Susceptibility of Rats to Corneal Lesions After Injectable Anesthesia

Patricia V. Turner, DVM, DVSc1,* and Mudher A. Albassam, DVM, PhD2

Corneal injury is not a commonly reported side effect after injectable or inhalation anesthesia in rats, but a number of surgery studies at our facility resulted in a high incidence of these injuries. To explore the potential association of various anesthetic protocols with the development of corneal lesions in rats, we retrospectively evaluated clinical records and sections of eyes from 215 male and 187 female Wistar rats used in eight intravenous infusion toxicology studies. None of the studied compounds was associated with eye toxicity. For placement of jugular vein vascular access ports, rats were anesthetized with enflurane, isoflurane, ketamine–xylazine, or Hypnorm–midazolam. Histo- logically, corneal changes were scored from 0 to 4 in light of degree of mineralization, leukocytic infiltrates, neovascularization, fibrosis, and ulceration. Prestudy (postsurgical) ophthalmic examination findings of corneal opacities were correlated with mineralization of the anterior limiting membrane and corneal ulceration. Corneal lesions were more severe in animals anesthetized with ketamine–xylazine, and minimal changes occurred after anesthesia with either enflurane or isoflurane. The results of further analysis suggest that corneal lesions can be observed within 24 h after injectable anesthetic administration and are not reversible. The severity of corneal changes was reduced when ketamine–xylazine anesthesia was reversed with yohimbine. Compared with Sprague-Dawley and Lewis rats, Wistar, Long-Evans, and Fischer 344 rats had increased incidence and severity of corneal lesions after anesthesia with ketamine–xylazine, suggesting that these three strains are at increased risk for developing postanesthetic corneal lesions with this regimen.

Provision of appropriate anesthesia and analgesia to labora- corneal lesions (for example, ulcerative keratitis, corneal dystro- tory animals undergoing invasive procedures is a necessary part phy, neovascularization, and mineralization) despite adequate of appropriate veterinary care. Corneal injury is not a commonly intraoperative corneal lubrication. Anesthetics used for instru- reported side effect after general anesthesia induced by either mentation studies have included ketamine–xylazine, pentobar- inhalational or injectable agents in the rat or the administration bital, Hypnorm (fentanyl-fluanisone)–midazolam, enflurane, of other substances. Corneal opacities have been reported in rats and isoflurane. Not only do these lesions contribute to animal after administration of acute or chronic doses of narcotic analge- discomfort and nonspecific clinical signs of distress, but also they sics (morphine, 1α-acetyl methadol) (8, 22), antidepressants (30), may potentially confound toxicological interpretation of inves- and capsaicin (25). In one study, keratoconjuncitivitis sicca with tigational drug effects on the eye. In an attempt to refine our secondary uveitis was reported in Lewis rats after ketamine–xy- perioperative technique, a retrospective study was conducted to lazine anesthesia and was not seen when either drug was given examine whether any correlation existed between the incidence alone (18). Similarly, corneal mineralization with neutrophilic of corneal lesion development and the type of anesthetic used for infiltrates and neovascularization was reported in 30 to 54% of surgical instrumentation. purchased shipments of hypophysectomized Sprague-Dawley Because historical clinical records suggested that corneal le- rats anesthetized with ketamine combinations. The pathogenesis sions developed shortly after surgery and prior to study initia- of the lesions in this study was interpreted to be related to lack tion, a time-course study of lesion development after anesthesia of use of intraoperative eye lubricants (17). The pathogenesis was conducted in Wistar rats. As part of this study, we also exam- of these lesions has not been further explored, although both ined whether mechanical abrasion of the cornea on cotton tether- narcotic analgesics and injectable combinations of ketamine–xy- ing jackets fitted postoperatively contributed to the pathogenesis lazine are used frequently for anesthesia and postoperative care of the corneal lesions. of rats and other laboratory animal species. In addition, age-related spontaneous corneal degeneration Recent in-house studies in which Wistar rats were anesthe- has been reported more commonly for some stocks and strains tized with injectable anesthetic agents for instrumentation with of rats than others and may be a predisposing factor for lesion jugular vein vascular access ports have resulted in postoperative development. In two reports, spontaneous corneal dystrophies re- sembling calcific band keratopathy in humans were observed in one or both eyes of aged male and female rats with the following Received: 11/30/04. Revision requested: 12/8/04. Accepted: 12/17/04. 1Pfizer Global Research and Development, Sheridan Park Laboratories, Missis- stock- or strain-associated incidences: 67 to 80% of Sprague-Daw- sauga, Ontario, Canada L5K 1B4; 2Pfizer La Jolla Laboratories, Safety Sciences, ley rats, 16 to 48% of Wistar rats, 10 to 15% of Fischer 344 rats, San Diego, California 92121. *Corresponding author: Department of Pathobiology, University of Guelph, and 0% of Lewis rats (1, 19). Other studies have suggested that Guelph, ON N1G 2W1. the incidence of spontaneous corneal dystrophy is decreased in

175 Vol 55, No 2 Comparative Medicine April 2005

Table 1. Anesthetic treatments for retrospective analysis of postanesthetic Table 2. Experimental groups and treatments for time-course study of post- development of corneal lesions in male and female Wistar rats anesthetic corneal lesion development in female Wistar rats No. of animals Group No. of animals Anesthetic agent(s)a Study Male Female Anesthetic agent(s)a Ophthalmic exam? 1 9 O2 control (without jackets)

