Evaluation of EMLA Cream for Preventing Pain during Tattooing of RabbitsChanges in Physiological, Behavioural and Facial Expression Responses

Stephanie Keating, Aurélie Thomas, Paul Flecknell, Matthew C. Leach•View original

OverviewBalancedalloy voice

If you’ve ever handled a rabbit, you know those soft ears feel almost like silk. In labs and on farms, those ears are also where identification takes place. A few seconds with a handheld tattoo device and the animal receives a permanent mark. It’s fast, it’s practical, and too often, it’s done with little or no pain relief. That’s the tension: a routine procedure that causes pain, and a welfare question that’s easy to overlook because the pain is brief and linked to a single sharp moment. Here’s the tricky part: the very things we usually measure in animals—heart rate, hormones, reflex thresholds—are clumsy tools for a pain that flares and fades in seconds. Reflex tests like Hargreaves and Von Frey inform us about nociception, the nerve signal, not the conscious experience. Operant tests are elegant but slow and hard to apply to a three-second tattoo. What you want at the cage side is something you can quickly read, without elaborate equipment, and something that reflects the actual pain the animal feels in that moment. That’s where two ideas converge. First, a practical analgesic that’s easy to apply. Second, a face-based readout you can score on the spot. EMLA cream—a topical mix of 2.5 percent lidocaine and 2.5 percent prilocaine—checks the first box. It’s available over the counter in the UK and North America, and unlike general anesthesia, it doesn’t carry the same risks for rabbits. For the second, there’s a growing family of "grimace scales" in rodents that quantify pain from subtle changes in the eyes, cheeks, nose, whiskers, and ears. Keating, Thomas, Flecknell, and Leach asked a simple question: can we combine both for rabbit ear tattooing, and will we see the pain disappear if we numb the skin first? They constructed a study you want when the signal is brief and the stakes are high. Eight New Zealand White rabbits participated. Four conditions in a crossover design allowed each animal to serve as its own control: a real tattoo or a sham, each with either EMLA or an aqueous placebo applied beforehand. Creams were applied 20 minutes before the procedure. The tattoo itself was quick—about three seconds with pliers—while the sham used the same device fitted with a flat plate. Everything else was standardized: handling, restraint, and even a catheter placed in the opposite ear for sampling. Blinding was integrated from start to finish. The person applying the mark never knew which cream had been used. The team collecting data didn't either. A separate handler managed the creams. On the measurement side, they gathered several time scales at once. Cardiovascular activity—systolic, mean, and diastolic pressures plus heart rate—was sampled every ten seconds for two minutes before and two minutes after the treatment. Hormonal output came from plasma corticosterone, with samples taken at baseline, then at 15 minutes, 30 minutes, and up to three hours after. And because so much of this pain is reflected on the face, they filmed the head and scored still images with a new Rabbit Grimace Scale. Let’s focus on that face scale, because it’s both clever and practical. The team adapted an approach that Langford and Sotocinal pioneered in mice and rats. They started with five facial "action units" that seemed to change with pain—orbital tightening, cheek flattening, nose shape, whisker position, and ear position—but then narrowed down to what was visible and uncontaminated by the procedure. Ears were being handled and pressed, so ear position had to be removed. Whiskers weren’t consistently visible in the best images either. This left three reliable features—eyes, cheeks, and nose—each scored as absent, moderate, or obvious. Add them together and you get a total score ranging from zero to six. Would different observers see the same thing? Mostly, yes. Ten raters, blinded to treatment and time, scored 64 images taken before and during the procedure. The overall agreement, measured as an intraclass correlation, was 0.91—on par with the mouse and rat scales. The individual features held up well too, with especially tight agreement for the eye score. When asked to make a global call—does this rabbit look like it’s in pain?—the group's accuracy averaged 83.6 percent. Not perfect, but solid enough for real-world use. So, what happened when the metal met the ear? Behavior provided the first indications, because behavior always does. In the tattoo without EMLA condition, every rabbit struggled and vocalized, and some did so for several seconds. With sham procedures, there was virtually none of that. With EMLA applied before tattooing, one rabbit vocalized briefly and didn’t struggle. The contrast was so stark that, statistically, it landed at the floor of the p-value scale. You didn’t need to be a behaviorist to see it on the screen. The faces mirrored the behavior. Intra-procedure grimace scores rose when the tattoo was done on untreated skin. The same motion on a numbed ear appeared, frankly, like nothing much had happened. Only the condition without EMLA and with a real tattoo showed a reliable increase from pre- to intra-procedure; the p-value there was below 0.001. Time alone didn’t drive the score changes. It was the interaction of time with what was actually done to the ear and whether the tissue had been numbed first. Now to the physiology. Heart rate climbed, but not uniformly. There was a clear treatment effect on the change from before to after the procedure, with the largest spike occurring when the tattoo happened without EMLA. The overall test for heart rate change across conditions reached