SEARCH

SEARCH BY CITATION

Given the recent series of events regarding the Food and Drug Administration (FDA), including issues associated with the withdrawal of rofecoxib and the subsequent removal of valdecoxib from the market, as well as issues surrounding some of the newer biologic agents for the treatment of multiple sclerosis, it seemed like a propitious time to provide an article about the FDA and how it works. I served as a Director of the division at the FDA that was dedicated to evaluating the safety and efficacy of new drugs in the treatment of pain and inflammation. This job was remarkable because it afforded me the opportunity to think about issues such as the relative safety of new and older medications; issues regarding trial design, recruitment of patients, and inclusion and exclusion criteria; and why some trials are successful whereas others fail. I remain grateful to have been afforded the opportunity to serve the community in this manner, and as a result, I am in a unique position to write an article describing how the FDA works.

A Brief History of the FDA

  1. Top of page
  2. A Brief History of the FDA
  3. The Components of the FDA
  4. The Review Process
  5. The Way the FDA Works
  6. Postmarketing Surveillance
  7. Conclusion
  8. REFERENCES

The issue of relative risk to relative benefit is what drives the work of the medical officer at the FDA. The fact that the FDA is a unionized shop that also serves as a Washington bureaucracy should describe for the reader a rather regimented workplace, with offices assigned according to seniority and pay grade, and with the work day defined to provide regulatory oversight of therapeutics in the US. Drugs are approved for interstate transport by authority of the Food Drug and Cosmetic Act. Although many managed care organizations and other groups such as formularies use FDA approval as an important threshold for deciding whether to reimburse for a therapeutic, the FDA does not regulate medical care in the US. This concept of approval was initially considered at the beginning of the 20th century with the development of the Pure Food and Drugs Act and was precipitated by a series of deaths due to certain contaminated foods (1). It was clear that something had to be done to protect Americans from impure foods, but what would be legislated should not be enough to restrict free trade in a democratic and capitalistic nation. This new law specifically “prohibited interstate commerce in misbranded and adulterated foods, drinks, and drugs” and required truth in labeling of substances covered by the law (2).

Then in the late 1930s, another scandal rocked the nation. This time dozens of children died when prescribed an antibiotic elixir that was contaminated. This led to the Food Drug and Cosmetic Act of 1938, which superseded the 1906 law and required that drugs and food be safe; the contents would be adequately described in the product label, but no explanation was offered as to how these products were to be proven safe. Clinical trials as understood in the modern sense were not required to provide this information. It was expected that the safety information would be extensively detailed in a product label, which would also describe what was known about the chemistry of a product (1, 3).

In 1951, the Durham-Humphrey Amendment defined “the kinds of drugs that cannot be safely used without medical supervision and restricts their sale to prescription by a licensed practitioner (3).” Therefore, the concept of prescriptive drugs was officially born and led to the concept of prescriptions, which was meant to distinguish over-the-counter or generally available drugs from therapeutics that needed a learned intermediary to describe how to use the product (1).

The requirement for adequate and well-controlled trials to prove efficacy and safety of products was not established until 1962. This modern approach was integrated into the normal daily routine at the FDA, and it continued in that manner until several changes took place in the 1990s. Prior to some modernization in the 1990s, a review of a newly submitted product could take a very long time, and the drug development companies had been lobbying for years to accelerate this process. The appropriation from Congress that supported the activities of the FDA was limited, and consideration was given to the development of a user fee that would be used to secure the appropriate staff to allow for a much tighter timeline for review. A sponsor of a new therapeutic would pay a user fee to the government to help support the services provided to the industry; this money was then funneled back to the FDA. A user fee system (Prescription Drug User Fee Act [PDUFA]) was implemented. This act required that the FDA perform to certain standards. Specifically, 90% of standard reviewed drugs had to be fully reviewed within 10 months of submission. In addition, drugs that were very important because of either a new mode of action or targeting of an unmet medical need would be reviewed more quickly as a priority, and in 90% of these cases the review would be finished in 6 months. This has led to a significant change in the length of time for approval for many products, making them available more quickly (1).