1 21 35 enﬂurane in O2 no 2 4 O2 control (with jackets) 2 18 20 ketamine–xylazine no 3 9 ketamine–xylazine 3 14 0 ketamine–xylazine no 4 9 pentobarbital 4 40 37 ketamine–xylazine yes 5 9 Hypnorm–midazolam

0 3 isoflurane in O2 yes 6 9 isoflurane (2%) in O2 5 40 40 ketamine–xylazine yes aAnesthetic drug dosages were as follows: ketamine, 60 mg/kg; xylazine, 12 mg/kg; 6 12 12 ketamine–xylazine no pentobarbital, 30 to 40 mg/kg; Hypnorm, 0.67 ml/kg; and midazolam, 3.4 mg/kg. 7 0 36 ketamine–xylazine yes Three animals from each group were euthanatized at 24 h, 4 days, or 9 days 40 4 Hypnorm–midazolam yes postanesthesia. Animals in Group 2 were euthanatized on Day 9. 8 30 0 Hypnorm–midazolam no aAnesthetic drug dosages were as follows: ketamine, 60 mg/kg; xylazine, 12 mg/kg; Hypnorm, 0.67 ml/kg; and midazolam, 3.4 mg/kg. ophthalmic ointment (chloromycetin ophthalmic ointment 1%, Pfizer, Groton, Conn.) was instilled into both eyes at induction.

The following protocols were used for anesthesia: enflurane in O2 Sprague-Dawley rats, with only 10% of rats older than 9 months (Enflurane, Abbott Laboratories, Toronto, Ontario, Canada) and affected (27). To address the issue of stock or strain predisposi- isoflurane in O2 (Aerrane, Janssen, Toronto, Ontario, Canada) tion to corneal lesion development, a final study was conducted by facemask; ketamine (Ketalar, Pfizer) with xylazine (Rompun, to determine whether the genetic background of male rats influ- Bayer, Toronto, Ontario, Canada) intramuscularly (i.m.); and enced the incidence of corneal lesions after anesthesia. Hypnorm (fentanyl–fluanisone, Janssen, Beerse, Belgium) with midazolam (Versed, Roche, Nutley, N.J. ) intraperitoneally (i.p.). Materials and Methods Buprenorphine (0.05 mg/kg s.c.; Buprenex, Reckitt and Colman, Animals. Barrier-raised Wistar (WI; Crl:(WI)BR), Sprague- Richmond, Va.) or butorphanol (2 mg/kg i.m.; Torbugesic, Fort Dawley (SD; Crl:CD[SD]IGS BR), Long-Evans (LE; Crl:[LE] Dodge, Fort Dodge, Iowa) was administered for postoperative BR), Fischer 344 (F344; F-344/CrlCrBR), and Lewis (LEW; LEW/ analgesia. The infusion studies were 4 to 28 days in duration. Ssn/Cr/CrlBR) rats (Charles River Canada, St. Constant, Que- The duration of time after animal instrumentation to euthanasia bec, Canada) weighing approximately 150 g (37 to 65 days old) for these studies ranged from 7 to 111 days. Prestudy (postsur- were housed individually in suspended stainless-steel cages, on gical) and terminal ophthalmic exams were conducted on ani- a 12:12-h light:dark cycle, and at constant temperature (22 ± mals in studies 4, 5, and 7 (Table 1). One drop of 1% tropicamide 4°C) and relative humidity (30 to 70%). Rats were given free ac- (Mydriacyl, Alcon Labs, Fort Worth, Tex.) was placed in each eye cess to food (Certified Rodent Chow 5002, PMI Feeds, Oakville, approximately 5 min prior to the indirect ophthalmic examina- Ontario, Canada) and water at all times. Vendor surveillance tion. The approximate size and shape of opacity was recorded. reports indicated that animals were free of Mycoplasma pul- At the conclusion of each study, rats were euthanized by carbon monis, cilia-associated respiratory bacillus, Salmonella spp., dioxide inhalation. Streptococcus moniliformis, Brucella bronchiseptica, Corynebac- Influence of anesthetic agent and time on corneal lesion terium kutscheri, Staphylococcus aureus, Klebisiella pneumoniae, development. We randomly assigned 49 female Wistar rats to Pseudomonas aeruginosa, Pasteurella spp., and common murine five experimental groups (Table 2). Four groups each received viruses including Sendai virus, Toolan’s H-1 virus, rat coronavi- one of the following anesthetic combinations: ketamine–xylazine rus/sialodacryoadenitis virus, reovirus type 3, Kilham rat virus, i.p.; pentobarbital (pentobarbital sodium, Ganes Chemical Co, pneumonia virus of mice, mouse adenovirus, and lymphocytic Pennsville, N.J.) i.p.; Hypnorm–midazolam i.p., and isoflurane choriomeningitis virus. In addition, animals were test negative in O2 by facemask. The remaining two groups served as controls, for fur mites and pinworms. Animals were acclimated 6 to 7 days either without or with jackets (Lomir Biomedical, Notre Dame prior to anesthesia and surgery. The facilities and procedures in- de I’lle Perrot, Quebec City, Canada) typically used with tether– volving animals are in compliance with the Animals for Research swivel assemblies. All animals received 0.05 mg/kg s.c. atropine Act of Ontario and the Guidelines of the Canadian Council on and 0.05 mg/kg s.c. buprenorphine as premedicants, except the Animal Care. The institutional animal care and use committee Hypnorm–midazolam group, which received buprenorphine after approved the study protocols. the period of anesthesia. Sterile ophthalmic ointment (Lacrilube, Retrospective analysis of eye tissues from infusion Allergan, Irvine, Calif.) was instilled into both eyes of all rats at studies. We evaluated the clinical records and hematoxylin and induction. After anesthetic induction, a surgical plane of anesthe- eosin-stained sections of eyes from 402 (215 male and 187 fe- sia was maintained for a minimum of 25 min. Control animals male) Wistar rats of eight intravenous infusion toxicology studies were placed in a cloth restrainer that permitted the faces of the (Table 1). None of the investigational compounds under study rats to be exposed to oxygen alone (0.5 liter/min) by facemask for was associated with eye toxicity. Animals from these studies had 25 min. Animals were observed and monitored throughout the been anesthetized and instrumented with vascular access ports period of anesthesia and subsequently until termination. Be- (jugular vein) 3 to 7 days prior to study initiation. In prepara- cause all animals receiving injectable anesthetic agents had tail tion for jugular vein catheterization, rats were premedicated arterial blood oxygen saturation readings of < 95%, they received with 0.05 mg/kg atropine (atropine sulphate, Lyphomed, Mel- supplemental O2 (0.5 liter/min) by facemask. Ophthalmic exami- rose Park, Ill.) subcutaneously (s.c.), and 1% chloramphenicol nations were conducted preanesthesia and at termination. 176 Susceptibility of rats to corneal lesions after injectable anesthesia

