FDA's Animal Rule: CBRN Approvals and Their Evidence Base

FDA's Animal Rule for CBRN medical countermeasures: how the approvals were built

When a chemical, biological, radiological, or nuclear (CBRN) threat cannot ethically or feasibly be studied in humans, FDA can approve a drug or biologic on the basis of animal efficacy data. This is the "Animal Rule," codified at 21 CFR 314.600 (Subpart I, drugs) and 21 CFR 601.90 (Subpart H, biologics). It has produced a small but instructive set of approvals: anthrax antitoxins and antibiotics, smallpox antivirals, hematopoietic growth factors for acute radiation syndrome, chemical antidotes, and a nerve-agent pretreatment. This write-up is not a complete census: botulism antitoxin heptavalent (BAT, approved in 2013 under 21 CFR 601.90) and several antibiotic CBRN indications are outside its scope. The common architecture is the same across every one of them: survival demonstrated in animals, human safety and pharmacokinetics demonstrated in healthy volunteers or borrowed from established use, and a pharmacokinetic bridge linking a human dose to the exposures that protected animals.

What the Animal Rule requires

FDA may treat adequate and well-controlled animal efficacy studies as substantial evidence of effectiveness only when all four criteria are met ²³³:

1. The pathophysiologic mechanism of the toxicity and the product's mechanism for preventing or substantially reducing it are reasonably well understood ²³³.
2. The effect is demonstrated in more than one animal species expected to react with a response predictive for humans, or in a single sufficiently well-characterized species ²³³.
3. The animal study endpoint is clearly related to the intended human benefit, generally survival or prevention of major morbidity ²³³.
4. Animal and human pharmacokinetic/pharmacodynamic data permit selection of an effective human dose ²³³.

Animal efficacy never substitutes for human safety work. Products approved this way must still be evaluated for safety under the ordinary requirements for new drugs, and FDA takes the position that safety is generally studiable in human volunteers ²⁴⁷²⁵⁵. Sponsors are expected to characterize safety and PK across a dose range, agree an upper dose limit with the review division, and define the size and composition of the human safety database, with additional safety studies where needed ^242254235. FDA also expects a plan for postmarketing studies of safety and clinical benefit if actual emergency use ever makes such studies feasible and ethical ²³²²⁴¹.

The sections below group the approvals by threat and set out, for each, the animal efficacy package and the human safety/PK evidence FDA actually relied on.

Anthrax (biological threat)

Anthrax medical countermeasures are the largest single cluster of Animal Rule approvals. Because deliberate inhalational-anthrax exposure of humans is unethical and natural cases are too rare and scattered for a trial, efficacy for the antitoxins rested on two-species survival studies using aerosolized Bacillus anthracis Ames spores, with treatment triggered by a clinical or biomarker sign of established disease.

Raxibacumab

Raxibacumab, an anti-protective-antigen (anti-PA) monoclonal antibody, was licensed under the biologics Animal Rule (21 CFR 601.90) for treatment of inhalational anthrax ⁵⁰⁶. Efficacy came from New Zealand White rabbits and cynomolgus monkeys challenged with a target of roughly 200 x LD50 aerosolized spores, with treatment started at first disease manifestations (sustained temperature rise or serum PA positivity) ⁵. A single 40 mg/kg IV dose produced 6/17 (35.3%) rabbit survival versus 0/13 with placebo (p=0.0237), and 9/13 (69.2%) monkey survival versus 0/10 with placebo (p=0.016) ¹². All surviving animals developed toxin-neutralizing antibodies ¹⁰.

Human data covered safety and PK only: 326 subjects received raxibacumab, PK was approximately linear with a terminal half-life of about 15 to 19 days, and a single 40 mg/kg IV infusion (with diphenhydramine premedication) was the proposed dose ¹¹²¹. The bridge to animal survival was exposure-based: the lowest human Cmax/AUC (589 µg/mL and 8,720 µg·day/mL) exceeded the lowest exposures seen in surviving challenged rabbits and monkeys, supporting the expectation that the human dose would reach protective exposures ²⁰.

