NIH Vertebrate Animals Section: Writing the Five Criteria That Prevent an Administrative Hold

Why the VAS Has Its Own Review Track

The Vertebrate Animals Section isn't reviewed once — it's reviewed twice, by different people asking different questions. Study section reviewers read it during peer review and consider it as part of the Approach evaluation under the Simplified Review Framework. They're asking whether your animal use is scientifically justified and whether the procedures are feasible given your aims.

NIH program and grants management staff review it again before issuing the Notice of Award. Their question is narrower: does the section comply with the PHS Policy on Humane Care and Use of Laboratory Animals and the Animal Welfare Act? These two audiences sometimes care about different things. A reviewer might be satisfied by a brief note that you'll use the minimum number of animals needed for statistical power. Grants management staff need to see explicit USDA pain category assignments, AVMA-compliant euthanasia methods, and a veterinary care statement — specific language that peer reviewers don't typically check word by word.

When grants management finds a deficiency, they issue an administrative hold: a formal notice that NIH cannot release your award until the problem is corrected. Holds are resolved by correspondence with your specialist, but the timeline depends on how quickly your institution responds and how clearly the VAS is revised. Depending on the severity and your institution's workload, a hold can add a few weeks or push you past a budget quarter boundary — which then complicates personnel start dates, equipment procurement, and everything downstream. Writing a clean VAS from the start is the better path.

The Five Criteria, One at a Time

The NIH VAS form instructions require you to address five distinct criteria, labeled A through E. An application that omits any criterion is deficient by definition, regardless of how strong the science is. Here is what each criterion actually demands.

Criterion A — Description of Proposed Use

Describe what you will actually do to the animals. Name the species and strains. List the experimental procedures — injections, surgeries, behavioral tests, imaging sessions, tissue collection — in enough detail that a program officer can confirm they match what your Approach section describes. This criterion connects directly to your Research Strategy; reviewers will notice if procedures named in the VAS don't appear in the Approach, and grants management will flag discrepancies between the two documents when preparing your award.

Criterion B — Justifications

Three justifications in one criterion: (1) why animals are necessary at all — why can't you answer the question with cell culture, organoids, or computational methods; (2) why you chose this particular species and strain; and (3) why you need exactly this number of animals. The number justification is where most VAS sections fail. You need a statistical rationale — a power calculation, a published effect size from similar studies, or a clear empirical basis for the group sizes proposed. "We will use 10 mice per group" with no supporting calculation is one of the most common deficiencies grants management flags.

Criterion C — Veterinary Care

Name the facility and describe the veterinary care arrangement. State that animals will be housed in an AAALAC-accredited facility if that's true, and that a veterinarian will be available for consultation and health monitoring. This criterion is brief but frequently omitted — and its absence constitutes a deficiency. Two or three sentences here is usually enough. If your facility is not AAALAC-accredited, state what oversight arrangement is in place and reference your institution's IACUC.

Criterion D — Minimization of Pain and Distress

Describe your procedures for limiting pain and distress to the minimum unavoidable in the conduct of sound research. Assign a USDA pain/distress category (B, C, D, or E) to each experimental group and justify the assignment. If you're using anesthetics, analgesics, or tranquilizers, name them, give approximate doses, and explain when in the protocol they're administered. If you need to withhold pain relief because it would compromise scientific validity — Category E — this is where you explain why, and this explanation receives close attention from program staff and your IACUC alike.

Criterion E — Method of Euthanasia

State the method and cite the current AVMA Guidelines for the Euthanasia of Animals as your authoritative source. If your euthanasia method deviates from AVMA guidelines, provide a scientific justification for the deviation. Cervical dislocation without prior anesthesia, for example, is acceptable under AVMA guidelines for rodents in specific circumstances, but you need to name those circumstances explicitly in the VAS. Program staff look at Criterion E carefully, and a deviation that isn't justified is a reliable hold trigger.

Number Justification: Where Most Sections Fail

Statistical justification for animal numbers surfaces at peer review as a feasibility question and again at the award stage as a regulatory one. Reviewers flag vague group sizes as evidence of poor experimental design. Grants management flags the same language as a potential waste of resources — which implicates the 3Rs principle of "Reduce" under the PHS Policy.

The standard approach is a power calculation: state the expected effect size (from your preliminary data or published literature), the expected standard deviation, the alpha level (usually 0.05), and the power you're designing for (usually 0.80). Then state the group size the calculation produces. If you need more animals than the calculation alone suggests — to account for attrition, technical failures, or collection of tissue at multiple time points — state that separately with a replacement estimate.

If you don't have preliminary data establishing a precise effect size, cite a closely related published study and acknowledge the uncertainty. Something like: "Based on a published effect size of X in a similar model (Author et al., year), a power analysis using G*Power indicates a minimum of N animals per group; we propose N + 20% to account for anticipated attrition." This is honest, transparent, and satisfies both audiences. Biostatistics core facilities at most research universities will run the power analysis for free for grant submissions — one consultation appointment is usually sufficient.

USDA Pain Categories and How to Assign One

USDA classifies animal use into pain/distress categories that your facility reports annually. The assignment you make in your VAS needs to be consistent with what your IACUC protocol records and what your institution reports to USDA.

