Esta referencia está disponible actualmente en inglés. La versión en español está en preparación.

Clinical Trial Phases: From IND to FDA Approval

Por Benn Bluestein-Veyra, M.Sc. en Química Orgánica · Actualizado el 21 de octubre de 2026

The phases of clinical trials are the regulated steps through which an investigational drug or biologic moves from first human use to marketing approval and beyond. Under 21 CFR 312.21, phase 1 studies metabolism and pharmacologic action, phase 2 evaluates effectiveness in controlled studies, and phase 3 gathers the expanded evidence for the overall benefit-risk evaluation.^[1]

An Investigational New Drug application (IND) opens the sequence, and a marketing application, an NDA or a BLA, closes it before phase 4 begins.^[2][3] This reference explains each gate, each phase and the record fields that let a reader place any study in the sequence.^[4]

Mapping the phases of clinical trials from preclinical research to phase 4

FDA describes drug development in five steps: discovery and development, preclinical research, clinical research, FDA review and postmarket safety monitoring. Its Step 3 page explains that preclinical research answers basic questions about safety and that clinical research means studies done in people.^[3]

The sequence runs from preclinical work through the IND gate, three investigational phases and the marketing application gate to phase 4 (Fig. 1).^[1][3] The regulation adds that the phases are generally conducted in order and may overlap.^[1]

Umscheid and colleagues trace the same path in their review of key trial concepts, from the preclinical stage to postmarketing surveillance, with ethical foundations and oversight throughout.^[4] Van Norman reported an average of about 12 years from preclinical testing to approval for a new drug in the United States.^[5]

Measuring progression between phases in large datasets

Wong, Siah and Lo estimated success rates and durations from 406,038 trial entries covering more than 21,143 compounds between 2000 and 2015, broken down by clinical phase, sponsor type and biomarker use. Trials that used biomarkers in participant selection had higher overall success probabilities.^[6]

FDA gives approximate shares of drugs that move on: about 70% after phase 1, about 33% after phase 2 and about 25 to 30% after phase 3.^[3] DiMasi and colleagues linked the costs of compounds discontinued during testing to approved ones and estimated a capitalized pre-approval cost of $2,558 million per approved compound in 2013 dollars.^[7]

Fogel reviewed literature from the past 30 years on what distinguishes successful trials. The review identified recurring patterns tied to the study site, the coordinator and investigator, and the participant experience.^[8]

Preclinical studies and the IND under 21 CFR 312 open human research

A sponsor submits an IND before starting a clinical investigation of an investigational new drug, and may begin only once the IND is in effect under 21 CFR 312.40.^[2] FDA summarizes the application as animal study and toxicity data, manufacturing information, clinical protocols, data from prior human research and information about the investigator.^[3]

Understanding the contents of an IND

21 CFR 312.23(a) lists the sections in order: a cover sheet on Form FDA 1571, a table of contents, an introductory statement and general investigational plan, the investigator’s brochure and a protocol for each planned study (Fig. 2).^[9] Chemistry, manufacturing and control information, pharmacology and toxicology information, previous human experience and additional information follow.^[9]

The cover sheet identifies the phase or phases to be conducted and commits the sponsor to IRB review under 21 CFR 56.^[9] Phase 1 protocols may be less detailed and more flexible, and phase 2 and 3 protocols describe all aspects of the study in detail.^[9]

Applying the 30-day review and clinical holds

An IND goes into effect 30 days after FDA receives it, or earlier on FDA notification, unless FDA notifies the sponsor of a clinical hold.^[10] FDA describes an IND review team of a project manager, medical officer, statistician, pharmacologist, pharmacokineticist, chemist and, for antimicrobial products, a microbiologist.^[3]

A clinical hold is an FDA order to delay a proposed investigation or suspend an ongoing one, and 21 CFR 312.42(b) lists the grounds for each phase.^[11] Those grounds address participant protection, investigator qualifications, the accuracy of the investigator’s brochure and whether the IND holds enough information to assess the proposed studies.^[11]

After a sponsor submits a complete response to a hold, FDA answers in writing within 30 calendar days, and the study resumes once FDA notifies the sponsor that it may proceed.^[11] FDA notes that a clinical hold is rare and that the agency often provides comments intended to improve the quality of a trial.^[3]

Phase 1, 2 and 3 answer distinct questions under 21 CFR 312.21

21 CFR 312.21 defines the three investigational phases by purpose, and FDA, NIH and ClinicalTrials.gov describe them with typical participant numbers (Fig. 3).^[1][3][12] The ranges differ between sources, so each is best read as an approximation.^[13]

