Review of Intraperitoneal Injection of Sodium Pentobarbital as a Method of Euthanasia in Laboratory Rodents

Abstract

Euthanasia is one of the most commonly performed procedures in biomedical research, involving tens of millions of animals in North America and Europe every year. The use of sodium pentobarbital, injected intraperitoneally, for killing rodents is described as an acceptable technique by the AVMA and CCAC euthanasia guidelines. This drug and route are recommended over inhalant anesthetics, carbon dioxide, and physical methods for ethical and aesthetic reasons as well as efficiency. However, a growing body of evidence challenges the efficacy and utility of intraperitoneal pentobarbital. This methodology has been described as inconsistent and may induce pain and stress. With these considerations in mind, a review of the literature is needed to assess the evidence surrounding this killing method, the associated welfare implications, and potential for refinement.

There are approximately 17 million rodents (rats and mice) used in research annually in the European Union, United Kingdom, and Canada.^19,37,97 An overwhelming majority of these rodents are euthanized either as part of a study protocol or at the end of a research project, making euthanasia one of the most commonly performed laboratory procedures. Despite the prevalence of euthanasia procedures and its conceptual simplicity, ensuring a ‘good death’ across a wide variety of laboratory settings and involving such large numbers of animals is difficult.

Killing methods in research consist of multiple—often conflicting—objectives: ideally, death must be rapid and painless yet aesthetic and compatible with research protocols.⁴⁹ Furthermore, killing methods must be simple to apply and yield consistent results. It may be difficult, even impossible, to completely satisfy all of these objectives in all instances, more so given the challenges in identifying important outcomes (e.g., distress, pain, aesthetic appeal).

In North America, the AVMA and Canadian Council on Animal Care (CCAC) provide guidance on euthanasia procedures. Both organizations publish guidelines regularly, reflecting the latest evidence on killing methods in a wide range of species. In rodents, the AVMA classifies the intraperitoneal or intravenous injection of barbituric acid derivatives (or barbiturate combinations), as well as injection of dissociative agent combinations, as the only acceptable methods,⁵ and the CCAC classifies intraperitoneal injection of buffered and diluted barbiturate as well as overdose of inhalant anesthetics (followed by another method to ensure death) as acceptable.¹⁸

Although the above methods are categorized as acceptable, there exists a secondary category for methods termed ‘acceptable with conditions’ (AVMA)⁵ or ‘conditionally acceptable’ (CCAC).¹⁸ This category is similarly defined by AVMA and CCAC and includes techniques that “may require certain conditions to be met to consistently produce humane death” (AVMA), “might not consistently produce humane death” (CCAC), are not well described or documented in the scientific literature (AVMA and CCAC), may have increased safety hazards or potential for operator error (AVMA and CCAC), or “may require a secondary method to ensure death” (AVMA).^5,18 The AVMA guidelines further state that “methods acceptable with conditions are equivalent to acceptable methods when all criteria for application of a method can be met.”⁵ Methods classified in this way remain common because acceptable methods may interfere with the collection of research data or are impractical for large numbers of animals. For example, overdose with CO₂ is one of the most commonly used killing methods.²⁷ CO₂ is practical, easy to use, inexpensive, relatively fast in action, can be applied to multiple animals simultaneously, and requires little or no direct handling.^4,27,54 However, many studies have demonstrated that CO₂ exposure elicits an aversive response in rodents and may induce distress and pain.^{22,27,51,61,76-79}

Inhalant anesthetics, such as isoflurane, induce aversive behavior in rodents.^{21,61,68,69,99,110} In addition, their use requires specialized equipment, and time to death can be slow.¹⁸ Therefore, the CCAC guidelines rate inhalant anesthetics as acceptable in rodents but recommend use of a secondary method after the animal is anesthetized to ensure death, whereas euthanasia by using inhalant anesthetics is ‘acceptable with conditions’ in the AVMA guidelines. Regarding dissociative agents (typically ketamine) or a combination of ketamine and an α2 adrenergic receptor agonist, such as xylazine, there is limited discussion of this euthanasia method in the current AVMA guidelines,⁵ with only one supporting study cited in which the primary objective was not qualitative evaluation of dissociative agents for killing.¹⁰⁰

Overdose with barbiturate is the only method classified as acceptable by both the AVMA and the CCAC guidelines.^5,18 Sodium pentobarbital (pentobarbital) is the most commonly used barbiturate for killing. It has a narrow safety margin, is potent, can be formulated as a concentrated solution so that relatively small volumes are needed, and has a rapid onset of action when given intravenously.^5,25 Furthermore, pentobarbital has a long shelf life, is stable in solution, and is inexpensive.⁵ Access and licensing requirements for purchase vary between countries.²⁵ When given as an intentional overdose to cause death, general anesthesia is induced, followed by depression of the respiratory and cardiovascular centers of the brainstem, leading to cardiorespiratory arrest.^25,27

Although the intravenous route of injection is preferred for pentobarbital, it is often impractical to perform on rodents, given that achieving consistently successful intravenous injections requires training, restraint and often a means to induce venodilation.^39,43 As such, pentobarbital for killing rats and mice is frequently administered via intraperitoneal injection.^25,75,96 Although the onset of anesthetic effect is not as quick after intraperitoneal injection as after intravenous injection, drug absorption and distribution occur much more rapidly after intraperitoneal dosage than after intramuscular or subcutaneous injections.^75,83 Furthermore, intraperitoneal injection is relatively simple and quick to perform, allows the administration of large volumes, and accommodates repeated injections.