Table 3. Corneal histopathology scoring system Score Histologic changes and distribution 0 Normal or minimal, locally extensive mineralization of the anterior limiting membrane 1 Minimal to mild, locally extensive mineralization of the anterior limiting membrane with focal corneal epithelial cell degeneration 2 Moderate locally extensive mineralization of the anterior limiting membrane and stroma with intraepithelial cell debris, hydropic epithelial cell degeneration, and/or multinucleated giant cells within the stroma 3 Moderate locally extensive mineralization of the anterior limiting membrane and stroma with stromal neovascularization and mixed stromal leukocytic infiltrates 4 Corneal erosion, ulcer, pustule, or fibrosis with or without moderate to marked mineralization of the anterior limiting membrane and stroma, mixed stromal leukocytic infiltrates, and/or stromal neovascularization

To study the time course of corneal lesion development, the Table 4. Retrospective summary of corneal opacities on pretest clinical a first three animals in each group were euthanized on the day examination of Wistar rats after anesthesia after anesthesia, the next three were euthanized 4 days after No. with focal No. with diffuse b anesthesia, and the final three animals were euthanized 9 days corneal opacities corneal opacities a after anesthesia. All control animals were euthanized 9 days Study Male Female Male Female after exposure to oxygen. 4c 7 9 27 13 Effect of rat stock or strain on susceptibility to corneal 5 1 1 39 39 7 11 10d 4 30d lesion development. We anesthetized 50 male rats (10 each a of Wistar, F344, SD, LE, and Lewis) were anesthetized with 40 male and 40 female rats for each study. bAnimals with single pinpoint to focal corneal opacities in one or both eyes after ketamine (60 mg/kg) and xylazine (12 mg/kg) i.p. In addition, surgical instrumentation but prior to study initiation. Animals were treated the effect of yohimbine reversal on corneal injury outcome was with sterile ophthalmic ointment at anesthetic induction. Refer to Table 1 for specific anesthetic protocols for studies. evaluated in 10 additional male Wistar rats, which received 2.1 cAll animals with corneal opacities in this study were anesthetized with ket- mg/kg yohimbine (yohimbine hydrochloride, A.G. Scientific, San amine–xylazine. Diego, Calif.) i.p. after the period of anesthesia. Lacrilube was in- dTwo female rats were anesthetized with Hypnorm–midazolam. stilled into the eyes of animals at anesthetic induction and after induction; surgical anesthesia was maintained for a minimum of did not resolve or diminish in size throughout the course of the 25 min. All animals received supplemental oxygen (0.5 liter/min) study, and their presence precluded detailed clinical evaluation by facemask. All rats received buprenorphine (0.05 mg/kg s.c.) af- of eye structures located posterior to the cornea during the stud- ter the period of anesthesia. Animals were monitored throughout ies or at termination. There were no differences in the number or anesthesia and recovery, and time to ambulation after recovery distribution of corneal lesions in control versus treated animals was recorded. Prestudy and terminal ophthalmic observations or between dose groups of treated animals (data not shown). were conducted on all animals, and animals were euthanized When pinpoint or focal, corneal opacities occurred on the cen- within 7 to 10 days of anesthesia. tral part of the cornea, along the lid aperture. In more severely Eye histopathology. Both eyes were collected at necropsy, affected animals with more diffuse lesions, corneal fibrosis and fixed in Zenker’s solution, 10% buffered formalin, 6% glutaral- neovascularization also were visible clinically (Fig. 1A). dehyde–60% ethyl alcohol, or cold 4% paraformaldehyde and Histologically, pinpoint or focal corneal opacities observed clin- processed, sectioned, and stained with hematoxylin and eosin for ically were correlated with a corneal score of 1 and were charac- light microscopy. terized by minimal to mild, locally extensive mineralization of Corneal lesion scoring. Two pathologists graded corneal the anterior limiting membrane of the cornea and rare corneal changes from 0 to 4 in light of extent and degree of mineraliza- epithelial cell degeneration (Fig. 1B through H). Depending upon tion, leukocytic infiltrates, neovascularization, stromal fibrosis the severity of the reaction, more diffuse corneal opacities were and ulceration (Table 3). Pathologists were unaware of animal correlated histologically with lesion scores of 2 to 4: moderate to treatment or group when viewing the slides. After a consensus marked mineralization of the anterior limiting membrane; mild decision was reached, the most severe score for either or both to marked mixed leukocytic infiltrates (lymphocytes, plasma eyes was used as the final corneal lesion score for the animal. cells, and neutrophils) within the epithelium and substantia pro- Statistics. Difference in lesion score after the use of different pria; stromal edema; and in some animals, multinucleated giant anesthetics in the retrospective study was evaluated using the cells within the substantia propria (Fig. 1H). Corneal erosion or Kruskal–Wallis test (one-way nonparametric analysis of vari- ulceration with or without intraepithelial pustules, stromal fibro- ance) followed by a post hoc Tukey least-squares test (5). sis, and neovascularization also was noted in the most severely affected animals (corneal score of 4). Results During retrospective comparison of the incidence of histo- Retrospective study. For those studies in which a prestudy, pathologic corneal lesion score severity between the different postsurgical ophthalmic examination occurred, 35 to 98% of ani- anesthetics used for these studies, for both male and female rats, mals anesthetized with ketamine–xylazine had diffuse corneal lesion score severity was significantly different between animals opacities of one or both eyes present prior to study initiation (Ta- anesthetized with ketamine–xylazine versus animals anesthe- ble 4). Of animals anesthetized with Hypnorm–midazolam, 10 to tized with Hypnorm–midazolam (P < 0.001 for both sexes) or 50% had diffuse corneal opacities in one or both eyes at prestudy enflurane alone (P < 0.001 for both sexes). Animals anesthetized examination. No corneal opacities were observed in the animals with ketamine–xylazine had the most severe lesion scores, with anesthetized with isoflurane. Once identified, corneal opacities 71% of animals having a score of 2 or greater and 33% of animals 177 Vol 55, No 2 Comparative Medicine April 2005

