Retro-Orbital Injection in Mice: When It Is Justified and What the Alternatives Are

Most arguments about retro-orbital injection in mice are actually arguments about retro-orbital bleeding, and the two are not the same procedure. The alarming welfare literature — globe rupture, corneal ulceration, optic nerve damage — describes a glass capillary tube or Pasteur pipette advanced into the orbit and manipulated until blood flows. Injection is a fine needle, one pass, bevel down, nothing aspirated and no tissue disrupted to establish flow. Conflating them is the single most common error in this decision, and it runs in both directions: IACUC members cite bleeding guidance to refuse an injection protocol, and investigators cite injection pharmacokinetics papers to justify serial bleeding.

This page separates the two bodies of evidence, states where the injection route is genuinely defensible, and is explicit about which claims are well supported and which are not.

What the comparative injection data actually show

Several head-to-head studies have compared retro-orbital sinus injection with lateral tail vein injection, and for systemic delivery of a soluble agent they broadly agree.

• Repeat dosing of a small molecule. MAFIA transgenic mice given five consecutive daily injections of the macrophage-depleting agent AP20187 by either route showed similar depletion across lung, spleen, bone marrow and peritoneal exudate cells. A parallel experiment in BALB/c mice indicated the retro-orbital route was the less stressful of the two.
• Antibody pharmacokinetics. A comparison of serum concentrations and PK parameters for a therapeutic antibody found no difference in absorption or pharmacokinetic activity between the lateral tail vein and the retro-orbital venous sinus.
• Bone marrow transplantation. In two transplant scenarios in P3B and B6D2F1 mice, donor engraftment did not differ significantly between routes, and haematological counts and leukocyte subpopulations were near-identical. Notably, engraftment was less homogeneous after tail vein injection, which the authors attribute to that technique’s higher failure rate.
• Contrast imaging. For dynamic contrast-enhanced MRI, the time course and magnitude of contrast enhancement in liver, kidney, lung and myocardium were similar by either route, with maximum contrast ratio 30 seconds after administration. Venous puncture plus mounting the injection system took roughly 10 minutes by tail vein versus about 4 minutes retro-orbitally.

That last figure is the honest reason many labs adopt the route. It is faster, it fails less often, and a failed tail vein attempt is not a welfare-free event — repeated sticks, restraint and tail warming carry their own cost, and they degrade your data through dose uncertainty.

Where the equivalence claim breaks down

Equivalence is payload- and timepoint-dependent, and it is not safe to assume. A 2025 study comparing the two routes for Streptococcus agalactiae (Group B Streptococcus) systemic infection found that retro-orbital inoculation produced an initial increase in bacterial dissemination to spleen and brain tissue, with increased brain burden also detected at a later timepoint. At experimental endpoints the routes converged: no significant difference in bacterial burden in blood, kidney, spleen, heart or lung, nor in health scores, brain cytokine abundance or survival.

The practical reading is narrow and important. If your readout is a survival curve or an endpoint organ burden, the route is unlikely to matter. If your readout is early biodistribution, first-pass kinetics, or anything CNS-adjacent, the retro-orbital sinus drains differently enough that you must validate the swap rather than assume it. Changing route mid-study, or pooling historical tail vein controls with new retro-orbital animals, is where this quietly becomes a reproducibility problem.

Jurisdiction decides this before the science does

Two respected bodies land in different places, and which one governs you is usually the first thing to establish.

United Kingdom. The NC3Rs guidance on mouse blood sampling is unambiguous: “Retro-orbital bleeding should only be performed under terminal anaesthesia because of the severity of adverse effects that can occur with this technique, even in skilled hands.” It is listed among the terminal, non-recovery techniques alongside cardiac puncture and decapitation, and the documented adverse effects include retro-orbital haemorrhage with excessive pressure on the eye, corneal ulceration, keratitis, pannus formation, rupture of the globe, micro-ophthalmia from proptosis, damage to the optic nerve and other intra-orbital structures leading to blindness, fracture of the fragile orbital bones, and penetration of the globe with loss of vitreous humour. This position shapes what a project licence will realistically authorise under the Animals (Scientific Procedures) Act 1986.