Obiltoxaximab (Anthim)

Obiltoxaximab, also an anti-PA monoclonal, was approved for inhalational anthrax on a much larger animal package: 22 monotherapy efficacy studies in NZW rabbits and cynomolgus macaques spanning treatment, post-exposure prophylaxis, and pre-exposure prophylaxis, with survival through end of study (usually day 28) as the primary endpoint ¹⁸⁰¹⁸⁶. The intended human regimen, 16 mg/kg IV, was effective after clinical signs emerged in two monkey and two rabbit studies, with survival falling as bacteremia/PA burden rose or treatment was delayed ¹⁸⁰¹⁸³.

Human studies were restricted to healthy volunteers across seven Phase 1 trials; the principal commercial-formulation studies exposed 320 subjects to obiltoxaximab and 70 to placebo, with more than 400 volunteers evaluated overall ¹⁷³¹⁸⁶. The defining safety signal was hypersensitivity, including anaphylaxis: any hypersensitivity sign in 10.6% of recipients, and clinically significant hypersensitivity requiring discontinuation in 3.1% ¹⁷³. There were no human data in children, pregnant women, or actual anthrax patients ¹⁷³. Dose selection used cross-species PK bridging, with 16 mg/kg in humans producing similar median Cmax and about twofold greater AUC than the fully effective 14.5 mg/kg animal dose; FDA judged higher doses would add little survival benefit while increasing hypersensitivity risk ¹⁸⁶¹⁷⁶.

Anthrax Immune Globulin (Anthrasil)

Anthrasil, a pooled polyclonal anti-PA immune globulin from AVA-vaccinated donors, was approved as a BLA in March 2015 for inhalational anthrax in combination with antibacterial drugs, with animal studies showing a survival increase when given without an antimicrobial ¹³²¹³³. The reviewed records here do not resolve the species, challenge design, survival proportions, or the human safety database in detail, so those specifics are best confirmed against the primary Anthrasil review ¹³²¹³³.

Anthrax vaccines

Anti-PA vaccines sit at the edge of this dataset. Anthrax Vaccine Adsorbed (AVA/BioThrax) is described as PA-based, with anti-PA antibody protecting susceptible animals against inhalational anthrax and adult seroconversion of 85% to 100% after subcutaneous or intramuscular dosing; a three-dose post-exposure schedule at 0, 2, and 4 weeks is cited as producing rapid, high-level antibody responses ¹⁵⁸¹³². BioThrax's post-exposure prophylaxis indication (2015) and AV7909 (Cyfendus, 2023) are both Animal Rule approvals; the records available here simply lack the study-level animal-efficacy results and immunogenicity bridge for them, so they are not characterized in the same detail as the products above.

Smallpox (biological threat)

Smallpox has been eradicated in nature, so human efficacy trials are impossible. Both approved antivirals used lethal surrogate orthopoxvirus models, with mortality (based on prospectively defined euthanasia criteria) as the endpoint directly tied to the intended human survival benefit.

Tecovirimat (TPOXX)

Tecovirimat's efficacy rested on two species: cynomolgus macaques challenged with monkeypox virus and NZW rabbits challenged with rabbitpox virus, treated only after clinical disease had developed ²⁴²³. In the primate model, oral tecovirimat begun on post-challenge day 4 or 5 produced statistically significant survival versus placebo, with survival of 75% to 100% at 10 mg/kg/day for at least five days and 100% at 20 mg/kg/day; the 14-day regimen did not show significant benefit when 10 mg/kg dosing began on day 6 ²⁹²⁵²³. In rabbits, doses at or above 20 mg/kg/day begun on day 4 each gave significant survival benefit, at least 80% per group ²⁵.

