Special Protocol Assessments: A Case Review for 2012 Regulatory Outcomes

A great deal of interest surrounds companies that have an active phase 3 clinical program underway and the assessment of their future potential regulatory risk when subsequent new drug applications (NDA) are reviewed. One measure which is designed to mitigate some regulatory risk is the Special Protocol Assessment (SPA) agreement between a sponsor and FDA, who first issued guidance on the process in 2002.  This guidance is not set to expire until 2/28/2014.

Protocols that can be submitted for SPA include animal carcinogenicity studies, final product stability studies and clinical protocols for phase 3 trials whose data will form the primary basis for an efficacy claim if the trials had been the subject of discussion at an end-of-phase 2/pre-phase 3 meeting with the review division, or in some cases, if the division agrees to such a review because the division is aware of the developmental context in which the protocol is being reviewed and the questions are being answered. The clinical protocols for phase 3 trials can relate to efficacy claims that will be part of an original new drug application or biologics license application (BLA) or that will be part of an efficacy supplement to an approved NDA or BLA clinical protocols for phase 3 registration studies.  For the purposes of this article, we will focus exclusively on the latter protocol type.

What does a SPA mean for regulatory risk?

A SPA is a written agreement between the FDA and sponsor, that a phase 3 study is conducted according the agreed upon protocol design, size, study end points and outcomes, will be sufficient from a regulatory stand point to file for an efficacy claim in a marketing application.  A SPA does not guarantee that a trial will be successful, nor does it guarantee regulatory success. However, if a trial is successfully conducted according to a SPA agreement, meeting all pre-specified efficacy end points, it is very unlikely that FDA would reject a marketing application on the basis of requiring more clinical information.   In essence, a SPA is measure to mitigate clinical efficacy risk, it does not however mitigate issues pertaining to preclinical toxicology, manufacturing, and other clinical safety issues that can result in a marketing application receiving a complete response letter.

How does a sponsor obtain a SPA?

Typically a sponsor must submit a written request to either Center for Drug Evaluation and Research (CDER) or Center for Biologics Evaluation and Research (CBER) detailing the goals of the protocol and overall development context at least 90 days prior to the projected start of study, following an end of phase 2 meeting.  The Agency will then review the protocol and provide comments related to protocol design, study conduct and execution, data analysis, and implications for labeling within 45 calendar days of the receipt of the initial request. The Agency's assessment will be based primarily on the questions posed by the sponsor, the underlying data, assumptions, information described by the sponsor, and relevant Agency policies and guidance documents. It is quite common for the Agency and sponsor to discuss revisions and changes to the protocol prior to the issuance of a formal response via a Type A meeting.

Type A meetings can be requested via a formal process after an applicant has submitted the initial request for SPA and should be scheduled to occur within 30 days of receipt of request for such a meeting. Typically the CBER or CDER division director who receives a meeting request will determine whether to hold the meeting and will respond to the sponsor or applicant by granting or denying the meeting within 14 days of receipt of the request for Type A meeting.  Sponsors will provide the Agency will specific objectives, goals and review materials prior to the meeting. During the live discussion, and prior to the conclusion both sides will summarize the important discussion points, agreements, clarifications, and action items.  FDA is the official record keeper of the meeting and will issue finalized minutes to all attendees within 30 days of the meeting. If new issues require clarifications that were not discussed at the initial meeting, this necessitates a new Type A request.

Following the receipt of the minutes, a sponsor can submit a revised protocol at which point the agency will consider this as a new SPA request and will act on the revised version within the prescribed 45 day timeline. Any subsequent amendments to the agreed upon protocol will render the SPA invalid, unless the agency has a chance to review and approve any such changes.

Hence, in a best case scenario, the process to obtain a SPA could take as little as 45 days, but when Type A meetings and revisions are required, it can take upwards of 120 days to reach mutual agreement.

What has been the regulatory success rate for SPA applicants that submitted an NDA or BLA in 2012?

In my estimation, there have been approximately 114 PDUFA decisions relating to NDA/BLA and sNDA/sBLA’s in 2012.  Of these, 97 applicants received drug approvals and 17 received complete response letters (CRL).

For the cohort of approvals, n=5 applicants submitted clinical data for studies that were completed pursuant to a SPA agreement.  Here is case summary for their regulatory reviews.

Protalix BioTherapeutics/Pfizer: Taliglucerase Alfa for Type 1 Gaucher disease

Protalix originally reached a SPA agreement for a pivotal study in 2007.  This was a randomized double blind study comparing two doses of active therapy in N= 32 untreated patients with Gaucher disease. The primary end point was change from baseline in spleen volume measured by MRI at 9 months. Secondary outcome measures were changes from baseline in liver volume, hemoglobin and platelet counts. The company reported positive topline results in October 2009 with the trial meeting its primary and secondary end points in both dose groups. Importantly, the primary end point was met at 6 months and maintained at 9 months. The drug received both Fast Track designation and Orphan Drug designation prior to the submission of the NDA in December of 2009. In February 2010, the company disclosed that FDA had requested further information pertaining to the Chemistry, Manufacturing and Controls (CMC) section of the NDA and that no further clinical data was requested.   In July, the NDA was formally accepted and a PDUFA goal of February 25, 2011 was set under standard review.   On the PDUFA date, the company subsequently received a complete response letter citing unresolved questions.  The main questions related to the CMC section of the NDA and the agency also requested additional data from switchover and long-term extension studies (which were immature at the time of original filing).   A response to the CRL was submitted in August, 2011, which was accepted as a Class 2 resubmission, for 6-month review. Further select data presentations were requested by the agency, resulting in a 3-month extension to the review and a PDUFA goal of May 1, 2012.  FDA approved ELELYSO(TM) (taliglucerase alfa) for injection as an enzyme replacement therapy for the long-term treatment of adults with a confirmed diagnosis of type 1 Gaucher disease on that date.