significance at a p-value of 0.002. Blood pressure showed a similar trend. Systolic pressure peaked higher in the untreated tattoo condition than in any other, with the overall effect having a p-value of 0.001. Mean arterial pressure followed suit, with a p-value of 0.004. Diastolic pressure did not separate the groups. Here’s the nuance: if you integrate heart rate over the entire two-minute window, the area under the curve showed no difference by treatment. The p-value was 0.81 for that comparison. In other words, the cardiovascular surge was a sharp peak linked to the event, not a prolonged elevation. Hormones add a slower layer and provide a complementary perspective. Corticosterone rose after the procedure, partly because being handled and sampled is a stressor. The time course was more critical than the label on the condition. Levels were elevated at 15 minutes and 30 minutes in most groups and moved back toward baseline by the one-hour mark and later. When the team specifically examined the 15-minute change, the increases were higher after tattooing—with and without EMLA—compared to sham with EMLA; those pairwise differences had p-values of 0.05 and 0.01. Sham without EMLA also nudged corticosterone up at 15 minutes but not at 30, suggesting that restraint and handling alone carry a short-lived hormonal cost. The key point is timing. The hormone signal was transient and did not align neatly with the seconds-long pain that EMLA eliminated. Back in the home pen, the rabbits looked like, well, rabbits. Grooming increased in some groups an hour after the procedure, but the researchers noted a confound: there might be residual cream on the fur. Movement and rearing dipped from baseline to one hour across conditions, with no robust differences between groups. These kinds of measures reassure you there isn’t a lingering welfare hit, but they’re too slow and too blunt to tell you about the pain of the moment. Put all of this together and you get a clear, convergent picture. Remove the topical anesthetic and a three-second ear tattoo triggers a cascade: overt struggling and squeals, a face that tightens around the eyes and nose, and rapid cardiovascular spikes. Apply EMLA 20 minutes beforehand and most of that disappears. Behavior and facial scores during the numbed tattoo resemble those of the shams. The hormone curves serve as a reminder that procedures have broader stress footprints, but the sharp pain—the aspect we care about most in this setting—can be reduced to near zero with a cream that’s easy to use. It’s worth recognizing the study’s design for that clarity. Treatments were semi-randomized and crossed within each rabbit, with at least two days between applications and a week between the sham block and the tattoo block to prevent any sensitization that might influence the sham. The operator and the assessment team were blinded to cream type, and a separate person conducted the applications to maintain that blinding. The cardiovascular sampling was detailed enough—every ten seconds for two minutes—to capture the peaks, and the corticosterone schedule—at 15 minutes, 30 minutes, and up to three hours—tracked the slow return to baseline. Even the ethics were a priority, with a rescue plan in place: if a rabbit displayed ongoing signs of pain, buprenorphine was available at a dose of 0.01 milligrams per kilogram, administered subcutaneously. There are practical takeaways here that extend beyond this specific lab. First, EMLA represents a low-friction change. It’s affordable, accessible, and in this research, it almost completely prevented the acute pain signals of ear tattooing. For a procedure that’s done widely and frequently, that alone is a significant win. Second, the Rabbit Grimace Scale provides a cage-side tool you can actually use. It doesn’t require a catheter or a telemetry unit. If you can see the rabbit’s eyes, cheeks, and nose, you can score it. The overall reliability was high, and while the original concept included whisker position and ear posture, those features were appropriately left out here because they were either obscured or manipulated. That kind of practicality makes a scale transferable. And the nuance is important. Keating and colleagues observed that treatment order and the interaction of time with treatment influenced the grimace scores; time alone did not. That’s a valuable reminder to design for within-subject contrasts and to be cautious about simple before-and-after readings in procedures where handling is itself a stressor. They also noted that corticosterone changes were modestly affected by cream application and restraint, highlighting that hormones are excellent for providing information about “the day” but not “the second.” If you're considering how this applies in your own context, the guidelines are straightforward. Apply EMLA about 20 minutes before the mark. Expect that most overt signs—the struggle, the cry, the facial tightening—won’t manifest. Be aware that you might still observe a short hormonal spike from handling, but it will settle by an hour. If you want to check yourself, use the face. Eyes narrowing, cheeks flattening, the nose pointing—those are the indicators. If they’re absent, you’re treating the animal well. Could this research go further? Certainly. Improved image capture would make whisker scoring more feasible. Testing the scale in different breeds, at different ages, or in other quick procedures—such as ear venipuncture—could address the limits of external validity. However, the core finding remains strong. As Leach and colleagues demonstrated through behavior, physiology, hormones, and facial scoring, ear tattooing without analgesia produces sharp, transient pain, while a thin layer of EMLA blocks that pain. That’s a simple change that yields a significant welfare benefit.