Human efficacy was never demonstrated; clinical studies were healthy-volunteer PK and safety trials ³⁰⁴⁰. The safety database comprised 788 subjects overall, with 437 on the proposed 600 mg twice-daily regimen (359 tecovirimat, 90 placebo in the randomized Study 246-008) ³⁸. Adverse reactions were predominantly mild, led by headache (12% vs 8% placebo), with a single serious event (a pulmonary embolism in a participant with prior DVT and concomitant Depo-Provera) 7 days after treatment ³⁸. The PK bridge was generous: the fed 600 mg twice-daily human regimen produced mean Cmax and AUC about twofold higher, and Cmin about fourfold higher, than exposures at the fully effective primate dose ²⁸³³. Pediatric dosing was derived entirely from population-PK modeling, with no pediatric clinical study ⁴¹.

Brincidofovir (Tembexa)

Brincidofovir was approved under the drug Animal Rule (21 CFR Part 314, Subpart I) for human smallpox, again using two lethal models ⁴⁷. In rabbitpox, the fully effective 20/5/5 mg/kg oral regimen (three doses every 48 hours) begun on day 4 gave 90% survival versus 29% with placebo ⁶¹. In the ectromelia (mousepox) model, the 10/5/5 mg/kg regimen begun on day 4 gave 78% survival versus 13% ⁶¹. The rabbit study's primary endpoint was survival through post-infection day 42 ⁶⁰.

Safety was never studied in people with smallpox. The supporting database came from randomized, placebo-controlled Phase 2/3 trials in other indications, chiefly CMV and adenovirus prevention after hematopoietic stem-cell transplant: 392 adults aged 18 to 77 exposed, 85% receiving 200 mg weekly for at least two weeks ⁶⁵⁷⁰. The principal safety findings were gastrointestinal toxicity and hepatotoxicity (ALT >3x ULN in 7%, bilirubin >2x ULN in 2%) ⁶⁴⁶⁵⁶⁷. A longer-duration CMV-prevention trial showed increased all-cause mortality at week 24 (16% vs 10% with placebo) under an extended dosing schedule, which drove boxed-warning/labeling restrictions to the short two-dose smallpox regimen ⁶⁵. FDA accepted the database with risk-mitigating labeling while noting the largely immunosuppressed safety population could differ from those exposed in a smallpox emergency ⁶²⁶⁷.

Acute radiation syndrome (radiological/nuclear threat)

The three hematopoietic growth factors detailed below carry indications to increase survival in the hematopoietic subsyndrome of acute radiation syndrome (HS-ARS); a fourth, pegfilgrastim (Neulasta), received an Animal Rule HS-ARS approval in November 2015 and is not detailed here. Each relied on irradiated nonhuman-primate survival, and, unusually, each borrowed most or all of its human safety evidence from long-established use in oncology and hematology rather than from ARS-specific trials.

For filgrastim, the human 10 mcg/kg/day dose was selected to give exposures exceeding those in the efficacious primate study, with safety inferred from other approved uses ¹⁰⁸¹¹⁷. Sargramostim added an exploratory higher-dose rhesus cohort (713 cGy, day-60 survival 61% vs 17%) and demonstrated benefit despite a 48-hour treatment delay and minimal supportive care, conditions chosen to mirror a mass-casualty environment ⁸⁶⁹³; its known safety concerns include capillary-leak syndrome, arrhythmias, and hypersensitivity ⁹¹⁸⁵. Romiplostim was supported by a second species (a mouse HS-ARS study, 60% vs 21% survival) and by human platelet-response data used to model a single 10 mcg/kg subcutaneous dose; longer-term romiplostim risks such as MDS-to-AML progression and thrombosis were carried over from its ITP experience ^145155151.

Internal radionuclide contamination (radiological/nuclear threat)

Two decorporation agents remove internalized radionuclides. Their approvals are worth distinguishing from the classic Animal Rule cases: rather than survival endpoints in animals, they relied on enhanced radionuclide elimination as a surrogate endpoint, supported by human observational data and animal studies, because clinically important outcomes (radiation-induced cancer, death) occur years later and controlled trials in exposed people were infeasible and unethical ^164167168.