Vivus: Qsymia for Obesity

Like Protalix, Vivus successfully negotiated a SPA agreement for key elements of the pivotal phase 3 clinical trials of Qsymia for the treatment of obesity and weight-related co-morbidities in 2007. The phase 3  program would include two pivotal, double blind, placebo-controlled, multi-center studies in distinct populations (EQUIP and CONQUER) comparing Qsymia to placebo during a 56-week treatment period. The program would also include a six-month confirmatory factorial study, known as EQUATE in obese subjects with BMI's from 30 to 45. This trial will evaluate two dose levels of Qsymia, compared to both placebo and the individual constituents of the combination. The co-primary endpoints for these studies would be to evaluate the differences between treatments in mean percent weight loss from baseline to the end of the treatment period, and the differences between treatments in the percentage of subjects achieving weight loss of 5% or more. All 3 trials were initiated in 2007. The company reported positive results from EQUATE in December 2008, and highly positive results from EQUIP and CONQUER in September 2009. These studies met all primary endpoints by demonstrating statistically significant weight loss with all three doses of Qsymia, as compared to placebo. Patients taking Qsymia also achieved significant improvements in cardiovascular and metabolic risk factors including blood pressure, lipid levels, and type 2 diabetes. The company submitted an NDA in December 2009 which was accepted for a standard review and PDUFA goal of October 28, 2010. The NDA was reviewed by a Endocrinologic and Metabolic Drugs Advisory Committee on July 15, 2010. However the Committee voted 10:6 against approval citing concerns with the adverse events observed with treatment, including depression, anxiety, sleep disorders, other cognitive disorders, metabolic acidosis, increased heart rate and an unclear potential for teratogenicity in both animals and humans. Subsequently a CRL was issued at the PDUFA goal date.

A number of meetings and communications took place between the sponsor and agency between December 2010 and September 2011. A new route to resubmission was agreed upon seeking initial approval for an indication restricted to obese men and women of non-child bearing potential, until further retrospective studies could be completed to fully evaluate teratogenic risk. Vivus resubmitted the NDA with a proposed contraindication in women of childbearing potential in October 2011, which was accepted as a class 2 resubmission, 6 month review cycle and a PDUFA goal was set for April 17, 2012.  The sponsor was notified that the resubmitted NDA would again be reviewed by an Advisory Committee. In January 2012, FDA requested that Vivus remove the Qsymia contraindication for women of childbearing potential contained in the proposed label. Qsymia would remain contraindicated for women who are pregnant. On February 22, 2012, the Endocrinologic and Metabolic Drugs Advisory Committee voted 20 to 2 in favor of approval for the drug contingent on an adequate Risk Evaluation and Mitigation Strategy (REMS) and with post-marketing safety studies.  Finalization of the REMS program necessitated a 3-month extension for the PDUFA goal taking it to to July 17, 2012. Qsymia was finally approved as an adjunct to a reduced-calorie diet and increased physical activity for chronic weight management in adult patients with an initial body mass index (BMI) of 30 or greater (obese), or 27 or greater (overweight) in the presence of at least one weight-related comorbidity, such as hypertension, type 2 diabetes mellitus or high cholesterol (dyslipidemia) on July 17, 2012.

 

Amarin: Vascepa for Very High Triglycerides

Amarin reached agreement with the FDA on a SPA for a pivotal registration study for Vascepa in 2009. The Phase 3 trial would be a multi-center, placebo-controlled, randomized, double-blind, 12-week study to evaluate the efficacy and safety of two doses of AMR101(Vascepa), a prescription grade Omega-3 fatty acid, in patients with fasting triglyceride levels of ≥500 mg/dL (the AMR101 MARINE Study). The primary endpoint in the trial was the percentage change in triglyceride level from baseline to week 12. Amarin also conducted a second phase 3 study- ANCHOR as a treatment for high triglycerides (≥200 and <500mg/dL) in 702 patients with mixed dyslipidemia (two or more lipid disorders) on background statin therapy at LDL-C goal who were at high risk of cardiovascular disease.  In November 2010, the company reported highly positive results for MARINE with the primary end point being met for both the 4 gram and 2 gram dose groups. All secondary end points were also met and the drug demonstrated a positive safety profile. The company reported highly positive results for the ANCHOR study in April 2011. The primary endpoint for triglyceride change was achieved at both 4 grams and 2 grams per day with median placebo-adjusted reductions in triglyceride levels of 21.5% and 10.1% for the 4 grams and 2 grams per day dose groups, respectively. Moreover, for the 4 grams per day AMR101 group, LDL-C decreased significantly by 6.2% from baseline versus placebo, demonstrating superiority over placebo.  

Amarin submitted an NDA on September 26, 2011, which was subsequently accepted for standard review with a PDUFA goal of July 26, 2012. On that date, the company announced that FDA had approved Vascepa  as an adjunct to diet to reduce triglyceride in adult patients with severe hypertriglyceridemia (TG greater than or equal to 500mg/dL).

 

 

Abraxis Biosciences/Celgene: Abraxane for the Non-Small Cell Lung Cancer (NSCLC)

Abraxis reached a SPA agreement with the FDA on the design of a pivotal study for the use of Abraxane in the first-line treatment of NSCLC in 2007. The Phase III pivotal trial was a randomized, open-label trial comparing weekly 100 mg/m² Abraxane (days 1, 8 and 15 of each cycle) and 200 mg/m² paclitaxel every three weeks. Carboplatin would be administered at AUC=6 on day 1 of each cycle repeated every three weeks in both treatment arms. The study would enroll approximately 1,000 patients with Stage IIIb and IV non-small cell lung cancer. The primary endpoint of the study was overall response rate (ORR) and final positive study results were presented by Celgene at ASCO in June 2011. The study met its primary end point in significantly improving ORR for patients receiving Abraxane in combination with carboplatin. An sNDA was submitted for this indication in December 2011, and was accepted for standard review with a PDUFA date of October 12, 2012. On that date, FDA approved Abraxane for the first-line treatment of locally advanced or metastatic non-small cell lung cancer, in combination with carboplatin, in patients who are not candidates for curative surgery or radiation therapy.