United States. The NIH Animal Research Advisory Committee guideline on survival blood collection permits the route as a survival procedure under conditions. In mice, general anaesthesia is recommended where compatible with experimental design, and if bleeding is conducted without general anaesthesia a topical ophthalmic anaesthetic such as proparacaine or tetracaine must be applied first. In rats, where the anatomy is a venous plexus rather than a sinus and orbital tissue damage is correspondingly greater, general anaesthesia must be used unless scientifically justified and approved; ARAC further states that retro-orbital bleeding in rats by a trained practitioner represents more than minimal or transient pain and distress and should be classified as a USDA Column D procedure.

Both statements are about bleeding. Neither body publishes an equivalent categorical position on retro-orbital injection, which is why local institutional policy carries so much weight here. Do not assume your IACUC protocol inherits the more permissive reading.

The welfare evidence, read honestly

The comparative welfare literature is genuinely conflicting, and anyone who tells you it points cleanly one way has not read it.

Complications are common even in trained hands. In a JAALAS study, two experienced instructors trained 40 novices in a single hands-on session; blood was then collected from anaesthetised mice via a capillary tube at the medial canthus, one insertion per animal, with the contralateral orbit as an unmanipulated control. Across 80 mice, 79 (98.8%) maintained normal body condition, posture and behaviour over 14 days — but ocular lesions occurred at least once in 43 of 80 animals (53.8%), and clinical and histopathologic scores did not differ between novice and experienced personnel. Two conclusions follow, and both are uncomfortable: training does not drive the complication rate to zero, and restricting the procedure to senior staff is not supported by this evidence either.

The obvious substitute is not reliably kinder. A PLoS ONE study comparing retro-bulbar sinus puncture with facial vein phlebotomy in C57BL/6J mice found that facial vein animals had significantly elevated plasma corticosterone at both timepoints while retro-bulbar animals did not, and lost more body weight — though histopathology revealed extensive tissue trauma after both. A separate study of serial sampling (up to 0.4 mL every second day for two weeks) reached the opposite conclusion, favouring the facial vein on collection time, corticosterone and tissue damage. The disagreement is real; sampling frequency and volume appear to be what separates the two findings.

No method wins overall. A JAALAS comparison of six sampling methods found that retrobulbar sinus puncture under isoflurane produced the largest effect on body weight loss and, along with sublingual puncture and isoflurane anaesthesia alone, the highest corticosterone. Yet retrobulbar, sublingual and submandibular puncture caused only minor, mostly fast-resolving local inflammation, whereas lateral tail incision, tail-tip amputation and saphenous puncture caused local tissue damage that resolved more slowly — inflammatory gene expression rising 10- to 1000-fold and not returning to baseline until day 6 or later. The authors concluded that none of the methods was superior. This is the finding most often omitted when the route is argued about, and it cuts directly against the reflex that any alternative is automatically a refinement in the sense meant by the 3Rs.

When retro-orbital injection is genuinely justified

Four situations survive scrutiny.

1. Neonates past the temporal vein window. The superficial temporal (facial) vein sits just anterior to the ear bud and is clearly visible for roughly the first two days after birth, accepting volumes up to 50 µL; by the third postnatal day it is difficult to visualise and the technique becomes unreliable. A published protocol describes retro-orbital venous sinus injection in neonatal mice and rats precisely for the window after the temporal vein disappears but before the eyes open — and notes that eye-opening actually helps, by letting the operator see they are not perforating the globe. This is a genuine gap with no obvious replacement.
2. Anaesthesia is already on the protocol. If the animal is going under for imaging or surgery anyway, the main welfare cost of the route is already being paid, and the argument shifts decisively.
3. Tail vein failure is itself the problem. Where the tail vein fails often enough to produce inconsistent dosing — pigmented strains, young or dehydrated animals, viscous or particulate vehicles, cell suspensions — the more homogeneous delivery seen in the transplant comparison is a data-quality argument, not just a convenience one.
4. Throughput at scale under time pressure. Where a cohort must be dosed within a narrow window, the several-minute per-animal saving is real.