The user fee act, which requires a payment of >$600,000 per application for a full review, has had significant impact on the FDA. The FDA has been able to hire an increased staff for review, as well as to perform other components of the daily work load. There are specific statements in the user fee act regarding access to meetings with the FDA and significant effort to increase the communication and interaction between a sponsor of the product and the FDA. Overall it provides for increased opportunities. However, there has been some suggestion recently that this increased opportunity has led the FDA to be less unbiased in its interpretation of evidence and perhaps to become too lax in review, allowing dangerous drugs with little benefit to reach the marketplace. This is an interesting proposition and sound bite for the media, but falls flat when one considers the issues. The user fee is paid to the government in general, and is still appropriated by Congress given the performance of the FDA. Furthermore, there is very good evidence demonstrating that these new safety issues observed in “real world” use are more reflective of the limitations of the review process than the coziness of industry and the FDA. Many of the safety signals recently identified emerged in longer-term trials of a large number of patients. Most approval data sets are usually representative of shorter-term trials, and most databases represent <3,000 exposed patients. To understand some of these safety signals, many more patients need to be exposed for far longer than is possible given the present constraints of preapproval drug development strategies of the industry, which are designed to meet the prevailing requirements at the FDA. Phase IV studies and postapproval pharmacovigilence are critical parts of the strategy for understanding the safety of all approved agents because a larger population of patients is typically observed in that arena, especially patients with comorbid illnesses and those who are subject to polypharmacy.

Prior to 1962, there were obviously many drugs that had been on the market. What was to become of those whose efficacy or safety had never been proven in any trials? In 1968 the FDA formed the Drug Efficacy Study Implementation (DESI) process to implement recommendations of the National Academy of Sciences investigation of effectiveness of drugs first marketed between 1938 and 1962. This then led to a series of drugs that by consensus underwent the DESI process and thus were considered acceptable and did not require any further study (1).

The Components of the FDA

  1. Top of page
  2. A Brief History of the FDA
  3. The Components of the FDA
  4. The Review Process
  5. The Way the FDA Works
  6. Postmarketing Surveillance
  7. Conclusion
  8. REFERENCES

The FDA consists of a highly motivated professional staff of physicians, chemists, pharmacists, nurses, statisticians (both mathematical and biostatisticians), lawyers, toxicologists, safety experts, mathematical modeling experts, and administrative staff that are organized somewhat by discipline to serve the public health as regulators of available drugs, foods, and cosmetics. The Code of Federal Regulations states that drugs are to be approved in the US for improvement in how a patient feels, functions, or survives (2, 3). All other ways to measure therapeutic benefit need to reflect these tenents in some fashion.

The organization chart reflects a hierarchical structure with responsibility being assigned down the line as appropriate to level of training and experience. This structure is generally reflected in the pay grade of the individual. The Commissioner of the FDA is appointed by the White House and is confirmed by the Senate. The appointee usually sits in that role through the tenure of the current denizen of the White House. The FDA is part of Health and Human Services (HHS), so the Commissioner reports to the Secretary of that Cabinet-level department and interacts significantly with HHS and the White House leadership (2). However, Congress controls the FDA budget, so there is a careful political balancing act between the overall health care agenda set by the White House and the public health needs as determined by Congress. Theoretically, the Commissioner should either be a professional as an MD, PhD, or DVM, or at least be affiliated with the health care field of either humans or animals. Our most recent Commissioner was a veterinarian by training, and his predecessor was an MD/PhD health economist.

The Commissioner serves to represent the FDA in front of Congress regarding important policy issues as well as the budget. There is a large supporting office consisting of staff as well as deputy commissioners. There are 5 centers at the FDA, each with a director (Table 1). The Center for Drug Evaluation and Research has been the important center for the approval of drugs and biologic therapeutics since 2002, while the Center for Devices evaluates the utility of joint implants, for example.

Table 1. Structure of the Food and Drug Administration
Comissioner's Office
 Center for Drug Evaluation and Research
  Office of New Drugs
  Office of Drug Safety
  Office of Compliance
 Center for Biologics
 Center for Cosmetics and Food
 Center for Devices
 Center for Veterinarian Medicines

The Review Process

  1. Top of page
  2. A Brief History of the FDA
  3. The Components of the FDA
  4. The Review Process
  5. The Way the FDA Works
  6. Postmarketing Surveillance
  7. Conclusion
  8. REFERENCES

The reviews are carried out in the review division, which consists of physicians, chemists, pharmacists, nurses, statisticians (both mathematical and biostatisticians), toxicologists, safety experts, mathematical modeling experts, and administrative staff, all of whom are within a team format (Table 2). The review division exists within the Office of New Drugs, and each division has a director. The head of the review team within a division is typically a medical officer who performs a review of the evidence for safety and efficacy along side the statistical reviewers who perform an in-depth analysis and reanalysis. Toxicologists review the nonclinical (preclinical) information, chemists review the manufacturing processes, and biopharmaceutists review the pharmaceutical nature of the drug, including the pharmacokinetics and pharmacodynamics as they are studied and understood. Interactions with the team are governed by a project manager who also maintains contact with the sponsor of a product and facilitates interactions with the developer of the product and the internal review team. These project managers are usually either pharmacists or nurses (4, 5).