 

Napo/Salix: Fulyzaq for HIV-Associated Diarrhea

Napo reached agreement with the FDA on the SPA process for Crofelemer in 2007. The ADVENT trial -- (Anti-Diarrhea therapy in HIV disease-Emerging treatmeNT concepts) was a randomized, double-blind, parallel-group, placebo- controlled, two-stage, adaptive design study to assess the efficacy and safety of crofelemer at 125mg, 250 mg and 500 mg oral doses twice daily for the treatment of HIV-Associated Diarrhea. The primary outcome measures were to determine the proportion of HIV-positive subjects experiencing relief of diarrhea with crofelemer compared to placebo during the placebo-controlled treatment phase of 28 days, and to determine how many patients experienced two or less watery bowel movements per week during at least two weeks of the treatment phase.  Positive results for the ADVENT study were reported by Salix in November 2010, with the trial meeting its primary end points, forming the basis of an NDA submission in December 2011. The submission was granted a Priority review status and a PDUFA goal date of June 5, 2012. In April 2012, the company was notified by the Agency that additional review time would be required, prompting a 3-month extension to the PDUFA date, reset to September 5, 2012.  On September 5, 2012 the company issued a release indicating that the NDA was still under review and that final action had not been taken by the Agency. “By taking no action at this time, the FDA has allowed for the currently ongoing dialogue between Salix and the FDA to continue…The primary topic is the production and control of the crofelemer active pharmaceutical ingredient, a complex mixture that is the first botanical product to be reviewed by the Agency for oral use. This focus is needed to ensure compliance with the manufacturing and product quality requirements of the Food, Drug & Cosmetic Act”.   On December 31, 2012, FDA approved Fulyzaq as an anti-diarrheal indicated for the symptomatic relief of non-infectious diarrhea in adult patients with HIV/AIDS on anti-retroviral therapy.

I have documented 17 complete response letters received in 2012, and in all cases bar 1, the FDA was seeking further clinical information or confirmatory studies prior to approval.  Only 1 applicant receiving a CRL in 2012 had a SPA agreement in place so it is worth revisiting this case.

Ariad/Merck:  Ridaforolimus for Advanced Soft Tissue Sarcoma

Ariad initially requested a SPA in March 2007 and finally reached agreement on a pivotal study design in August 2007.  The Phase 3 randomized, placebo-controlled, double-blind SUCCEED (Sarcoma Multi-Center Clinical Evaluation of the Efficacy of Ridaforolimus) study examined the use of oral ridaforolimus as maintenance therapy in patients with metastatic soft-tissue or bone sarcomas who previously had a favorable response to chemotherapy. The primary end point was progression free survival (PFS) with overall survival (OS) and other measures of tumor response as secondary outcome measures. Importantly, in the agreed to SPA protocol, the applicant estimated that the median time to PFS was 6-9 months and projected a 25% improvement in PFS. This translated into an increase in median PFS from 6 to 8 months or from 9 to 12 months with ridaforolimus. In communications to the sponsor, FDA stated that “Of note, PFS is not proven to be a surrogate for survival in this disease setting. A statistically significant difference in PFS may not necessarily represent a clinically meaningful difference. Whether PFS supports approval will be a review issue and would depend on the magnitude of the improvement and the risk/benefit ratio.” In January 2011, Ariad reported positive results from the SUCCEED study. However, on close examination of the results, ridaforolimus treatment resulted in a statistically significant 21% (3.1 week) improvement in median PFS (ridaforolimus, 17.7 weeks vs. placebo, 14.6 weeks). This was short of the 25% projected improvement in the SPA agreement, and the absolute magnitude of PFS benefit was only 3 weeks compared with the projected 2-3 month benefit assumption, with no significant improvement in overall survival. Nonetheless, Merck submitted an NDA in August 2011, which was accepted for standard review in October, 2011. The NDA was reviewed by the Oncologic Drugs Advisory Committee on March 20, 2012, resulting in a vote of 13:1 against approval. The committee voiced concerns regarding the relative efficacy toxicity profile, citing the modest PFS benefit and high rate of treatment discontinuations and serious adverse events in the context of a maintenance therapy setting. On June 5, 2012, Merck received a CRL stating that additional clinical trial(s) would need to be conducted to further assess safety and efficacy.

 

There are some key lessons that can be garnered from this analysis.  Firstly, sponsors who have successfully conducted studies that have met the predefined outcomes in a SPA agreement are highly unlikely to be rejected on the grounds that more clinical data/studies are required.  In 2012, applicants with successful  SPA studies forming the basis of a submission had a 100% approval rate (5/5). Unlike conventional sponsors, there is de minimis risk that they will be required to perform a further confirmatory study to gain approval.  It is extremely important to understand the study design and underlying assumptions that have been agreed upon when determining if a study has in fact met the conditions for the SPA or not. Statistically significant does not always equal clinically meaningful.

Secondly, a successful SPA-backed NDA does not guarantee approval on the first regulatory review cycle. Companies submitting a clinical dossier with a SPA can still receive complete response letters.  Just because a drug is efficacious, does not mean that it is safe, the overall safety risk profile is the primary focus from a regulatory perspective and must be considered when evaluating positions.

Thirdly, even the safest and most efficacious drugs can be held up if there are issues with the CMC component of the NDA.  This aspect can be particularly challenging for novel biologic agents or drugs with non-oral routes of administration.  It is extremely important to look into CMC hurdles when conducting due diligence.