Phase 1 studies metabolism, pharmacology and dose

Phase 1 includes the initial introduction of an investigational drug into humans. It studies metabolism and pharmacologic actions, the side effects associated with increasing doses and, if possible, early evidence on effectiveness, in a total that is generally 20 to 80 subjects.^[1]

Storer compared the standard dose-escalation design with simple alternatives in small samples, and found that two-stage designs reduced bias in maximum likelihood estimates of the maximum tolerable dose in less than ideal dose-response settings.^[14] ClinicalTrials.gov describes phase 1 trials as usually conducted with healthy volunteers and a small number of participants.^[13]

Phase 2 gathers controlled evidence on effectiveness

Phase 2 comprises controlled clinical studies that evaluate effectiveness for a particular indication and determine common short-term side effects, usually in no more than several hundred subjects.^[1] NIH gives a typical range of 100 to 300 participants, and FDA describes phase 2 data as the basis for refining research questions and designing phase 3 protocols.^[3][12]

Simon published two-stage phase 2 designs that minimize the expected sample size, or the maximum sample size, within set limits on type 1 and type 2 errors. In these designs a first stage enrolls a small group, and the study continues to the second stage when the observed activity meets a prespecified threshold.^[15]

Phase 3 provides the expanded evidence for approval

Phase 3 trials are expanded controlled and uncontrolled trials, performed after preliminary evidence of effectiveness, that gather the information needed to evaluate the overall benefit-risk relationship and to support physician labeling.^[1] The regulation gives several hundred to several thousand subjects, and FDA calls these studies pivotal.^[1][3]

Downing and colleagues characterized 448 pivotal efficacy trials behind 188 novel agents approved from 2005 to 2012. Most were randomized and double-blinded, the median enrollment per indication was 760 participants, and 36.8% of indications rested on a single pivotal trial.^[16]

Reports of randomized phase 3 trials are read against the CONSORT 2025 statement, a 30-item checklist with a diagram of participant flow developed by Hopewell and colleagues.^[17] Its companion, SPIRIT 2025, sets the minimum content of the trial protocol.^[18]

Phase 0 and adaptive designs extend the classic sequence

FDA’s 2006 guidance on exploratory IND studies describes trials conducted early in phase 1 with very limited human exposure, before the traditional dose escalation studies, such as screening and microdose studies.^[19] ClinicalTrials.gov records these as Early Phase 1, formerly listed as Phase 0.^[13]

Burt and colleagues reviewed phase 0 and microdosing approaches, whose uses now extend beyond pharmacokinetics to mechanism of action and pharmacodynamics. They proposed that phase 0 approaches be considered in most drug development programs.^[20]

Adaptive and seamless designs follow prespecified rules

Pallmann and colleagues define adaptive designs as trials that use accumulating results to modify the trial’s course under prespecified rules, and note that they apply across all phases, from early dose escalation to confirmatory trials. Their tutorial centers on the interpretation and transparent reporting of such trials.^[21]

Seamless phase 2/3 designs select a dose or arm at an interim analysis and compare it with control in the same study. Stallard and Todd reviewed the group-sequential, combination test and adaptive Dunnett approaches to these designs in one framework.^[22]

The Adaptive designs CONSORT Extension (ACE), developed by Dimairo and colleagues through a Delphi process and consensus meeting, adds seven new items and modifies nine for reporting randomized adaptive trials.^[23] Park and colleagues identified 83 master protocols (49 basket, 18 umbrella and 16 platform trials), and four of the platform trials used a seamless phase 2/3 design.^[24]

IRB review and informed consent protect participants at every phase

21 CFR 56.103 requires that a clinical investigation needing prior FDA submission be reviewed and approved by an IRB, and remain under its continuing review, before it starts.^[25] Each IRB has at least five members with varying backgrounds, and it may approve, require modifications in or disapprove the research.^[25]

Under 21 CFR 50.20, no investigator may involve a person as a subject without legally effective informed consent, sought under circumstances that minimize coercion or undue influence.^[26] Section 50.25 lists the basic elements, including the purposes, procedures and expected duration of the research and whom to contact with questions.^[26]

The IND links the two parts: 21 CFR 312.40 permits an investigational drug in a clinical investigation when sponsor and investigators comply with parts 50 and 56.^[10] Emanuel, Wendler and Grady proposed seven requirements for ethical clinical research, among them scientific validity, independent review and informed consent.^[27]

The protocol is the document the IRB and FDA review. SPIRIT 2025, led by Chan and colleagues, sets 34 minimum protocol items, with a new open science section and an item on public involvement in trial design, conduct and reporting.^[18]

