Fallout from failed clinical trials poses challenges to companies

So far in 2015, four different ophthalmic products in phase 3 clinical trials failed to meet their primary endpoints. Ocular Surgery News spoke with three of these companies to find out what happened and how they responded. Their stories offer a telling glimpse at the challenges and opportunities companies face when endpoints are missed in clinical trials as part of the overall innovation cycle.

“Private companies often simply learn from their experience and make appropriate adjustments to their research and business plan. They continue on as long as their investors are optimistic they will eventually create value,” OSN Chief Medical Editor Richard L. Lindstrom, MD, said. “Public companies must disclose any material event in a timely fashion and usually see a significant fall in their share price, usually in the 20% to 50% range, but it can be even more if it is an FDA trial not approved by a panel.”

As for how companies regroup after failure, “they study the data, learn, negotiate with the FDA and move forward, usually with an adjustment in the clinical trial plan and additional studies,” according to Lindstrom, who over the past 35 years has been an angel investor, consultant and/or medical advisory board member for more than 60 companies, mostly in the ophthalmic field. He currently serves on the board of directors for three publicly traded ophthalmic companies and six private ophthalmic businesses. Lindstrom also serves as a board member for three private non-ophthalmic entities.

Image: Shari Fleming Photography

Failure to meet endpoints

Sometimes with failure, the product or the indication is changed, Lindstrom said. “Basically, one does R&D and learns every day, applying those learnings and iterating the product forward to eventual regulatory approval and commercialization. Even after commercialization, learnings continue, and one adjusts the product indications and method of use,” he said.

In a privately funded company, much of this revamping occurs under the radar, according to Lindstrom. But in a publicly traded company, all is disclosed and subjected to broad public evaluation, “with resultant large swings in share price.”

Take the example of Ocular Therapeutix, whose stock fell 30% within a few days after the company announced in April that its second of two phase 3 studies of Dextenza, a sustained-release dexamethasone intracanalicular depot in development for the treatment of post-surgical ocular inflammation and pain, failed to meet one of its two primary endpoints: reduction of ocular inflammation after cataract surgery. As of mid-July, the stock had rebounded some, due to communicating a clear path forward.

“Dextenza met the endpoint for ocular pain in both of our phase 3 studies and our phase 2 study; however, due to an unusually high placebo response rate, there was not a statistically significant absence of inflammatory cells in the anterior chamber compared to a placebo vehicle control punctum plug in this second phase 3 study,” Amar Sawhney, PhD, chairman, president and CEO of Ocular Therapeutix, said, noting that the inflammation endpoint was attained in the other two trials.

After the depot is inserted, Dextenza elutes drug over a 4-week period. The endpoint was day 8 for pain and day 14 for inflammation.

Sawhney said that the study’s shortcoming in treating inflammation “was not a big miss, a matter of several patients” in the 240-patient trial. Still, “we were disappointed that we did not meet both endpoints.”

Some patients were rescued prematurely, before the designated time for evaluating the two endpoints.

“Although this was technically allowed, if you do not have a justifiably high level of inflammation or pain to rescue the patient, rescue should not take place. But if you do, these patients are automatically counted as failures.” Sawhney said. “We believe this reduced the power of our study.”

Eleven Biotherapeutics also had a miss when it announced in May that its pivotal phase 3 OASIS study designed to assess the safety and efficacy of its lead drug candidate, EBI-005, in patients with moderate to severe dry eye disease failed to meet either of the co-primary endpoints: a reduction in the mean score for total corneal fluorescein staining and a reduction in ocular pain and discomfort. The 699-patient study evaluated the mean change from baseline to week 12 with EBI-005, a protein therapeutic topically administered three times daily for 12 weeks. The study compared EBI-005 with a vehicle control for total corneal fluorescein staining score and patient-reported measurement for ocular pain and discomfort.

“We were disappointed that our phase 3 study, which was designed and powered based on our phase 2 study results, did not meet the co-primary efficacy endpoints,” Abbie C. Celniker, PhD, president and CEO of Eleven Biotherapeutics, said. “But we were encouraged that we continue to see a favorable tolerability profile for EBI-005 as we pursue additional plans to advance EBI-005 into a phase 3 study for allergic conjunctivitis.”

EBI-005 is “formulated in a vehicle that is comfortable to the eye and maintains stability of the protein,” Celniker said. The control group received the vehicle alone.

Aerie Pharmaceutical failed to reach clinical study endpoints in its Rocket 1 phase 3 study of 370 patients with open-angle glaucoma or ocular hypertension. In total, 182 patients received the eye drop Rhopressa 0.02%, dosed once daily, and 188 were given timolol 0.5%, dosed twice daily. Duration of therapy for both groups was 90 days. The primary efficacy endpoint was non-inferiority to timolol.