Finally, the path from SPA to approval can typically take upwards of 5 years.  Investors looking to initiate positions in a company holding a SPA agreement should consider holding stocks over an extended time horizon. 

Disclosure: I have no positions in any companies mentioned in this article

Ziopharm Palifosphamide: Evaluating the prospects for the PICASSO-3 study

Small-cap biotech investors love the thrill and drama surrounding the release of phase 3 study results. ZIOPHARM Oncology is about to provide just that in the run up to publishing topline results for their PICASSO-3 study in the last week in March.

Ziopharm is developing palifosfamide for the treatment of advanced soft tissue sarcomas (STS) (a rare form of cancer affecting ~10,000 people in the US annually). Palifosfamide is an active form of an existing chemotherapy drug called ifosfamide which is frequently used in the treatment of STS. Unfortunately, ifosfamide has serious side effects when used at high doses, or when combined with other forms of chemotherapy such as doxorubicin, a drug that is considered standard of care. Palifosfamide appears to lack such toxicities and can be given safely in combination with other chemotherapies without incremental side effects, thus making it an attractive agent in the treatment of STS.

Unfortunately, in the majority of cases, STS is not cured and eventually spreads to other organs or parts of the body (metastatic disease). Once STS has spread, the prognosis for patients is very poor, and the expected survival is typically about one year. New treatments are needed that will help control the disease and improve survival outcomes for these patients.

Ziopharm began testing palifosfamide in cell models and animal models of cancer and found that it was active in combination with doxorubicin. They then tested the drug in humans in a dose finding phase 1 study where it showed promising activity, and in the phase 2 PICASSO study. The results of the phase 2 PICASSO study were presented at ASCO 2010. PICASSO was a randomized, multi-center, open-label study of patients with unresectable or metastatic soft-tissue sarcoma randomized to receive doxorubicin (standard of care) or doxorubicin plus palifosfamide. The two arms were well matched for age and tumor type. The primary end point was progression free survival (PFS) and was highly positive: median PFS was 7.8 months in the active treatment arm vs.  4.4 months for doxorubicin alone HR (0.427) [95% C.I. 0.191, 0.951]. The partial response rate for palifosfamide was 23% vs. 9% for doxorubicin.  There were no major differences in safety between the two arms indicating that the combination of palifosfamide and doxorubicin was safe.

Obviously based on these highly statistically significant data, it is hard not to be very optimistic about the prospects for a positive outcome to PICASSO-3, the phase 3 registration study. However, here are some aspects that bulls should think about before declaring this a slam dunk positive outcome.

1)      Promising phase 2 signals rarely pan out to the same extent in phase 3 studies, especially in cancer studies.

I can barely think of a single example where a drug with promising phase 2 data replicated compelling outcomes in larger phase 3 studies. Sadly, there is a reason that doxorubicin and ifosfamide have been the standard of care in STS for the past 20 years.

2)      Why would the phase 2 results not be replicated in phase 3?

In order to answer this, we need to look at differences between the two studies.

Study populations: PICASSO enrolled a mixture of patients who had previously received chemotherapy (~1/3) and those who had not received prior chemotherapy (~2/3), those who had both unresectable and metastatic disease, predominantly from US centers. All these patients were treated open-label (meaning everyone on the study knew which patients received which medications).  Open label studies can introduce a level of bias to the results. Patients who have received prior chemotherapy for metastatic disease and who have subsequently progressed, have in general, a worse prognosis than those who are naïve to treatment so this would also potentially influence the results if there was an imbalance in exposure to prior chemotherapy between the two studies.  Similarly, if phase 2 data were used for the assumptions to power phase 3 for effect size, this may lead to a mis-match as the two patient populations are quite different.

PICASSO-3 is a randomized, multi-center, multi-national, double-blind study, that has enrolled only patients with metastatic STS who have never received chemotherapy, a large proportion of whom are from outside the US.  PICASSO was predominantly conducted in the US, however PICASSO-3 is a multi-national study enrolling patients from areas of the world where access to care and population health in general may influence survival outcomes. This is also a consideration when trying to extrapolate results between phase 2 and 3. Just based on the mix of 1^st line and 2^nd line patients in PICASSO, I would anticipate that the median PFS in both arms of PICASSO-3 would be longer as this group is all treatment-naïve, and they more recently may have had exposure to newer chemotherapies in a neoadjuvant/adjuvant  setting when they initially underwent surgery to remove their tumors. This too will influence the natural history of their disease and you would potentially expect them to live slightly longer.

3)      The control arm “did better than expected”

Sometimes phase 3 studies fail to show a statistically significant difference in effect, not because the drug didn’t have an effect, but because the control group did a lot better than expected, “closing the gap” so to speak between the two study arms.  In the phase 2 study, the response rate for doxorubicin was only 9%.  However, most studies indicate that ORR with doxorubicin in STS is usually 20-30%, so did the underperforming control arm exaggerate an effect size in phase 2? Does PICASSO actually show a true absolute benefit with the addition of palifosfamide over what we would expect with standard of care? The response rate for the palifosfamide+ doxorubicin was only 23% which is consistent with most single agent doxorubicin studies.

Most oncologists will agree that the best standard of care for a patient with metastatic STS is a clinical trial. By the nature of being in the study itself and the intensive review and monitoring patients receive, you would expect all the patients including those on the control arm to do slightly better than the overall general population, another reason why control arms can fare better than expected in blinded studies.

4)      No drugs have ever shown an improvement in overall survival outcomes when added to doxorubicin.

Hundreds of studies have been conducted over the past decades exploring the optimal dosing and combinations of chemotherapy to treat STS. Unfortunately none have ever translated into an improvement in overall survival when added to doxorubicin. A complete response (CR) is known to improve survival rates, but sadly treatment with palifosfamide does not demonstrate CRs.

