IND exemption: the five 21 CFR 312.2(b) conditions and how to document them

At a glance

• FDA does not grant an IND exemption. A clinical investigation of a drug lawfully marketed in the United States is exempt from Part 312 only if all five conditions in 21 CFR 312.2(b)(1) apply.
• The five conditions cover intent to change the labeling, intent to change prescription drug advertising, added risk from the route, dose, population or another factor, IRB review and informed consent, and the ban on promoting investigational drugs.
• An exempt study still has to meet the IRB requirements of 21 CFR Part 56 and the informed consent requirements of 21 CFR Part 50.
• FDA will not accept an IND for an investigation that is exempt under 312.2(b)(1), and a study that uses the exception from informed consent in 21 CFR 50.24 is never exempt.
• When the answer is unclear, FDA will advise on request whether Part 312 applies to a planned investigation.

What 312.2 does: applicability first, exemption second

Section 312.2 answers two questions in order. Paragraph (a) sets the scope: Part 312 applies to all clinical investigations of products subject to section 505 of the Federal Food, Drug, and Cosmetic Act or to the licensing provisions of the Public Health Service Act. Paragraph (b) then lists the investigations that are exempt from it.

Two uses of a drug are outside Part 312 before the exemption question arises. Part 312 does not apply to the use of an approved drug or licensed biologic for an unlabeled indication in the practice of medicine (312.2(d)). And the definition of an experiment in 312.3 is “any use of a drug except for the use of a marketed drug in the course of medical practice.” A physician prescribing off label for a patient is practising medicine. The same physician randomising patients to answer a research question is running a clinical investigation.

If an investigation is in scope and not exempt, the sponsor has to submit an IND and may not begin the investigation until the IND is in effect (312.20). The 21 CFR 312 overview covers the rest of the part.

The marketed-drug exemption: five conditions, all required

Section 312.2(b)(1) exempts the clinical investigation of a drug product lawfully marketed in the United States “if all the following apply.” The test is cumulative. A study that passes four conditions and fails one needs an IND.

Condition	Citation	What to check	Where studies fail
No intent to support a new indication or other significant labeling change	312.2(b)(1)(i)	Will anyone report this study to FDA as a well-controlled study for a new indication, or use it to support a significant labeling change?	A manufacturer supplies the drug or funds the study and plans to use the data in a supplement
No intent to support a significant change in advertising (prescription drugs)	312.2(b)(1)(ii)	If the drug is marketed as a prescription product, is the study meant to support new advertising claims?	Marketing interest in the result is treated as irrelevant because the investigator has none
No significant increase in risk	312.2(b)(1)(iii)	Does the route, dose, population or any other factor significantly increase the risks, or decrease the acceptability of the risks, compared with the marketed use?	A new route, a dose above the label, or a population the label does not cover, such as children or pregnant women
IRB review and informed consent	312.2(b)(1)(iv)	Will the study comply with Part 56 (IRBs) and Part 50 (informed consent)?	Teams assume an exempt study needs no IRB
No promotion	312.2(b)(1)(v)	Will the study comply with 312.7, which bars promoting the drug as safe or effective for the use under study?	Study materials or press activity make efficacy claims for the investigated use

Condition (iii) is where most of the judgment sits. The regulation does not define “significantly,” so someone has to weigh the study against the approved labeling and write down why the added risk is or is not significant. The other four are closer to yes or no, but conditions (i) and (ii) turn on intent, which means the sponsor’s plans and any funder’s plans both count.

Three illustrations of how the conditions interact:

• An academic team compares two labelled dosing schedules of an approved oral drug in the labelled adult population and plans only to publish. If no factor adds significant risk, the study can meet all five conditions.
• The same team proposes twice the maximum labelled dose, or intravenous use of a drug approved only as a tablet. Condition (iii) is now in doubt, and the team has to justify why the risk is not significantly increased or file an IND.
• The manufacturer offers to fund the first study in exchange for the data, which it plans to submit for a new indication. Condition (i) fails, whatever the risk profile.

The other exempt categories in 312.2(b)

The marketed-drug exemption is the one most readers need, but 312.2(b) lists others:

• A clinical investigation of blood grouping serum, reagent red blood cells or antihuman globulin is exempt if the product is used to confirm a diagnosis made by another medically established diagnostic product or procedure and is shipped in compliance with 312.160 (312.2(b)(2)).
• A drug intended solely for tests in vitro or in laboratory research animals is exempt if shipped in accordance with 312.160 (312.2(b)(3)).
• A clinical investigation involving a placebo is exempt if the investigation does not otherwise require an IND (312.2(b)(5)). A placebo does not trigger Part 312 by itself, but a placebo-controlled trial of an investigational drug still needs an IND for that drug.

One category is excluded outright. A clinical investigation that involves an exception from informed consent under 21 CFR 50.24 is not exempt (312.2(b)(6)).

In vivo bioavailability studies follow their own rule in 21 CFR 320.31 (312.2(c)). Studies of medical devices fall under FDA’s investigational device regulations in 21 CFR Part 812, which FDA’s 2013 guidance on studies without an IND points to and which this guide does not cover.

FDA has also proposed new exemptions. In December 2022 it published a proposed rule (Federal Register document 2022-26728) for certain studies of products lawfully marketed as conventional foods, dietary supplements or cosmetics when studied for a drug use. It is a proposal. Part 312 as it stood on 14 April 2026 contains no such exemption.

Decision flow: does this study need an IND?

Work through the steps in order and stop at the first answer.

