Step 1: discovery and development
Research starts in the laboratory. The FDA lists several routes to a new medicine: new insight into how a disease works, large screens of molecules, unexpected effects of existing treatments, and new technologies. Thousands of compounds may be candidates at first; after early testing only a few look promising [1].
For those few, researchers study how the body absorbs, distributes, breaks down and removes the compound, how it works, the best dose and route, its side effects, how it behaves in different groups of people and how it interacts with other treatments [1].
Step 2: preclinical research
Before a medicine is given to people, researchers must find out whether it could cause serious harm. Preclinical studies are done in vitro (in the lab) and in vivo (in animals), under the FDA’s good laboratory practice rules. They are usually not large, but they must give detailed information on dosing and toxicity [2].
Preclinical results stay in the label for good. Section 13.1 of the Ozempic® label, for example, describes two-year studies in mice and rats, and those rodent findings are the reason GLP-1 labels carry their thyroid boxed warning [13].
Step 3: clinical research in phases
Before any study in people, the developer files an Investigational New Drug application (IND) with animal and toxicity data, manufacturing details and study plans. The FDA has 30 days to review it and can place a clinical hold if participants would face unreasonable or significant risk; the FDA says holds are rare, and it more often comments on how to improve a trial [3]. The FDA describes the phases this way [3]:
- Phase 1: 20 to 100 healthy volunteers or people with the condition, over several months, to study safety and dosage. About 70% of medicines move on.
- Phase 2: up to several hundred people with the condition, from several months to 2 years, for efficacy and side effects. About 33% move on.
- Phase 3: 300 to 3,000 volunteers, over 1 to 4 years, to show benefit and monitor adverse reactions. These larger, longer studies are more likely to reveal rare or long-term side effects. About 25% to 30% move on.
- Phase 4: studies of several thousand people, carried out after approval.
GLP-1 trials, by name
Section 14 of a label is where the phase 3 trials appear. The Mounjaro® label says its blood sugar effect was established in five trials, named SURPASS-1 to SURPASS-5, which compared tirzepatide with placebo, semaglutide and insulins [12]. The Ozempic® label names SUSTAIN 6 for heart outcomes and FLOW for kidney outcomes [13].
According to the FDA, a developer needs adequate data from two large controlled trials before it can apply for marketing approval [3]. The Foundayo™ label rests on two 72-week, randomised, placebo-controlled trials in adults [15]. It also notes that those trials used an investigational formulation of orforglipron, and presents the results as the equivalent doses of the marketed tablet [15].
Step 4: FDA review
With enough evidence, the company files a New Drug Application (NDA), which the FDA calls the full story of a drug: every study from the lab to phase 3, proposed labeling, safety updates and directions for use. Once the FDA accepts the application as complete, its reviewers have 6 to 10 months to reach a decision [4].
Medical officers and statisticians review the clinical data, pharmacologists the animal studies, and FDA inspectors visit trial sites to look for fabricated or withheld data. Once the evidence shows the medicine works and is safe enough for its proposed use, the FDA and the company settle the wording of the prescribing information [4]. When questions need more consideration, the FDA may ask an independent advisory committee, which includes a patient representative, for advice [4].
The Foundayo™ approval letter shows these pieces in a real file. Lilly’s application was received on 20 January 2026 and approved on 1 April 2026; the letter states that it was not referred to an advisory committee because there were no controversial issues needing discussion [11]. The pages checked for this article do not explain why that review was shorter than the usual window.