Category B applies to animals used for breeding or holding with no experimental procedures. Category C applies to procedures that cause no more than momentary pain or distress — blood draws from the tail vein, brief isoflurane anesthesia for imaging, ear tagging. Category D applies to procedures involving more than momentary pain or distress where appropriate pain relief is provided — surgeries, tumor implantation, chronic infection models with analgesia on protocol. Category E applies to procedures involving pain or distress for which pain-relieving agents cannot be used without compromising the science. Category E use requires specific scientific justification and receives the most scrutiny from both your IACUC and NIH program staff.

Misclassifying a Category D procedure as Category C is one of the most common VAS errors — especially in infection models, metabolic disease studies, and behavioral paradigms that involve restraint or repeated procedures. When in doubt, use the higher category and note that you're erring toward the more conservative classification. Upgrading a category rarely causes a problem; a downward classification that doesn't match your procedures can trigger a hold or an adverse finding during an NIH site visit.

How the VAS Connects to Your IACUC and What Happens After the Award

A common misconception: your IACUC protocol and your VAS are separate documents with separate approval processes, and they don't need to match word for word. That's partly right — the VAS is your grant document and the IACUC protocol is your institutional document — but they do need to be consistent in species, procedures, and scope. NIH does not require IACUC approval before submission; a pending protocol is acceptable at the time you apply. But you do need IACUC approval before NIH issues the Notice of Award.

During Just-In-Time, NIH typically requests your IACUC protocol number and approval date. Grants management staff compare that protocol against your VAS. If the protocol covers rhesus macaques and your VAS describes cynomolgus macaques, that discrepancy goes on hold. Review both documents side by side before submission — and again when you respond to the JIT request. The few minutes this takes are well worth it given the timeline consequences of a late-stage hold.

After the award, significant changes to your animal work generally require prior approval from NIH. Adding a new species, substantially increasing animal numbers beyond what was approved, or introducing new invasive procedures not covered in the original VAS are the clearest examples. Small deviations within the scope of your approved VAS and IACUC protocol often don't need NIH-level approval separately — but when you're uncertain, ask your grants management specialist before proceeding with the work. Conducting animal procedures under a NIH award before receiving prior approval for out-of-scope changes can create compliance problems that are harder to resolve than the underlying scientific question.

Pre-Submission Checklist

After drafting the VAS, run through this verification pass before you send the application to your sponsored programs office.

Every experimental group in your Approach section appears in the VAS

Read through each aim and make sure every animal group, time point, or cohort described in the Research Strategy has a corresponding entry in the VAS. Missing groups are the most frequently flagged discrepancy.

Every animal number has a statistical or empirical basis

For each group size, confirm that a power calculation, a published effect size, or a clear rationale is stated in the VAS. Attrition estimates should be listed separately from the minimum required number.

USDA pain category assigned and justified for each group

Every experimental cohort should have an explicit Category B, C, D, or E assignment with one or two sentences explaining the basis for the classification.

Euthanasia method named and referenced to AVMA guidelines

If deviating from AVMA guidelines, a scientific justification is in place. If following AVMA guidelines, cite the current edition explicitly — don't just say "standard methods."

Veterinary care statement is present

Even two sentences naming the facility and confirming veterinary oversight satisfies Criterion C. This is the most commonly omitted criterion, and also the easiest one to fix before submission.

Frequently Asked Questions

Do I need a VAS if I'm using human tissue or cell lines?

No. The Vertebrate Animals Section only applies to applications involving live vertebrate animals. If your study uses human subjects, you address that in the Human Subjects section and the Inclusion Enrollment Report. If you're using commercially obtained human cell lines or cadaveric tissue with no connection to identifiable individuals, you typically need neither a VAS nor Human Subjects protections. If you're uncertain about your specific materials, consult your institution's IACUC or IRB coordinator before submission.

Do zebrafish and Xenopus require a VAS?

Yes. All live vertebrate animals, including zebrafish, Xenopus, and other non-mammalian vertebrates, require a Vertebrate Animals Section. Some researchers assume the VAS applies only to mammals. It applies to any live vertebrate. USDA pain category assignments for fish and amphibians are less standardized than for mammals, but you still need to address all five criteria. Consult your IACUC and the NIH OLAW guidance on non-mammalian vertebrates for current expectations.

Can I submit with a placeholder if I don't know exact animal numbers?

You can estimate, but you need to state the basis for the estimate. "Approximately 120 mice total across three experimental groups, based on a power analysis assuming 30% attrition and a published effect size of X" is acceptable. A bare number with no supporting rationale is a deficiency. Estimates with acknowledged uncertainty are generally fine — NIH understands that preliminary data won't always produce a perfectly powered study design for a first submission.

What if I need to add animal work not in the original VAS after the award?

Changes that go beyond the scope of your original VAS — adding a new species, substantially increasing animal numbers, introducing a new category of invasive procedure — typically require prior approval from your NIH program officer and grants management specialist. The process involves submitting a prior approval request through your institution's sponsored programs office. Small within-scope variations generally don't require NIH-level approval separately from your IACUC, but when the line isn't clear, ask your grants management specialist before conducting the work.