Thus, intraperitoneal injection of sodium pentobarbital (intraperitoneal pentobarbital) is one of the most widely accepted methods of rodent euthanasia that does not have additional conditions attached to its use. This rating is shared not only by the CCAC and AVMA but also by animal care guidelines in Europe,⁶ India⁵⁸ and Australia.⁹ Importantly, however, despite the many advantages and widespread use of intraperitoneal pentobarbital, this killing method is not without drawbacks. Here we aim to provide a narrative review of the literature regarding intraperitoneal pentobarbital as a killing method, evaluate its current designation as ‘acceptable’ (by the AVMA and CCAC), identify gaps in knowledge that may be pertinent to the refinement of this method, and suggest potential alternatives.

Materials and Methods

A literature search was performed by using combinations of 4 key words (euthanasia, rodents, pentobarbital, and intraperitoneal, creating 11 search terms) in 3 databases (PubMed, Cab Abstracts, Web of Science) in English or French from 1950 to the search date (15 May 2019). Document types were set as journal article, book, and government report. Search results were screened for inclusion by reading titles and abstracts. After initial screening, full articles were obtained for review. Data, results, and conclusions from these downloaded articles were then analyzed and reviewed for relevant content. Additional articles, books, theses, and government reports were identified and included after a manual search of the reference lists of articles identified during the database search.

Results

A total of 247 articles, government reports, or books passed initial screening and were read for the purpose of this review; 113 of the 247 publications were used as cited references in the writing of the review. The complete list of all postscreening articles is available in an online repository (https://doi.org/10.7910/DVN/DLSLDF).

Intraperitoneal injection technique.

Several techniques for intraperitoneal injection have been described, the most common variations are described here. Few studies specifically evaluate the outcomes of different injection techniques. Important factors to be considered—but seldom described—are the proficiency of the investigators, behavior of the animals (habituated compared with nonhabituated),^33,86 and inherent variability associated with drug absorption. A 2-person technique with one holder and one injector, rather than a one-person technique, is considered more efficient when working with mice, given that the 2-person method led to a reduction in misinjection rate from 11% to 13% to 1% to 2%.⁷ The holder grips the upper limbs and head of the animal and maintains the animal in a horizontal position. It is often recommended that the holder tilts the animal slightly downward so that its head is lower than its abdomen. This positioning supposedly creates more space in the caudal quadrants of the abdominal cavity by moving organs cranially.¹⁰³ However there is little evidence that the tilting of the animal has any effect on the success of the injection.⁷⁵ Indeed, similar misinjection rates have been obtained when injecting rats vertically²⁸ or horizontally with a head down tilt.^75,113 The injector, using one hand to hold the right hindlimb of the animal, visualizes the abdomen as though divided into 4 quadrants and injects into the caudal left quadrant (the animal's right side).^11,28,98,113 Within the caudal left quadrant, the injection should be made at the level of the coxofemoral joint, approximately halfway between midline and the lateral abdominal wall.⁹⁸ The goal of the injection is to deposit injectate into the peritoneal cavity without piercing any of the abdominal organs. To avoid doing so, the injection angle should be approximately 10 to 20 degrees relative to the body wall in mice and 20 to 45 degrees in rats, with the needle directed cranially.⁹⁸ Before injection, it is often suggested to aspirate the needle to assure its correct placement in the peritoneal cavity, although there is no evidence supporting the usefulness of this practice.¹³

Various sizes of needles and gauges can be used, but a 3/8-in. (9.5 mm) needle is sufficiently long and may be the least likely to puncture the organs in the abdominal cavity.^13,28 For a successful intraperitoneal injection, the needle does not need to exceed a depth of 4 to 5 mm beyond the skin, so a long needle is unnecessary. The maximal suggested injectable volume is 0.5 to 1.0 mL in mice and 5 to 10 mL in rats.^13,75

Pharmacokinetics.