Figure 1. Gross photograph of rat cornea and representative photomicrographs of corneal lesion scores as described in Table 3 in rats after injectable anesthesia. (A) Photograph of right eye from a female control rat 15 days after anesthesia with ketamine–xylazine (Study 7; Table 1). A large, irregular corneal opacity containing blood vessels is visible over the central part of the eye. (B, C) Corneas corresponding to a lesion score of 0 (normal cornea). There is minimal mineralization (arrows) of the anterior limiting membrane in (C). (D) Cornea corresponding to a lesion score of 2. There is locally extensive mineralization of the anterior limiting membrane with epithelial cell degeneration (arrow). (E) Cornea corresponding to a lesion score of 3. There is moderate mineralization of the anterior limiting membrane with multifocal stromal neovascularization, moderate mixed stromal leukocytic infiltrates and occasional multinucleated giant cell. (F, G, H) Corneas corresponding to lesion score of 4. In (F) and (G), there is locally extensive ulceration (arrows in F) with stromal neovascularization and moderate mixed leukocytic infiltrates. In (H), a subepithelial pustule is present with occasional multinucleated giant cells (arrow). Hematoxylin and eosin; magnification: ×40 (G), ×100 (B, D through F, and H), ×250 (C). 178 Susceptibility of rats to corneal lesions after injectable anesthesia

Table 5. Summary of corneal histopathology scores after anesthesia in Wistar rats Summary of corneal scoresa Male rats Female rats Combined sexes Anesthesia method (total no. animals) 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 Enflurane or isoflurane (59) 18 3 0 0 0 27 10 1 0 0 45 13 1 0 0 Ketamine–xylazineb (268) 25 12 13 31 44 30 11 14 45 44 55 23 27 76 88 Hypnorm–midazolam (67) 43 9 5 2 4 2 0 0 0 2 45 9 5 2 6 aCorneal sections from seven male and one female rat could not be evaluated because of inappropriate tissue orientation. Refer to Table 3 for the corneal lesion grading system. All eye sections (left and right) were evaluated by two pathologists and scored by consensus. The higher of the lesion scores for either the left or right eye was used as the final corneal lesion score for the animal. bCorneal lesion scores for animals (males and females) treated with ketamine–xylazine were significantly (P < 0.001) more severe than those from animals anesthetized with either enflurane, isoflurane, or Hypnorm–midazolam.