5)      FDA does not recognize PFS as being a predictor of clinical outcomes in STS

Ziopharm intitiated the pivotal Phase III trial without reaching a special protocol assessment (SPA) with FDA.  FDA did not consider the endpoints as designated for the proposed pivotal trial as supportive of SPA in this disease setting, although they would have granted a SPA with modified endpoints. Instead Ziopharm communicated external experts and chose to retain PFS as the primary endpoint with a prespecified method for radiologic evaluation of PFS (independent central review). The company estimates that a median increase in PFS of 3 months or greater over the estimated 4.3 month median PFS for the control arm (doxorubicin only) could achieve the targeted hazard ratio (HR = 0.60; p = 0.0005, one-tailed) and also predicts a demonstrable improvement in overall survival (OS). I have already mentioned the concerns I have with the basis for these assumptions based on extrapolating the phase 2 results and it is quite possible that the control arm performs a lot better than the expected 4.3 months, diminishing any treatment difference between the two arms.

Ultimately, FDA has indicated that regulatory acceptability of the trial’s results for approval would depend on the magnitude of the difference between the trial study arms, as well as a risk-benefit analysis.  ZIOPHARM could potentially file for accelerated approval based on positive PFS (whereby they would be required to conduct a second confirmatory study), or if they hit the secondary end point of an improvement in overall survival, they could file for regular approval.

It’s not all bad news, as recently GSK was able to receive approval for Votrient for the treatment of patients with advanced soft tissue sarcoma who have received prior chemotherapy on the basis of a 3 month improvement in PFS. This was a second line study versus placebo and despite the 3 month PFS benefit, this did not translate to a benefit in overall survival. However, the NDA was reviewed by ODAC, and the FDA advisory committee voted in favor of approving the drug.

I think the chance of palifosfamide showing some incremental benefit over standard of care is fair, but I am dubious that it will hit the primary PFS (0.6 HR) or secondary OS end point for PICASSO-3. Given the lack of additional options for patients with STS, any NDA will invariably be reviewed by ODAC, and an absolute benefit in PFS of even 2 months coupled with no additional toxicity, could potentially be enough to swing a positive ODAC opinion.

Disclosure:  I have no positions in any of the companies mentioned in this article.

An edited version of this article was published on TheStreet.com

Sarepta Eteplirsen: Accelerated Approval Unlikley

Less than a year ago, AVI Bioscience (AVII) was a penny stock. Yet a couple of NASDAQ non-compliance letters, a one-for-six reverse stock split, one ticker change to SRPT and one complete reinvention to Sarapta Therapeutics  later, here we are with a soaring stock price and market cap > $700 MM.

Investors are simply smitten with the company’s lead compound, eteplirsen, a novel investigation drug for Duchenne Muscular Dystrophy (DMD), which aims to help patients produce a normal amount of dystrophin, a protein required for correct muscle function. So smitten are they, Bulls seem confident that Sarepta will be allowed to successfully file for accelerated approval for the drug on the basis of their single Phase 2b study.  Sarepta is meeting with the FDA in Q1 13 to discuss this.

Accelerated Approval is a track by which FDA can allow early approval of drugs to treat serious diseases, that fill an unmet medical need based on a surrogate endpoint.   According to the FDA’s website, “the FDA bases its decision on whether to accept the proposed surrogate endpoint on the scientific support for that endpoint. The studies that demonstrate the effect of the drug on the surrogate endpoint must be “adequate and well controlled” studies, the only basis under law, for a finding that a drug is effective”. 

Unlike the bulls, I think it is extremely unlikely that Sarepta will convince the FDA that their data warrant accelerated approval. Instead, I think it is much more likely that they will be required to perform a new pivotal study prior to submitting the drug for FDA review.  Here are some of the reasons why I think this will happen.

The “pivotal” phase 2 data are systematically biased

AVI Biopharma conducted their phase 2b study at a one hospital in Columbus, Ohio. This means that FDA has no idea whether or not these results could ever be reproduced at another institution. This is a problem in my opinion, most pivotal studies are multi-center to ensure the results are seen consistently across multiple environments.

At 24 weeks, the study was unblinded at which point bias is inherently introduced. This means for all measures reported after this time point, investigators and raters know exactly which patients have received how much treatment and at what dose.  Whether or not bulls agree, it places limitations on what can be inferred from the results. 

In July, Sarepta reported results from 36 weeks of follow-up. They decided to exclude 2 boys from the lower dose treatment arm as they had experienced significant worsening of their condition and could not complete the assessments.  This means all subsequent results were reported as modified intent to treat (mITT) for 10 patients and adds further bias in favor of treatment effect. The results of the mITT group was from only 10 patients which is a tiny sample size compared with most pivotal studies.

Bulls argue that FDA will find dystrophin a compelling biomarker as a surrogate end point to support accelerated approval.  Let’s take a look at the data. In October 2012, Sarepta shared 48 week results.  At this time point, we have n=4 patients receiving  50mg/kg and n=2 patients receiving 30 mg/kg for 48 weeks, and the delayed treatment group (n=4) who have been treated for 24 weeks in total. Biopsies were repeated for all patients at this time point.  After 48 weeks of eteplirsen treatment, there was a statistically significant increase in dystrophin-positive fibers to 47% of normal.  The delayed treatment group, also showed significant increase of dystrophin-positive fibers to 38.3% of normal after just 24 weeks of therapy.  This is surprising as it’s nearly double the increase observed after 24 weeks of treatment for the original cohort (which was only 22.5%).

