A Model for Clinical Evaluation of Perioperative Analgesia in Rabbits (Oryctolagus cuniculus)

Abstract

Assessment of pain in rabbits is challenging, and studies of effective surgical analgesia are lacking for this species. Seeking potential indicators of postoperative pain, we performed ovariohysterectomy and telemeter placement as a form of moderate surgical injury in 20 female rabbits. Rabbits were assigned to 1 of 4 treatment groups (5 per group): buprenorphine (0.02 mg/kg SC every 12 h for 3 d); fentanyl (25-μg patch placed 24 h preoperatively); ketoprofen (1 mg/kg SC every 24 h for 3 d), and control (no treatment given). Various physiologic and behavioral variables were recorded by blinded observers, including food and water consumption, fecal output, and remotely recorded behaviors during daily exercise in 1.2 × 1.8 m floor pens. Compared with preoperative values, significant declines occurred in: food consumption (days 1 to 7), water consumption (days 1 to 4), fecal output (days 1 to 2), mean travel distance, and rearing (days 1 to 3 and day 7). No single treatment proved significantly better than another. Our results demonstrate substantial inappetance and reduction of normal activity levels in rabbits after surgery. Although results from rabbits treated with empirical doses (those typically recommended) of analgesics did not appear substantially better than those from the untreated control group, comparison of other doses and multimodal analgesic techniques by using these behavioral monitoring strategies may prove useful in future studies aimed at optimizing postoperative analgesia in rabbits.

Rabbits are widely used as a research animal model in areas ranging from infectious disease to orthopedic surgery. In 2007, the most recent year for which data are available, 236,511 rabbits were used in research in the United States, representing 23% of the total number of animals reported to the US Department of Agriculture³¹ and the single largest category of reported species. Approximately 41% of the rabbits used were reported in category D, meaning that they underwent procedures involving pain or distress and received drugs intended to relieve the pain or distress associated with the procedure. However, control of pain and distress can be assured only when valid methods to recognize them are used and evidence for effectiveness of analgesics exists. Despite the common use of rabbits in a wide variety of research areas, methods to assess clinically relevant pain in this species are lacking in the literature.^{8, 17,20} The current dogma is that rabbits, as a prey species, usually do not display obvious behaviors when in pain; therefore, it is difficult to know how much pain they experience and whether analgesic methods are effective.^1,15,21,29

Methods of assessing pain in animals have included objective measurement of pain threshold or latency to respond to a noxious stimulus and physiologic variables (for example, blood pressure, heart rate, or respiratory rates).^{2,10,16,18,26,28} These outcome measures often are used in studies of pain biomedicine but may not be suitable for assessing the clinical state of an animal. Clinically, a combination of objective (for example, body weight, food or water consumption) and subjective (for example, general temperament, activity, response to handling, appearance) assessments can be combined to generate a ‘score’; this methodology is the basis for composite measurement pain scoring systems.^{9,12,14,22-25} A number of clinical pain scoring systems for animals have been developed but relatively few of these have been validated, which means they may not be sensitive to detecting pain.^25,34 Two previous attempts at developing a consistent pain scoring system for rabbits have been unsuccessful.^27,29 To develop a clinical pain scoring system, the relevant components must first be established.

To this end, we sought to evaluate changes in several variables in rabbits after surgery (ovariohysterectomy with concurrent telemeter implantation) by using assessment of activity conducted alongside more traditional variables, such as food and water consumption and body weight. Preliminary video data collected by our group indicated that rabbits display a severely limited behavioral repertoire in traditional cages, a concept noted by others in the literature.^1,4 To encourage normal behaviors such as exploration and foraging, we provided rabbits with a 4-h daily exercise period in 1.2 × 1.8-m pens. Our hypothesis was that behavior changes would occur in rabbits after moderate surgical injury. The extent of changes from baseline variables, if known, could serve as a means of assessing postoperative pain or distress. If these changes were known, then the same variables could serve as a basis for clinical evaluation and future assessment of methods of perioperative analgesic efficacy and support.