Sodium pentobarbital is an oxybarbiturate ligand of the γ-aminobutyric acid subtype A (GABA_A) receptor. The drug increases chloride conductance through the receptor channel, causing neuronal hyperpolarization and consequent CNS depression. After successful administration of a lethal dose of pentobarbital, animals become ataxic and sedated, after which they experience loss of consciousness, apnea, cardiac arrest, and death. The aim of an intraperitoneal injection is to deposit pentobarbital into the peritoneal cavity, from where it is absorbed into the circulation. The peritoneal cavity is defined as the potential space between the visceral and parietal peritonea of the abdomen. The visceral peritoneum adheres to most of the abdominal organs, whereas the parietal peritoneum adheres to the interior of the abdominal wall. The blood vessels of the visceral peritoneum connect to the mesenteric, colic, and intestinal veins to converge into the anterior mesenteric vein and then carry blood to the liver via the portal system.¹⁰⁵ Thus intraperitoneal substances may undergo first-pass metabolism if absorbed via the visceral peritoneum.^3,89 After the liver, venous blood flows to the heart via the caudal vena cava and then on to the pulmonary circulation before returning to the left side of the heart for distribution to the systemic circulation.

Although intraperitoneal administration is considered a parenteral route, the pharmacokinetics of substances administered via intraperitoneal injections have similarities to substances administered orally because of the potential for hepatic metabolism.^74,96 However, the very large surface area provided by the peritonea and omentum, as well as the abundant blood supply, allow for absorption that is more rapid than after oral or intramuscular routes.^75,83 Absorption across the visceral peritoneum and omentum is the predominant—but not the only—route of absorption after intraperitoneal injection. Other routes to the systemic circulation include absorption via the parietal peritoneum and lymphatic drainage. These different absorption routes contribute variability to the pharmacokinetics and pharmacodynamics of intraperitoneal injections.²³ Vessels draining the parietal peritoneum do not connect to the portal veins but instead empty directly into systemic veins,⁵⁹ thereby providing a route to bypass first-pass metabolism.

In addition to peritoneal absorption, lymphatic absorption across the diaphragm can affect the fate of substances delivered intraperitoneally and further contribute to variability in intraperitoneal drug absorption and distribution.¹ Access to the lymphatic system occurs through stomatas (small openings on the surface of the peritoneal mesothelium) that allow the passage of fluids from the abdominal cavity into subperitoneal lacunae, which drain directly into the lymphatic system.⁹⁵ These stomatas are located on the muscular portion of the diaphragm in rats.

Lymphatic drainage can be rapid and effective in delivering substances from the peritoneal space to the systemic circulation.^62,91 Indeed, within minutes, large amounts of drained fluid from the abdominal cavity are transported to the subclavian vein, which feeds directly into the cranial vena cava.⁹¹ However the speed of lymphatic drainage varies and is affected by the stretching of the diaphragm during respiration.^14,102 Furthermore, the posture of the animal affects the rate of drainage.¹² Indeed, a slower rate of absorption was measured when rats were held vertically (head up) compared with a head-down position, whereas a sternal posture produced intermediate absorption.

Although lymphatic drainage of the peritoneal cavity has been widely studied in multiple species, this system is often undescribed in discussions of intraperitoneal injection of pentobarbital.^28,75,98,109 These studies tend to focus entirely on peritoneal absorption, thereby omitting what may be a major component of intraperitoneal injections. Many authors have suggested that lymphatic drainage is more important quantitatively than absorption from either the peritoneal or visceral peritoneums.^{1,15,29,42,73} Compelling evidence demonstrating the relative importance of lymphatic drainage is the observation that a 59% decrease in absorptive capacity from the peritoneal space resulted after closure of the diaphragmatic stomata owing to fibrous tissue (induced by abrasion).⁶⁴ In addition, ligation of the principle lymphatic ducts draining the lymphatic vessels emanating from the diaphragm markedly reduced the amount of dye (delivered via intraperitoneal injection) absorbed from the peritoneal cavity.^30,31

These varied routes of absorption alter the pharmacokinetics of substances delivered through intraperitoneal injection by providing different means of attaining the systemic circulation. These undoubtedly play a role in affecting the variability in timing of physiologic responses after successful intraperitoneal injections.^23,4,113 Table 1, which presents data from various euthanasia studies, illustrates this variability. Successful intraperitoneal injections are those that administer injectate completely into the peritoneal cavity, as compared with misinjections, which fail to do so. Misinjections are a further source of variability, as described later. Importantly, unless behavioral outcome measures (for example, loss of righting reflex) are accompanied by confirmation of a successful injection through necropsy examination, it is impossible to differentiate variability resulting from partial misinjection (some, but not all, injectate failing to be deposited in the peritoneal cavity) compared with inherent variability associated with absorption routes and rates. In addition, different operational definitions of measured endpoints and methodologies for their determination are apparent and contribute to variability (Table 1).

Table 1.

Timing of the various effects during intraperitoneal sodium pentobarbital euthanasia at various doses in rats and mice.