However, when we look at the relationship between % dystrophin-positive fibers and mean change in 6MWT, there is no clear correlation between drug dosage,  % dystrophin-positive fibers and physical performance. Patients treated at the lower dose of drug produce more dystrophin-positive fibers, but this does not translate to improved physical performance on the 6MWT.  Similarly, Prosensa and GSK, published results of a phase I/II dose-escalation study demonstrating treatment with their drug, PRO051 (a direct competitor to eteplirsen) resulted in 20-100% dystrophin-positive fibers after only 2 weeks of treatment in some patients, with levels maintained >70% at high doses after  7 weeks of treatment.  As with eteplirsen, there was no correlation between % dystrophin-positivity and 6MWT performance. Obviously the argument can be made that the dystrophin is not fully functional as early as 7 weeks, which is why GSK is in the process of conducting a phase 3 multi-center, double-blind, randomized controlled study to investigate the effects of treatment for 48 weeks in 180 patients.  The results of that study should report 2H13.

The point I am making is that there is insufficient clinical evidence from human studies that suggest dystrophin-positive fiber counts predict clinical outcomes in patients with DMD.  To that end, it is entirely unknown what the correct dose or dosing frequency should even be for this eteplirsen based on early study results. I doubt FDA will be persuaded otherwise. Furthermore, if GSK are required to demonstrate efficacy under blinded, controlled conditions so why should the regulatory hurdle be any lower for Sarepta?  Bulls argue that GSK is conducting a phase 3 study due to their drug’s safety profile. FDA requires a drug demonstrate both safety and efficacy. A review of PRO051’s phase1 /2data do not suggest there are any concerning safety signals compared with eteplirsen’s safety early data.

The fact remains that eteplirsen’ s open-label “placebo-adjusted” benefit  has never been demonstrated by the sponsor under blinded, controlled conditions. In fact, if Sarepta were to repeat this trial with adequate power for 48 weeks under double-blind conditions, it is possible that absolute benefit could be modest.

Bulls argue that FDA will allow for accelerated approval because the drug is safe and tolerable. I’m the first to agree that we need to get more medications to market where there are no other viable options, but it would simply be unprecedented for FDA to award Sarepta accelerated approval on the basis of this open-label signal. The regulatory precedent for accelerated approval is very clear:

Orphan drugs are held to the same statutory requirements for demonstrating effectiveness and safety as regular applications.

Orphan drugs must: Demonstrate substantial evidence of effectiveness/clinical benefit (21CFR 314.50)

Substantial evidence of benefit requires: Adequate and well-controlled clinical study(ies) (§314.126)

“Study has been designed well enough so as to be able “to distinguish the effect of a drug from other influences, such as spontaneous change…, placebo effect, or biased observation” (§314.126)”.

This study is neither well-controlled, nor adequate given that its single-center with only 10 patients completing assessments at 48 weeks. Being entirely open-label past 24 weeks, it is riddled with potentially biased observations.

But wait, Sarepta bulls feel FDA will be swept away by patient advocates demanding quicker approval of treatments.  When did Federal agencies become so nimble and responsive to emotional demands from the public? They didn’t. And while I have immense sympathy for families living with loved ones suffering with rare diseases, I look at the situation objectively, as will FDA.

As an example of FDA’s stoic position on ensuring regulatory standards are upheld under all circumstances, Bulls should take a look at the regulatory history for Vyndaqel for the treatment of the rare and universally fatal disorder, transthyretin familial amyloid polyneuropathy (TTR-FAP) for which there are no approved drugs. The US prevalence of TTR-FAP is only 2,500, yet Pfizer still managed to submit a single multi-center, global, randomized, controlled study enrolling N=128 patients as the basis for their submission. The drug was extremely well tolerated in the pivotal study with a placebo-like side effect profile. During the open public hearing session of an FDA Advisory Committee meeting to review the application in May 2012, countless  patients, caregivers and advocates begged, wept and pleaded with FDA directly to approve this drug, especially on the basis of its’ benign safety profile.  They wanted hope where there was none.  Sadly, hope wasn’t enough.  FDA issued a complete response letter to the sponsor requesting the completion of a second efficacy study to establish substantial evidence of effectiveness prior to an approval.  More recently, patients and advocates suffering with chronic fatigue syndrome have extensively lobbied the FDA to approve Hemispherx’ Ampligen, a potential first in class treatment for a chronic, debilitating condition.  The agency received “hundreds of letters, emails, and testimonies” in support of approval. One patient even went on a hunger strike, but to no avail. On February 5, 2013, Hemispherx was issued a complete response letter citing that there were insufficient safety and efficacy data to approve the product. 

In any future regulatory meeting to discuss accelerated approval, FDA will be presented with single-center data on 12 patients. Blinded observations were available up to 24 weeks of therapy with  no significant differences between treatment and placebo at this time point. FDA then has to evaluate open-label observations for the 10 patients who could actually complete assessments, 6 of whom were treated for 48 weeks and the other 4 patients who received treatment for 24 weeks. In my opinion there is nothing adequate and well-controlled in this data and the evidence of effectiveness is questionable due to bias, and there are simply not enough patients treated at the dose the company is seeking an indication for, to establish comprehensive safety. In the extremely unlikely event the FDA granted accelerated approval, an FDA Advisory Committee would likely shoot this application down and demand more robust data.

Disclosure: The author owns May puts $SRPT

Depomed Serada: odds stacked against positive Advisory Committee Vote


Thank-you to Adam Feuerstein for publishing this article on the Serada bear thesis.  For those of you who are interested in a bit more depth, the unedited version is featured here.

I don’t disagree with recent reports suggesting Serada, Depomed’s proprietary extended release formulation of gabapentin represents significant upside potential to the stock. Although, I think the probability of a positive panel vote (and subsequent approval) is slim, this does however, provide interesting trading opportunities into the run up to the panel on March 4, 2012 and the release of the FDA briefing documents expected no later than 2/28/2012.

1)      What’s the rationale for using gabapentin to treat menopausal hot flashes?