It is correspondingly hard to justify when the study is UK-regulated, when the injection must be repeated frequently in the same animal, when the readout is early CNS or splenic biodistribution, or when a tail vein injection would have worked and the only argument for switching is operator preference.

The alternatives, and what each one costs

• Lateral tail vein. The default IV comparator and the route against which everything above was benchmarked. Technically demanding, benefits from warming to induce vasodilation, higher failure rate, and repeated failed attempts carry real welfare and dosing costs.
• Intraperitoneal. Far easier and lower-risk, but it is not intravenous — absorption is slower and subject to a first-pass effect, so it is a substitute only where the kinetics genuinely permit. Our guide to routes of administration in laboratory rodents covers gavage, IP, SC and IV technique and volume limits in detail.
• Superficial temporal / facial vein (neonates). The preferred neonatal IV route inside its narrow window, taking larger volumes than neonatal retro-orbital injection and completed in one to two minutes.
• Vascular cannulation. The recognised refinement for repeated access, trading a surgical procedure and its recovery for the elimination of repeated punctures.
• Sublingual vein. Described in the literature specifically as an alternative to the retrobulbar technique that provides large volumes while minimising tissue damage — though note that in the six-method comparison it also sat among the highest for corticosterone.

Writing a protocol that will survive review

The procedural controls below come from the NIH ARAC survival blood collection guideline. They are written for sampling, not injection, but committees routinely apply them to injection protocols by analogy, and a protocol that pre-empts them is far less likely to come back with questions.

• Approach the medial canthus, angled toward the back of the orbit; ARAC specifies insertion under the nictitating membrane at 45 degrees for sampling.
• Use one eye. Alternating orbits should not be attempted until the operator is proficient on the side most readily accessed by their dominant hand.
• Allow at least 10 days for tissue repair before returning to the same orbit; earlier repetition means the healing process itself interferes.
• Apply topical ophthalmic anaesthetic where the animal is not under general anaesthesia, and protect the contralateral eye with ophthalmic lubricant while avoiding pressure damage.
• Train on terminal animals. ARAC requires general anaesthesia during training, with animals euthanised immediately afterwards.
• Ensure haemostasis afterwards, and use sterile equipment to avoid periorbital infection and long-term eye damage.
• Justify exceptions explicitly. Anything outside the guideline — including retro-orbital procedures without topical anaesthesia — must be scientifically justified in the approved protocol.
• Report the route. Route, volume, anaesthesia and operator training all belong in the methods section under the ARRIVE guidelines; “injected intravenously” is not a reproducible description when the routes differ in early biodistribution.

Frequently asked questions

How much can I inject retro-orbitally into a mouse?

The most-cited protocol reports routinely administering volumes up to 150 µL in the adult mouse without incident, and up to 10 µL in neonates, where the technique is more challenging. Treat that as an author-reported operational figure rather than a regulatory ceiling — institutional caps vary, and your own IACUC-approved limit governs.

Does retro-orbital injection require general anaesthesia?

There is no single federal rule for injection. The NIH ARAC guideline covers blood collection, where general anaesthesia is recommended in mice if compatible with the experimental design and a topical ophthalmic anaesthetic is mandatory if the animal is conscious; in rats, general anaesthesia is required absent approved justification. Most institutions require brief general anaesthesia for injection as well, but this is set locally — check your own SOP rather than reasoning from the sampling guidance.

Is retro-orbital injection allowed in the UK?

NC3Rs classifies retro-orbital bleeding as a terminal, non-recovery technique to be performed only under terminal anaesthesia. It publishes no equivalent categorical statement on injection, but that position strongly shapes what UK establishments and Home Office licensing will accept. Assume you will need to argue the case, and expect the tail vein to be treated as the default.

Can I inject into the same eye more than once?

