A site coordinator is screening a 52-year-old candidate for a Phase 3 multi-receptor agonist trial. BMI is 29.6 — just under the protocol's 30 kg/m2 threshold. Fasting glucose is borderline. There's a distant family history of thyroid cancer that isn't confirmed as medullary thyroid carcinoma. None of this disqualifies the candidate outright, but each data point changes what the consent conversation has to cover, and each one becomes a line item an auditor will eventually check against the source documents. This is the unglamorous mechanics of informed consent in GLP-1 and GLP-3 research — and it is where data integrity is either built or quietly compromised, months before a single dose is administered.
"GLP-3" is not a formally recognized receptor class; it is used in this piece, consistent with how the term circulates in research contexts, to describe next-generation multi-receptor agonists — GIP/GLP-1/glucagon triple agonists like retatrutide and amylin-GLP-1 combinations like cagrilintide/semaglutide — that remain investigational. That distinction matters for informed consent specifically, because disclosure obligations differ for a compound with an FDA-approved safety label versus one still generating its first long-term dataset.
The Regulatory Framework Governing Consent
Informed consent for U.S.-based trials is governed primarily by 21 CFR 50.25, which specifies eight required elements: a statement that the study involves research, its purposes and expected duration, a description of procedures, foreseeable risks or discomforts, expected benefits (if any), disclosure of alternative treatments, confidentiality limits, and compensation or treatment available if injury occurs. A ninth cluster of "additional elements" applies when relevant — unforeseeable risks, circumstances for involuntary termination, additional costs, and consequences of withdrawal.
Internationally, ICH E6(R2) Good Clinical Practice guidance layers on requirements for documentation, monitoring, and the qualifications of the person obtaining consent — it cannot be delegated to unqualified staff, and the process must be documented contemporaneously, not reconstructed after the visit. Under 45 CFR 46 (the Common Rule), an Institutional Review Board or Independent Ethics Committee must approve the consent form before enrollment and re-review it whenever the protocol or risk profile changes materially.
These aren't procedural formalities. FDA Bioresearch Monitoring (BIMO) inspection findings routinely cite consent-related deficiencies — missing IRB-approved versions, consent obtained after a study procedure rather than before, or forms missing a required element — as major findings that can affect data acceptability for the entire site. A consent defect isn't just an ethics problem; it is a data integrity problem, because any data collected under a deficient consent process is subject to exclusion from the primary analysis.
Patient Selection Criteria in GLP-1/GLP-3 Trials
Eligibility criteria in pivotal obesity and metabolic trials are narrower than the population that will eventually use an approved therapy, and that gap is precisely what the consent process has to make explicit. Typical inclusion criteria across STEP and SURMOUNT-class trials require BMI ≥30 kg/m2, or ≥27 kg/m2 with at least one weight-related comorbidity (hypertension, dyslipidemia, obstructive sleep apnea, cardiovascular disease). Exclusion criteria are where most of the clinical judgment lives.
Standard exclusions across the drug class include personal or family history of medullary thyroid carcinoma or Multiple Endocrine Neoplasia syndrome type 2 (MEN2), history of pancreatitis, severe gastrointestinal disease, and renal impairment below defined eGFR thresholds — commonly 15-30 mL/min/1.73m2 depending on the specific protocol. The FLOW trial's kidney endpoint analysis illustrates why renal function thresholds are treated as a hard boundary rather than a relative contraindication: the pharmacokinetics of these compounds and the downstream monitoring burden change substantially outside defined renal function ranges.
Age boundaries (typically 18-75), pregnancy and lactation exclusion, and washout periods for prior GLP-1 or GIP exposure round out the standard eligibility framework. Each exclusion criterion has to appear in the consent form's risk-disclosure language in some form — not as a checklist, but as a plain-language explanation of why the person sitting across from the coordinator either qualifies or doesn't, and what that boundary is protecting against.
Consent for Investigational Multi-Receptor Agonists
Consent language for an FDA-approved compound being studied in a new indication differs meaningfully from consent language for a molecule still in Phase 2 or Phase 3 with no approved use anywhere. Retatrutide, a GIP/GLP-1/glucagon triple agonist, remains investigational; its Phase 2 dataset showed weight reductions up to 24.2% at the highest dose at 48 weeks, but long-term cardiovascular outcomes, thyroid safety, and bone density effects are still being characterized. The Phase 2 data on weight and glycemic endpoints is informative, but it is not the same evidentiary tier as a completed Phase 3 program, and consent language has an obligation to say so explicitly rather than imply equivalence to approved therapies.