Vasomotor symptoms (VMS) of menopause  or hot flashes are experienced by the majority of women during the menopausal transition, with prevalence rates varying by racial/ethnic group and being influenced by risk factors such as age, BMI and smoking status . Although the pathophysiology is not fully understood, a reduction in endogenous reproductive hormones and thermoregulatory changes are both thought to contribute to the etiology.  Temperature information is processed within the anterior hypothalamus in the preoptic area (POA), where it is hypothesized that heat-sensitive neurons control the release of heat.  Serotonergic neurons and noradrenergic pathways have both been identified as projecting into the POA and hence, given data linking estrogen fluctuations to a reduction of serotonin and norepinephrine, it has been proposed that pharmacologic agents such as serotonin reuptake inhibitors and serotonin–norepinephrine reuptake inhibitors may help to reestablish levels of these neurochemicals within the thermoregulatory centers, reducing the occurrence of VMS.  The role (if any) of GABA transmission in temperature regulation is undefined.

2)      Why use gabapentin to treat hot flashes if there is no sound physiologic basis?

Good question, maybe the role of gabapentin in this setting is tenuous at best.  The results of the Serada phase 3 program would tend to support this.  BREEZE-1 and BREEZE-2 were two randomized, double-blind, placebo-controlled studies of approximately 540 patients per study, receiving treatment at either 1200mg or 1800mg daily or placebo.  Total treatment duration in BREEZE-1 was six months, with primary efficacy endpoints assessed after 4 and 12 weeks of stable therapy. Persistence of efficacy was to be assessed at 6 months as a secondary endpoint. Treatment duration in BREEZE-2 was three months, also with assessment of efficacy at 4 and 12 weeks. In both studies, the coprimary end points were to assess the mean change in frequency and severity of symptoms from baseline to week 4 and 12.  The results were reported in 2009 and both studies failed to meet all of the primary end points for statistical significance.  There are many issues with the data which only serve to raise questions as to the true clinical benefit for this drug in a VMS setting.  In BREEZE-1, 36% of patients receiving treatment at 1200mg, and 30% of those receiving 1800mg failed to complete the study.  These are very high drop-out rates for any trial and raise concerns as to why patients were leaving the study. Some women can experience VMS for years, so any treatment needs to be well tolerated, safe, and to demonstrate persistent benefit. Serada did not consistently exhibit persistency of effect for even 12 weeks let alone 24. The 24 week data were never reported fully reported but will hopefully surface in March. At best, the results of BREEZE-1 and BREEZE-2 seem to indicate that treatment is associated with ~1 less hot flash per day compared with placebo treatment for a short period of time. Historically, the placebo effect in VMS studies has always been ~50% response without explanation, so how “real” is the Serada treatment effect and how meaningful is it to women who would be taking this especially in light of the overall gabapentin adverse event profile?  The FDA’s responder analyses will surely shed light on some of this. Of course, there is a bull argument in favor of statistically manipulating these data in order to get a positive outcome which has been summarized here by @JNapodano.  I have previously quoted a statistical adage that “if you torture the data long enough, it will confess to anything”. Depomed certainly should be commended on their statistical prowess in explaining why both BREEZE-1 and BREEZE-2 failed, but I am certainly not convinced that removing 2 outliers from a study enrolling >500 patients is enough to make it statistically meaningful.

However, statistics aside, it’s the absolute efficacy benefit that will be a review issue and it will be interesting to see the placebo adjusted frequency and severity scores debated at the Reproductive Health Drugs Advisory Committee. It will also be interesting to see how the lack of any meaningful persistency of effect at 24 weeks will be perceived, this is after all a regulatory benchmark for approval.

Aspiring to seize victory from the jaws of defeat,  Depomed went back to FDA and reached agreement on a Special Protocol Assessment (SPA) for BREEZE-3, a double-blind, placebo-controlled study of approximately 600 patients, randomized to receive either placebo or a total dose of 1800mg of Serada. The co-primary efficacy endpoints were reductions in the mean frequency of moderate-to-severe hot flashes, and the average severity of hot flashes, measured after 4 and 12 weeks of stable treatment. As in the prior BREEZE-1 trial, the treatment duration of the study was 24 weeks, to establish the regulatory requirement for persistence of efficacy. The results were reported in October 2012: Using the pre-specified non-parametric statistical analysis, the trial results for the co-primary endpoints were statistically significant for the reduction in average frequency of hot flashes at 4 weeks (p=0.0003), the reduction in the average severity of hot flashes at 4 weeks (p<0.0001) and 12 weeks (p=0.0102), but were not statistically significant for the reduction in average frequency of hot flashes at 12 weeks (p=0.10). The results were not statistically significant at 24 weeks for the reduction of hot flash frequency (p=0.2351) or severity (p=0.1510).  Oh dear.

At the end of the day, the conditions of the SPA were not fully met in this study, the drug still doesn’t show a significant reduction in frequency of hot flashes beyond 4 weeks, and given that the two supporting trials failed to hit a single end point, it casts major doubts as to why Serada would provide any meaningful clinical benefit in VMS, especially when women experience symptoms for months to years.

3)      Bulls argue that we need safer alternatives to hormone replacement therapy(HRT)

Firstly we know a lot more today about HRT and the long term risk profile in women treated for early menopausal symptoms and late menopausal symptoms. The 2012 Hormone Therapy Position Statement of The North American Menopause Society clarifies the benefit-risk ratio of estrogen therapy versus estrogen-progestogen therapy for both treatment of menopause-related symptoms and disease prevention at various time intervals since menopause.  In many cases, estrogen therapy with or without progesterone therapy (the most effective treatment for VMS) and can be used safely for prolonged periods of treatment (e.g. 3-7 years) without any increased risk of malignancy. Obviously there will be women whose risk profile is not appropriate for HRT and there are effective alternatives in the pipeline, which brings me to a major reason I feel Depomed will not get a positive Committee Panel vote.  

On March 4, 2013, during the afternoon session, the Committee will discuss the NDA submitted by Noven Therapeutics, LLC, for Low-dose Mesylate Salt of Paroxetine (LDMP) for the treatment of VMS associated with menopause. Paroxetine is a potent and highly selective serotonin-reuptake inhibitor and the putative mechanism for efficacy in VMS would relate to regulation of serotoninergic transmission in the POA/thermoregulatory center.