Pentetate calcium/zinc trisodium (Ca-DTPA / Zn-DTPA)

Ca-DTPA and Zn-DTPA were approved to increase elimination of plutonium, americium, and curium, and were not expected to work for uranium, neptunium, or radioiodine ¹⁶⁹¹⁷². FDA accepted rapid, continued urinary radionuclide elimination as a clinically meaningful surrogate for reduced long-term radiation burden and cancer risk ¹⁶⁴¹⁶⁸. The reviewed records describe the package as a human observational database supported by animal data rather than an explicitly labeled Animal Rule approval ^164168171. Animal support came from plutonium studies: chelation lengthened the latency to death in dogs and reduced osteosarcoma incidence in mice, and Ca-DTPA gave roughly 10-fold higher early urinary elimination than Zn-DTPA in rodents ¹⁶⁸¹⁶⁹. The principal human evidence was the DOE/ORISE REAC/TS registry of 685 accidental transuranic-exposure cases over 42 years, with each patient's pre- versus post-treatment urinary radiation as an internal control; among 18 cases with matched measurements, mean urinary radiation rose 39.3-fold after the first dose ^168167169. Safety across the 685 exposures was mostly minor and transient, with rare serious events (an allergic reaction, respiratory distress with Ca-DTPA, microscopic hematuria in 13 patients) ¹⁶⁷¹⁷⁰.

Prussian blue (Radiogardase)

Insoluble Prussian blue was approved to increase elimination of radioactive cesium and radioactive or non-radioactive thallium by binding them in the gut, interrupting enterohepatic recirculation, and shifting elimination to feces to shorten biologic half-life ²⁰⁷¹⁹⁹. FDA's review explicitly notes that prospective human safety/efficacy trials had never been performed; support was animal studies plus anecdotal clinical, epidemiologic, and mechanistic data ²⁰¹. Animal work in rats, pigs, and dogs showed the elimination route shifted from renal to fecal, with a rat dose-response through 50 mg/day and porcine data showing 99% fecal excretion of a labeled dose ^207198210. The central human evidence was the 1987 Goiânia cesium-137 accident: among 33 evaluable patients, treatment reduced mean whole-body effective half-life by 69% in adults, 46% in adolescents, and 43% in children versus untreated values ²¹³. For thallium, 34 literature-reported poisoning cases showed mean serum biologic half-life falling from 8 days to 3 days ¹⁹⁹.

Chemical threats

Hydroxocobalamin (Cyanokit)

Cyanokit was evaluated under 21 CFR 314.600 for cyanide poisoning, with FDA finding all Animal Rule criteria met: a well-understood toxic mechanism (cyanide binding cytochrome oxidase) and antidotal mechanism (hydroxocobalamin binding cyanide to form cyanocobalamin), a sufficiently characterized dog model, survival as the endpoint, and adequate PK/PD for dose selection ⁴⁹⁷⁸⁷⁹. The pivotal study (N106342) was a randomized, vehicle-controlled 14-day beagle study in which dogs received potentially lethal IV potassium cyanide, then saline or hydroxocobalamin 75 or 150 mg/kg; both doses improved survival (about 60% of vehicle dogs died in the first 4 hours versus 5% at 75 mg/kg and none at 150 mg/kg), with the 75 and 150 mg/kg doses corresponding to roughly 5 g and 10 g human doses ⁷⁸⁷³⁸⁰.

Human safety came from a placebo-controlled healthy-volunteer study of 102 subjects dosed at 2.5, 5, 7.5, or 10 g plus 34 placebo recipients, with the 2.5 g and 5 g doses generally well tolerated and no deaths or serious adverse events ⁷⁷⁷². Characteristic effects were chromaturia, red skin discoloration, and a transient pressor response (systolic increases up to 90 mmHg, generally resolving within 4 to 8 hours), which could even be beneficial in hypotensive poisoned patients ⁷²⁷⁴. Uncontrolled French case series in suspected poisonings offered limited supportive context ⁷⁷⁷².