Previous behavioral studies of postoperative pain have indicated sufficient power to detect differences in baseline and postoperative values with 5 to 10 subjects. However, we also wanted to collect pilot data to compare 3 analgesic regimens; our experience is that rabbits usually receive a single analgesic medication. This study also included an untreated control group that received no analgesic unless predetermined criteria were reached for rescue.

Materials and Methods

This study was approved by the Institutional Animal Care and Use Committee at Tufts University Cummings School of Veterinary Medicine (North Grafton, MA) and was conducted in an AAALAC-accredited facility in compliance with the Animal Welfare Act³ and the Guide for the Care and Use of Laboratory Animals.¹⁵

Animals.

Female New Zealand white rabbits (Oryctolagus cuniculus; n = 20; weight, 3.0 to 3.5 kg) were purchased from a commercial vendor (Millbrook Breeding Labs, Amherst, MA). The rabbits were serologically negative for Treponema cuniculi, Encephalitozoon cuniculi, cilia-associated respiratory bacillus, and Pasteurella multocida. They were culture-negative for Pasteurella multocida, Pasteurella pneumotropica, Bordetella bronchiseptica, Treponema cuniculi, Clostridium pilifomis, and Salmonella, Klebsiella, and Citrobacter spp. The rabbits were free of ectoparasites on gross examination and endoparasites by fecal examination. Upon arrival, rabbits were housed individually in stainless steel cages (61 × 76.2 × 43.2 cm) under standard environmental conditions (18 to 21 °C, 30% to 70% humidity, 12:12-h light:dark cycle, and 15 to 18 air changes per hour). Rabbits were fed a high-quality commercial pelleted food (Laboratory Rabbit Diet HF 5326, Lab Diet, PMI Nutrition International, St Louis, MO), and water was provided ad libitum in bottles. The diet was supplemented with hay (mixed grass and timothy blend) and CarrotSlims (Vitakraft Pet Products, Bound Brook, NJ). One week after arrival, rabbits were introduced to individual exercise pens (1.2 × 1.8 m) for 4 h each day (starting at 0800).

Surgery.

Ovariohysterectomy and telemeter implantation was used as a surgical model of pain. Rabbits were not fasted prior to surgery. On the day of surgery, rabbits were anesthetized with ketamine (35 mg/kg IM; Ketaset, Fort Dodge Laboratories, Fort Dodge, IA) and xylazine (5 mg/kg IM; Xylaject 2%, Phoenix Pharmaceuticals, St Joseph, MO). The rabbits were blind-intubated by using 3.0- to 3.5-mm, uncuffed endotracheal tubes and were maintained on isoflurane (1% to 5% in oxygen; IsoFlo, Abbott, Laboratories, Abbott Park, IL). A 20-gauge over-the-needle intravenous catheter was placed in an ear vein, and lactated Ringers solution (Baxter, Deerfield, IL) was administered at 10 mL/kg hourly throughout the procedure. The abdomen, flank, and groin areas of the rabbit were shaved and surgically prepared by using chlorhexidine and alcohol. Aseptic technique was adhered to throughout the procedure. The ovariohysterectomy was performed by means of a ventral midline incision. The blood vessels supplying the uterus and ovary were ligated by using hemoclips, and the uterine stump was oversewn by using 3-0 polydioxanone suture (Ethicon, Somerville, NJ) in a simple continuous pattern. The uterus and ovaries were removed. The body wall incision was closed by using 3-0 polydioxanone suture in a simple continuous pattern.