In early phase 3 results, released in April, “we were surprised that Rhopressa did not meet the primary endpoint of IOP less than 27 mm Hg, since the eye drop had shown in phase 2 trials that it was effective at even higher baseline IOPs,” Vicente Anido Jr., PhD, CEO and chairman of the board for Aerie, told Ocular Surgery News.

Upon analyzing the data, the company found a number of issues associated with patients enrolled in the uppermost 1 mm Hg cohort (baseline IOP of less than 27 mm Hg and 26 mm Hg or greater).

“We identified quite a few discrepancies in IOPs and a significant number of compliance issues in this small group, so we were able to determine that influences unrelated to Rhopressa kept us from reaching significance on our primary endpoint,” Anido said.

Aerie was pleased, however, that Rhopressa demonstrated non-inferiority to timolol at IOP below 26 mm Hg, 25 mm Hg, 24 mm Hg and 23 mm Hg; IOP less than 24 mm Hg was a prespecified secondary endpoint in the trial. Additionally, Rhopressa showed numerical superiority to timolol at four of nine time points for baseline pressures less than 26 mm Hg, seven of nine time points for pressures less than 25 mm Hg and all nine time points for pressures below 24 mm Hg.

Ohr Pharmaceutical announced in March that it failed to meet its primary endpoint in a phase 2 study of 70 patients. No meaningful difference was found in the mean number of injections of OHR-102 (0.2% squalamine lactate ophthalmic solution) plus as-needed Lucentis (ranibizumab, Genentech) compared with as-needed Lucentis monotherapy when treating wet age-related macular degeneration, as previously reported in Ocular Surgery News. As a result, the company’s stock dropped nearly 70% in 1 day. The company declined an interview request to discuss the result.

Recovering from disappointing results

Among the lessons that can be learned from a failed clinical trial is “how a product works or does not work, who are the best patients, and what are the best techniques and methods of use,” Lindstrom said. “These learnings continue ad infinitum. For example, we are still learning ways to perform phacoemulsification and IOL surgery better, even after 100 million-plus procedures and 30-plus years of global use.”

Failure tarnishes the images of these companies temporarily.

“The critical issue, however, is response to ‘failure,’” Lindstrom said. “Temporary ‘failure’ is how we learn. No innovation happens without failure: learning, learning applied, failure again, more learning, learning applied, failure again, and again and again.”

“As is often told to children, it is OK to get knocked down,” Lindstrom said. “However, it is the person or company with character that gets back up and tries again.”

In retrospect, Sawhney wishes Ocular Therapeutix’s trial had been designed differently.

“However, hindsight is always 20/20,” he said. “A few criteria could have been modified that might have made a real difference.” This includes inclusion and exclusion patient criteria, such as prohibiting systemic NSAID use, and trial execution, such as ensuring site-to-site adherence to study protocol, for example.

Going forward, Sawhney does not believe Dextenza itself needs to be redesigned in any fashion. “But we are modifying the design of the next phase 3 study,” he said.

In April, the company met with the FDA to share study results and seek the agency’s advice, including discussions regarding filing for a pain-only indication. “This is what we are doing — in other words, separating the two indications of pain and inflammation,” Sawhney said.

A third phase 3 study for evaluating both pain and inflammation is expected to recruit later this year and complete late next year.

“However, the primary endpoint of the new trial will be inflammation,” Sawhney said. “We are optimistic that we can achieve this endpoint.”

Among the lessons learned is that “we should ensure we are not taking anything for granted and that we pay close attention to the design criteria for our next clinical study, as well as its execution,” Sawhney said. “It is important that training for investigators is appropriate and that the conduct of the trial is closely monitored.”

For studies in general, a drug that fails to perform is much more problematic than a flaw in clinical design, according to Sawhney. The latter may result in a delay in time and money, “but it is not a knockout punch,” he said. Ocular Therapeutix is incorporating only FDA-approved molecules and proven mechanisms of action.

“Dexamethasone is a well-known steroid that has previously been shown to have an effect on ocular inflammation,” Sawhney said.

Similarly, if Aerie Pharmaceutical were able to redesign the phase 3 trial, “knowing what we now know, we would try to overcome potential problems seen in the original trial in the uppermost 1 mm Hg baseline cohort by enrolling patients with starting pressures up to 30 mm Hg, but confining our analysis to those under 27 mm Hg,” Anido said.

Because patients are still being treated in a second phase 3 study, Rocket 2, for which the database has not been locked, the FDA has allowed Aerie to adjust the primary endpoint analysis to include subjects with baseline IOP of above 20 mm Hg and to below 25 mm Hg.