Pyridostigmine bromide (soman pretreatment)

Pyridostigmine bromide 30 mg (NDA 020414) was approved under the Animal Rule for pretreatment against the lethal effects of soman nerve-agent poisoning, used together with post-exposure atropine and pralidoxime ⁴⁶¹²⁶. The efficacy case comprised 15 well-controlled animal studies (2 rhesus monkey, 5 guinea pig, 3 rabbit, 3 mouse, 2 rat), most using soman because it resists atropine/2-PAM therapy, with survival at 24 or 48 hours and efficacy expressed as a protective ratio measuring pyridostigmine's contribution beyond atropine plus 2-PAM ¹²⁷. Rhesus dose-survival data combined with human and simian PK were used to calculate an equivalent human dose ¹²⁸. The mechanism is protective enzyme sequestration: reversible occupation of peripheral acetylcholinesterase preserves a reserve that irreversible soman cannot inhibit, later reactivated with atropine and pralidoxime ¹²⁶. The reviewed records confirm prior FDA safety and efficacy findings for PB30 but do not itemize the original human safety database; the older Mestinon 60 mg product supplied prior safety evidence for a later extended-release formulation ¹²⁹¹³⁰.

Plague (biological threat)

Levofloxacin (pneumonic plague)

Levofloxacin was approved under the Animal Rule for treatment of pneumonic plague due to Yersinia pestis in adults and pediatric patients ²²²²²⁹. Efficacy came from African green monkeys with established pneumonic plague after aerosol exposure to Y. pestis CO92: a humanized IV levofloxacin regimen started within 6 hours of sustained fever and given for 10 days produced day-28 survival of 16/17 (94.1%) versus 0/7 with placebo (p<0.001) ^222229231. There were no human pneumonic-plague efficacy studies; support came from established levofloxacin safety and efficacy in other indications plus exposure matching, with recommended adult and pediatric regimens reaching or exceeding the plasma concentrations associated with primate survival (adult steady-state peak ~5.7 to 6.4 mcg/mL at 500 mg once daily) ²²²²²³. This approach, an antibacterial with a large existing human safety record earning a new CBRN indication purely on animal survival plus PK matching, is the same template FDA has applied to other fluoroquinolone plague and anthrax indications.

How the pieces fit together

Across these approvals a consistent evidentiary logic emerges, and it is worth naming for anyone planning an Animal Rule submission:

• Efficacy is survival in the most human-predictive models available. Two species is the default (rabbit plus nonhuman primate for anthrax and smallpox; rhesus plus mouse for radiation), with a single well-characterized species accepted only where the model is exceptionally predictive, as with the beagle cyanide model ²³³⁴⁹. Treatment is triggered by a clinically realistic sign of established disease, not given prophylactically at the moment of challenge, and delayed-treatment and minimal-supportive-care designs are valued because they mirror a real mass-casualty response ¹⁸³⁹³.
• The human dose is chosen by a pharmacokinetic bridge, not a human efficacy readout. In every case FDA selected a human dose predicted to produce exposures at least as high as, and usually well above, those associated with animal survival, and this exposure margin is the load-bearing element of the approval ^2033186117.
• Human safety is real, human, and separate. Sponsors either ran healthy-volunteer safety/PK programs (tecovirimat, obiltoxaximab, raxibacumab, hydroxocobalamin) or leaned on established clinical experience in other indications (the ARS growth factors, levofloxacin, brincidofovir) ^381738665. Where the borrowed safety population differed materially from the emergency-use population, as with brincidofovir's largely immunosuppressed database, FDA compensated with boxed warnings and regimen restrictions ⁶⁵⁶².
• Decorporation agents are a different animal. Ca/Zn-DTPA and Prussian blue rest on enhanced radionuclide elimination as a surrogate endpoint, drawing heavily on human contamination registries and accident data rather than animal survival, and are best understood as surrogate-endpoint approvals adjacent to, rather than squarely within, the classic Animal Rule survival paradigm ^164201213.

The through-line is that the Animal Rule does not lower the safety bar; it relocates the efficacy bar from humans to well-characterized animal models while insisting that dose selection, mechanism, and human safety all remain anchored in human data.