Table 6. Summary of corneal histopathology scores 24 h, 4 days, and 9 days after anesthesia of female Wistar rats Corneal scoresb 24 h 4 days 9 days Anesthesia method 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 Control 2 1 0 0 0 3 0 0 0 0 1 2 0 0 0 Control (jacket) nd nd nd nd nd nd nd nd nd nd 3 1 0 0 0 Ketamine–xylazinec 0 1 1 0 1 2 0 1 0 0 0 0 0 0 3 Pentobarbital 0 1 1 1 0 1 1 1 0 0 2 1 0 0 0 Hypnorm–midazolam 0 1 1 1 0 1 1 1 0 0 1 1 1 0 0 Isoflurane 1 2 0 0 0 2 0 1 0 0 1 2 0 0 0 nd, not done. an = 9/group, except for the control (jacket) group, in which n = 4. Animals were euthanized 24 h, 4 days, or 9 days after a 25-min period of surgical anesthesia. All eye sections (left and right) were evaluated by two pathologists and scored by consensus. The higher of the corneal scores for either the left or right eye was used as the final corneal score for the animal. bRefer to Table 3 for the corneal grading system. cRefer to Table 2 for anesthetic drug dosages used. having a lesion score of 4 (Table 5). Of animals anesthetized with of the original dose of the ketamine–xylazine cocktail. There was Hypnorm–midazolam, 19% had histologic corneal lesion scores no correlation, however, between animals that received a top-up of 2 or greater and 9% had a lesion score of 4. Only one animal dose of ketamine–xylazine and prolonged anesthetic recovery anesthetized with either enflurane or isoflurane had a lesion times or to severity of corneal histopathology score at termina- score of 2 (< 2%). No sex-associated differences were noted be- tion (data not shown). Between 50 to 90% of Wistar, LE, and tween distributions of corneal lesion scores for any of the anes- F344 animals developed postanesthetic corneal changes with thetics used. lesion scores of 2 or greater. This finding compares with 20% of Time-course study on corneal lesion development after SD rats and 20% of Wistar rats reversed with yohimbine with anesthesia in Wistar rats. Indirect ophthalmic examinations lesion scores of 2 or greater and no corneal scores exceeding 1 in were normal for all animals prior to undergoing anesthesia. We Lewis rats (Table 7). Lesions were most severe in F344 rats, with noted that the eyelids did not completely close for any animals two animals developing corneal ulcers after anesthesia. Corneal during anesthesia. Corneal scores for oxygen-alone control ani- ulcers (lesion score 4) also occurred in one SD rat and one Wistar mals, either fitted with or without cotton tether jackets, never rat that had been reversed with yohimbine. Lesion development exceeded 1. Postanesthetic corneal changes with lesion scores of was not associated with time to recovery from anesthesia, as rats 2 or greater were noted in 66% (6 of 9) of animals anesthetized from all treatment groups had similar recovery durations with with ketamine–xylazine, 33% (3 of 9) of those anesthetized with animals ambulating within 60 to 120 min of anesthesia (data pentobarbital, 44% (4 of 9) of rats anesthetized with Hypnorm– not shown). midazolam, and 11% (1 of 9) of those anesthetized with isoflurane (Table 6). Four animals anesthetized with ketamine–xylazine de- Discussion veloped corneal ulcers in one or both eyes (corneal score of 4), one The results of this study demonstrate that anesthesia of rats animal within 24 h of anesthesia. Clinically, these animals de- by injectable agents predisposes them to development of cor- veloped blepharospasm, photophobia, and grossly visible corneal neal lesions, despite appropriate intraoperative use of sterile eye opacities within 18 h of anesthesia. Signs of discomfort resolved lubricants. The lesions are visible clinically as focal to diffuse within 48 h of development, although corneal opacities did not central corneal opacities and are correlated histologically with regress with time. No corneal erosions or ulcers developed in mineralization of the anterior limiting membrane and keratino- animals from any of the other anesthetic groups. cyte degeneration to leukocytic infiltrates, neovascularization, Effect of rat stock or strain on susceptibility to corneal stromal fibrosis, intraepithelial pustules and ulcers (corneal le- lesion development. Indirect ophthalmic examinations were sion scores of 2 or greater, according to the criteria in Table 3). normal for all animals prior to undergoing anesthesia. To main- Compared with those of animals anesthetized with other proto- tain general anesthesia for a 25-min period, two to six animals cols, corneal lesions were more severe in animals anesthetized in each group required a “top-up,” consisting of an additional 50% with ketamine–xylazine and developed after brief periods of 179 Vol 55, No 2 Comparative Medicine April 2005