After body wall closure, a radiotelemetry transmitter (D70 PCT, Data Sciences International, St Paul, MN) was placed subcutaneously in the flank region, with electrocardiography leads tunneled subcutaneously to the thoracic region in a lead II configuration and the arterial pressure line tunneled subcutaneously to the groin for surgical placement. Bupivacaine (3 mg of 0.5% solution; Hospira, Lake Forest, IL) was distributed intraperitoneally and subcutaneously, and the overlying skin layer was closed with 3-0 polydioxanone suture in a continuous subcuticular pattern. The groin area then was surgically prepared, and a skin incision was made over the femoral artery. The blood pressure catheter was placed into the femoral artery and secured by using 4-0 polydioxanone suture, and the femoral artery was ligated distal to the catheter insertion site. Bupivacaine (2 mg; Hospira) was placed subcutaneously, and the groin incision was closed in 2 layers. The subcutaneous layer was closed by using 3-0 polydioxanone suture in a simple continuous pattern, and the skin layer was closed by using 3-0 polydioxanone suture in a continuous subcuticular pattern.

Postoperative care.

Prior to surgery, rabbits were assigned randomly to 1 of 4 analgesic groups (5 rabbits per group) by using a random numbers table: buprenorphine (Reckitt and Colman, Richmond, VA) 0.02 mg/kg SC every 12 h for 3 d; fentanyl 25-μg patch (Duragesic; Janssen, Titusville, NJ) placed above the interscapular region 24 h prior to surgery; ketoprofen (Ketofen, Fort Dodge Animal Health, Fort Dodge, IA) 1 mg/kg SC every 24 h for 3 d; and control (no treatment given). At the time this study was designed, there was no evidence to support efficacy of any analgesic intervention in the rabbit. Therefore, we used commonly cited doses from empirical suggestions in textbooks and the literature.^11,13,19 Thermal support (infrared heat lamps and water blankets) was provided to warm rabbits until they could maintain sternal recumbency after recovery from anesthesia. When fully awake after surgery, the rabbits were returned to their cages and the telemetry devices activated to record physiologic data (blood pressure, heart rate, and body temperature).

Telemetry.

A radiotelemetry transmitter (Data Sciences International) was used to collect heart rate, arterial blood pressure, and temperature readings. A receiver was mounted to the side of the rabbit cage or pen. These data were used in real time to check physiologic status and will be presented in a subsequent report.

Data collection.

Assessment of body weight and food and water consumption and fecal output scoring was performed daily over a 4-d period after acclimation and continued for as long as 7 d after surgery. Between 0730 and 0800 each day, rabbits were weighed; water consumption was evaluated by weighing the water bottles, and food consumption was evaluated by weighing pellets remaining in the food hopper. Rabbits then were transferred to the exercise pen to record behavior.

A digital photo of the cage tray was taken for fecal output scoring each day. A fecal output scoring system (Figure 1) was established to roughly quantify fecal output. The photographs of the cage trays were evaluated by a blinded investigator, and fecal output scores were assigned for each day of the study.

Figure 1.

System for scoring fecal output in rabbits.

The baseline values for food and water consumption and fecal output were established by averaging the values obtained on each of the 4 d prior to surgery. The body weight on the day of surgery was considered the baseline weight, and the weight of the telemeter unit (49 g) was subtracted from all body weights obtained after surgery.

After 3 d of acclimation to pens, videorecording of behavioral data was done on each of the 2 d prior to surgery to establish average baseline activity values and on postoperative days 1 through 4 and 7. After the morning examination, each rabbit was placed in the exercise pen (on a white vinyl flooring surface marked with a grid pattern) with a handful of hay, a treat stick (CarrotSlims), a jingle ball (BioServ, Frenchtown, NJ), a maple wood log (BioServ), and a polycarbonate rabbit hut (BioServ). Water bottles were hung on the pen and the receiver was placed nearby to gather physiologic data. Rabbit behavior was videotaped for 15 min, once each hour for 4 h, by using standard surveillance cameras (model SG6941, Lorex Technology, Markham, ON, Canada), which transmitted wireless signals to standard VHS videocassette recorders in the hallway so that behavior could be assessed without the presence of a person in the room. Entry into the room was prohibited during the time rabbits were in pens.