“The trial is still adequately powered with this modification, and no additional patient enrollment will be necessary,” Anido said. “Thus, what we learned in Rocket 1 has guided study design adjustments in Rocket 2 that increase the chances of that trial, and the Rhopressa program, being successful.”

The best lesson Aerie has learned is that “our team can move very quickly,” Anido said. “We unmasked Rocket 1 data at the end of April and had already achieved resolution on a path forward for the Rhopressa program, with the FDA granting our request to adjust the primary endpoint analysis of Rocket 2 by the end of June.”

Rocket 1 data revealed that previous prostaglandin use enhanced the IOP-lowering effect of Rhopressa, showing a “unique synergistic effect with prostaglandins that is not present with timolol,” Anido said in a company-sponsored conference call to report findings from the Rocket 1 program, Healio.com/OSN previously reported.

Aerie’s stock price fell after the initial announcement of the failed primary endpoint, dropping by about 70%.

“However, based on the significant progress we have made since, our valuation is recovering nicely and our consensus rating from stock analysts covering the company is a ‘buy,’” Anido said. And from a clinical development standpoint, “we do not anticipate that our regulatory timelines for Rhopressa will be lengthened or delayed.”

If the Rocket 2 trial is successful, the company anticipates filing a new drug application for Rhopressa in the second half of 2016.

“Our strong working relationship with the FDA allowed us to reach agreement rapidly about the path forward for Rhopressa,” Anido said. “We also have collected important information that gives us the confidence to initiate the phase 3 program for quadruple-action Roclatan (a fixed-dose combination of Rhopressa and latanoprost).”

Unlike Ocular Therapeutix and Aerie Pharmaceutical, Celniker said Eleven Biotherapeutics would not have designed its trial differently.

“We believe the study was well-powered, well-designed and well-executed and the endpoints appropriately based on the clinical results from our phase 1b/2a trial,” she said. “This study provided a clear answer on time and on budget.”

The trial results also “reinforced our belief that drug development is challenging, and early results may not always translate to late-phase results,” Celniker said. “This is why it is critical to have a broad technology platform and portfolio of product candidates in order to create value beyond a single drug asset. Our company was well-positioned prior to initiating this phase 3 dry eye trial to evaluate other opportunities with EBI-005, for example, allergic conjunctivitis.” Eleven Biotherapeutics also has new molecules in the pipeline, including EBI-031 for a potential diabetic macular edema treatment.

Despite the setback, the company “is still viewed as a highly talented team of individuals with sufficient resources to fund our near-term plans,” Celniker said.

The phase 3 study of EBI-005 for allergic conjunctivitis is expected to start by the end of the year. “Allergic conjunctivitis is a mechanistically different ocular disease from dry eye disease because in allergic conjunctivitis there is a single central mechanism of allergen stimulation which initiates the ocular surface inflammation, as opposed to dry eye disease which has diverse underlying causes of the disease,” Celniker said.

Results from a phase 2 study of EBI-005 for treating allergic conjunctivitis were released late last year and “demonstrated statistically significant improvements in ocular itching, tearing and nasal symptoms in the late-phase allergy response in patients exposed to allergen,” Celniker said.

Eleven Biotherapeutics also plans to move its novel interleukin-6 inhibitor for diabetic macular edema into investigational new drug studies later this year.

“Our strategy remains steadfast to develop novel protein therapeutics to address unmet needs of patients suffering from debilitating ocular conditions,” Celniker said. “We continue to believe there is a significant untapped opportunity to apply protein therapeutics for ophthalmic diseases by using proteins to modulate inflammatory cytokines, which are key targets in ocular diseases that have been difficult to access with conventional therapeutic approaches.”

Although failed clinical studies eventually impact the ability of companies to raise money, “all sophisticated investors plan on failures with their associated learnings,” Lindstrom said, noting that venture capitalists plan on at least three rounds of investing often totaling more than $100 million and 10 or more years to bring a new device to market, whereas a breakthrough drug can cost anywhere from $500 million to $2 billion.

“No meaningful innovation, especially disruptive innovation, occurs without repeated failure, each followed by learning and an appropriate response,” Lindstrom said. Although this process is “tougher on everyone in the fishbowl of a public company, the public marketplace often provides the lowest-cost access to capital, so it is a necessary evil.” – by Bob Kronemyer

• For more information:
• Vicente Anido Jr., PhD, can be reached at Aerie Pharmaceutical, 2030 Main St., Suite 1500, Irvine, CA 92614; email: vanido@aeriepharma.com.
• Abbie C. Celniker, PhD, can be reached at Eleven Biotherapeutics, 215 First St., Suite 400, Cambridge, MA 02142; email: abbie.celniker@elevenbio.com.
• Richard L. Lindstrom, MD, can be reached at Minnesota Eye Consultants, 9801 DuPont Ave. South, Suite 200, Bloomington, MN 55431; email: rllindstrom@mneye.com.
• Amar Sawhney, PhD, can be reached at Ocular Therapeutix, 34 Crosby Drive, Bedford, MA 01730; email: asawhney@ocutx.com.