Table 7. Summary of corneal histopathology scores for male rats of various nist, and xylazine, an α2-adrenergic receptor agonist, frequently stocks and strains after ketamine–xylazine anesthesia are combined in laboratory animal medicine to provide general Corneal scoresb anesthesia to rodents and rabbits. Both xylazine and another po- Stock or straina 0 1 2 3 4 tent α2-adrenoreceptor agonist, detomidine, induce marked and Wistar 4 0 5 1 0 persistent vasoconstriction without increasing systemic vascular Sprague-Dawley 8 0 1 0 1 Long-Evans 5 0 2 3 0 resistance, which in conjunction with their potent negative chro- Fischer 344 0 1 1 6 2 notropic effects, may lead to reduced tissue perfusion. In dogs Lewis 9 1 0 0 0 Wistar + yohimbinec 7 1 1 0 1 and horses, these effects may persist long after any apparent sedative effects have passed and may be a contributing factor for an = 10/group. Animals were anesthetized with ketamine (60 mg/kg) and xylazine (12 mg/kg) administered intraperitoneally, and sterile ophthalmic ointment was post-anesthetic myopathy in horses (6, 13, 16, 28). In rabbits, the instilled into both eyes at anesthesia induction. Animals were euthanized 7 to combination of ketamine and xylazine (or detomidine) given in- 10 days after the 25-min period of surgical anesthesia. All eye sections (left and tramuscularly for anesthesia can induce myocardial necrosis and right) were evaluated by two pathologists and scored by consensus. The higher of the corneal scores for either the left or right eye was used as the final corneal fibrosis, leading to death (15, 21). The pathogenesis of the cardiac score for the animal. lesions in rabbits is thought to be due to prolonged xylazine-in- b Refer to Table 3 for the corneal grading system. duced hypoxemia and coronary vasoconstriction, in conjunction cTreated with yohimbine (2.1 mg/kg) intraperitoneally after anesthesia. with limited collateral myocardial circulation and subsequent reperfusion injury (21). surgical anesthesia (25 min in duration). That the lesions are in- In pharmacologic studies, systemically administered xylazine duced by the use of injectable anesthetics is supported by a lack has been demonstrated to have a marked, persistent mydriat- of pronounced mineralization of the anterior limiting membrane ic effect in rats, due to a potent α2-adrenoreceptor mechanism of the cornea or other pathologic changes in animals anesthe- that suppresses parasympathetic tone to the iris (2). Xylazine tized with inhalant agents such as enflurane or isoflurane or alone induces acute reversible cataracts in the lens of mice and in nonanesthetized control animals. Further, the incidence of rats, at doses similar to those used in our study, and this effect corneal lesions was reduced when Wistar rats anesthetized with is enhanced by combination with ketamine. The pathogenesis of ketamine–xylazine were treated with yohimbine after anesthe- these cataracts has been postulated to be due to drug-induced sia. Lesion development also is temporally related to anesthesia, alterations of intraocular fluid dynamics or a change in aqueous with focal to diffuse corneal opacities developing in some animals oxygenation and osmolality (4). Aqueous humor oxygenation in within 18 h of anesthesia. Once present, the corneal lesions did the rat occurs primarily via arterioles of the ciliary body and not regress, even during as long as 111 days postsurgery. iris and to a much lesser extent from transcorneal absorption We interpreted corneal histologic scores of 0 to 1 for these (9). Ketamine–xylazine anesthesia of rats may cause persistent studies to be within the range of expected background degenera- and profound changes in the partial pressure of oxygen within tive corneal changes in laboratory rats and scores of 2 or greater the aqueous humor. Oxygen partial pressures measured within as pathologic alterations of corneal architecture. Only one eye the anterior eye chamber of spontaneously breathing rats anes- from a single rat anesthetized with either enflurane or isoflu- thetized with ketamine–xylazine were 25% of the partial pres- rane in each of the retrospective (118 eyes) and time-course (18 sures obtained for the same anesthetized rats supplemented eyes) studies had a corneal score of 2, suggesting that inhalant with 100% O2 by facemask (9). This finding is consistent with the anesthesia does not induce clinically significant corneal lesions profound hypoxemia noted systemically in our studies (cyanosis in rats. In addition, it was tempting initially to speculate that and tail-cuff SpO2 readings < 65%) in spontaneously breathing corneal lesions were induced or aggravated when rats were fitted rats anesthetized with injectable agents. with cotton tether jackets postoperatively, because of direct cor- We hypothesize that the ketamine–xylazine-induced corneal neal abrasion on the jackets; however, no corneal changes were lesions in rats seen in this study and reported by others result noted in control rats fitted with jackets. from a combination of profound and persistent drug-induced va- Although no differences were noted in corneal lesion suscepti- soconstriction of ciliary and iridial vessels and local and systemic bility between sexes, there may be stock- or strain-associated dif- hypoxemia, leading to prolonged corneal hypoxia and subsequent ferences in susceptibility to anesthetic-induced corneal lesions. cell injury. The injury may be exacerbated by subsequent re- Wistar, Long-Evans, and Fischer 344 rats may be at increased perfusion and oxygen radical damage. After corneal injury, the risk for developing postanesthetic corneal lesions after the use corneal endothelium and anterior limiting membrane of the epi- of ketamine–xylazine combinations, whereas Sprague-Dawley thelium rapidly express epidermal growth factor (EGF), vascular and Lewis rats were more resistant. Spontaneous corneal degen- endothelial growth factor (VEGF), interleukin-1 (IL-1) and other eration characterized clinically as punctate superficial corneal cytokines to stimulate corneal epithelial apoptosis and wound opacities without evidence of accompanying inflammation or healing (14, 24, 31). In the conscious animal, extracellular cal- irritation and mineralization of the anterior limiting membrane cium is supersaturated in the matrix and fluids of the eye and of the cornea occurs as a background lesion in laboratory rats (1, other tissues (23). It may be that either hypoxemia-induced local 19, 27). The pathogenesis of this degenerative lesion is unknown, pH reduction, or cytokine release by the anterior limiting mem- but it has been reported previously to occur in Wistar, Sprague- brane with or without subsequent cellular damage, act to precip- Dawley, and Fischer 344 rats (1, 19, 27). Rats infected with si- itate extracellular calcium leading to dystrophic mineralization alodacryoadenitis virus and subsequently immunosuppressed of the anterior limiting membrane. Prolonged release of IL-1 and with dexamethasone also demonstrate a stock- or strain-induced upregulation of growth factors (e.g., EGF and VEGF) by the an- susceptibility to keratitis development (3). terior limiting membrane may initiate further inflammation and Ketamine, a potent N-methyl-D-aspartate receptor antago- neovascularization (9, 10, 23, 32). The very large corneal surface 180 Susceptibility of rats to corneal lesions after injectable anesthesia

of the rat eye compared with its volume may predispose this spe- ies involving anesthesia, in which evaluation of the cornea may cies to both induced and spontaneous injuries of this nature. be important. Injectable agents and, in particular, ketamine–xy- The lesion may take several hours to initiate because immedi- lazine anesthesia may induce severe corneal lesions, despite the ate postanesthetic treatment with yohimbine decreased the in- use of adequate intraoperative ophthalmic ointment. Corneal cidence of ketamine–xylazine-induced corneal lesions, although lesions occurred after periods of surgical anesthesia as short as it had no effect on time to ambulation after anesthesia. Others 25 min. If these agents must be used, lesion incidence may be also have noted the protective effect of yohimbine on the rat cor- reduced by supplementing rats with 100% O2 by facemask dur- nea after ketamine–xylazine anesthesia (18). Providing oxygen ing anesthesia or reversing xylazine with an appropriate dose of by facemask also seemed to confer corneal protection. In the yohimbine immediately after surgery. Using inhalant agents for early surgical studies that were examined retrospectively, rats anesthesia may minimize the incidence of corneal lesions. Wistar, anesthetized with injectable agents were not supplemented with Lewis, and Fischer 344 rats may be at increased risk for develop-