It’s fast, it’s practical, and too often, it’s done with little or no pain relief. That’s the tension: a routine procedure that causes pain, and a welfare question that’s easy to overlook because the pain is brief and linked to a single sharp moment.

Here’s the tricky part: the very things we usually measure in animals—heart rate, hormones, reflex thresholds—are clumsy tools for a pain that flares and fades in seconds. Reflex tests like Hargreaves and Von Frey inform us about nociception, the nerve signal, not the conscious experience. Operant tests are elegant but slow and hard to apply to a three-second tattoo.

What you want at the cage side is something you can quickly read, without elaborate equipment, and something that reflects the actual pain the animal feels in that moment.

That’s where two ideas converge. First, a practical analgesic that’s easy to apply. Second, a face-based readout you can score on the spot.

EMLA cream—a topical mix of 2.5 percent lidocaine and 2.5 percent prilocaine—checks the first box. It’s available over the counter in the UK and North America, and unlike general anesthesia, it doesn’t carry the same risks for rabbits. For the second, there’s a growing family of "grimace scales" in rodents that quantify pain from subtle changes in the eyes, cheeks, nose, whiskers, and ears.

Keating, Thomas, Flecknell, and Leach asked a simple question: can we combine both for rabbit ear tattooing, and will we see the pain disappear if we numb the skin first?

They constructed a study you want when the signal is brief and the stakes are high. Eight New Zealand White rabbits participated. Four conditions in a crossover design allowed each animal to serve as its own control: a real tattoo or a sham, each with either EMLA or an aqueous placebo applied beforehand.

Creams were applied 20 minutes before the procedure. The tattoo itself was quick—about three seconds with pliers—while the sham used the same device fitted with a flat plate. Everything else was standardized: handling, restraint, and even a catheter placed in the opposite ear for sampling.

Blinding was integrated from start to finish. The person applying the mark never knew which cream had been used. The team collecting data didn't either.

A separate handler managed the creams. On the measurement side, they gathered several time scales at once. Cardiovascular activity—systolic, mean, and diastolic pressures plus heart rate—was sampled every ten seconds for two minutes before and two minutes after the treatment.

Hormonal output came from plasma corticosterone, with samples taken at baseline, then at 15 minutes, 30 minutes, and up to three hours after. And because so much of this pain is reflected on the face, they filmed the head and scored still images with a new Rabbit Grimace Scale.

Let’s focus on that face scale, because it’s both clever and practical. The team adapted an approach that Langford and Sotocinal pioneered in mice and rats. They started with five facial "action units" that seemed to change with pain—orbital tightening, cheek flattening, nose shape, whisker position, and ear position—but then narrowed down to what was visible and uncontaminated by the procedure.

Ears were being handled and pressed, so ear position had to be removed. Whiskers weren’t consistently visible in the best images either. This left three reliable features—eyes, cheeks, and nose—each scored as absent, moderate, or obvious. Add them together and you get a total score ranging from zero to six.

Would different observers see the same thing? Mostly, yes. Ten raters, blinded to treatment and time, scored 64 images taken before and during the procedure.

The overall agreement, measured as an intraclass correlation, was 0.91—on par with the mouse and rat scales. The individual features held up well too, with especially tight agreement for the eye score. When asked to make a global call—does this rabbit look like it’s in pain?—the group's accuracy averaged 83.6 percent. Not perfect, but solid enough for real-world use.

So, what happened when the metal met the ear? Behavior provided the first indications, because behavior always does. In the tattoo without EMLA condition, every rabbit struggled and vocalized, and some did so for several seconds.

With sham procedures, there was virtually none of that. With EMLA applied before tattooing, one rabbit vocalized briefly and didn’t struggle. The contrast was so stark that, statistically, it landed at the floor of the p-value scale. You didn’t need to be a behaviorist to see it on the screen.

The faces mirrored the behavior. Intra-procedure grimace scores rose when the tattoo was done on untreated skin. The same motion on a numbed ear appeared, frankly, like nothing much had happened.

Only the condition without EMLA and with a real tattoo showed a reliable increase from pre- to intra-procedure; the p-value there was below 0.001. Time alone didn’t drive the score changes. It was the interaction of time with what was actually done to the ear and whether the tissue had been numbed first.

Now to the physiology. Heart rate climbed, but not uniformly. There was a clear treatment effect on the change from before to after the procedure, with the largest spike occurring when the tattoo happened without EMLA.

The overall test for heart rate change across conditions reached significance at a p-value of 0.002. Blood pressure showed a similar trend. Systolic pressure peaked higher in the untreated tattoo condition than in any other, with the overall effect having a p-value of 0.001.