Disclosures: Anido is CEO and chairman of the board of Aerie Pharmaceutical. Celniker is president and CEO of Eleven Biotherapeutics. Lindstrom reports he has a financial interest in Alcon Laboratories, Abbott Medical Optics, Bausch + Lomb, Aerie Pharmaceutical and Ocular Therapeutix. Sawhney is chairman, president and CEO of Ocular Therapeutix.

What are the advantages and disadvantages of adaptive design clinical trials?

Adaptive design may enhance certain trials

Randomized clinical trials provide the highest level of evidence-based medicine. The randomization process seeks to create balanced treatment groups that differ only in their assigned treatment. Randomized trials maximize study validity and minimize bias.

Clinical trials are hypothesis-driven and address specific questions. The endpoint of the trial should be clinically meaningful and is established during its design. Changing outcome measures during the course of a clinical trial is not appropriate and is expected to introduce significant bias into the study.

In select cases, though, an adaptive design may enhance a clinical trial. What happens when one treatment begins to look superior to the other as data accrues? A data and safety monitoring committee would generally decide to stop enrollment in the trial, if the treatment benefit or risk is so high that randomization is deemed unethical. Formal statistical testing is used to guide this decision.

However, adaptive design is an alternative that has been used. For this approach, randomization is adjusted so that new participants are more likely to be assigned to the treatment that appears superior at the time they are recruited. The FDA advises, “Adaptive randomization should be used cautiously in adequate and well-controlled studies, as the analysis is not as easily interpretable as when fixed randomization probabilities are used.”

Adaptive design has been used in less controversial areas of clinical trials. In the nonrandomized phase 1 dose escalation study, the dose received by the next patient is determined by the incidence and severity of adverse events in previously treated subjects. Another use of adaptive design is to adjust the mix of covariates between treatment groups, such as substantial gender or race imbalances.

I would like to acknowledge William J. Feuer, MS, a senior scientist and biostatistician at the Bascom Palmer Eye Institute, who assisted with this comment.

• Reference:
• Brannath W, et al. J Biopharm Stat. 2010;doi:10.1080/10543406.2010.514453.

Steven J. Gedde, MD, OSN Glaucoma Board Member, is a professor of ophthalmology and the John G. Clarkson Chair at Bascom Palmer Eye Institute. Disclosure: Gedde reports no relevant financial disclosures.

Better guidelines need to be established

In traditional clinical trials, key elements such as primary endpoint, clinically meaningful treatment difference and measure of variability are prespecified during planning of the design of the study and are not altered during the course of the trial. The success of the trial then depends on the accuracy of the original assumptions. In adaptive design clinical trials, the study design allows planned adaptations or modifications to the trial after its initiation, based on accumulated data, without undermining the validity and integrity of the trial.

Adaptive design methods are attractive due to their flexibility to change various aspects of the trial (eg, drug dosage, treatment arms, superiority vs. non-inferiority, sample size, endpoint, inclusion/exclusion criteria), which can result in more efficient therapy development. Efficiencies can include a smaller sample size, a shorter treatment development process and possibly lower trial costs. Proper use of adaptive designs can increase the chance of correctly answering the clinical question of interest and the probability of success in clinical trial due to its flexibility.

However, improper adaptation can lead to a biased study. Therefore, proving lack of bias and advantages of using adaptive designs may require extensive planning and validation. Other disadvantages to the use of adaptive designs include concerns about whether the P value or confidence interval regarding the treatment effect obtained after the modification is reliable or correct; in addition, the use of adaptive designs may lead to a totally different trial that is unable to address scientific questions that the trial is intended to answer. Also, there is a possibility that the actual patient population after the adaptations could deviate from the originally targeted patient population, and the overall type I error (to erroneously claim efficacy for an ineffective drug) rate may not be controlled.

The use of adaptive designs in clinical trials makes statistical analyses more complicated and interpretation of trial results more difficult; hence, guidelines regarding the use of adaptive designs need to be better established so that appropriate statistical methods and statistical software packages can be developed accordingly. Finally, shorter trial durations allowed with adaptive designs may not reveal longer-term risks or benefits of a drug or therapy.

Judy E. Kim, MD, OSN Retina/Vitreous Board Member, is a professor of ophthalmology at Medical College of Wisconsin, Milwaukee. Disclosure: Kim reports no relevant financial disclosures.