100% O2 by facemask, whereas in the time-course and stock/ ing postanesthetic corneal lesions after the use of ketamine–xy- strain studies they were. It is noteworthy that the incidence and lazine for anesthesia. severity of corneal lesions decreased overall after implementa- tion of this measure. Ketamine induces mild proptosis of the globe, and one may Acknowledgments The authors would like to thank Michelle Tyssen for expert technical speculate that loss of blink reflex and exposure of the eye during assistance during the anesthesia studies and Tim Sullivan (Information ketamine anesthesia may predispose corneas of animals anes- Technology Services, Ontario Veterinary College) for assistance with thetized with this drug to drying. We think this theory unlikely photomicrographs. as the sole cause underlying corneal lesion development because the eyelids of all anesthetized animals remained partially open during and after anesthesia with inhalant or injectable agents, References 1. Bellhorn, R. W., G. E. Korte, and D. Abrutyn. 1988. Spontane- and the time to return to consciousness and ambulation was ous corneal degeneration in the rat. Lab. Anim. Sci. 38:46-50. similar for all groups. In humans, the blink reflex is markedly 2. Berrige, T. L., B. Gadie, A. G. Roach, and I. F. Tulloch. 1983. α-2 depressed after general anesthesia with both injectable and in- adrenoreceptor agonists induce mydriasis in the rat by an action halant agents, and patients often recover consciousness prior to within the central nervous system. Br. J. Pharmacol. 78:507-515. 3. Boivin, G. P. and S. A. Theus. 1996. Ophthalmic lesions in dexa- re-establishing the blink reflex (20). It is possible that the blink methasone-treated rats naturally infected with sialodacryoadenitis reflex is depressed to a greater degree in rats anesthetized with virus while concurrently used for Pneumocystis carinii propagation. ketamine–xylazine versus other injectable and inhalant drugs Lab. Anim. Sci. 35(1):73-75. and that this effect contributes to lesion development. 4. Calderone, L., P. Grimes, and M. Shalev. 1986. Acute reversible Although opioid and α -adrenoreceptor agonists can have cataract induced by xylazine and by ketamine-xylazine anesthesia 2 in rats and mice. Exp. Eye Res. 42:331-337. similar pharmacologic effects, we do not believe that the use of 5. Edgington, E. S. 1995. Randomization tests. Marcel Dekker, New buprenorphine in these studies contributed to the corneal lesions York. seen. Buprenorphine is only a partial opioid agonist and neither 6. Edner, A., G. Nyman, and B. Essen-Gustavsson. 2002. The re- exacerbates nor diminishes the ocular inflammation and later lationship of muscle perfusion and metabolism with cardiovascular variables before and after detomidine injection during propofol-ket- keratoconjunctivitis in guinea pigs used in the Sereny test (12, amine anaesthesia in horses. Vet. Anaesth. Analg. 29:182-199. 26). Corneal lesions did occur in rats anesthetized with pentobar- 7. Eke, T., D. A. Morrison, and D. J. Austin. 1999. Recurrent bital and Hypnorm–midazolam but with much lower incidence symptoms following traumatic corneal abrasion: prevalence, sever- and severity than with ketamine–xylazine. Barbiturates and ity, and the effect of a simple regimen of prophylaxis. Eye 13(Pt opioids induce severe respiratory depression and hypoxemia in 3a):345-347. 8. Fabian, R. J., J. M. Bond, and H. P. Drobeck. 1967 Induced the rat but are free of the direct vascular effects of xylazine. The corneal opacities in the rat. Brit. J. Ophthalmol. 51:124-129. hypoxemia induced by these agents may be profound enough to 9. Fitch, C. L., S. H. Swedberg, and J. C. Livesey. 2000. Measure- precipitate dystrophic mineralization in some animals. ment and manipulation of the partial pressure of oxygen in the rat Whether animals might benefit from repeated application of anterior chamber. Curr. Eye Res. 20(2):121-126. 10. Giachelli, C. M. 1999. Ectopic calcification: gathering the hard sterile or antibiotic-containing eye lubricant either during or facts about soft tissue mineralization. Amer. J. Pathol. 154(3):671- after anesthesia is unknown. Rats treated repeatedly with chlor- 675. amphenicol eye ointment after orbital sinus puncture had in- 11. Grover, V. K., K. V. Kumar, S. Sharma, N. Sethi, and S. P. creased incidence of corneal lesions compared with those not Grewal. 1998. Comparison of methods of eye protection under receiving eye ointment, suggesting that repeated application to general anaesthesia. Can. J. Anaesth. 45(6):575-577. 12. Hanson, C. E., G. R. Ruble, I. Essiet, and A. B. Hartman. 2001. a damaged cornea may exacerbate injury (29). Similarly, human Effects of buprenorphine on immunogenicity and protective efficacy patients with traumatic corneal injury had a poorer therapeutic in the guinea pig keratoconjunctivitis model (Sereny test). Comp. outcome with prolonged nightly use of Lacrilube ointment, com- Med. 51(3):224-229. pared with patients that did not have further treatment beyond 13. Hennig, G. E., M. H. Court, and V. L. King. 1995. The effect of xylazine on equine muscle surface capillary blood flow. J. Vet. initial topical therapy (7). In addition, the use of atraumatic tape Pharmacol. Ther. 18:388-390. to keep eyelids closed during anesthesia does not appear to con- 14. Hoppenreijs, V. P. T., E. Pels, G. F. J. M. Vrensen, and W. F. fer further corneal protective advantages over a single applica- Treffers. 1996. Corneal endothelium and growth factors. Surv. tion of eye lubricant (11). Ophthalmol. 41(2):155-164. In conclusion, our studies have important implications for pro- 15. Hurley, R. J., R. P. Marini, D. L. Avison, J. C. Murphy, J. M. Olin, and N. S. Lipman. 1994. Evaluation of detomidine anes- viding optimal anesthesia and postoperative care for rats and for thetic combinations in the rabbit. Lab. Anim. Sci. 44:472-477. minimizing confounding factors during toxicologic or ocular stud- 181 Vol 55, No 2 Comparative Medicine April 2005

