Cervical Dislocation in Mice: Acceptability, Operator Competence, and Confirming Death

Cervical dislocation is the method people assume they already know how to do. That assumption is the problem. In the only published study that scored it directly, trained operators working on anaesthetised mice under laboratory conditions failed 21% of the time — and the better of the two operators still failed in 8.1% of attempts. A method that fails one time in twelve in the hands of your best person is not a skill you certify once at induction and never look at again.

Three things decide whether cervical dislocation is defensible for a given mouse in a given protocol, and they are separable. First, whether the animal falls inside the weight and technique limits your jurisdiction actually sets — the numbers are not the same in the US, Canada and the UK, and people routinely quote the wrong one. Second, whether operator competence is evidenced rather than asserted. Third, whether death was confirmed by a listed method, not inferred from the fact that the animal stopped moving. Getting the first right and the third wrong is the failure mode that produces an animal recovering in a carcass bag.

Every major framework calls it conditionally acceptable — and the conditions differ

There is no jurisdiction in which cervical dislocation of mice is simply “acceptable”. A peer-reviewed comparison of five guidance documents — the AVMA guidelines, the American Society of Mammalogists guidelines, the CCAC euthanasia guidelines, EU Directive 2010/63/EU and ANZCCART — classifies it as conditionally acceptable in every one of them. What varies is the attached condition, and the differences are large enough to matter when you write a protocol.

• United States. There is no federal weight threshold in statute. The operative document is the AVMA Guidelines for the Euthanasia of Animals (2020 edition is current), applied through your IACUC. Tabulated against that guidance, the conditions are that personnel must be trained and their proficiency validated, that a secondary method must be available if the first attempt is unsuccessful, and that the technique is intended for rodents under 200 g. In practice the binding constraint is whatever your approved IACUC protocol says, because that is what an inspector reads.
• Canada. The CCAC guidelines are explicit and stricter than most people remember. Manual cervical dislocation “should only be performed on birds <3kg, rodents <200g and rabbits <1kg”, and “commercial cervical dislocators/luxators must be used on heavier rats (>200g)”. The CCAC also caps throughput on welfare grounds: manual dislocation “should only be performed where the number of animals is relatively low to prevent human error due to fatigue”. And unless it would interfere with the scientific outcome, animals “should be anesthetized prior to cervical dislocation”.
• United Kingdom. Schedule 1 of the Animals (Scientific Procedures) Act 1986 permits “dislocation of the neck” in rodents up to 500 g, but requires “the prior use of a sedative or anaesthetic in the case of rodents and rabbits over 150 g”. For a typical adult laboratory mouse at 20–35 g the sedation trigger does not bite, but the confirmation-of-death requirement below always does. See our guide to the Home Office animal research licence under ASPA 1986 for how Schedule 1 sits alongside personal and project licences.

Note the shape of the discrepancy: 200 g in Canadian guidance, 500 g with a 150 g sedation trigger in UK legislation, and secondary tabulations of EU Annex IV that report a figure different again. If your protocol cites a threshold, cite the one from the instrument that actually governs your facility and read the primary text rather than a summary table — this is one of the places where second-hand numbers have visibly diverged in the literature.

It is neither instantaneous nor anatomically precise

Two pieces of evidence should shape how you talk about this method in a protocol, because both cut against the usual claim that dislocation causes immediate death.

Cortical activity persists for seconds. Cartner and colleagues recorded electroencephalograms and visual evoked potentials in mice during the first 30 seconds after euthanasia by several methods. EEG activity was significantly decreased after cervical dislocation at 5 to 10 seconds — faster than decapitation (15 to 20 s) or intracardiac potassium chloride (20 to 25 s) in the same study, but demonstrably not zero. “Instantaneous” is the wrong word to put in a protocol; “rapid disruption of cortical function” is the claim the data support.