	Species	Dose (mg pentobarbital/kg body weight)	Effects	Time (s) (mean [range] or mean ± 1 SD)
Reference 2	Mice	5400a	Loss of righting reflex	156
	Rats	200a	Respiratory arrest	45
		Cardiac arrest		276 ± 30
		Loss of righting reflex		75
		Respiratory arrest		270
Reference 4	Mice	330b	Ataxia	52.0 ± 8.1
			Cessation of movement	80.1 ± 21.2
			Cessation of breathing	317.0 ± 150.5
Reference 16	Mice	150a	Ataxia	35.6 ± 11.0
			Death	343.3 ± 110.3
Reference 22	Rats	200	Recumbency	174.6 ± 125.4
			Loss of righting reflex	272.1 ± 204.8
			Quiescent EMG	259 ± 201
Reference 36	Mice	250c	Loss of righting reflex	98.5 ± 4.7
		250d	Cessation of heartbeat	619.6 ± 359.6
			Loss of righting reflex	74.3 ± 4.9
		1300–1680e	Cessation of heartbeat	508.3 ± 303.4
			Loss of righting reflex	66.4 ± 4.5
			Cessation of heartbeat	253.8 ± 118.1
Reference 54	Rats	150	Induction of unconsciousness	152 (105–195)
	Mice	150	Time to respiratory arrest	676 (510–815)
			Induction of unconsciousness	80 (45–120)
			Time to respiratory arrest	482 (315-720)
Reference 88	Mice	100	Death	235
Reference 101	Rats	667	Ataxia	40.6 ± 4.4
			Cessation of movement	63.0 ± 8.2
			Loss of pedal reflex	175.00 ± 6.52
			Heart rate < 150 bpm	444
Reference 113	Rats	200	Loss of righting reflex	111.6 ± 19.7
		200f	Cessation of heartbeat	485.8 ± 140.7
		800	Loss of righting reflex	104.2 ± 19.3
		Cessation of heartbeat	347.7 ± 72.0
		Loss of Righting reflex	139.5 ± 29.6
		Cessation of heartbeat	283.7 ± 38

^aPentobarbital–phenytoin combination product (390 mg/mL sodium pentobarbital and 50 mg/mL sodium phenytoin)

^b50:50 mixture of 100 mg/mL sodium pentobarbital with 10 mg/mL of lidocaine hydrochloride

^cDiluted to 5 mg/mL with USP sterile water from a 50mg/mL solution

^d50 mg/mL

^e390 mg/mL with no phenytoin added

^fDiluted 1:3 with PBS

Although many different doses, volumes, and concentrations of pentobarbital have been used for euthanasia, a dose of 800 mg/kg IP in rats is associated with greater consistency and speed of effect than 200 mg/kg.¹¹³ Doses in the range of 150 to 200 mg/kg IP are commonly used and are based on being approximately 5 times the dose required to induce general anesthesia.^25,43 Furthermore, at the 800-mg/kg dose, the decreased variability of effect facilitates identifying misinjections based on the time taken to achieve loss of righting reflex and apnea. At this dose, it has been suggested that the time to loss of righting reflex and apnea should not exceed approximately 2.5 to 4.5 min, respectively (calculated from mean + 2 SD of study population).¹¹³ If these times are exceeded, misinjection is likely to have occurred.

Little information is available regarding optimal dose and volume for euthanasia with pentobarbital in mice, although 150 mg/kg IP has long been a suggested dose.⁴³ The anesthetic dose for pentobarbital is 40 to 50 mg/kg when given intraperitoneally,^13,39 so by applying the same principle as described earlier, the minimum dose to cause death should be 200 to 250 mg/kg IP, but doses for mice described in the literature range from 150 to 5400 mg/kg IP (Table 1). A recent study found that increasing the concentration of the pentobarbital solution to 390 mg/mL, which resulted in a dose greater than 1300 mg/kg for males and 1680 mg/kg for females, significantly shortened the time to unconsciousness and death compared with a dose of 250 mg/kg (concentration of 50 mg/mL).³⁶ Therefore, similar to what was observed in rats, increasing the dose of pentobarbital creates a more efficacious killing method when using intraperitoneal pentobarbital in mice.

The extent of current scientific knowledge regarding the pharmacokinetics of intraperitoneal pentobarbital is still limited in many areas.^94,96 Indeed, some of the basic principles of this route of administration, such as absorption from the peritoneal cavity, have yet to be fully elucidated. Determining the effect of varying the placement of an injectate within the abdominal cavity on absorption rates and quantifying the various absorption routes could lead to greater consistency and predictability as well as improve our ability to detect misinjections.

Disadvantages of intraperitoneal injections of sodium pentobarbital.

Variability in effect.

A major disadvantage of intraperitoneal injections—and arguably the most significant—is its variability. This variability can be divided into 2 categories: inherent variability and misinjection. Inherent variability refers to the existence of different pathways of absorption and distribution, as described earlier, which tend to produce a wide range of responses after successful intraperitoneal injection and thus create variability in euthanasia procedures.