16. Ilback, N.-G. and T. Stalhandske. 2003. Cardiovascular effects 26. Swearengen, J. R., R. A. Cockman-Thomas, J. A. Davis, of xylazine recorded with telemetry in the dog. J. Vet. Med. A. and P. J. Weina. 1993. Evaluation of butorphanol tartrate and 50:479-483. buprenorphine hydrochloride on the inflammatory reaction of the 17. King, K. L., W. L. West, and W. L. Hunter. 1995. Diagnostic Sereny test. Lab. Anim. Sci. 43:471-475. exercise: ocular lesions in hypophysectomized rats. Contemp. Top. 27. Taradach, C. and P. Greaves. 1984. Spontaneous eye lesions in Lab. Anim. Sci. 34(3):103-104. laboratory animals: incidence in relation to age. CRC Crit. Rev. 18. Kufoy, E. A., V. A. Pakalnis, C. D. Parks, A. Wells, C.-H. Yang, Toxicol. 12(2):121-147. and A. Fox. 1989. Keratoconjunctivitis sicca with associated sec- 28. Teng, B., W. W. Muir, and D. E. Mason. 1995. Segmental response ondary uveitis elicited in rats after systemic xylazine/ketamine of canine cornary arteries to alpha-2 agonists, xylazine and me- anesthesia. Exp. Eye Res. 49:861-871. detomidine. Vet. Surg. 24:537. 19. Losco, P. E. and C. M. Troup. 1988. Corneal dystrophy in Fischer 29. van Herck, H., V. Baumans, C. J. Brandt, A. P. Hesp, J. H. 344 rats. Lab. Anim. Sci. 38:702-710. Sturkenboom, H. A. van Lith, G. van Tintelen, and A. C. 20. Marelli, R. A. and A. D. Hillel. 1989. Effects of general anesthesia Beynen. 1998. Orbital sinus blood sampling in rats as performed on the human blink reflex. Head Neck 11:137-149. by different animal technicians: the influence of technique and 21. Marini, R. P., X. Li, N. K. Harpster, and C. Gangler. 1999. Car- expertise. Lab. Anim. 32(4):377-386. diovascular Pathology possibly associated with ketamine/xylazine 30. Vonvoigtlander, P. F., G. J. Kolaja, and E. M. Block. 1982. anesthesia in Dutch belted rabbits. Lab. Anim. Sci. 49:153-160. Corneal lesions induced by antidepressants: a selective effect upon 22. Roerig, D. L., A. T. Hasegawa, G. J. Harris, K. L. Lynch, and young Fischer rats. J. Pharmacol. Exper. Therap. 222:282-286. R. I. H. Wang. 1980. Occurrence of corneal opacities in rats after 31. Wilson, S. E., Y.-G. He, J. Weng, Q. Li, A. W. McDowall, M. Vital, acute administration of l-α-acetylmethadol. Toxicol. Appl. Phar- and E. L. Chwang. 1996. Epithelial injury induces keratocyte macol. 56:155-163. apoptosis: hypothesized role for the Interleukin-1 system in the 23. Schuller, S., M. Knorr, K.-P. Steuhl, and H.-J. Thiel. 1996. modulation of corneal tissue organization and wound healing. Exp. Lacrimal secretion of human epidermal growth factor in perforat- Eye Res. 62:325-337. ing keratoplasty. Ger. J. Ophthalmol. 5:268-274. 32. Zheng, M., M. A. Schwarz, S. Lee, U. Kumaraguru, and B. T. 24. Schultz, G., N. Chegini., M. Grant, P. Khaw, and S. MacKay. Rouse. 2001. Control of stromal keratitis by inhibition of neovas- 1992. Effects of growth factors on corneal wound healing. Acta cularization. Amer. J. Pathol. 159(3):1021-1029. Ophthmal. 70(Suppl.):60-66. 25. Shimizu, T., S. Fujita, K. Izumi, T. Koja, and T. Fukuta. 1984. Corneal lesions induced by the systemic administration of capsaicin in neonatal mice and rats. Naunyn Schmiedelbergs Arch. Pharmacol. 326:347-351.

182