The same applies to combination products. CagriSema (cagrilintide plus semaglutide) has produced Phase 3 REDEFINE-1 data showing substantial weight reduction versus semaglutide alone, but as with any fixed-dose combination in development, the interaction profile and long-term tolerability are still accumulating. Details from the REDEFINE-1 Phase 3 results underscore why consent documents for combination agonists need a separate risk section rather than borrowing language from single-agent semaglutide trials — the adverse event profile of a combination is not simply additive, and consent language that treats it as such misrepresents the state of the evidence.
Oral formulations add a further layer: orforglipron's ACHIEVE program is testing a non-peptide oral GLP-1 agonist, and its Phase 3 weight and glycemic endpoint data reflects a different absorption and tolerability profile than injectable agents, which changes both the informed consent risk narrative and the practical burden disclosed to participants (fasting requirements, absorption interactions).
Vulnerable Populations and Equipoise
Obesity trials sit in an unusual position within research ethics: the condition under study is stigmatized, and candidates may have prior negative experiences with clinicians dismissing weight concerns, which can create pressure to enroll regardless of full risk comprehension. IRBs increasingly require documented "teach-back" — the participant restates the risks and purpose in their own words — specifically to counter consent obtained under therapeutic misconception, where a participant believes trial enrollment guarantees personal clinical benefit rather than contributing to generalizable knowledge.
Populations with obstructive sleep apnea, chronic kidney disease, or prior bariatric surgery require additional consent elements because monitoring burden and risk profile shift. The SURMOUNT-OSA findings are a useful reference point: apnea-hypopnea index reductions were substantial, but participants with severe OSA required closer monitoring during dose titration, and that monitoring schedule has to be disclosed as a participation burden, not buried in the protocol appendix.
Equipoise — genuine uncertainty about whether the investigational arm is better than standard care or placebo — is the ethical foundation that justifies randomization at all. Where equipoise has eroded, such as in populations where cardiovascular benefit is now well-established (per SELECT trial data, NCT03574597), placebo-controlled designs face additional ethical scrutiny, and consent forms must be explicit that a participant randomized to placebo is foregoing an intervention with an established evidence base in that population.
Data Integrity: Source Documentation and Monitoring
Data integrity begins with source data verification (SDV) — the process by which monitors confirm that data entered into case report forms matches the original source documents (clinic notes, lab reports, patient diaries). ICH E6(R2) formalized risk-based monitoring, which shifted many sponsors away from 100% SDV toward targeted verification of critical data points: primary endpoint measurements, serious adverse events, and eligibility criteria confirmation.
Consent documentation is itself a critical data point subject to this scrutiny. Monitors check that the consent form version matches the IRB-approved version active at the time of signature, that the date of signature precedes any study procedure, and that the person who obtained consent was listed on the delegation log with appropriate training documented. A mismatch on any of these — an outdated consent version used because a site missed an amendment notification, for example — generates a protocol deviation that has to be reported and, depending on severity, can affect whether that participant's data is included in the per-protocol analysis population.
Query rates (the number of data clarification requests per case report form) are a commonly tracked site-quality metric, and sites with poor consent documentation practices tend to generate disproportionate queries in exactly this domain — evidence that consent rigor and broader data quality are correlated, not independent workstreams.
Attrition, Dropout, and the Consent-Retention Link
Discontinuation is where consent quality and data integrity intersect most visibly. In the STEP 1 trial (semaglutide 2.4 mg, NEJM 2021, PMID 33882218), gastrointestinal adverse events were the most common reason for treatment discontinuation, occurring in a minority of participants but skewed toward the early titration period. Long-term follow-up data matters more than single-trial dropout figures: the STEP 4 long-term follow-up findings showed a 6.9% adverse-event-related discontinuation rate in the group continuing semaglutide versus 3.8% in the group switched to placebo, a differential that has to be accounted for in any efficacy interpretation.