Table 2. Other review groups involved in new drug application review
Review groups
Chemistry Review Team (manufacturing)
Microbiology Review Team (sterility assurance)
Statistical Review Team (clinical studies)
Biopharmaceutics Review Team (bioavailability)
Office of Drug Safety (safety profile)
Division of Drug Marketing, Advertising and Communication (labeling)
Compliance (manufacturing site inspections)
Division of Scientific Investigations (clinical site inspections)
Division of Medication Errors and Technical Support (trademark review)

Because the FDA approves drugs for interstate transport that have been determined, through the performance of adequate and well-controlled trials, to be relatively safe for the relative benefit of the approved indication determined through this aforementioned review process, one might ask when the first contact between a drug developer and the FDA occurs. Once a company decides to develop a new chemical or biologic entity, the company will request a preinvestigational new drug (pre-IND) meeting. At this meeting, plans are made to move forward with applying for a pre-IND, which allows for interstate transport of therapeutics that have not been approved but will need to be shipped to allow studies to be conducted on the drug. If there is little accumulated animal work using the drug to be studied, then the timeline for all steps could be up to 14 years. If some of this work has been done, then the timeline until approval for some therapies might be as short as 9 years; however, some drugs might need up to 2- or 3-year studies to be able to prove efficacy and safety. It might take much longer to approve these drugs because a 2–3-year study might actually take 1–2 years to fully recruit patients, and then the study might last for up to 6 years until the last patient finishes the assigned protocol (5).

The drug development process is high risk and quite expensive. Therefore, the drug developer is encouraged to avail themselves of the opportunity to meet with the respective review division at specific time points to ensure an orderly progression of accruing evidence of safety and efficacy. The various sections of the process towards drug development are listed in Table 3. The initial section is the pre-IND meeting. At some point the animal exposure data need to be discussed, including the maximally tolerated dose, whether the drug's mechanism of action can be evaluated in the animal or animal model of disease, and a review of long-term animal exposure to those drugs that would be predicted to be used chronically in humans. These data are submitted to the final new drug application of the biologic licensing application, but are also reviewed as they are accrued. This is required because it is impossible to administer a study drug at any dose to a human, single dose or dose escalating, prior to an extensive exposure experience in animals.

Table 3. Typical development program*
  • *

    POC = proof of concept; SAD = single ascending dose; MAD = multiple ascending doses; PK = pharmacokinetic; PD = pharmacodynamic; NDA = New Drug Application; BLA = Biologic Licensing Application.

Nonclinical
 Animal data usually surrounding safety but efficacy   might be added in POC
Clinical
 Phase I
  Usually safety, SAD, MAD, PK/PD, very short term
 Phase II
  Efficacy signal studied, dose ranging, may include   PK/PD
  Longer than in phase I, some safety data accumulated
 Phase III
  More robust, larger numbers, and longer duration,  would serve as pivotal trials for approval
NDA/BLA

It is not unusual that much of this information will be presented at a pre-IND meeting, but if it is not accrued at that time then there will likely be another meeting arranged to discuss these data. Therefore, there are very few drugs in development that are not known to the FDA. Once an IND is established and a unique identifier assigned to the product, the review division reviews each study as it is initially designed and information about both animal experiences and human exposures are submitted periodically, for example, after studies are completed and safety data are accrued and submitted in periodic safety update reports. Once this period of study is completed, which usually concerns both phase I and II studies of dose, dose duration, and bioavailability and metabolism of the new drug in humans, the developer of the drug then returns to the review division for an end of phase II meeting. At this time the developer discusses with the FDA the plans for the phase III pivotal trials to allow for the accrual of data, which will be the primary data used for the New Drug Application (NDA) or Biologic Licensing Application (BLA) allowing for subsequent registration and marketing of the product if approved. This discussion is critical in that the phase III program should be large enough for the FDA to ultimately have enough data to understand relative risk and relative benefit, as well as enough information accrued from the potential completed data set to describe how to use the new drug, its risks, its expected benefits, and any approaches to manage the risk that would be learned along the way. This information is what is printed within the product label, which reflects the adjudicated, unbiased evaluation of the totality of the evidence accrued in the development program.

The total amount of expected patient exposure is defined in a consensus document that is agreed to by many of the regulatory bodies around the world. This “harmonization” document suggests that minimum exposure for a drug meant for chronic use would be defined as a total of 1,500 patients with 100 patients exposed for 1 year, and 300–600 exposed for 6 months (6).