The second source of variability is misinjection, or incorrect placement of administered substances. Misinjection has long been described as an issue with intraperitoneal injections and occurs when the injection fails to introduce the drug into the peritoneal cavity.^63,93 The principal consequence of this failure is a marked delay in the onset of drug action. Indeed, misinjection caused an increase in the time from intraperitoneal pentobarbital injection (667 mg/kg IP) to loss of the pedal reflex in rats from 175 s (successful injection) to 588 s (misinjection).¹⁰¹ In one study, approximately 40% of misinjections were classified as failed euthanasia procedures because time to death exceeded 20 min.¹¹³ Furthermore, there is the possibility of complications arising out of misinjection, which may include local irritation and inflammation, perforation of abdominal organs, hemorrhage, and respiratory distress.⁹⁶

Misinjection can be difficult to identify given the inherent variability of successful intraperitoneal injections. Aspirating before injecting is often suggested, but cecal and intestinal content may not easily be aspirated through a small-gauge needle.⁴ Therefore, misinjection can only be quantified with any confidence at necropsy, through the addition of dye to the injectate.

The most common sites of intraperitoneal misinjection in rats, ranked by frequency of occurrence, are into the cecum, into the small intestine, subcutaneously, retroperitoneally, and into the urinary bladder.⁵⁹ In mice, the most frequent locations of misinjections are: stomach, small intestine, uterine horn, and subcutaneous.^72,93 Reported rates of misinjection based on necropsy findings are variable, ranging from 6% to 20% in rats and 10% to 20% in mice.^{2,4,11,23,28,63,87,93,103,113} Although the application of appropriate technique and training can reduce these rates,^7,23,72 the documented incidence of misinjection rarely falls below 6%, particularly in rats. One study did achieve a reduction in misinjections in mice from 11% to 13% to 1% to 2% (both values obtained after injecting 250 mice).⁷ This reduction occurred after the injections were performed by using a 2-person technique rather than by a single person.

Two studies examined the position of the cecum in rats and mice and found that it was predominantly in the left caudal abdominal quadrant.^28,98 This finding is consistent with the practice of giving injections in the right caudal quadrant to avoid penetrating the cecum. However, the lateral dominance of the organ never exceeded 80%. Moreover, the percentage varied among strains, sexes, and even colonies. So the cecum, which is the main site of intraperitoneal misinjections in rats, lies in the middle or on the right (directly in the target of an intraperitoneal injection) in as many as 20% of cases. This unpredictability appears unavoidable and is an important factor contributing to the high misinjection rate in rats. Cecal position may be less variable in mice, and the organ is indeed less frequently involved during misinjections.^72,93,98

The peritoneal space is best understood as being a potential space rather than actual space.⁴ Therefore, the shifting and changeable position of the abdominal organs, such as the cecum, within this space makes it impossible to completely avoid unsuccessful injections. This drawback becomes a concern especially when considering the number of animals killed via this technique each year on a global scale. Misinjection prolongs the time to loss of consciousness, and injection of pentobarbital into the peritoneal space may induce pain (discussed later); therefore, it is reasonable to consider the welfare implications of intraperitoneal pentobarbital and to develop methods to refine the method.

Histopathologic and physiologic changes.

Pentobarbital can damage local tissue that it contacts;^8,31 significant damage to superficial cells of the organs near or at the injection site have been observed,⁴⁵ thus suggesting the injectate was causing this damage. The same study⁴⁵ also noted splenomegaly after intraperitoneal pentobarbital, most likely caused by relaxation of smooth muscle, causing splenic engorgement with blood. Other histologic changes affecting organs farther from the injection site may include focal congestion of intestinal serosa, congestion in pulmonary veins, necrosis in subcapsular levels of the liver and pancreas, lung emphysema and edema, and hyperemic kidneys (see reference 7 for review).

In light of these physiologic disruptions, it is appropriate to consider alternative euthanasia options where blood or tissue samples are required for a research protocol. Alternative methods should be justified, weighing the balance between research outcomes and the risk of failing to achieve a humane death.

Pain associated with intraperitoneal pentobarbital.

