More on New Developments at the FDA in the Era of Precision Medicine

This week the FDA approved the expanded use of the cancer drug Keytruda for any solid tumor that expresses a particular molecular marker. It is the first time that the FDA has approved new uses of a cancer drug in cancers that are defined by a molecular marker independent of the site in the body where the tumor originated. While the expanded use of the drug was based on an accelerated approval supported by surrogate markers, this is another demonstration of the FDA’s continuing adaptation to the new era of precision medicine, in which the appropriate use of a drug is defined by particular genetic and molecular markers rather than the broad traditional disease categories.

Postscript to last week’s post–– I wrote that the approval of Kalydeco for cystic fibrosis patients with 23 additional mutations was, to my knowledge, the first time ever that the FDA had approved new indications for a drug without any human clinical data. The Kalydeco label change was based entirely on an in vitro assay of the drug’s effect on the function of the cystic fibrosis transmembrane receptor. The in vitro assay had correlated well with clinical trial data from the original Kalydeco studies and provided a reasonable degree of confidence that the drug would work in the additional mutations tested. Later reports and an additional statement from the FDA confirmed my initial conclusion that this approval was indeed the first of its kind and a real breakthrough in drug approval. That is good news for the approximately 900 cystic fibrosis patients who have one of those 23 mutations. Of course the 900 patients covered by the 23 additional mutations means that there is, on average, fewer than 40 patients with any one of those mutations, which is why the FDA’s flexibility is so important. Precision medicine will need that kind of regulatory innovation to move forward when, as the FDA’s Janet Woodcock stated, there is a solid understanding of the disease’s biology, a drug with a known safety (and pharmacokinetic) profile, and a good assay to predict the drug’s effect in target populations.

A milestone for precision medicine and for the FDA

Precision medicine is much in the news­­––and the need to target the right drug to the right patient using genomic and other information is clearly going to be major part of healthcare in the future. Today's news included an announcement from the FDA that it had expanded the approved indications for Kalydeco,  a breakthrough drug for the treatment of cystic fibrosis.  The previous label indication described Kalydeco as a 

cystic fibrosis transmembrane conductance regulator (CFTR) potentiator indicated for the treatment of cystic fibrosis (CF) in patients age 2 years and older who have one of the following mutations in the CFTR gene: G551D, G1244E, G1349D, G178R, G551S, S1251N, S1255P, S549N, or S549R …[or] R117H

The new label, according to the FDA, expands the use of Kalydeco from the 10 mutations above to include an additional 23 mutations, for a total of 33.  This is great news for cystic fibrosis patients, as the clinical studies in the original population showed that treatment with Kalydeco resulted in a significant increase in patients’ lung function. From a precision medicine perspective the Kalydeco expanded label could be viewed simply as yet another drug label with genomic information about the drug’s appropriate use—after all, there are already well over 200 drugs approvedby the FDA that include pharmacogenomic information.  However, it isn’t just the dramatic increase in the number of mutations in the Cystic Fibrosis Transmembrane Receptor gene that makes the FDA’s announcement remarkable, it is the evidentiary basis on which the dramatic expansion in the use of Kalydeco was approved.  The FDA’s announcement also had this particularly noteworthy statement:

Results from an in vitro cell-based model system have been shown to reasonably predict clinical response to Kalydeco. When additional mutations responded to Kalydeco in the laboratory test, researchers were thus able to extrapolate clinical benefit demonstrated in earlier clinical trials of other mutations. This resulted in the addition of gene mutations for which the drug is now indicated.

In other words, the indication for Kalydeco was approved solely on the basis of in vitro studies without additional human clinical data (or animal data for that matter).  Vertex Pharmaceuticals, the maker of Kalydeco, was able to persuade the FDA that because there was a sufficiently high correlation between the in vitro assays and clinical data for the previously studied mutations, it was more than reasonable to allow the drug to be used for patients with 23 additional mutations that showed a significant gain in function in cell-based assays.  I am unaware of any similar action by the FDA.  So the Kalydeco announcement is more than another step in the targeting of drugs based on individual biomarkers, it is a major step by the FDA in regulatory flexibility in streamlining drug development as the science of predicting drug response evolves.  We are indeed at the beginning of an interesting journey on an uncharted road to the future of drug development.

A Better Balance Between Accelerated Access And High-Priced New Drugs: A New Conditional Approval Option

This entry is reposted from Health Affairs Blog on March 20, 2017.

There are two different storms brewing in the pharmaceutical world. On the one hand there is increasing opposition to the very high prices of drugs. On the other hand there is ever more pressure to accelerate access to drugs for seriously-ill patients. President Trump expressed both of these very different concerns in his January 31st meeting with pharmaceutical executives, during which he called drug prices “astronomical” while also vowing to “streamline” the process of drug approval. It is vital that the cost of new drugs not overwhelm patients and the health care system. It is also important to get drugs to desperate patients as quickly as possible.