The dislocation often is not where you think it is. In the Carbone study, 63 of 64 successfully euthanised mice had radiographically visible lesions in the high cervical or atlanto-occipital region — but 50 of those 64 (78%) also had thoracic or lumbar lesions. The technique routinely transmits force well beyond the target segment. The control condition makes the point sharply: deliberate thoracic dislocation failed in 18 of 18 animals. Hand placement too far caudal is not a degraded version of the method, it is a total failure of it.

Both findings support the same operational conclusion. Failure is not rare, it is not always visually obvious, and the animal must be checked rather than assumed dead.

What “trained and competent” has to mean in practice

Nearly every framework requires validated proficiency, and almost none of them define how to measure it. The literature does offer a usable answer.

Carbone and colleagues proposed training on anaesthetised mice with time to respiratory arrest as the pass criterion: an attempt after which the mouse never breathes (time to respiratory arrest of 0 s) is a success; any attempt after which it breathes at all is a failure. That is a binary, observable, recordable metric that a trainer can score in real time and a compliance office can audit. It is a substantially better assessment than the usual “observed by a senior colleague, deemed competent” sign-off, and it lets you detect skill drift by re-testing. The same paper found post-mortem radiography had little value as a quality-control tool — the imaging does not tell you what the breathing already told you.

The scale of the standardisation problem was quantified in a 2026 UK survey of 317 people who kill laboratory rodents. Among the 312 working with mice, 69.4% used manual methods and 25.9% mechanical. The tools reported for mice included metal rods (27.1%), closed scissors (20.2%) and pens (15.8%), and 32.5% did not know which tool their facility used. The authors describe “a wide range of improvised tools (i.e. not designed for killing)” including cage scrapers. Institutional SOPs existed for mice in only 63.5% of responses; 22.1% did not know whether one existed. Most striking, there was “little or no consensus regarding which physical actions are essential for a successful dislocation (i.e. a stretch and/or a twist)” — only 3.7% of mechanical-method users for mice regarded a lateral twist or roll as important, while 67.1% cited a forceful downward push.

Two practical conclusions follow. Write your SOP so it specifies the physical actions, not just the method name; and specify the tool, because “a pen” is what fills the gap when the SOP is silent. If your institution runs structured IACUC training, this is a technique that warrants hands-on assessment with a scored endpoint rather than a completion certificate.

Technique variants: tail tension is not obligatory

The classic description applies traction to the tail while separating the skull from the vertebral column. A 2025 evaluation tested three variants that apply no tension to the tail — a push, a pinch and a roll — and reported no gasping across the observed animals, with CT imaging at 118 µm resolution confirming dislocation without collateral vertebral damage in the 66 animals scanned. The study was single-centre and the authors themselves call for a larger multicentre comparison, so treat it as promising rather than settled. But it removes the assumption that tail traction is intrinsic to the method, which matters if your local technique has been built around it.

Anaesthesia first: the objection that does not survive the evidence

The standard reason given for dislocating a conscious mouse is that anaesthesia would confound the science. Sometimes that is true and the justification belongs in the protocol. Often it is inherited folklore.

The clearest example is euthanasia of a pregnant or lactating dam, where anaesthesia is commonly omitted for fear of harming the pups. A 2023 study compared time-mated C57BL/6J dams given 5% isoflurane before cervical dislocation (n=11) with conscious controls (n=11). Of 169 pups born, 97 (57%) survived to weaning: 48 (59%) from anaesthetised dams versus 49 (58%) from conscious dams. Female pups from anaesthetised dams were in fact heavier at weaning (9.6 g versus 8.5 g). The authors conclude that using isoflurane-induced harm as a justification for omitting anaesthesia “should be questioned”.

If you claim a scientific justification for conscious dislocation, name the specific analyte, tissue or endpoint that anaesthesia would compromise. “Anaesthesia affects the pups” is now a claim with evidence against it, and a reviewer who knows the literature will ask. This is a routine point of tension between scientific convenience and welfare obligation covered more broadly in our guide to animal research ethics.

Confirming death is a separate step, and palpation is not it

This is where compliance findings actually come from. Two distinct checks are involved and they are frequently collapsed into one.