Mean arterial pressure followed suit, with a p-value of 0.004. Diastolic pressure did not separate the groups. Here’s the nuance: if you integrate heart rate over the entire two-minute window, the area under the curve showed no difference by treatment.

The p-value was 0.81 for that comparison. In other words, the cardiovascular surge was a sharp peak linked to the event, not a prolonged elevation.

Hormones add a slower layer and provide a complementary perspective. Corticosterone rose after the procedure, partly because being handled and sampled is a stressor. The time course was more critical than the label on the condition.

Levels were elevated at 15 minutes and 30 minutes in most groups and moved back toward baseline by the one-hour mark and later. When the team specifically examined the 15-minute change, the increases were higher after tattooing—with and without EMLA—compared to sham with EMLA; those pairwise differences had p-values of 0.05 and 0.01. Sham without EMLA also nudged corticosterone up at 15 minutes but not at 30, suggesting that restraint and handling alone carry a short-lived hormonal cost.

The key point is timing. The hormone signal was transient and did not align neatly with the seconds-long pain that EMLA eliminated.

Back in the home pen, the rabbits looked like, well, rabbits. Grooming increased in some groups an hour after the procedure, but the researchers noted a confound: there might be residual cream on the fur. Movement and rearing dipped from baseline to one hour across conditions, with no robust differences between groups.

These kinds of measures reassure you there isn’t a lingering welfare hit, but they’re too slow and too blunt to tell you about the pain of the moment.

Put all of this together and you get a clear, convergent picture. Remove the topical anesthetic and a three-second ear tattoo triggers a cascade: overt struggling and squeals, a face that tightens around the eyes and nose, and rapid cardiovascular spikes. Apply EMLA 20 minutes beforehand and most of that disappears.

Behavior and facial scores during the numbed tattoo resemble those of the shams. The hormone curves serve as a reminder that procedures have broader stress footprints, but the sharp pain—the aspect we care about most in this setting—can be reduced to near zero with a cream that’s easy to use.

It’s worth recognizing the study’s design for that clarity. Treatments were semi-randomized and crossed within each rabbit, with at least two days between applications and a week between the sham block and the tattoo block to prevent any sensitization that might influence the sham. The operator and the assessment team were blinded to cream type, and a separate person conducted the applications to maintain that blinding.

The cardiovascular sampling was detailed enough—every ten seconds for two minutes—to capture the peaks, and the corticosterone schedule—at 15 minutes, 30 minutes, and up to three hours—tracked the slow return to baseline. Even the ethics were a priority, with a rescue plan in place: if a rabbit displayed ongoing signs of pain, buprenorphine was available at a dose of 0.01 milligrams per kilogram, administered subcutaneously.

There are practical takeaways here that extend beyond this specific lab. First, EMLA represents a low-friction change. It’s affordable, accessible, and in this research, it almost completely prevented the acute pain signals of ear tattooing.

For a procedure that’s done widely and frequently, that alone is a significant win. Second, the Rabbit Grimace Scale provides a cage-side tool you can actually use. It doesn’t require a catheter or a telemetry unit.

If you can see the rabbit’s eyes, cheeks, and nose, you can score it. The overall reliability was high, and while the original concept included whisker position and ear posture, those features were appropriately left out here because they were either obscured or manipulated. That kind of practicality makes a scale transferable.

And the nuance is important. Keating and colleagues observed that treatment order and the interaction of time with treatment influenced the grimace scores; time alone did not. That’s a valuable reminder to design for within-subject contrasts and to be cautious about simple before-and-after readings in procedures where handling is itself a stressor.

They also noted that corticosterone changes were modestly affected by cream application and restraint, highlighting that hormones are excellent for providing information about “the day” but not “the second.”

If you're considering how this applies in your own context, the guidelines are straightforward. Apply EMLA about 20 minutes before the mark. Expect that most overt signs—the struggle, the cry, the facial tightening—won’t manifest.

Be aware that you might still observe a short hormonal spike from handling, but it will settle by an hour. If you want to check yourself, use the face. Eyes narrowing, cheeks flattening, the nose pointing—those are the indicators. If they’re absent, you’re treating the animal well.

Could this research go further? Certainly. Improved image capture would make whisker scoring more feasible.

Testing the scale in different breeds, at different ages, or in other quick procedures—such as ear venipuncture—could address the limits of external validity. However, the core finding remains strong. As Leach and colleagues demonstrated through behavior, physiology, hormones, and facial scoring, ear tattooing without analgesia produces sharp, transient pain, while a thin layer of EMLA blocks that pain. That’s a simple change that yields a significant welfare benefit.

Related lectures