Nyquist AI

1. Intended Audience Expanded

The draft's stated audience was sponsors and/or IRBs.
The final guidance explicitly adds clinical investigators as a third intended audience, alongside sponsors and IRBs, and reflects this in the title and introduction throughout.

Throughout the document, the final guidance replaces "trial participants"/"patients" (used in the draft) with "trial subjects" or "subjects." This is a consistent wording change, not a substantive policy shift, but industry should update internal templates/checklists that quote the guidance's language.

The draft framed the goal as improving the diversity of clinical trial populations (e.g., "may improve the diversity of clinical trial populations," "accrual and diversity for clinical trials").
The final guidance reframes this goal as improving representativeness/generalizability and applicability of trial data to real-world U.S. populations (e.g., "may improve the applicability of clinical trial data to U.S. populations").
The draft's footnote cross-referencing FDA's separate guidance *Enhancing the Diversity of Clinical Trial Populations – Eligibility Criteria, Enrollment Practices, and Trial Designs* (November 2020) is removed in the final version.

The final guidance's footnote listing companion cancer-trial eligibility guidances now includes two additional titles not present in the draft's list:
Washout Periods and Concomitant Medications (July 2026)
Performance Status (July 2026)
Sponsors should be aware this guidance is now part of a broader, more complete suite of eligibility-criteria guidances than existed at the time of the draft.

The draft mentioned generically that lab values "may vary among healthy individuals across different racial and ethnic populations."
The final guidance adds a specific, citable example: the Duffy null phenotype, common in African Americans, is associated with lower absolute neutrophil counts without an associated increase in infection risk (new citation: Hibbs et al., 2024, *JAMA Network Open*).
This example is also newly added to the "Demographic differences" recommendation bullet in Section B, directly instructing sponsors to consider broadening ANC eligibility criteria for patients with the Duffy null phenotype.

The final guidance adds that laboratory values may also vary "across different age groups (e.g., pediatric patients)," broadening the background discussion beyond race/ethnicity alone (the draft only discussed racial/ethnic variation in this passage).

Section A (Scientific Justification) in the final guidance adds a new sentence not found in the draft: eligibility criteria defined by broad laboratory reference-based ranges rather than specific numerical values may help address inter-laboratory variation. This gives sponsors an additional, explicit tool for criteria design.

Section B in the final guidance adds an entirely new recommendation: hematologic laboratory criteria should be carefully considered and broadened where appropriate in trials of products intended for patients with hematologic malignancies. This bullet does not appear in the draft at all.

The draft's list of data sponsors should use to justify loosening criteria in later-phase trials included only renal/hepatic impairment studies and drug-drug interaction studies.
The final guidance broadens this list to explicitly include pharmacokinetic, safety, and exposure-response data from early-phase trials, giving sponsors more explicit bases for adjusting criteria.

The draft allowed sponsors to consider "a single repeat test" within a certain period for abnormal but non-clinically-significant lab values.
The final guidance loosens this to "repeat test(s)" (plural), allowing for more than one retest where appropriate.

The final guidance clarifies that hepatic entry criteria should be broad when a drug is not expected to cause hepatic toxicity "for the intended patient population," tying the justification more explicitly to the specific trial population rather than a general statement.

The differences are generated by AI models and can contain mistakes. Please double-check the content.