The act of intraperitoneal injection, regardless of the chemical agent used, can induce distress and pain. Behaviors associated with pain, including vocalization, increased locomotion, and flinching, have been observed immediately after intraperitoneal injections.^2,36,104 In addition, pentobarbital is highly alkaline, with a pH of 11 to 12, whereas the range of pH that is said to be nonirritating to local tissue at the site of intraperitoneal injections is approximately 4.5 to 8.0 in rats.^75,101 It follows that intraperitoneal injection of pentobarbital may result in irritation to the peritoneum or surfaces of visceral organs, as well as pain. Indeed, one study¹⁰¹ noted signs of local redness and swelling after intraperitoneal pentobarbital at a dose of 667 mg/kg with no additives in the solution. In addition, writhing behavior was observed, beginning approximately 11 ± 2 s after injection. Writhing was defined as “an abnormal posture in which the rats contract their abdomen and extend their hindlegs backward.”¹⁰¹ Writhing has been reported in both rats and mice after intraperitoneal injection,^2,4,57,113 and it is recognized as a behavioral response to abdominal pain, given that it is commonly observed after the injection of a known irritant into the abdominal cavity⁹² and after abdominal surgeries, such as laparotomy and vasectomy.^85,111 The reported incidence of writhing after intraperitoneal pentobarbital varies widely, from approximately 30% to 100%.^4,101,113 The reason for this discrepancy is unclear but may stem from difficulties in assessing motor behaviors as pentobarbital induces general anesthesia.

Many of these data, however, should be interpreted cautiously given the difficulty in accurately quantifying behavioral responses dependent on gross movement in the presence of an agent that depresses motor function. Some authors have attempted to circumvent this issue through means of indirect quantification of nociception. One study used the presence of electrical brain activity in anesthetized piglets to infer the presence of nociceptive input after intraperitoneal pentobarbital injection.⁵⁶ Another injected pentobarbital into a hindpaw of mice and quantified the response to various pentobarbital concentrations by using a paw-lick test.³⁶

Additional evidence for the occurrence of pain after intraperitoneal injections can be obtained via the study of the neuronal markers c-fos and Fos. c-fos is an early-response protooncogene that is rapidly activated and expressed in specific nociceptive neurons of the dorsal horn after noxious or sensory stimulation.^26,48,52 The c-fos gene encodes for the protein Fos, which acts as an intermediary between extracellular events and long-term intracellular adaptations. Although much is still unknown about the physiologic roles of c-fos and Fos, including during nociceptive processes, these markers are nonetheless commonly used in research to measure the activity of nociceptive neurons.⁴⁸ Indeed, immunohistochemical staining is a practical and reliable way to detect Fos activity, therefore leading to the use of FLI (fos-like immunoreactive) neurons in research.⁴¹

The quantification of FLI neurons has thus been used to infer pain after intraperitoneal pentobarbital. A study of FLI neurons has shown an increase in neuronal activity after intraperitoneal pentobarbital administration in areas of the dorsal horn related to visceral nociception.⁹⁴ A 4-fold increase in the number of FLI neurons was observed in rats administered intraperitoneal pentobarbital compared with rats receiving saline intraperitoneal injections, although the saline-treated group also showed increased FLI neurons when compared with basal levels. Furthermore, the addition of lidocaine (10 mg/mL) to intraperitoneal pentobarbital lowered the number of FLI neurons in the spinal cord compared with intraperitoneal pentobarbital administered alone. The FLI neurons were present in laminae I, II, V, and X of the spinal dorsal horn; these laminae receive visceral nociceptive input.^46,71

The relationship between FLI neurons and writhing behavior provides further evidence that intraperitoneal pentobarbital produces pain. There exists a positive correlation between the number of FLI neurons in the laminae that receive visceral nociceptive input and the number of writhing behaviors, or stretches, observed after the intraperitoneal administration of acetic acid, a known irritant.⁴⁶ The administration of analgesics (including morphine) produced dose-dependent inhibition of the writhing behavior and concurrently a reduction in the number of FLI neurons.

However, much debate remains surrounding the use of FLI neurons in pain studies. The relationship between FLI neuron activity and the conscious, cortical perception of pain is unclear. FLI neurons did not significantly increase in the ventroposterolateral nucleus of the thalamus, a region implicated in the perception of pain,^108,109 after a noxious stimulus.³ In addition, rats treated with morphine (as high as 10 mg/kg SC) still showed Fos expression in the dorsal horn yet displayed few to no behavioral signs of pain, although Fos activity did diminish with increasing doses of morphine (from 1 to 10 mg/kg SC).⁸¹ Therefore, the quantification of FLI neurons to infer pain is probably oversimplistic. Moreover, stress is a confounding factor, given that it has been reported to increase FLI neuron activity.⁸⁰ Given the lack of specificity, converging lines of evidence (pain-related behaviors, biomarkers, gross and histologic tissue changes) should be used to draw inferences about the likely presence of pain.