The applicable published interval — from the sampling guideline — is a minimum of 10 days before returning to the same orbit, so that healing does not interfere. Alternating eyes is not a way around this for an inexperienced operator: guidance is explicit that proficiency on the dominant-hand side must come first.

Is it really equivalent to tail vein injection?

For endpoint measures — PK parameters of a therapeutic antibody, macrophage depletion, bone marrow engraftment, contrast enhancement, infection survival curves — the published comparisons find no meaningful difference. For early timepoints it is not safely equivalent: retro-orbital inoculation increased initial bacterial dissemination to spleen and brain in a Group B streptococcal model even though endpoints matched. Validate the swap for your specific readout.

Are complications just a matter of operator skill?

No, and this is the most consistently misunderstood point. In the JAALAS training study, ocular lesions occurred at least once in 53.8% of mice after a single capillary insertion, and clinical and histopathologic scores did not differ between novices trained in one session and the experienced instructors who taught them. Skill matters for success rate; it did not eliminate ocular findings in that study.

Should I switch to the facial vein instead?

Not automatically. One controlled comparison found facial vein phlebotomy produced higher corticosterone and greater weight loss than retro-bulbar sinus puncture, with extensive tissue trauma after both; another, using more frequent serial sampling, favoured the facial vein. A six-method comparison concluded that no sampling method was superior overall. Pick on the specifics of your volume, frequency and readout, not on a general reputation.

References

• NIH Office of Animal Care and Use, ARAC. Guidelines for Survival Blood Collection in Mice and Rats (B2, revised 10 December 2025).
• NC3Rs. Blood sampling: Mouse.
• Yardeni T et al. Retro-orbital injections in mice. Lab Animal 2011;40(5):155–60. doi:10.1038/laban0511-155
• Comparison of the lateral tail vein and the retro-orbital venous sinus as routes of intravenous drug delivery in a transgenic mouse model. Lab Animal 2008. PMID 18094699
• Comparison of the lateral tail vein and the retro-orbital venous sinus routes of antibody administration in pharmacokinetic studies. Lab Animal 2015. doi:10.1038/laban.481
• Comparison of haematopoietic stem cell engraftment through the retro-orbital venous sinus and the lateral vein. Laboratory Animals 2015. doi:10.1177/0023677214567915
• Assessment of MRI contrast agent kinetics via retro-orbital injection in mice: comparison with tail vein injection. PLoS ONE 2015. doi:10.1371/journal.pone.0129326
• Comparison of lateral tail vein and retro-orbital venous sinus as routes of inoculation to study Group B streptococcal systemic infection. Microbiology Spectrum 2025. doi:10.1128/spectrum.02104-24
• Fried JH et al. Type, duration, and incidence of pathologic findings after retroorbital bleeding of mice by experienced and novice personnel. JAALAS 2015;54(3):317–27. PMC4460946
• Sørensen DB et al. Time-dependent pathologic and inflammatory consequences of various blood sampling techniques in mice. JAALAS 2019;58(3):362–72. doi:10.30802/AALAS-JAALAS-18-000064
• Teilmann AC et al. Physiological and pathological impact of blood sampling by retro-bulbar sinus puncture and facial vein phlebotomy in laboratory mice. PLoS ONE 2014;9(11):e113225. doi:10.1371/journal.pone.0113225
• Jo EJ et al. Comparison of murine retroorbital plexus and facial vein blood collection to mitigate animal ethics issues. Laboratory Animal Research 2021;37(1):12. doi:10.1186/s42826-021-00090-4
• Intravenous injections in neonatal mice. JoVE 2015. doi:10.3791/52037
• A neonatal rodent model of retroorbital vein injection. JoVE 2024. doi:10.3791/65386
• Diehl KH et al. A good practice guide to the administration of substances and removal of blood, including routes and volumes. J Appl Toxicol 2001;21(1):15–23. PMID 11180276
• Heimann M et al. Blood collection from the sublingual vein in mice and hamsters: a suitable alternative to retrobulbar technique that provides large volumes and minimizes tissue damage. Laboratory Animals 2009;43:255–60. doi:10.1258/la.2008.007073