Limitations of Use Statements in FDA Prescribing Information: Legal Basis and Review Precedents

Limitations of Use in FDA prescribing information: legal basis, drafting conventions, and review precedents

A "Limitations of Use" (LOU) statement sits inside Section 1, Indications and Usage. It tells prescribers where the evidence behind an approved indication stops. FDA uses it to flag reasonable concern or uncertainty about safety or effectiveness in a particular population or clinical situation. It is not meant to list every gap in the data. Review documents show the agency adding LOUs at approval, moving them in and out of other sections during labeling negotiations, and removing them once new outcome or efficacy data resolve the uncertainty. This article covers the regulatory basis, FDA's drafting framework, and the review precedents that explain each type of action.

Key takeaways

• The legal hook is 21 CFR 201.57. Indications and Usage must describe limitations of usefulness and uncertainty about anticipated clinical benefit in certain situations, and Highlights must briefly note "major limitations of use" ²⁸³²⁸⁵.
• FDA's test is clinical relevance, not missing data. The 2018 draft Indications and Usage guidance reserves an LOU for situations where practitioners need to know about a reasonable concern or uncertainty for safe and effective use. A population that simply was not studied generally does not qualify ⁴⁰⁶⁴¹⁵.
• LOUs are added for uncertain effectiveness in a subset, safety signals in an off-label or longer-duration use, class benefits the product has not shown, biomarker-negative disease, statutory scope requirements, and to head off promotion for an unapproved use ^{59193190692226183}.
• LOUs are removed when new data close the gap. Examples in the review records include cardiovascular-outcome LOUs on lipid-lowering drugs (Zetia, Leqvio) ⁵⁷⁷⁶⁶⁴. "Reserve for limited options" language has also been removed after confirmatory trial data ⁵³⁹⁵⁴⁰.
• Placement is negotiated. FDA has moved statements from Warnings and Precautions into an LOU, kept restrictions in Section 5 instead of creating an LOU, and declined LOUs proposed by its own review team ^191602728.

Regulatory basis

Under 21 CFR 201.57(c)(2), when the applicable conditions are met, Indications and Usage must include "a succinct description of the limitations of usefulness of the drug and any uncertainty about anticipated clinical benefits, with reference to the 'Clinical Studies' section for a discussion of the available evidence." This applies when evidence supports safety and effectiveness only in selected subgroups, or when an indication is approved on a surrogate endpoint ²⁸³. If the limitation bears on dosing intervals, treatment duration, or dosage modification, the section must give a concise description and cross-reference Dosage and Administration ²⁸⁴.

For Highlights, 21 CFR 201.57(a)(6) requires a concise statement of each indication and says "Major limitations of use (e.g., lack of effect in particular subsets of the population, or second line therapy status) must be briefly noted" ²⁸⁵.

FDA's drafting framework

Core guidance documents

Two CDER/CBER labeling guidances set out the framework:

• Indications and Usage Section of Labeling for Human Prescription Drug and Biological Products: Content and Format (draft, July 6, 2018) ⁴⁶⁶⁶³
• Labeling for Human Prescription Drug and Biological Products: Implementing the PLR Content and Format Requirements (final, February 25, 2013) ⁶⁸

The draft guidance describes an LOU as a way to clarify the scope of an approved indication when reviewers have a reasonable concern or uncertainty about the risk-benefit profile. An LOU generally discourages use in a particular population or setting. That separates it from a contraindication, which applies where risk clearly outweighs any possible benefit ⁴⁰⁶.

When an LOU may be appropriate

The draft guidance identifies five circumstances:

1. Reasonable concern or uncertainty about safety or effectiveness in a clinical situation, including populations outside the approved age range. Where evidence supports use only in selected subgroups, an LOU may help if prescribers need to understand the concern outside the approved group. It should cross-reference Clinical Studies ⁴⁰⁷³⁸¹.
2. A common use that the weight of evidence shows is ineffective, or where benefits generally do not outweigh risks. This matters most when proven alternatives exist ⁴⁰⁸.
3. An important benefit shown by other drugs in the class has not been shown for this drug ⁴⁰⁹.
4. Important dose, interval, duration, or dose-modification considerations, including conditions that must be met before long-term use (for example, confirming individual response first) ⁴¹⁰⁴¹¹.
5. Statutorily required scope statements. For certain single-enantiomer products approved under FD&C Act section 505(u), the required statement that the product is not approved for a condition of use of the previously approved racemate should appear as an LOU ⁴¹².

When an LOU generally should not be used

The guidance advises against using an LOU to restate a contraindication or repeat what the indication already says. An example is noting that a combination-only product is not for monotherapy. Information that defines who should receive the drug (approved subgroups, prior-therapy requirements, adjunctive use, companion tests) belongs in the indication itself. Most warnings and precautions should not be repeated as LOUs ^41338290414.