1. Check the dislocation worked. The CCAC guidance is direct: “It is essential to check that the neck is broken at the end of the procedure by palpation of the vertebrae. If adequate separation is not observed, a backup method, such as decapitation or exposure to high concentrations of CO2, should be used immediately.” Palpating a gap tells you the anatomy separated. It does not tell you the animal is dead.
2. Confirm death by a listed method. Under UK Schedule 1 the process must be completed by one of six methods: confirmation of permanent cessation of the circulation, destruction of the brain, dislocation of the neck, exsanguination, confirming the onset of rigor mortis, or instantaneous destruction of the body in a macerator. The CCAC states the principle generally — “Death must be verified following euthanasia and prior to disposal of the animal” — and specifies that for methods including cervical dislocation, death “should be ensured without it regaining consciousness through application of a second step”. US practice mirrors this from the other direction: the NIH ARAC guideline on CO2 euthanasia requires that “death must be confirmed by an appropriate method” and, failing that, names cervical dislocation itself as an acceptable secondary method.

The practical rule: something must happen after the dislocation. Absence of respiration observed over a defined interval, absence of a heartbeat, fixed and dilated pupils, or an applied second physical step. Write which one into the SOP, and write it as a step with an owner rather than as a sentence of principle. Bodies going into a carcass cooler is the point at which an unconfirmed death becomes an incident.

Deciding whether to use it at all

Cervical dislocation persists for good reasons: it leaves no chemical residue, it is immediate, it requires no equipment, and it avoids the aversiveness of rising-concentration carbon dioxide. Those are real scientific and welfare advantages, particularly where tissue chemistry, oocyte quality or blood gases would be altered by an inhaled or injected agent.

Against that, the UK National Centre for the Replacement, Refinement and Reduction of Animals in Research states plainly that physical methods such as cervical dislocation “raise welfare concerns both because they can have high failure rates and because there is some limited evidence that they may not immediately abolish brain activity or consciousness”. The honest position for a protocol is that this is a trade-off with a named beneficiary, not a default. If the answer to “why not an anaesthetic overdose?” is “it is what we have always done”, that is not a justification a well-run committee should accept.

Where the science genuinely requires it, the defensible package is: an SOP that names the physical actions and the tool; a competence assessment with a scored endpoint on anaesthetised animals; anaesthesia before dislocation unless a specific endpoint precludes it; a backup method physically present at the bench; and a named confirmation-of-death step before disposal.

Frequently asked questions

Is cervical dislocation of mice AVMA-approved?

It is classified as conditionally acceptable rather than unconditionally acceptable. A published comparison of five major guidance documents finds the same classification in all of them, with the AVMA conditions tabulated as validated operator proficiency, availability of a secondary method if the attempt fails, and use in rodents under 200 g. Check the current 2020 edition of the AVMA guidelines for the exact wording, and remember that in the US the binding document is your approved IACUC protocol.

What is the weight limit for cervical dislocation in mice?

It depends on the jurisdiction, and the numbers genuinely differ. Canadian CCAC guidance limits manual dislocation to rodents under 200 g and requires a commercial dislocator for rats over 200 g. UK Schedule 1 permits it in rodents up to 500 g but requires prior sedation or anaesthesia above 150 g. Adult laboratory mice sit well below all of these thresholds, so weight is rarely the limiting factor for mice — competence and death confirmation are.

Do I have to anaesthetise a mouse before cervical dislocation?

Under UK Schedule 1 it is mandatory for rodents over 150 g, which most adult mice are not. CCAC guidance says animals should be anaesthetised first unless it would interfere with the scientific outcome, and that conscious use requires scientific justification. So for a typical mouse it is usually a protocol-level decision rather than a legal one — but the justification for skipping it has to be specific, and the common justification about harming pups has evidence against it.

How often does cervical dislocation fail in mice?

In the published study that measured it, 17 of 81 attempts (21%) resulted in the mouse continuing to breathe and being scored as unsuccessful. Of those 17, 11 (65%) were still alive at the 180-second cut-off. The more successful of the two operators still failed 8.1% of the time. Treat a non-trivial failure rate as expected, and design the backup step accordingly.