The phase III trials are then implemented, and once they are completed and analyzed, the company returns to the review division to once again gain insight into moving forward with submitting all of the accumulated evidence for review within their NDA or BLA. This pre-NDA or pre-BLA meeting is a crucial piece of the pie.

Once submitted, the product is then reviewed as described below. Once the review takes place, the new product could be approved or not approved. In addition, the product might be thought to be approvable, meaning it did work but more information needs to be accrued or other issues need to be clarified, or, as shown in Table 4, the drug is not approvable.

Table 4. Action letters*
  • *

    CFR = Code of Federal Regulations; NDA = New Drug Application.

Not approvable
 Lists deficiencies why application can not be approved
  21 CFR 314.120
  21 CFR 314.125
Approvable
 NDA can be approved, often with labeling changes and   phase IV study commitments
  21 CFR 314.110
Approval
 NDA is approved
  21 CFR 314.105

It is poorly understood by the general community what an actual review entails. A submission in electronic format could be up to the equivalent of 1,000,000 pages of information. It is split into the appropriate sections according to discipline, thus the chemists review the manufacturing information as well as the proposed way in which the product will be stored and shipped, etc. The safety and efficacy data are sent to the appropriate statisticians who actually dissect the information using the submitted raw data and often both check the submitted evidence and redo the analyses using both the method described by the drug developer and other methods. Sensitivity analyses may also be performed to check the integrity of the data. The components of the review outcome and the corresponding sections of the Code of Federal Regulations are outlined in Table 5.

Table 5. New Drug Application review (focus)
Substantial evidence of effectiveness
 “…Evidence consisting of adequate and well-controlled investigations, including clinical investigations, by qualified scientific experts, that proves the drug will have the effect claimed by its labeling…” (Kefauver-Harris Drug Amendments to Section 505[d] of Federal Food, Drug and Cosmetic Act, 1962)
Safety (Federal Food, Drug and Cosmetic Act of 1938)
Labeling (Original Food & Drug Act of 1906)

Because the organization of the FDA is hierarchical, each discipline involved in a review team is represented by the individual who is assigned the primary review for that discipline (Table 2). This individual reports to a team leader in that discipline who supervises the work. The team leader then reports to either the deputy director of the division or the division director, either of whom is responsible for the review process of the team as a whole. Not only is the team required to perform the review within the parameters defined by the user fee process, but the review is expected to be honest and quite rigorous to withstand outside review; in my experience, this was always the case. These documents are then placed within the permanent historical files of a division.

During the review process, the clinical research forms are randomly evaluated for completeness and inspections of the clinical sites where the studies were carried out are performed to ensure that the appropriate review boards were apprised of the studies, informed consent from the patients was completed, and charts of the progress were kept according to the expectations of good clinical practice. In multicenter trials as well as multicountry trials, analyses are performed to be sure that there are no unusual effects of the drugs in specific study centers or study countries that might have skewed the results. While this type of work is pursued, the medical officer then proceeds to dissect the safety analyses to ensure there is clear understanding of the safety signals. Monthly progress meetings of the review group are held with a general discussion as well as specific reports on the progress of each discipline.

Once a decision is reached, the sponsor of a potentially new therapeutic will receive a letter from the FDA. Table 4 outlines the types of letters. An approval letter will lead to the subsequent marketing of a product, and will be issued when the FDA has determined that the relative risk to benefit warrants the marketing of a product and when the label for that product is finished and approved by both the sponsor and the FDA. Another answer might be that the product is approvable, which suggests that there remain some questions but overall the therapeutic could be marketed if these questions are answered. Finally, there is the nonapproved letter, which fundamentally is issued when there is clear evidence that the relative risk to relative benefit as shown in the submitted database does not warrant marketing of the product at that time. These are predicated on what is known about the product at the time. Further study might allow that information to be updated, and with subsequent submitted information the decision might be changed.

At times, the FDA turns to outside experts who serve for specified terms on Advisory Committees (7). These committees are constructed to be diverse as far as demographics, race, and sex so that they represent the expertise throughout the country. There are 4-year terms of service, and the individual who joins such a committee becomes a special government employee and is governed in that role by oversight from the Ethics groups at the FDA. The ethics rules during the tenure of the committee member are not terribly dissimilar from those that full employees have to follow. The advisory committee might meet to discuss particular drugs under review; to review policy or deal with the development of guidances, which serve as roadmaps delineating a path for drug development; or even to meet behind closed doors to review proprietary information about a product or sponsor.