However, accelerated access allows a drug to reach the market quickly, based on clinical trials that measure surrogate endpoints, for instance time to cancer progression or observed tumor response rates, rather than survival rates. These surrogate endpoints frequently fail to predict whether patients will actually live longer or have a better quality of life, but are used because they can be measured within a relatively short period of time, while evidence as to the real effectiveness of a drug can take years to collect. At the same time, any reasonable approach to drug pricing requires substantial knowledge of a drug’s effectiveness for its value to be considered when evaluating the drug’s price and that knowledge simply is not available when a drug is approved before its performance is known on truly meaningful endpoints such as overall survival (in cancer) or long-term ability to function (in diseases such as Parkinson’s or Muscular Dystrophy. With Congress and the President both seeking to change the current system, there may be an opportunity to pursue a different approach to accelerated approval and, at the same time, take at least a small step towards reducing the costs of new, potentially life-saving drugs. One way to strike a better balance between accelerated access and limiting drug prices until their value is known might be a new form of “conditional approval” with prices discounted until full approval is warranted.

The Issues with Accelerated Approval

            The FDA’s accelerated approval of eteplirsen, a new antisense drug for Duchenne Muscular Dystrophy (DMD), is a clear example of both problems——“astronomical prices” and accelerated access. DMD is a devastating illness that affects children and causes muscle weakness and eventual death. There are no effective treatments. Understandably there was enormous pressure from patient groups to approve the drug despite the lack of evidence that the drug actually works, beyond a change in a surrogate marker.

The drug is now available at a price of $300,000 per patient per year, but it may be years before the data from additional clinical trials can provide substantial evidence of whether or not the drug is effective. If the drug turns out not to provide meaningful clinical benefit, then the $300,000 per year cost of providing patients the drug is a terrible waste of our health care dollars. However, if insurers do not pay for the drug and it actually would provide significant therapeutic benefit to Duchenne’s patients, then there would be even more terrible unnecessary suffering and death among DMD patients.

            Eteplirsen is not by any means the only drug approved before the real risks were known or, in some cases, a lack of real efficacy was demonstrated. One useful model for accelerated access and controlled pricing was developed in response to an earlier era of crisis in pharmaceutical policy when the HIV epidemic first caused a public outcry for accelerated access. Prior to the AIDS crisis of the 1980s, major patient advocacy groups, such as the American Cancer Society and the American Heart Association, focused their efforts on raising money for research and paid virtually no attention to the FDA. With AIDS, and particularly with ACT UP (AIDS Coalition to Unleash Power), the world changed. For the first time there was enormous pressure on the FDA to do something—anything—to get drugs out to patients before all of the safety and efficacy data was in.

            The FDA responded to the AIDS crisis with a number of efforts to expand early access. One that has the most relevance for today was the 1992 parallel track initiative. The parallel track initiative, was used only once for stavudine, a still-experimental drug that was made widely available to physicians treating AIDS patients.  The drug sponsor could seek to charge for the drug but only in an amount sufficient to recover its costs for the trial, which required financial disclosures to the FDA. Treating physicians providing the parallel track drug were required to provide the sponsor with basic data on their patients and patients’ responses to treatment.

How Conditional Approval Could Work

            A more balanced approach to the current cry for accelerated access and lower prices could adapt and build on the 1992 HIV-only parallel track approach. The two key components that could be revised for today’s use are: first, an accelerated approval that permits wide distribution before final approval; and, second, a mechanism that limits the price of the conditionally-approved drugs while more data is collected. Conditional approval based on surrogate endpoints would allow a drug’s sponsor to distribute the drug to all of the desperately ill patients who have no alternative, while maintaining the price restrictions until additional data on actual clinical benefit and risk is provided. This would provide a 21^st Century update of the 1992 HIV-only Parallel Track.

            Under the conditional approval proposed here any physician treating a patient with the targeted indication could prescribe the drug and would agree to collect and report basic data on the duration of treatment, responses to the drug, and any other changes in their patients’ conditions. Such longer-term single arm trials are likely to provide evidence of real effectiveness and safety when the target is an untreatable serious disease. While the drug is being widely used and the data collected, the sponsor could charge for the drug based on a set price formula until the sponsor provides further data and the FDA completes its review. There would be no need for case-by-case negotiations over costs and pricing as was required by the 1992 Parallel Track policy. This differs from other accelerated access programs that use the term “conditional approval” (such as that used by the European Medicines Evaluation Agency, which does not limit prices and must be reviewed annually).

Determining Prices for Conditionally Approved Drugs

            How might the predetermined discounted price provision work? One possible mechanism would require a pharmaceutical company seeking conditional approval to specify its intended initial market price for the new drug. The conditional approval distribution price could be limited to 25 percent of the specified initial market price. Alternatively, the conditional approval price formula might be a predetermined percentage of the average introductory price of breakthrough drugs approved during the prior two years. Unlike the prior Parallel Track provisions, either price formula would avoid the need for the sponsor to disclose its costs and negotiate with the FDA to justify charging during the conditional approval period. Given the strong demand by patients for a potentially life-saving drug when there is no effective alternative, marketing expenses should be low.