How do I confirm death after cervical dislocation?

Palpating the separated vertebrae confirms the dislocation, not death. Death must be confirmed separately — UK Schedule 1 lists six acceptable methods including permanent cessation of the circulation, destruction of the brain, exsanguination and onset of rigor mortis. In US practice, confirming cardiac and respiratory arrest or noting fixed and dilated pupils is the standard check, with a second physical method applied if death cannot be confirmed.

Can I use a mechanical device instead of doing it by hand?

Yes, and CCAC guidance requires a commercial dislocator for rats over 200 g. For mice, mechanical methods are widely used — but the 2026 UK survey found the “devices” in use are frequently improvised objects such as metal rods, closed scissors and pens rather than validated tools, and that respondents disagreed about which physical actions the device is meant to deliver. Specify the tool by name in your SOP.

Is cervical dislocation acceptable as a secondary method after CO2?

Yes. The NIH ARAC guideline on CO2 euthanasia of rodents states that if death cannot be confirmed, narcosis must be followed by a secondary method such as cervical dislocation, decapitation or bilateral thoracotomy. Applied to an already-unconscious animal, most of the welfare objections to the technique fall away.

References

• Carbone L, Carbone ET, Yi EM, et al. Assessing cervical dislocation as a humane euthanasia method in mice. J Am Assoc Lab Anim Sci. 2012;51(3):352–356. pmc.ncbi.nlm.nih.gov/articles/PMC3358985
• Martin JE, Leach MC, Clarkson JM. Inconsistency in cervical dislocation: a UK survey of techniques and tools utilised for laboratory rodents. Animal Welfare. 2026;35:e32. pmc.ncbi.nlm.nih.gov/articles/PMC13126198
• Davidge A, Bulat F, Vernet A. Evaluating cervical dislocation methods, without using tension on the tail, for humanely killing adult laboratory mice. Lab Anim. 2025;59(5):570–577. pmc.ncbi.nlm.nih.gov/articles/PMC12583621
• Cartner SC, Barlow SC, Ness TJ. Loss of cortical function in mice after decapitation, cervical dislocation, potassium chloride injection, and CO2 inhalation. Comp Med. 2007;57(6):570–573. pubmed.ncbi.nlm.nih.gov/18246869
• Hunter CL, Yang P, Renner DM, Kennedy L. Effects of isoflurane anesthesia on C57BL/6J pups after cervical dislocation of dams. J Am Assoc Lab Anim Sci. 2023;62(5):449–452. pmc.ncbi.nlm.nih.gov/articles/PMC10597329
• Shomer NH, Allen-Worthington KH, Hickman DL, et al. Review of rodent euthanasia methods. J Am Assoc Lab Anim Sci. 2020;59(3):242–253. pmc.ncbi.nlm.nih.gov/articles/PMC7210730
• Hawkins P, Prescott MJ, Carbone L, et al. A good death? Report of the second Newcastle meeting on laboratory animal euthanasia. Animals. 2016;6(9):50. pmc.ncbi.nlm.nih.gov/articles/PMC5035945
• Canadian Council on Animal Care. CCAC guidelines on: euthanasia of animals used in science. 2010. ccac.ca/Documents/Standards/Guidelines/Euthanasia.pdf
• Animals (Scientific Procedures) Act 1986, Schedule 1 — appropriate methods of humane killing. legislation.gov.uk/ukpga/1986/14/schedule/1
• NIH Animal Research Advisory Committee. Guidelines for euthanasia of rodents using carbon dioxide, updated 2024. oacu.oir.nih.gov — ARAC B5
• NC3Rs. Euthanasia — selection of killing technique. nc3rs.org.uk/3rs-resources/euthanasia
• American Veterinary Medical Association. AVMA guidelines for the euthanasia of animals: 2020 edition. avma.org/resources-tools/avma-policies/avma-guidelines-euthanasia-animals