FDA Draft Guidance: Rare Diseases: Common Issues in Drug Development Guidance for Industry August 2015
In August 2015, the FDA released this new guidance on rare diseases. Its aim is to aid sponsors and manufacturers in the development of drugs and products for rare diseases (defined more or less as fewer than 200,000 Americans suffering from this disease).
This is a sixteen page document covering the regulatory background of orphan drugs and rare diseases, natural history studies, disease pathophysiology and biomarkers, nonclinical studies, efficacy endpoints, evidence of effectiveness and safety and CMC.
This document is an important read for anyone developing such products or handling their safety (in clinical trials and during marketing). I would like to go through this document looking at the safety issues that are addressed (or not addressed).
There are several drugs that reach the market based on a very small number of patients studied in clinical trials. For example, in 2010, a drug for refractory chronic gout was approved based on, what appears in the approved US package insert, less than two hundred patients studied on the drug. The standard table of adverse reactions occurring in 5% or more of patients is based on 85 patients treated with the drug and 43 with placebo. The drug has a black box warning and a REMS. Although this drug appears to be a significant addition to the rheumatologist’s armamentarium, the safety profile clearly cannot be fully known with such a small number of patients treated.
This is not to say that such drugs should not be on the market without more safety data. Rather, it is necessary to understand that the safety profile is not well described at launch with only the very frequent or a few very severe/serious adverse reactions known. The rarer ones will only be picked up after marketing and only if there is a sufficient level of attention paid by both prescribers and patients.
Let’s walk through the document and look at the safety discussions and comments. Much of the document discusses the understanding of the disease, its natural progression and studies thereof, its (often unknown) pathophysiology, the drug’s method of action (also often unknown) and effectiveness and endpoint assessment. FDA acknowledges that often little is known about these diseases and that the usual requirements of one or more adequate and well-controlled studies in an identified patient population that are required for NDA/BLA approval are often not attainable with rare disease drugs.
FDA notes that natural history studies (often without comparators) are then used as historical controls in clinical studies which have no direct simultaneous controls. They acknowledge that this is not necessarily a happy situation: “In general, studies using historical controls are credible only when the observed effect is large in comparison to variability in disease course (e.g., substantial improvement in outcome is observed with treatment in a disease that does not naturally remit).” Although not stated directly, this is true both for efficacy and safety judgments.
FDA emphasizes several times the need for customization of each drug’s development plan with frequent meetings between FDA and the sponsor.
Nonclinical studies, as always, should be done to understand (where possible) the mechanism of the drug’s action and to have an idea of early stage doses, routes of administration, dose escalation etc. This data “will often determine some important safety monitoring procedures.” Safety testing in animals should be evaluated against the standard, approved guidances of “nonclinical safety studies relative to clinical trials in drug development.” FDA also notes that “Safety evaluation in an animal model also may be particularly valuable when it is suspected that drug toxicity may be more severe in the presence of disease pathophysiology.” But FDA also notes that FDA usually does not require testing for safety in an animal model of the disease though in rare diseases, if such a model exists, it might be useful to do safety testing there.
There is a long discussion of clinical trial efficacy endpoints and biomarkers that may be used as endpoints if clinical endpoints and outcomes are not fully available.
Finally, FDA addresses clinical safety in the “Evidence of Effectiveness and Safety” section. The guidance is high-level and non-specific:
“Assessment of the safety of the drug should use “all tests reasonably applicable” to establish safety for its intended use. Clinical trials should also include a monitoring plan adequate to ensure the safety of clinical trial patients. The elements and procedures of the monitoring plan should be based upon what is known about the drug, including nonclinical toxicology and chemistry, manufacturing, and controls (CMC) information, and, if available, previous human experience.
“There is no specific minimum number of patients that should be studied to establish effectiveness and safety of a treatment for any rare disease. The number of patients to establish effectiveness and safety is determined on a case-by-case basis…
“When conducting a benefit-risk assessment for a drug or a serious or life-threatening illness, FDA also recognizes that greater risks may be accepted for a treatment that is an advantage over available therapy. This reflects FDA’s commitment to expediting the availability of drugs for serious diseases as soon as it can be concluded that the benefits of the drugs exceed their risks, while preserving appropriate standards for safety and effectiveness, especially when these patients have unmet needs, as is often the case with patients with rare diseases.”
That’s basically all there is on safety. It is a clear recognition that the safety profile may not be well known and that “greater risks may be accepted.” There is no guidance here on numbers or type of studies, drug interaction studies etc. That is left for discussions between the sponsor and the FDA for each product.
POINTS OF VIEW
From the pharmacovigilance point of view, this is what keeps us up at night! It is not at all a happy situation, particularly if there is aggressive marketing focusing on the effectiveness without really noting that little is known about the safety. It often puts the drug safety group on the defensive when new, serious/severe, previously unseen SAEs occur. They will often be attributed to the disease or other factors rather than the drug.
If only dozens or hundreds of patients are studied, under the so-called statistical Rule of Three, the ADRs with a frequency of even 1/100 will likely not be seen in the small clinical trials. This may give a false sense of security to the prescribers and patients who are not entirely aware of the limited knowledge obtained from small studies.
On the hand, taking a clinical, public health and personal view, patients with rare and dreaded diseases for which there are no treatments, few treatments or where known treatments have failed, there is no other recourse than to try the less well clinically substantiated drugs. The argument here is that there is nothing else and there may indeed be increased risk but we have little else to offer.
This is a well-known phenomenon in the history of medicine from the use of blood-letting and leeches for many diseases in the more distant past to the use of corticosteroids for chronic hepatitis in the not at all distant past. These treatments either did not work or even made the patients worse. Who knew?!
Until we have something better, most agree that offering a patient a not well- established treatment can be justified. This must be done transparently, with consent and, hopefully, oversight by FDA, the prescriber and some way to track safety more intensively than in the usual post-marketing situations. This might mean a REMS, a registry, post approval studies or a detailed Risk Management Plan (in the EU).
What can the folks in the “trenches” of drug safety do? If there is no REMS or RMP in place, at the very least, all SAEs and non-serious AEs should be very carefully reviewed, full information gotten (even on non-serious cases) and a very high level of suspicion should be kept that the AE may actually be due to the drug and not the disease. MedDRA coding should be careful, precise and complete.
There are examples of drugs worsening the very disease they are treating (e.g. Fialuridine/FIAU), producing disastrous results. Company management should understand that there are significant risks to the patient (and the company too) if the safety profile is not carefully tracked with changes made as needed to dose, prescribing, monitoring, labeling, etc. This is even more challenging in the post-marketing setting where there are no investigators reporting SAEs immediately and where medical validation may not always be possible. Frequent and careful signaling must be done and records kept. Aggregate data and “outlier” analysis with case series must be done and kept up to date.
Communication with the regulators (FDA and elsewhere) must be done. Cases should be obtained from regulators (e.g. FDA’s MedWatch to Manufacturer program), Freedom of Information and on-line review of those databases available for searching (e.g. Health Canada, Vigibase, EudraVigilance, FAERS) and, of course, the medical literature must be followed (weekly per EU requirements after marketing).
This is still a draft guidance. If you have comments and wish to submit them to FDA, see the title page of the draft guidance.
So my bottom line: This is a good start by FDA and is encouraging sponsors to develop and study drugs for rare diseases. The implication, though not directly stated, is that FDA will customize and be sympathetic to the developer and perhaps “cut some slack” allowing more rapid market access with post-approval requirements. But, of course, good science, careful trials and attention to detail are always obligatory.