The AVMA and CCAC euthanasia guidelines acknowledge that intraperitoneal pentobarbital may cause pain.^5,10,18,23 The CCAC guidelines suggest the concurrent use of a local anesthetic, such as lidocaine, with a buffered and diluted barbiturate and recommend that steps should be taken to ensure that the pH is within a nonirritating range.^18,23 However, the pH of pentobarbital cannot be lowered below 10 without risk of precipitation, and a pH of 10 remains irritating to tissue.^31,75 The addition of lidocaine to pentobarbital (in a 50:50 mixture) reduces pain, as measured by writhing behavior.⁴ In addition, one study noted a decrease (but not absence) in writhing when either lidocaine (10 mg/mL of pentobarbital solution) or bupivacaine (2.5 mg/mL of pentobarbital solution) was added to pentobarbital (pH 10.1 and 10.2, respectively).⁵⁷ This outcome suggests that the addition of lidocaine represents a refinement of the intraperitoneal pentobarbital technique. Although the lidocaine–pentobarbital mixture was beneficial, it still significantly increased the number of FLI neurons 3-fold compared with a saline control injection.⁹⁴

The potential for pain during intraperitoneal pentobarbital is compounded by the time to achieve loss of consciousness with this killing method, especially in rats. In addition to implications for tissue harvesting and sample quality, this is an important concern for animal welfare. Development of a technique that induces a consistently shorter time to unconsciousness and death is desirable.

In light of the limited available evidence, various authors and the AVMA guidelines both raise the need for additional studies.^5,94 Different local anesthetic agents could be explored, as could alternative markers of nociception. Alternative markers, such as phosphorylated extracellular signal-regulated kinases, may offer an advantage in that they are more rapidly expressed than c-fos, reaching a peak 2 to 3 min after stimuli^41,55 (c-fos induction and expression in the spinal cord take at least 30 min).³⁵ Indeed, phosphorylated extracellular signal-regulated kinases could be used to establish a timeline of potential nociception and pain after intraperitoneal pentobarbital injection.

Stress.

The stressful nature of intraperitoneal injections has been inferred via the measurement of hormonal markers. In one study, ACTH levels increased as much as 2-fold compared with basal levels in Sprague–Dawley rats after intraperitoneal injections of saline.³³ However, the concentrations of ACTH measured varied substantially. A similar increase was observed in mice.¹⁰ In addition, plasma corticosterone levels in both mice and rats can be increased through intraperitoneal injections of saline. Two studies in mice reported significant increases in corticosterone after intraperitoneal injection of saline.^10,86 In rats, results have varied: one study reported an increased level of corticosterone in Sprague–Dawley rats after intraperitoneal saline injections, whereas another reported no such increase.^33,112

Intraperitoneal injection of saline has been associated with hyperthermia and tachycardia in both rats^34,47,90 and mice.^24,60 Another study found that intraperitoneal injection of saline or sham injection (needle insertion with no fluid administered), increased the heart rate of mice over basal levels for up to 30 min following injection.⁷⁰ The authors also showed that changes in heart rate parallel those of plasma corticosterone and may, therefore, be another useful indicator of stress.⁷¹

Stress related to laboratory procedures varies significantly between strains of laboratory rodents. One study reported plasma levels of corticosterone increasing from 3 to 7 fold across different strains of mice.⁸⁶ If ACTH physiology is similar among rat strains, Lewis rats may be less susceptible to stress; this strain had a minimal increase in plasma ACTH concentrations after handling or injection, compared with Sprague–Dawley rats.³³

Thus, there is evidence that the handling necessary for intraperitoneal injections can be stressful. Handling stress can be reduced. Repeated exposure—i.e., habituation—can reduce the stress response associated with laboratory procedures such as intraperitoneal injections. After habituation to saline intraperitoneal injections, there was a reduced expression of immediate early response genes and corticosterone levels.^33,86

Characterizing the relationship between stress and distress cannot be made on hormonal markers alone. Other markers, such as behavioral observations, should be assessed in parallel. Behavioral assessments of pain and stress are both challenging because they lack consistency and tend to be fairly subjective.¹⁰⁶ However, although quantifying the effects of intraperitoneal injections in laboratory rodents is difficult, establishing that there is an effect is less so: in the majority of studies, intraperitoneal injections are identified as a potential source of pain and stress.

Alternatives to intraperitoneal pentobarbital.

There are few, if any, well-described alternatives to pentobarbital as an injectable killing method for rats and mice. One chemical agent that has shown promising results is ethanol. In many countries, government drug regulations require accounting of barbiturate drugs,⁴ which makes ethanol an attractive alternative. Ethanol is currently described in the AVMA guidelines as ‘acceptable with conditions,’ owing to a series of studies that explored its use as a killing method.^65-67 Interest in this agent has recently resurged, and it has been determined that 100% ethanol at a dose of 15.3 to 15.8 g/kg results in similar rates of onset of respiratory and cardiac arrest as a similar volume (approximately 0.5 mL) of intraperitoneal pentobarbital in mice.^2,32 The injection of ethanol induced pain-related behaviors such as vocalization² and kicking at the needle.³² In both studies, these behaviors were not more frequent than those seen with intraperitoneal pentobarbital, and vocalization is not a specific indicator of pain in mice.^106,107 Therefore, ethanol appears at least similarly effective to pentobarbital, but with the advantage of not being a controlled agent. A study of the number of FLI neurons after injection could be used to quantify neuronal activity associated with ethanol injections compared with intraperitoneal pentobarbital injections. In contrast to that in mice, intraperitoneal injection of ethanol in rats as a killing method was unsuccessful due to the large volumes of ethanol required (more than 7.1 mL for a dose of 20.1 g/kg) and because the time from injection to respiratory arrest was slow (8 ± 5 min).²