Consent forms are required to describe the right to withdraw "at any time, for any reason, without penalty" — but the operational question for data integrity is what happens to that participant's already-collected data and how withdrawal timing is documented. A participant who withdraws mid-trial due to an adverse event that goes undocumented as a formal protocol deviation creates a gap in the safety dataset that intention-to-treat analysis has to address through sensitivity analyses or multiple imputation, both of which weaken the precision of the final effect estimate.
Long-term maintenance data adds another dimension: the SURMOUNT-4 withdrawal and regain findings demonstrate that consent conversations at the withdrawal decision point — not just at enrollment — materially affect whether a participant contributes complete follow-up data or becomes lost to follow-up, which is one of the more stubborn threats to long-duration trial data quality.
Representativeness and Trial Diversity
Consented populations in pivotal GLP-1/GLP-3 trials have historically skewed toward specific demographic profiles, which limits how confidently efficacy and safety findings generalize. Enrollment in STEP and SURMOUNT trials has been predominantly White and female, with Black and Hispanic participants underrepresented relative to U.S. obesity prevalence in those populations. This isn't a footnote — it directly affects the external validity of consented risk-benefit disclosures given to future patients who don't match the trial demographic.
Sponsors have begun addressing this through targeted site selection and enrollment quotas in newer protocols, but the consent process itself carries an obligation that is often underappreciated: participants from underrepresented groups should be informed, where relevant, that safety and efficacy data in populations similar to theirs may be more limited than the aggregate trial results suggest. This is a nuance that a boilerplate consent form rarely captures, and it is an area where trial-level data integrity depends on consent language being tailored rather than templated.
Age representation is a related gap — most pivotal trials cap enrollment around age 75 and require minimum ages of 18, leaving both adolescent and geriatric populations with substantially thinner evidence bases, a limitation that consent documents for any future trials in those age brackets will need to disclose explicitly given the absence of extrapolated safety data.
Adverse Event Disclosure and Ongoing Consent
Informed consent is not a single event completed at enrollment; it is a continuing obligation. Data Safety Monitoring Boards (DSMBs) conduct interim analyses throughout a trial's duration, and when a safety signal crosses a pre-specified threshold, sites are required to re-consent actively enrolled participants using an IRB-approved updated form before continuing their participation. This happened at scale in cardiovascular outcomes research: the SELECT trial's four-year follow-up data generated updated risk-benefit information that fed back into consent language for related ongoing studies in the same drug class.
Serious adverse event (SAE) reporting timelines are strict — sponsors typically require site reporting to the IRB within 24-48 hours of awareness, and to regulatory authorities on an expedited basis for unexpected, drug-related SAEs. A site that fails to update its consent process after a labeled safety signal (e.g., a boxed warning addition) is operating with a document that no longer reflects the known risk profile, which is both an ethical and regulatory deficiency.
Post-market pharmacovigilance mechanisms — REMS programs where applicable, FDA's FAERS database — feed back into ongoing trial consent revisions for compounds in the same class, meaning consent documents for newer multi-receptor agonists increasingly reference real-world safety signals from earlier-approved GLP-1 therapies as contextual risk information, even when the investigational compound itself has a distinct molecular target.
Practical Standards for Sites
Sites running GLP-1/GLP-3 protocols reduce both ethical exposure and audit risk by treating consent as a monitored process rather than a one-time signature event. Concrete practices that show up favorably in FDA inspection outcomes include: maintaining a version-control log cross-referenced against the IRB approval letter for every consent form in circulation, requiring documented teach-back at the time of signature, and flagging any protocol amendment for mandatory re-consent within a fixed window (commonly 10 business days) rather than at the next scheduled visit.
Delegation logs should specify exactly which staff members are authorized to obtain consent, with training documentation dated before their first consent encounter — a gap here is one of the most frequently cited BIMO findings. Eligibility criteria confirmation (renal function, MEN2/thyroid history, pancreatitis history) should be documented with the specific lab value or clinical note reference used to make the determination, not just a checkbox, since that specificity is what source data verification checks against.
For sites and sponsors reviewing their current consent workflow, the concrete next step is an internal audit comparing every active consent form version against the current IRB-approved version and the delegation log, before the next scheduled monitoring visit identifies the gap first.
This article summarizes research and does not constitute medical advice. Consult a licensed clinician for diagnosis, treatment, or any decisions about medications or supplements.