The Way the FDA Works

  1. Top of page
  2. A Brief History of the FDA
  3. The Components of the FDA
  4. The Review Process
  5. The Way the FDA Works
  6. Postmarketing Surveillance
  7. Conclusion
  8. REFERENCES

The daily work plan typically includes some aspect of work dedicated to the evaluation of a new drug application. The work plan also includes reviewing periodic safety updates that are submitted for drugs previously approved for marketing by the medical officer, which then have to be followed for the life of the drug; reviewing new investigational new drug applications; answering queries and consults regarding products under review from other divisions; reviewing policy and procedural changes; participating in educational programs ranging from statistical analysis programs to how to write reviews; and participating in organizational meetings of the division as well as the office. In addition, the work day is ∼8 hours, with an ∼40-hour work week. The employee is not able to hold another job or “moonlight” without permission, regardless of when that work might take place. All outside income must be vetted and reviewed by the ethics group at the FDA to ensure that there are no perceived conflicts of interest. In fact, there are limited opportunities to actually participate in outside professional activities, and the medical officers are urged to develop some liaison with an outside hospital or clinic so that they can participate in periodic medical care as might be practicable. This development time is critical to allow the medical officer to “keep their hands in” and allow them to maintain some perspective regarding the practice of medicine. This time is usually for half a day per week, but at times that can be significantly increased. The medical officer donates this time to the practice site while continuing to be paid by the FDA.

Much has been said about potential conflicts of interest, which is really unrealistic. The PDUFA user fee system has allowed for increased numbers of medical officers to perform the review function and participate in all of the other duties. The fact that a fee is paid for a review is not appreciated directly by a division or reviewer. Furthermore, the reviewers are all very dedicated to their jobs and the work that they do. There is a very rigorous ethical review of each employee on a regular basis that evaluates all sources of income and investment holdings. No employee or anyone in that employee's family can receive any type of income from any source whose activities are overseen by the FDA.

Postmarketing Surveillance

  1. Top of page
  2. A Brief History of the FDA
  3. The Components of the FDA
  4. The Review Process
  5. The Way the FDA Works
  6. Postmarketing Surveillance
  7. Conclusion
  8. REFERENCES

Although a significant amount of time and energy is spent reviewing new therapeutics before approval for relative risk to relative benefit, it should not be forgotten that this benefit-to-risk balance is truly understood by observing more real world use of a drug than use in randomized controlled trials. This is accomplished once a drug is marketed. At that point more heterogeneous groups of patients are exposed, including patients with comorbid illnesses who often are using other concomitant therapies. This type of information is gained by facilitated postmarketing reports that are submitted by the sponsor of a product, or through the voluntary MedWatch or Adverse Event Reporting System (AERS) maintained by the FDA. There are ∼300,000 MedWatch reports submitted about all therapies approved in the US per year. These are reviewed and correlated with other international reports, and lead to complex risk analyses that are considered by the Office of Drug Safety (FDA) along with the individual and appropriate Review Divisions. There is an attempt to understand the overall risk and benefit by reviewing both premarketing and postmarketing experiences. The premarketing experiences often allow hazard rates to be determined in that trials have denominators of patients exposed, whereas in the postmarketing realm, the number of prescriptions of a certain product is known but how many patients may take a therapeutic needs to be approximated. Therefore, the postmarketing data are somewhat more inferential than the clinical studies.

Other data sets can be reviewed to help supplement the accumulated data from the AERS data set. These sources can be from registries, cohort studies, and managed care organizations. Unfortunately, these approaches provide signals of potential problems; however, due to a lack of randomization as to how a patient receives one versus another therapeutic, channeling bias may decrease the utility of these observations. Nonetheless, these data along with the previously accumulated pre- and postmarketing data may provide a robust window into the risk of any particular therapeutic. All of these data are used to determine the importance of continued marketing of a product. This level of scrutiny is constant throughout the life cycle of any product in the US.

Conclusion

  1. Top of page
  2. A Brief History of the FDA
  3. The Components of the FDA
  4. The Review Process
  5. The Way the FDA Works
  6. Postmarketing Surveillance
  7. Conclusion
  8. REFERENCES

The process of regulatory oversight is one that needs to be detail oriented and rigorous. Presently, despite recent media coverage, the FDA is quite well constructed to provide robust and careful regulatory oversight of the development of therapeutics. I urge everyone interested in providing the best possible care to their patients to take advantage of the FDA Web site (2) as well as the product labels to learn the best and unbiased appropriately adjudicated evidence published about therapies.

REFERENCES

  1. Top of page
  2. A Brief History of the FDA
  3. The Components of the FDA
  4. The Review Process
  5. The Way the FDA Works
  6. Postmarketing Surveillance
  7. Conclusion
  8. REFERENCES