1. Is it a clinical investigation? If the drug is used in the course of medical practice, including off-label use for an individual patient, Part 312 does not apply. If the protocol assigns treatment to answer a research question, continue.
2. Is the product subject to section 505 of the FD&C Act or to biologics licensing? If not, Part 312 does not apply. If it is, continue.
3. Does a specific category apply? An in vitro diagnostic biological product used as 312.2(b)(2) describes, a drug used only in vitro or in laboratory animals, or a placebo in a study that needs no IND for anything else, is exempt. Otherwise continue.
4. Does the study use the 50.24 exception from informed consent? If yes, it needs an IND. If no, continue.
5. Is the drug lawfully marketed in the United States? If no, it needs an IND. If yes, continue.
6. Do all five 312.2(b)(1) conditions hold? One “no” means an IND. Five documented “yes” answers mean the study is exempt.
7. Still unsure? Ask FDA (see below) before enrolling anyone.

Who decides, and how to document it

Part 312 places the duty on the sponsor. A sponsor who intends to run a clinical investigation that is subject to 312.2(a) must submit an IND (312.20(a)), so the person who would have filed is the person who decides that filing is not required. In an investigator-initiated study that person is often a sponsor-investigator, an individual who both initiates and conducts the investigation and who carries the obligations of both roles (312.3). The investigator-initiated trials guide covers that role in more depth.

FDA’s 2013 guidance on whether studies can be conducted without an IND says the sponsor, or the sponsor-investigator, should in most cases be able to make this determination. If the sponsor is uncertain, the guidance recommends contacting the FDA review division for the therapeutic area. The regulation backs that route: on request, FDA will advise on the applicability of Part 312 to a planned clinical investigation (312.2(e)). The guidance describes informal inquiries, which get a nonbinding answer that FDA does not track, and formal written inquiries with enough detail, such as a protocol, which get a written response limited to the investigation described. A pre-IND meeting is a separate route, suited to a program that probably needs an IND anyway.

The determination does not need FDA’s signature, but it needs a record. A file that would satisfy a monitor, an auditor or your own IRB contains:

• The protocol version and date the determination covers.
• The product’s marketed status and the labeling version you compared the study against (route, dose, population, indication).
• A written answer to each of the five 312.2(b)(1) conditions, with the reasoning for condition (iii).
• A statement of intent from the sponsor, and from any manufacturer or funder, that the data will not support a new indication, a significant labeling change or, for a prescription drug, a significant advertising change.
• The IRB approval reference and the approved consent form version.
• Any correspondence with FDA, marked as informal or formal.
• The name of the person who made the determination and the date.
• A list of changes that will trigger a fresh review (see the last section).

If the answer is no, the IND goes in under Form FDA 1571, and the IND application guide covers what the dossier has to contain. Filing anyway “to be safe” is not an option for a study that qualifies: FDA will not accept an application for an investigation that is exempt under 312.2(b)(1) (312.2(b)(4)).

IND exemption is not IRB “exempt” review

Exempt from Part 312 does not mean exempt from IRB review. Condition (iv) makes compliance with Part 56 and Part 50 part of the exemption itself (312.2(b)(1)(iv)). Part 56 sets the standards for IRBs that review clinical investigations regulated by FDA (56.101(a)), and under Part 50 an investigator may not involve a human subject without legally effective informed consent, except as Part 50 itself provides (50.20). The informed consent guide covers what that consent requires.

FDA’s own exemptions from IRB review are short and specific. Part 56 exempts investigations that began before 27 July 1981, emergency use of a test article reported to the IRB within 5 working days, and certain taste and food quality studies (56.104). An IND exemption is not on that list.

The “exempt” label many IRB staff use day to day is a different thing: a category of IRB review, not the Part 312 IND exemption. When an IRB letter says “exempt”, check which rule the word comes from. For an FDA-regulated drug study claiming the IND exemption, Part 56 review applies. The IRB approval guide explains the review tiers.

When an exempt study stops being exempt

The exemption attaches to the investigation as designed, so it lasts only while all five conditions still hold. Part 312 has no separate procedure for an exempt study that changes. If an amendment breaks a condition, the investigation is subject to 312.2(a), and 312.20(b) bars a sponsor from beginning an investigation subject to 312.2(a) until an IND is in effect for it.

Re-run the five-condition test whenever you change any of these:

• The dose, route, schedule or duration, beyond what the original determination assessed.
• The population, especially toward children, pregnant women or patients with organ impairment.
• The funder, drug supplier or data-sharing terms, because a new party may want the results for labeling.
• The endpoints or design, in a way that makes the study look like a well-controlled study for a new indication.

FDA’s 2013 guidance adds a warning for anyone who relied on a formal FDA answer: if the protocol or other aspects of the investigation change significantly after FDA responds, that response may no longer be valid. If a change does require an IND, hold the change until the IND is in effect. From then on the study sits under the full IND rules, including FDA’s clinical hold procedures.

Sources

• 21 CFR Part 312, Investigational New Drug Application, eCFR text up to date as of 14 April 2026
• 21 CFR Part 50, Protection of Human Subjects, 2024 edition
• 21 CFR Part 56, Institutional Review Boards, 2024 edition
• FDA guidance: Investigational New Drug Applications (INDs), Determining Whether Human Research Studies Can Be Conducted Without an IND, September 2013
• FDA proposed rule: Exemptions for Clinical Investigations To Evaluate a Drug Use of a Product Lawfully Marketed as a Conventional Food, Dietary Supplement, or Cosmetic, 9 December 2022