Lack of data in an unstudied population generally does not justify an LOU unless there is a reasonable safety or effectiveness concern. The guidance gives three examples: wild-type disease when the indication is mutation-specific, lung transplant when the indication is heart transplant, and age groups outside a vaccine's approved range ⁴¹⁵³⁸⁶. Short-term trials in a chronic condition likewise do not call for an LOU when long-term use raises no known concern. Trial duration belongs in Clinical Studies ⁴¹⁶⁸⁸.

Format and wording

LOUs appear separately from the indication, under a "Limitations of Use" heading, usually without a numbered subsection. A numbered subsection may be better when one limitation covers several indications, and the format should make the scope clear ⁴¹⁷⁴¹⁸. In Highlights, major LOUs must be noted with clear scope. Limitations that apply to every indication may be grouped under a "Limitations of Use" subheading ⁵⁶²⁹².

The guidance does not prescribe one formula such as "not indicated," "not recommended," or "has not been established." Its model language is matched to what is actually known ⁴¹⁹⁴²⁰:

Accelerated approval and product-area guidance

FDA's final guidance on labeling for accelerated approval products (February 4, 2019) says that once clinical benefit is verified, Indications and Usage should be revised to reflect the supported population, including any new or remaining LOUs. Statements about limitations of usefulness and continued approval should be removed or revised as appropriate ¹²⁰. If an accelerated approval indication is withdrawn, the labeling must not suggest it is still approved. FDA may require an LOU about the withdrawn use, or place the associated risks in Warnings and Precautions or Adverse Reactions ^121122123. At initial accelerated approval, the basis for approval and its endpoint limitations should immediately follow the indication rather than sit under a separate heading ¹²⁴.

Product-area guidance makes related points:

• Biosimilars. The 2023 draft guidance on biosimilar and interchangeable labeling calls for labeling to address any LOUs relevant to the product's indications ²⁵¹.
• Antibacterials for unmet need (2017 guidance; 2025 Q&A). Indications and Usage should summarize the limits of the supporting data. FDA's example restricts use to patients with limited or no alternative treatment options ²⁵²²⁵³.

Precedents: when FDA adds a Limitation of Use

The rationales in these reviews match the categories in the 2018 draft guidance: subset uncertainty (PEDMARK, TPOXX), common or off-label use (VYVANSE, OMONTYS), unproven class benefit (WELCHOL, AVEED), and the statutory 505(u) statement (FETZIMA). They also show FDA using an LOU as a middle option between silence and a contraindication, as in the LINZESS pediatric discussion ¹⁵².

Precedents: narrowing, relocating, or declining an LOU

Labeling negotiations often turn on where a restriction belongs rather than whether it exists.

• Highlights and FPI must match, and lack-of-indication statements belong in Section 1. For PERFOROMIST (NDA 022007), FDA added LOU language to the full Indications and Usage section that had already been accepted in Highlights. FDA stated that limitations in Highlights must also appear in the FPI and that statements about lack of indication do not belong in Warnings and Precautions ¹⁵³¹⁵⁴.
• Moving into an LOU. For WELCHOL, a cardiovascular-outcome statement moved from Warnings and Precautions to an LOU ¹⁹¹.
• Keeping the LOU in Section 1. For NUEDEXTA (NDA 021879), the reviewer recommended keeping limitations material in Indications and Usage rather than moving it to Warnings and Precautions, because separating it from the indication could obscure the approved use. The reviewer also refined language distinguishing pseudobulbar affect from other emotional lability ¹⁸⁶¹⁸⁷.
• Using Section 5 instead of an LOU. For TAVNEOS (NDA 214487, 2021), FDA considered an LOU against use in active or uncontrolled chronic liver disease and cirrhosis because of hepatotoxicity concerns. It then accepted placing the restriction in Warnings and Precautions ⁶⁰.
• An LOU instead of a narrower indication. For XENOVIEW (NDA 214375, 2022), FDA agreed not to restrict the indication to pre-operative evaluation for lung surgery. It added "XENOVIEW has not been evaluated for use with lung perfusion imaging" instead ¹⁸⁴¹⁸⁵.
• Rewriting LOU bullets. For ARANESP (BLA 103951/S-5173, 2011), FDA replaced LOU bullets with statements that Aranesp is not indicated "In patients with cancer receiving myelosuppressive chemotherapy when the anticipated outcome is cure" and "has not been shown to improve quality of life, fatigue, or patient well-being." The related indication wording was revised to reflect the population in whom benefit had been shown ²²⁴²²⁵. For VALCYTE (NDA 021304/S-007), FDA recommended aligning the section 1.3 presentation with the four-bullet Highlights format and clarifying "in other solid organ transplants" ⁶¹⁴.
• An LOU declined by the review division. For LENMELDY (BLA 125758), the CBER clinical review team proposed two LOUs: possible accelerated motor progression in early-symptomatic early-juvenile patients, and lack of effect on gallbladder disease. Office of Therapeutic Products leadership did not agree. The lack of motor benefit was described in Section 14 instead ²⁷²⁸²⁹.

