	5.36	7.98
Obese	35%	6.42	8.88