Phase 3 studies are expanded in controlled and uncontrolled trials (Fig. 5.4). They are intended to gather the additional information about effectiveness and safety that is needed to define the overall benefit-risk relationship of the drug in the target population. Phase 3 studies should provide an adequate basis for extrapolating to the general population and transmitting that information in the product labeling for health care providers. 
r Confirm effectiveness in larger studies ("pivotal studies") r Typically need 2 positive, well-designed studies r Monitor adverse events c Over a longer period (12-24 weeks)
o More patients for more accurate frequencies and presentation r Design of study = package insert for marketing r Usually 1,000-3,000 patient volunteers at many sites with many investigators r Overall Duration ~2.5 - 5 years
Fig. 5.4. Phase 3 Studies r Scale-up in manufacturing r Heterogeneous patient sample r Competitive patient enrollment r Role/use of central IRB-advantages, limitations r Investigator & site training for product, protocol, & processes
Fig. 5.5. Phase 3 Studies: Issues
The decision to move ahead with phase 3 studies is a major one because the costs are considerably higher than for the two earlier phases combined. A drug identified as effective and safe in phase 2 studies may not enter phase 3 clinical trials for a number of reasons including insufficient efficacy when compared with its competitors, the expense and difficulty of drug formulation, especially when scaling up production, and side effects that exceed the risk profile needed to proceed.
Submission of a NDA requires at least two well-designed phase 3 studies that demonstrate both efficacy and safety in a large number of patients with the target disease, typically 1,000 to 3,000 patient volunteers. This large patient sample necessitates using many sites and investigators with their staffs, often 100 to 200 or more. The treatment period in phase 3 studies is longer than in phase 2 studies for drugs used chronically (12 to 24 weeks vs. 4 to 6 weeks). This provides an opportunity to monitor and detect adverse events and tolerance over a longer time.
A third phase 3 study may be incorporated into the initial drug development plan for at least two reasons. The risk exists that one of the two phase 3 studies may not be sufficiently positive for a new drug to be approved by the FDA. Also, procedural problems could occur in, for example, patient monitoring consistency or data collection, which results in the regulatory authority, after auditing sites and finding such a serious procedural problem, discarding an entire study from the NDA package. Disadvantages of this approach include the substantial increased costs and the longer time needed to complete the phase 3 studies. The advantage is the equivalent of an insurance policy for a more timely NDA submission, instead of waiting several years to conduct an added follow-up phase 3 study after the standard 3 studies were done. The usual time required to complete the phase 3 studies is about 2.5 years but could last 5 years. Then, data analysis and study reports are done, reviewed, and the NDA is filed, which can take 6 to 12 months or more at the company.
Phase 3 studies have several unique characteristics relative to the earlier phase studies (Fig. 5.5). They include the need to scale-up manufacturing to guarantee an adequate supply of drug for the duration of the studies, the more heterogeneous patient population than that studied in phase 2, the competitive nature of patient enrollment, the use of private as well as academic sites, and data reliability. Scaling up in manufacturing is not a trivial matter for some small molecules and even more so for biologics. For example, scaling up monoclonal antibody production may inadvertently change the characteristics of the antibody such that it is no longer identical to the antibody used in the early phase studies. Further, at this phase the formulation must be the same as what will be marketed and sold.
The more heterogeneous patient population than that studied in phase 2 creates a desirable, more representative patient sample but increases variability, the range of patient responses, provides more opportunity for side effects, and requires more patients to demonstrate a statistically significant benefit. The more heterogeneous patient population can reduce the signal to noise ratio such that a drug that was effective in phase 2 studies is no longer as effective in the pivotal phase 3 studies. Subject enrollment has increasingly become a rate-limiting step in the drug development process. Substantial efforts and funds must be expended to recruit appropriate research subjects. The need to recruit patients from a large number of sites makes it more efficient to use a central IRB, if possible. A variety of systems are used to assist in this recruitment, such as health care networks of hospitals and clinics, advertisements on the radio and in local media, Internet ads, and recruitment companies.
Academic sites are usually adds to use a central IRB, because of institutional policy and ethical concerns; however, they constitute a small percentage of the sites in phase 3 studies, which contrasts with their larger representation in phase 2 studies. Central IRBs are used by companies to expedite study approval at many sites, which may not have routine IRB access as in private physician offices. Cautions with such IRBs are their independence, sufficient expertise and appropriate representation, sufficient oversight of protocols, and appropriate oversight of the many investigative sights over wide geographic areas. Data reliability becomes a major issue in phase 3 with so many sites and people involved, which requires a significant investment in training of the site staff as well as investigators about the drug, protocol, and procedures, especially patient inclusion and exclusion, drug administration, monitoring requirements, and data collection requirements.