Videotapes were viewed in random order for a variety of behaviors (Figure 2) by an investigator who was blinded to treatment groups. Counts of episodes of vertical exploring, grooming, interacting with a toy, foraging, and frolicking as well as duration of sprawling (a potential indicator of relaxation) and distance traveled (number of grid lines traversed) during each session were scored. Incidents of monitored behaviors were denoted by using tick marks. Any episode of behavior with a duration longer than 30 s received a tick for each 30-s interval observed. For grooming, foraging, sitting, and sprawling, only behavior durations of greater than 5 s were counted. If the behavior stopped for less than 2 s and then resumed, it was considered continuous behavior.

Figure 2.

Rabbit behaviors observed.

Many of the behaviors occurred too infrequently to be useful, and of all behaviors scored, travel distance and rearing were affected most by surgery. Therefore, these behavioral variables underwent further statistical analysis.

Statistics.

All statistical analyses were conducted by using SigmaStat for Windows (version 3.1, SysStat Software, San Jose, CA). Treatment groups were compared by using one-way repeated-measures ANOVA and, when data failed normality tests, and Kruskal–Wallis 1-way ANOVA on ranks was performed. There were no significant differences between analgesic treatment groups. Therefore, baseline and postoperative body weight, food and water consumption, fecal output, and behavior data were analyzed for all rabbits combined, by using 1-way repeated measures ANOVA on ranks, with day as factor and differences tested by an all-pairwise multiple comparison using the Holm–Sidak method. Differences were considered significant only if P values were 0.05 or less.

Results

Dehiscence of telemeter incisions within the first 5 d after surgery occurred in 3 rabbits, including 2 rabbits from the ketoprofen group (days 1 and 5) and 1 rabbit from the fentanyl group (day 2). Data points from each of these 3 rabbits were excluded from the day on which the complication occurred onward. This exclusion resulted in 19 data points for day 1, 18 each for days 2 through 4, and 17 each on days 5 through 7.

No significant differences were found between analgesic treatment groups for any variable. Baseline and postoperative variables subsequently were analyzed for all rabbits combined. All rabbits experienced loss of body weight postoperatively, with weight remaining significantly (P < 0.05) below baseline through day 7 (Figure 3). Food consumption was decreased significantly (P < 0.05) when compared with baseline values for each of postoperative days 1 through 7 (Figure 4). The mean change on day 1 was −88.5% (range, −62% to −100%) and by day 7 was −27.3% (range, 3% to −89%). Water consumption was decreased significantly (P < 0.05) when compared with baseline values for postoperative days 1 through 4 (Figure 5). The mean change on day 1 was −92.7% (range, −74% to −100%). By day 7, the mean change was −16.6% (range: 23% to −49%). Fecal output scores were decreased significantly (P < 0.05) compared with baseline on postoperative days 1 and 2. Median fecal output scores for the 4 d prior to surgery were 2. On day 1 the median fecal output score was 0, and on days 2 and 3 the median score was 1. Median fecal output did not return to baseline until day 4. Travel distance was decreased significantly (P < 0.05) from baseline on days 1, 2, 3, and 7 (Figure 6). A mean change of −74.8% (range, −24.8% to −100%) occurred on day 1. Travel distance remained decreased on day 7, with a mean change of −38% (range, 60.9% to −83.4%). Rearing activity was decreased significantly (P < 0.05) on days 1, 2, 3, 4, and 7 postoperatively (Figure 7). A mean change of −97.6% (range, −81.4% to −100%) occurred on day 1. Rearing remained decreased on day 7, with a mean change of −77% (range, −30.7% to −100%).

Figure 3.

The percentage change (mean ± 1 SD) in body weight for laboratory rabbits for 7 d after surgery. *, Significantly different (P < 0.001) compared with baseline value.

Figure 4.

The percentage change (mean ± 1 SD) in food consumption of laboratory rabbits for 7 d after surgery. *, Significantly different (P < 0.001) compared with baseline value.

