Sedation or Inhalant Anesthesia before Euthanasia with CO₂ Does Not Reduce Behavioral or Physiologic Signs of Pain and Stress in Mice

Dear Editor,

Valentine and colleagues¹³ tested the effects of anesthetic induction with isoflurane on behavioral and physiologic signs of pain and stress in mice euthanized with CO₂, and concluded that induction with isoflurane prior to euthanasia with CO₂ is worse for the animals’ welfare than euthanasia with CO₂ alone. This conclusion seems to contradict a growing body of literature⁹,¹¹,¹⁴ that shows that exposure to CO₂ is strongly aversive to rodents, likely due to feelings of dyspnea¹¹ (that is, “air hunger”) or anxiety¹⁴ as reported in humans, and that induction with isoflurane is a less aversive alternative. For example, a key study⁹ (not cited by Valentine and colleagues) showed that mice would sometimes choose to stay in a chamber filling with isoflurane until they were recumbent rather than abandoning a sweet food reward, but always abandoned the reward when the chamber was filling with CO₂.These results correspond with those from an earlier preference study⁷ showing that mice tolerate longer periods of exposure to isoflurane than to CO₂.

The conclusion of Valentine and colleagues¹³ rests on 4 results. We argue below that each of these is based on problematic methods or interpretation.

1) 5 of the 10 mice anesthetized with isoflurane recovered consciousness while the cage was being filled with CO₂. The criterion the authors used to determine when mice were unconscious was “cessation of voluntary movement,” but this is not an appropriate proxy for unconsciousness. Isoflurane has sedative properties, and after an initial excitatory phase, mice appear ‘sleepy’ and settle down. At this stage mice are not unconscious and will withdraw in response to touch. Unconsciousness occurs only after the animal becomes recumbent (loss of muscular tone)⁵ and breathing is deep and slow. A simple method to ensure the mice have lost consciousness is to check for the absence of the righting reflex when the mice are rolled onto their backs by tilting the chamber; this measure shows a strong correlation with measures of loss of consciousness in humans.² At UBC, where standard practice is to anesthetize with isoflurane and switch to CO₂ only after mice are recumbent, we have never had a report of a mouse recovering consciousness during the procedure.

2) Mice had highest agitation and dyspnea scores with isoflurane. The ‘agitation’ score used was likely not appropriate for comparing distress between isoflurane and CO₂.¹⁰ Isoflurane induces an excitatory phase,⁸ but there is no evidence that this behavior is reflective of aversion or distress. In contrast, mice often respond to CO₂ by gasping at the bottom of the cage, a response associated with low levels of activity. The use of the term ‘dyspnea’ is another source of confusion; in the veterinary literature this term is generally defined as “labored breathing” while the medical literature defines this as a feeling of “air hunger.” Feelings of air hunger (often extreme and distressing) are reported by humans exposed to CO₂¹ but not isoflurane. Valentine and colleagues defined dyspnea as “increased respiratory effort,” and likely measured increased breathing rates associated with activity during the excitation stage of isoflurane induction rather than air hunger.

3) Mice exposed to isoflurane produced calls with a peak frequency of 26.5 kHz, potentially indicative of stress. The paper cited in support of this claim⁴ discusses only vocalizations in rats in response to pain. However, in contrast to rats, there is no evidence that vocalizations in mice are indicative of negative or positive affect.¹² The observed peak may be an artifact of increased activity rather than actual vocalizations.

4) c-fos expression was highest in the sedative and isoflurane groups. Valentine and colleagues failed to detect an increase in c-fos after exposure to CO₂ alone, but a previous study⁶ found that a brief exposure to CO₂ caused a specific, localized expression of c-fos in brain areas involved in panic and defensive reactions in rodents, including the hypothalamic-pituitary axis. Valentine and colleagues evaluated global levels of c-fos mRNA in a 2-mm brain slice whereas the previous study⁶ examined local expression of c-fos using immunostaining; Valentine and colleagues’ global measure may have ‘averaged out’ localized increases in expression of c-fos in specific brain areas that would have been detected by a more selective assay.

Moreover, Valentine and colleagues examined c-fos expression only 4 min after the onset of CO₂ whereas the previous study examined c-fos levels 90 min after exposure when c-fos expression in response to a stressor is likely to be maximal.⁶ Previous work has shown that c-fos levels are not significantly elevated 5 min after induction and maintenance of anesthesia with isoflurane but are elevated at 30 min.³ Consequently, the elevated c-fos levels obtained by Valentine and colleagues likely reflect preeuthanasia handling in the case of mice receiving premedication (explaining why the sedative and the saline control groups also had elevated c-fos levels) or increased locomotor activity in the case of animals receiving isoflurane.

In summary, we suggest that the conclusion from Valentine and colleagues¹³ should be treated with caution. Our reading of the literature suggests that CO₂ is aversive to rodents, and current evidence indicates that isoflurane is less aversive than CO₂.

Sincerely,

Joanna Makowska MSc
PhD candidate, Animal Welfare Program University of British Columbia
Huw Golledge PhD
Senior Research Associate, Institute of Neuroscience University of Newcastle
Nicole Marquardt Veterinarian
Research Assistant, Institute of Pharmacology and Toxicology Freie Universität Berlin, School of Veterinary Medicine
Daniel M Weary D Phil
Professor, NSERC Industrial Research Chair in Animal Welfare, Animal Welfare Program University of British Columbia