Therefore, pentobarbital remains the only practical injectable killing agent in rats. Alternative injectable techniques that show potential include retroorbital and intrahepatic injection. In mice, retroorbital injection of an overdose of ketamine and xylazine resulted in rapid death (cessation of heartbeat occurring in approximately 5 s), because the pharmacokinetics of retroorbital injection closely resemble those of intravenous injection.^50,82,88 However, neither the risks and prevalence of misinjection with this route of administration nor the potential for pain have been studied. This technique has not been evaluated in rats.

Another potential route of pentobarbital administration is intrahepatic. One article has reported use of this injection technique for euthanasia of shelter cats.⁴⁴ The technique proved more accurate and faster than intraperitoneal injections, given that most cats became recumbent almost immediately after injection of pentobarbital. A few cats did respond negatively to the injection; a negative response was defined as vocalization or turning toward the injection site. However, these behaviors occurred with similar frequency during intrahepatic and intramuscular injections. The feasibility and consistency of the intrahepatic injection technique remain to be evaluated in laboratory rodents.

These alternative methods of drug administration require further study to facilitate comparison with intraperitoneal injection, particularly of inherent variability in effects, likelihood of misinjection and speed of action. As with all proposed killing methods, the criteria listed in the AVMA guidelines⁵ should be applied to evaluate their suitability.

Best-practice recommendations.

The following recommendations are based on the available literature and are intended as suggestions to promote best practice.

• 1. Within an institution, establish a standard dose, volume, and formulation of sodium pentobarbital. Standardized dosing will allow for derivation of reference ranges for key outcomes, such as loss of righting reflex, respiratory arrest, and cardiac arrest.¹¹³ If dose, volume, and formulation match those in the literature (Table 1), published reference ranges can be used as a benchmark for performance, provided that the methods of identifying key outcomes is the same. Establishing reference ranges also allows auditing (see point 3). A backup plan should be available for when key outcomes are not achieved within the expected time. This plan may involve redosing or the use of a secondary killing method. When a standard drug solution volume is used for a range of animal weights, the lower and upper limits of the acceptable body mass for each injectate volume should be available.

• 2. Ensure that users are proficient in the steps required to perform intraperitoneal injection. These include drug dose calculation and preparation, handling and restraint, injection, and identification of signs of general anesthesia and death. Technical skills follow a learning curve, and learners achieve proficiency at different rates.^17,20,58 Should time to achieve key outcomes after intraperitoneal injection of pentobarbital deviate substantially from reference ranges, investigating the source of this deviation should include examining the steps involved in administering intraperitoneal injections.

• 3. Consider auditing key outcomes as a means of tracking consistent practice and facilitating early identification of deviation from best practice.⁸⁴ For example, times to loss of righting reflex, apnea, and cardiac arrest, and rate of failure to achieve key outcomes all merit tracking.

Conclusion

In light of the data cited in this review, it is legitimate to question whether intraperitoneal pentobarbital as a killing method always meets the criterion of euthanasia. According to the AVMA guidelines, euthanasia procedures should be consistent, easy to perform, reliable, and predictable. Despite the relative simplicity and widespread use of intraperitoneal pentobarbital, there is a distinct possibility that this methodology causes distress and pain. The likelihood of misinjection only serves to exacerbate these problems. Finally, many important gaps exist in scientific knowledge related to this procedure. The following data remain to be established:

• 1. The optimal dose of pentobarbital in mice and predictability of misinjection.

• 2. Criteria for early identification of misinjection and remedial procedures to minimize the potential for pain.

• 3. The optimal dilution of pentobarbital that minimizes pain associated with its alkali pH yet maintains efficacy.

• 4. The refinement of the addition of lidocaine (dose and volume) to pentobarbital.

• 5. The potential role of other local anesthetics for mixing with pentobarbital.

• 6. The elucidation of the pharmacokinetic parameters after intraperitoneal administration, most notably the relationship between injection site and peritoneal absorption.

• 7. The potential for alternative injectable routes of administration to replace intraperitoneal injection.

• 8. The role of training in minimizing misinjection.

Clearly, numerous aspects of intraperitoneal pentobarbital give cause for concern. Novel approaches, such as the use of intraperitoneal ethanol and alternative routes of injection, are promising but require further research to establish their strengths and weaknesses before they can be proposed as suitable alternatives for, or improvements upon, intraperitoneal pentobarbital. The important limitations described for intraperitoneal pentobarbital, current lack of suitable alternatives, and the large number of animals killed underline the importance of further research in this field.