Precedents: when FDA removes or relaxes an LOU

Outcome data resolve a class-benefit LOU

The lipid-lowering class provides the clearest removal precedents:

• ZETIA (NDA 021445/S-042, July 20, 2023). FDA removed "The effect of ZETIA on cardiovascular morbidity and mortality has not been determined." The review cited the IMPROVE-IT cardiovascular outcomes results ⁵⁷⁷⁵⁷¹. Earlier Zetia labeling also said it had not been studied in Fredrickson Types I, III, IV, and V ⁶⁶².
• LEQVIO (inclisiran). FDA removed the cardiovascular morbidity and mortality LOU and simplified the indication, including dropping "maximally tolerated" before statin therapy. The review cited positive cardiovascular outcomes trials with other PCSK9 inhibitors together with inclisiran safety and effectiveness data ^664665666. As described in the review, the removal drew partly on class-level outcome evidence, which makes Leqvio an unusual case.
• VYTORIN (NDA 021687/S-061, February 8, 2024), NEXLETOL/NEXLIZET, and ZYPITAMAG (NDA 208379/S-005, January 31, 2024). Approval letters record LOU removal. For Zypitamag, the change was made under FDA's July 30, 2021 prior approval supplement request letter, alongside removal of "dyslipidemia" language and non-LDL-C parameters ^{578667658625626}. The letters do not give product-specific rationales for these removals, and the July 2021 request letter concerned indication wording broadly rather than LOUs specifically ⁶⁵⁸.

New indication or efficacy data accompany removal

• RYBELSUS (NDA 213051/S-024, October 17, 2025). FDA approved removal of the type 1 diabetes LOU in the same supplement that added a cardiovascular risk reduction indication. The approval letter does not state a rationale for the removal ^541542543.
• AVYCAZ (NDA 206494/S-002, June 22, 2016). FDA approved removal of "As only limited clinical safety and efficacy data for AVYCAZ are currently available, reserve AVYCAZ for use in patients who have limited or no alternative treatment options" from the complicated intra-abdominal infection indication. The same supplement added Phase 3 cIAI data, although the letter does not expressly tie the two together ^538539540. This follows the antibacterial guidance logic that such language reflects the limits of the data available at approval ²⁵².

Conversion from accelerated approval

The accelerated approval labeling guidance expects LOU and continued-approval statements to be removed or revised once benefit is verified ¹²⁰. The approval letters cited here show this happening to the restrictive indication language rather than to a statement headed "Limitations of Use." For KEYTRUDA (BLA 125514/S-4 and S-6, December 18, 2015), FDA removed the melanoma restriction to disease progressing after ipilimumab (and a BRAF inhibitor if BRAF V600-positive), along with the continued-approval statement ⁷²². For ALUNBRIG (NDA 208772/S-008, May 22, 2020), FDA removed the restriction to patients who had progressed on or could not tolerate crizotinib, along with the confirmatory-trial statement ⁷²³.

Practical implications for labeling strategy

• Tie every proposed LOU to a specific concern. Reviewers ask whether prescribers need the information for safe and effective use. Proposing "not studied in X" without a clinical concern runs against the draft guidance ⁴⁰⁶⁴¹⁵.
• Expect an LOU where the data invite off-label use. VYVANSE, TEMBEXA, and OMONTYS show FDA adding LOUs where observed effects or class history create a foreseeable unapproved use ^183380378.
• Class labeling cuts both ways. Class precedent has supported adding LOUs (REZZAYO, COTELLIC, AVEED) and removing them (LEQVIO) ^189697328664.
• Plan the evidence that will lift the LOU. Removals in these records follow outcome trials (ZETIA, LEQVIO), new indication data (RYBELSUS), or confirmatory efficacy data (AVYCAZ). They are processed as prior approval or efficacy supplements, sometimes at FDA's request ^{577664543540626}.
• Negotiate placement early. Section 1 versus Section 5 versus Section 14 is a common labeling-review discussion. FDA has moved statements in both directions ^1916018629.

Open questions for further research

Several follow-ups remain open. These include the specific rationale for the Rybelsus type 1 diabetes LOU removal and for the Vytorin, Nexletol/Nexlizet, and Zypitamag removals. Others are whether CDER has finalized the 2018 Indications and Usage guidance, and how LOUs are being handled for biosimilars and interchangeables relative to their reference products ^54357863251.