With the very low cost of small molecule manufacture or even the higher costs of manufacturing biologic drugs, the 25 percent pricing formula should cover the costs of manufacture, distribution, a limited marketing outreach, and the process of data collection and still provide a modest profit. This conditional approval update of Parallel Track would provide the needed balance between access to potential breakthrough drugs and substantial evidence to support their unrestricted entry into the marketplace. Limiting profits would motivate drug companies to distribute the drug widely enough to provide the needed evidence as expeditiously as possible. And in an era of skyrocketing prices for new drugs it would avoid imposing even greater costs to consumers and insurers for what are actually experimental drugs such as eteplirsen.

A Way Forward with Conditional Approval

            New proposals are being discussed that would even further accelerate access to new drugs and members of both parties in the House and Senate are moving forward with a variety of approaches to the high cost of new drugs. Now is the time to take a new approach to accelerating access and limiting drug prices. Of course, it would also be necessary to require insurers and government payers to cover the drugs during the conditional approval period in the same way that they currently cover drugs approved under the accelerated access and breakthrough drug procedures. If the data confirms the benefit of the treatment, full approval would be granted and the drug sponsor could charge whatever price it can justify in the marketplace, but with much better evidence as to what the drug’s real worth actually is. Patients desperate for treatment would get access to drugs, insurers would be paying less than under the current system, and patients, providers and insurers would get the data they need on the drug’s efficacy. It is time for a new approach to accelerated access that is good for patients and good for us all.

The 2st Century Cures Act: Patient Experience in Drug Approval

On Thursday, May 19, 2015, the Energy and Commerce Committee of the House of Representatives voted 51 to 0 to approve the 21^st Century Cures Act (“the Cures Act”) which would, if passed by the full House and Senate and signed by the President bring about a number of significant changes in NIH research funding, the process of drug development, and the FDA review of New Drug Applications (“NDAs”). Because the current draft is 300 double-spaced pages in length the focus in this post will be on just six of those pages--“Sec. 2001. Development And Use Of Patient Experience Data To Enhance Structured Risk-Benefit Assessment Framework.” While there are likely to be some changes to the current draft, given the overwhelming bipartisan support in the House, it seems that much, if not all, of the current draft will become law.

Section 2001 of the Curs Act consists of two subsections, currently denominated (x) and (y).  Section 2001 would appear to require very little change to the current approach of the FDA to drug approval.  Section 2001 Subsection x (1) states that “The Secretary shall implement a structured risk-benefit assessment framework in the new drug approval process [A] to facilitate the balanced consideration of benefits and risks; and [B] to develop a consistent and systematic approach to the discussion of…the benefits and risks of new drugs.” The current draft of Sec. 2001 Subsection (x) 2 contains the proviso that “Nothing in Paragraph (1) “shall alter the criteria for evaluating an application for premarket approval of a drug.” The apparent focus of this subsection (x) is to bring greater clarity and consistency to the FDA’s weighing of risks versus benefits. This is an unobjectionable goal, if somewhat difficult to implement. For example, one cancer drug may cause severe skin reactions while increasing overall survival by one month, while another drug for the same form of cancer may cause severe gastric effects while increasing overall survival by 1.2 months. Comparing the risks and benefits of those two drugs is obviously difficult to do.  It is that difficulty that may be the objective of the much more interesting Subsection (y).

Subsection (y) is entitled “Development And Use Of Patient Experience Data To Enhance Structured Risk-Benefit Assessment Framework.”  Subsection (y) 1 requires the Secretary of HHS, over a period of 2 to 3 years, to a promulgate a number of guidances and establish procedures by which non-NDA sponsor entities may submit patient experience research proposals for feedback from the Secretary as well as patient experience data and data analyses.  Patient experience data, or, as it is also known, patient-reported outcomes data has been a field of interest in clinical research for a number of years. Patient-reported outcomes was the subject of an FDA Guidance issued in 2009 (“2009 Guidance”).  The 2009 Guidance recognized the value of assessing patient experience endpoints in clinical trials and encouraged patient input in developing the instruments to be used in assessing patient reported outcomes or experience:

Item generation should include input from the target patient population to establish the items that reflect the concept of interest and contribute to its evaluation. The population will help generate item wording, evaluate the completeness of item coverage, and perform initial assessment of clarity and readability.

However, the approach to patient experience data in the Cures Act goes well beyond the 2009 Guidance.  Subsection (y) provides the Cures Act definition of patient experience data:

In this subsection, the term ‘patient experience data’ means 
data collected by patients, parents, caregivers, patient advocacy organizations, disease research foundations, medical researchers, research sponsors or 
other parties determined appropriate by the Secretary that is intended to facilitate or enhance the 
Secretary’s risk-benefit assessments, including information about the impact of a disease or a therapy on patients’ lives.

The statutory inclusion of data collected by “patients, parents, caregivers, patient advocacy organizations, [and] disease research foundations” is radically different from the drug-sponsor driven approach that was addressed in the 2009 Guidance. Bringing data collected by persons or entities other than the drug sponsor into the New Drug Approval process (or the sNDA process of approving additional indications or major label changes) would be a fundamental change in the concept of data to be used by the FDA in new drug review and would significantly increase the role of patients and patient organizations in the drug approval process itself.  If so, the 21^st Century Cures Act’s effect on drug development might be as dramatic as the title of the Cures Act suggests.