First approvals of GLP-1 medicines in the FDA record
Drugs@FDA lists each application with its type and review priority. A “new molecular entity” is an active ingredient never before marketed in the US in any form; other first approvals can be a new dosage form or a new indication filed as a separate application [9][10]. The table shows the original approvals checked for this article.
| Product (application) | Approved | Application type | Review |
|---|---|---|---|
| Ozempic® injection (NDA 209637) | 5 December 2017 | New molecular entity | Standard [9] |
| Rybelsus® tablets (NDA 213051) | 20 September 2019 | New dosage form | Priority [9] |
| Wegovy® injection (NDA 215256) | 4 June 2021 | New indication as a separate application | Priority [9] |
| Mounjaro® (NDA 215866) | 13 May 2022 | New molecular entity | Priority [9] |
| Zepbound® (NDA 217806) | 8 November 2023 | New indication as a separate application | Priority [9] |
| Wegovy® tablets (NDA 218316) | 22 December 2025 | New dosage form | Standard [9] |
| Foundayo™ (NDA 220934) | 1 April 2026 | New molecular entity | Standard [9][11] |
Faster routes and approval with conditions
Priority review is a designation the FDA gives to applications that, if approved, would significantly improve the safety or effectiveness of treating a serious condition. Its goal is action within 6 months instead of 10, and the FDA says it does not change the scientific or medical standard for approval [7].
Accelerated approval works differently. Since 1992 it has allowed medicines for serious conditions with an unmet need to be approved on a surrogate endpoint, a marker thought to predict clinical benefit, instead of waiting for the benefit itself; the company must still confirm the benefit, and the approval or indication can be withdrawn or changed if confirmatory trials fail [6]. The Wegovy® label says its MASH indication was granted this way, based on improvement in liver disease and scarring, and that continued approval may depend on a confirmatory trial [14].
The Highlights page of every label also shows an Initial U.S. Approval year. The FDA explains that this is the year the new molecular entity was first approved [8]: 2017 for semaglutide, 2022 for tirzepatide and 2026 for orforglipron [12][13][15].
Step 5: after approval
The FDA says it is impossible to know everything about a medicine’s safety at approval, so the picture keeps developing for years. Significant changes, such as a new formulation, labeling or strength, need a supplemental application, and the FDA inspects manufacturing sites, regulates advertising and reviews reports of problems through MedWatch [5][10].
The FDA can also require studies as a condition of approval. In the Foundayo™ letter it deferred studies in children aged 6 to 17 under the Pediatric Research Equity Act and, finding that spontaneous reports would not be enough, required a medullary thyroid carcinoma registry case series of at least 15 years, pregnancy studies, a lactation study, an ultrasound study of food left in the stomach, and a report from ACHIEVE-4, a trial still running at approval, with more data on heart events and liver injury [11].
Labels change in both directions. The Mounjaro® label was first approved in 2022 for adults; later supplements added children aged 10 and older with type 2 diabetes, based on SURPASS-PEDS, and a cardiovascular indication, based on SURPASS-CVOT, a trial of 13,299 adults that compared tirzepatide with dulaglutide [12]. The Wegovy® label’s Recent Major Changes box shows a warning on suicidal behaviour and ideation removed in 01/2026 and a new warning never to share a pen between patients added in 06/2026 [14]. The FDA defines that box as the record of substantive recent changes to the main sections of a label [8].
Generic copies and tentative approval
A generic medicine follows a shorter route. Its application, an Abbreviated New Drug Application, generally does not need new animal or human studies of safety and effectiveness; instead, the applicant must show that its product is bioequivalent, meaning it performs in the same way as the original [10].
If a generic is ready before the original’s patents or exclusivities run out, the FDA can issue a tentative approval, which does not allow the product to be marketed until those issues are resolved [10]. Drugs@FDA lists one such case for semaglutide: a tentative approval on 7 April 2026 for a generic semaglutide injection in three strengths [9]. A tentative approval is not a generic you can buy.
Where you fit in
Post-approval monitoring depends partly on people who use the medicines. Reports of side effects, from patients as well as clinicians, are one of the ways new safety information reaches the label [5]. Our MedWatch article explains how reporting works.
Mounjaro® and Zepbound® are registered trademarks and Foundayo™ a trademark of Eli Lilly and Company. Ozempic®, Rybelsus® and Wegovy® are registered trademarks of Novo Nordisk A/S. PepLabo is not affiliated with or endorsed by Eli Lilly and Company or Novo Nordisk.