The typical study design used to demonstrate efficacy is randomized and placebo-controlled, but an active comparator control group can be considered (Fig. 5.6). In recent years, especially in Europe, there has been increasing concern about the safety and ethics of performing placebo-controlled trials. The most recent International Council on Harmonization/ Good Clinical Practice (ICH/GCP) guidelines recommend against doing placebo-controlled trials except under specific circumstances [4, 5]. This concern and guidelines have led to greater use of active comparator trials in which an approved, generally accepted therapy is the control arm and compared with the new therapy. When the goal of the study is to demonstrate superiority of the new therapy, the issues are the same as when the comparator therapy is placebo. However, if the goal is to show statistically equivalent benefit, it is called an equivalence or noninferiority trial (Fig. 5.6).
The major problem with noninferiority trials is the assumption that the active control treatment is effective in the trial
(e.g., the trial has an assay sensitivity). Unfortunately, this situation is not always true for effective drugs and is not directly testable from the data collected, because there is no placebo group . There are ways to maximize the value of noninferi-ority trials, such as determining from historical trials that the active control group reliably has an effect of at least a certain size, planning the trial design to be similar to that of prior trials (e.g., stage of disease, concomitant therapy, and end points), setting a noninferiority margin to be smaller than the total active control effect, and ensuring appropriate trial conduct (e.g., concomitant medications, study drug compliance). Nonetheless, because one cannot formally establish a minimal effect size, noninferiority cannot be per se taken as evidence of efficacy, and the interpretation of the trial must be based on the totality of the data, including additional analyses.
A number of additional studies may be performed during the time (2-3 years) from completion of the phase 3 studies to drug approval, also known as phase 3b studies (Fig. 5.7). This time is required to analyze and prepare the large amount of data for submission to the FDA, as well as the actual FDA review. The studies performed during this time serve to expand the adverse event database and dosing and efficacy data before approval, provide marketing support, increase physician participation (e.g., those in practice-based settings), institution familiarity with the drug prior to its approval and release, and increase the number of publications. Although the phase 2 and 3 programs may include 100 sites and 1,000 patients (smaller numbers for accelerated approvals), only 10 or 20 major universities may have participated, leaving many specialists without direct experience with novel products under study. Phase 3b allows for expansion of experts at more universities as well.
r Also called "equivalence" trial r Uses "active" comparator r Increasingly used due to concerns regarding placebo-controlled trials r Non-inferiority cannot be taken as evidence of efficacy!! Fig. 5.6. Non-Inferiority Trials r Studies performed during time (2-3 years) from completion of phase 3 studies to approval r Goals:
o Expanded adverse event database and dosing and efficacy data before approval e Marketing support e More physicians (e.g., those in practice based settings) and institutions become familiar with drug c Additional publications r Large size - may have sub-studies in certain populations k Requires FDA approval (file under IND)
Fig. 5.7. Phase 3b Studies r Total time from beginning phase 1 studies to product launch is usually 8-10 years r Many questions unanswered at end of phase 3 studies (e.g., longer term toxicity, use inspecial populations, role of genetic factors)
r Role of regulatory affairs department r Requires plans to do these studies integrated phase 3b and 4 studies
Many of these studies are quite large and may have sub-studies that utilize specific populations. All studies performed during this time require FDA approval and are filed under the original IND application. As noted earlier, some smaller studies may be performed at the time of the phase 2 studies.
In an summary of phase 1 to 3 studies (Fig. 5.8), the time from beginning phase 1 studies to product launch averages about 7 years (a range of 5-10 years). Even at the end of this lengthy period, many questions will remain to be answered such as longer term toxicity (withdrawal of Vioxx is one such example), use in special populations, and the role of genetic factors. Studies to answer these and other questions must be planned before the phase 3 studies are completed and conducted during phase 3b and 4 trials. Often, the FDA and company negotiate which of such studies need be done further as a contingency for approval. Many of such studies will enhance product use in patients, create good publications, and even improve sales.
All throughout the phases of clinical drug development, the regulatory affairs department should maintain an open, ongoing dialogue with the FDA. Reasons for doing this include (1) approval from the FDA is needed for each study (goals and designs) performed under the IND; (2) the FDA is privy to data from clinical trials of related drugs or of unrelated drugs in the same disease that may influence study design; and (3) any surprises with the FDA are avoided, which can slow the approval process. For example, the FDA may be aware of a possible toxic effect, not have anticipated from the preclinical pharmacology and toxicology. Suggestions from the FDA to incorporate additional measurements or modify other aspects of the study design should be considered very carefully.
Was this article helpful?