Noven has completed 2 studies as part of the clinical development program.  The first trial was a 24-week, multicenter, double-blind, randomized, placebo-controlled, Phase 3 study of LDMP in postmenopausal women with moderate to severe VMS. Eligible subjects entered a 12-day, single-blind run-in period, and were randomized 1:1 to receive LDMP 7.5 mg or placebo for 24 weeks. Unlike Serada which is dosed twice a day, LDMP is taken as a single nightly dose. Primary efficacy end points were mean changes in frequency and severity of moderate to severe VMS from baseline to Week 4 and Week 12.  Persistence of effect at week 24 was evaluated with a responder analysis, where a responder was defined as a subject with ≥50% reduction in VMS frequency from baseline to Week 24. Mean weekly reductions in VMS frequency were significantly greater for LDMP than placebo at Week 4 (–28.9 and –19.0, respectively; P<0.0001) and at Week 12 (–37.2 and –27.6, respectively; P=0.0001).  Of course, this also needs to be translated into absolute placebo-adjusted efficacy benefit for patients, but overall the study met all end points.

Importantly, unlike Serada, significantly more subjects treated with LDMP than placebo were responders at Week 24 (47.5% vs. 36.3%, respectively; P=0.0066), demonstrating persistence of treatment benefit.  Most commonly reported treatment emergent adverse events were nasopharyngitis (5.1% vs 4.9%), headache (4.3% vs 3.7%), nausea (3.8% vs 1.4%),fatigue (3.4% vs 1.5%), diarrhea (2.9% vs 2.5%), and dizziness (2.0% vs 0.8%).

Noven also reached SPA agreement with the FDA for the second study in the clinical development program. This was a 12-week, multicenter, double-blind, randomized, placebo-controlled, phase 3 study of LDMP in postmenopausal women with moderate to severe VMS (>7-8 moderate to severe hot flashes daily).  This study was identical in design to the first, with the same 12-day run in period prior to randomization. Overall, 614 subjects were randomized, and importantly, 549 subjects(271/306 [88.6%] LDMP; 278/308 [90.3%] placebo) completed the study (a stark contrast to the completion rates observed in BREEZE-1). Mean weekly reductions in VMS frequency from baseline were significantly greater for LDMP than placebo at week 4 (–33.0 and –23.5, respectively; P < 0.0001) and at week 12 (–43.5 and –37.3, respectively; P = 0.0090). Mean weekly reductions in the frequency of moderate to severe hot flashes were significantly greater for the LDMP group than for placebo beginning at week 1 and continuing until the end of treatment (P ≤ 0.0259 for all comparisons). Mean weekly reductions in VMS severity from baseline were significantly greater for LDMP than placebo at Week 4 (–0.09 and –0.05, respectively; P = 0.0048) but not at week 12 (–0.10 and –0.09, respectively; P = 0.2893). Mean weekly reduction in VMS hot flash composite scores (moderate and severe) from baseline were significantly greater for LDMP than placebo at week 4 (–85.51 and –60.83 respectively; P < 0.0001) and at Week 12 (–111.9 and –96.85 respectively; P = 0.0063).  There were no new safety signals and the drug was well tolerated.

Now given that the Advisory committee will be reviewing both NDA’s for the exact same indication on the same day, it is very likely that the same panel members (or at least a large overlap) who review Depomed Serada in the morning will be reviewing LDMP in the afternoon, and hence will be reviewing both sets of briefing documents ahead of time.

Let’s don our committee member hats and consider the two packages for a moment:

AM Session Serada :       Minimal mechanistic plausibility as to why it should work in VMS, dosed twice a day, high patient dropout rates in clinical studies, failed to meet majority of prospectively defined end points in two large randomized controlled studies, failed to demonstrate significant reduction in frequency of hot flashes beyond 4 weeks in third confirmatory study, failed to demonstrate persistency of benefit at 24 weeks in every study, overall side effect profile and tolerability issues with gabapentin

PM Session LDMP:          Mechanistic rationale for efficacy in treating VMS, dosed once nightly, met 8/9 prospectively defined end points in two large, randomized controlled studies, significantly reduced the frequency of moderate to severe hot flashes at 12 weeks, and maintained persistence of effect at 24 weeks, excellent overall safety and tolerability profile

Hmmm….I know which way I would vote.

On a final note, in the extremely unlikely event that FDA were ever to approve Serada for the treatment of moderate to severe VMS, realistically, what could the drug potentially do in terms of peak sales? Here is my liberal sales estimate with assumptions. (The conservative version is a lot lower).

10 MM per year seeking treatment for hot flashes

75% agreeing to try pharmacotherapy following consultation (high estimate)

30% requiring alternative to HRT

5% peak market share for Serada

90% patients given a prescription choosing to fill it

50% likelihood that prescription filled with Serada first as opposed to generic substitution with gabapentin. (Note: this drug is “extended release” yet needs to be dosed twice daily, there is no evidence that this confers any benefit over generic gabapentin from a managed care perspective so it is  likely (if ever approved) Serada will have a low tier formulary status and will frequently be subject to generic substitution)

Average course of treatment 3 months (Note: 30% of patients failed to complete the studies so there is no reason to believe real world experience would be different, in fact I would predict a high drop off rate after the first script. The drug does not demonstrate persistent effects beyond 4 weeks under clinical trial conditions, so it is extremely unlikely the average course of therapy will be as high as 9 months).

Average cost per script $225

Peak Sales Estimate: 10MM x 75% x 30% x5% x 90% x50% x3 x 225 = $34 MM

 

Disclosure:  I am long in March puts

This article is not intended as, nor should it be taken as investment advice. Do you own due diligence before trading securities.