Marketing applications, FDA review and phase 4 complete the sequence

After phase 3, a sponsor files a New Drug Application under 21 CFR 314.50 or a Biologics License Application under 21 CFR 601.2.^[28][29] A BLA for a license under section 351 of the PHS Act includes nonclinical and clinical data demonstrating that the product meets requirements of safety, purity and potency.^[29]

FDA’s goal is to act on an application within 10 months under standard review and within 6 months under priority review, and priority designation leaves the scientific standard for approval unchanged.^[30]

Darrow, Avorn and Kesselheim reported that FDA review times fell from more than 3 years in 1983 to less than 1 year in 2017, while total time from authorization of clinical testing to approval stayed at about 8 years.^[31] The share of approvals supported by at least 2 pivotal trials moved from 80.6% in 1995 to 1997 to 52.8% in 2015 to 2017.^[31]

Phase 4 studies, postmarketing requirements and commitments

Concurrent with approval, FDA may seek agreement from a sponsor to conduct phase 4 studies that add information about a drug’s risks, benefits and optimal use, for example other doses, other populations or longer use.^[32] FDA separates postmarketing requirements, which a statute or regulation mandates, from postmarketing commitments that a sponsor agrees to conduct.^[33]

Wallach and colleagues characterized 437 postmarketing requirements for 97 drugs and biologics approved from 2009 to 2012, and 30.7% were prospective cohort studies, registries or clinical trials. Of 134 such studies, 76.1% were registered on ClinicalTrials.gov.^[34]

Cell and vesicle products developed as biologics follow the 351 pathway

Human cells and tissues regulated solely under section 361 of the PHS Act must meet all the criteria of 21 CFR 1271.10.^[35] Under 21 CFR 1271.20, a product outside those criteria is regulated as a drug, device or biological product, and a biological product reaches the market through a license under section 351.^[36][37]

The HCT/P and biologic classification reference on this site explains why isolated vesicle preparations fall outside the 361 pathway. For a cell or vesicle product developed as a biological product, the IND, the three phases and the BLA described above form the development path, with CBER as the reviewing center.^[29][38]

Remestemcel-L shows the endpoint of that path for a cell product. CBER licensed it under section 351(a) of the PHS Act on December 18, 2024, under STN BL 125706/0, in an approval letter that carries the address of FDA’s White Oak campus in Silver Spring, Maryland (Fig. 4).^[39] The remestemcel-L reference on this site walks through that approval record.

Expedited programs support development and review

FDA’s 2014 guidance on expedited programs describes fast track designation, breakthrough designation, accelerated approval and priority review.^[40] Fast track and breakthrough designation support the development program, and priority review sets a shorter goal for the review of the application itself.^[30][40]

Section 506(g) of the FD&C Act, added in 2016, created the RMAT designation for regenerative medicine products outside the 361 pathway that are intended for a serious or life-threatening condition, when preliminary clinical evidence indicates potential to address unmet medical needs. A request accompanies or amends an IND.^[38][41] FDA’s 2019 guidance describes RMAT designation alongside the other programs, including accelerated approval for designated products.^[41]

Kesselheim and colleagues found that priority review applied to 43% of 774 drugs approved from 1987 to 2014, and that the number of programs per new drug rose 2.6% per year.^[42] Vaggelas and Seimetz reviewed the RMAT designation and the FDA tools available to developers of cell and tissue products.^[43]

Reading a ClinicalTrials.gov record’s phase, status and enrollment fields

ClinicalTrials.gov assigns a phase from FDA definitions, using Early Phase 1, Phase 1, Phase 2, Phase 3 and Phase 4, and labels studies without FDA-defined phases, such as device or behavioral trials, Not Applicable.^[13] Phase 4 records include postmarketing requirement and commitment studies.^[13]

Recruitment status runs from "Not yet recruiting" through "Recruiting" and "Active, not recruiting" to "Completed", with "Suspended", "Terminated" and "Withdrawn" marking studies that stopped.^[13] Enrollment is the number of participants, and an estimated enrollment is the target number the researchers need.^[13]

Zarin and colleagues described the results database added in 2008 and found that 61% of a sample of registered outcome measures lacked specificity about the planned metric.^[44] Tse, Fain and Zarin set out 10 issues to consider when using the database for research, from its built-in capabilities to the reporting policies that shape which studies appear.^[45]

Where to go from here

The randomized controlled trials reference explains the design most phase 3 studies use, and the sham and placebo controls reference covers the comparators behind them. The hierarchy of evidence, risk of bias assessment and systematic reviews references show how trial results are graded and combined.

The reading the evidence guide applies these methods to clinic decisions, and the exosome clinical trials reference follows registered vesicle studies through their phases. The remestemcel-L reference, the extracellular vesicle potency assays reference and our quality references continue the thread, and our team is glad to discuss any of them.