Figure 5.

The percentage change (mean ± 1 SD) in water consumption of laboratory rabbits for 7 d after surgery. *, Significantly different (P < 0.001) compared with baseline value.

Figure 6.

The percentage change (mean ± 1 SD) in distance traveled during 4 remotely videotaped sessions (15 min each) on days 1 through 4 and 7 after surgery in laboratory rabbits. *, Significantly different (P < 0.001) compared with baseline value.

Figure 7.

The percentage change (mean ± 1 SD) in rearing observed during 4 remotely videotaped sessions (15 min each) on days 1 through 4 and 7 after surgery in laboratory rabbits. *, Significantly different (P < 0.001) compared with baseline value.

Discussion

Despite the common use of rabbits in research studies, little is known about clinical diagnosis of pain or distress in this species. This lack of data hampers efforts to identify and evaluate the effectiveness of analgesic regimens designed to alleviate the pain that occurs as a result of research procedures. The goal of this study was to identify variables consistently associated with pain or postoperative morbidity in rabbits for use in the development of a clinically relevant postsurgical assessment scheme. Such an assessment method would enable scientists and veterinarians to better address the optimal treatment of rabbits after surgery and other painful manipulations. The results of our study indicate that in addition to commonly used variables such as food and water consumption and body weight, behavioral variables may be important measures to consider in the postoperative rabbit.

The loss of body weight can be explained by the decreased food and water consumption but may also be affected by postoperative catabolism. The degree to and duration for which the values were depressed could be in part consequences of inadequately controlled postoperative pain. Similar decreases in body weight have been documented in other studies in rabbits.^8,16,17 However, in our study the rabbits appeared to have a slower return to baseline body weight. This finding may indicate that the degree of surgical injury, whether due to the addition of the telemeter implantation, surgical expertise, or some other factor, was greater than that in the earlier studies. ^8,16,17 Numerous data document the association of decreases in body weight and food consumption with pain in a variety of species.^16,32 These reports vary in terms of degree of weight loss and extent of effect by analgesic treatments. Despite the inconsistencies, body weight is an easy, objective measure that should routinely be assessed in all subjects after surgery, regardless of the species or procedure.

Food consumption was depressed in the immediate postoperative period with a steady return toward baseline values by day 7. This finding has also been documented by others.^8,17,20 In addition, water consumption was depressed in the immediate postoperative period, with a steady return to baseline by day 5. The delay in return of food consumption to baseline levels may be due to pain but also in part to the fact that rabbits were given hay, the consumption of which was not quantified. Beside the obvious effects food and water consumption can have on bodyweight, reduced oral intake can contribute to gastrointestinal ileus, which is of considerable concern in the postoperative rabbit. Clinical signs associated with ileus include decreased or absent fecal output and abdominal distention and pain. Factors such as pain and environmental and dietary changes can result in gastrointestinal stasis in rabbits.⁶ We provide hay in addition to the pelleted diet because the hay may promote gastrointestinal motility and prevent development of clinical signs of ileus.

In addition, fecal output declined sharply compared with baseline values during the first 3 d after surgery. This finding is similar to that seen in another study.⁸ All rabbits had a dramatic reduction in postsurgical food intake, which likely was responsible for the lack of fecal output. That the decrease in fecal output was only statistically significant on days 1 and 2 may again be the result of hay consumption, which was not quantified. This finding may further support the practice of providing hay during the postoperative period to maintain gastrointestinal motility and prevent ileus. Although it is widely claimed that opioids are problematic due to their effects on reduction of propulsive motility of the gastrointestinal tract of rabbits, we found no difference in fecal output between any of the treatment groups, which included rabbits receiving 2 opioid treatments, transdermal fentanyl and buprenorphine (data not shown). Although fecal output can be quantified objectively by weighing waste products, we opted to develop a fecal output scoring system (Figure 1), which facilitated rapid quantification of fecal output by all staff members regardless of the level of training or experience.

Decreases in overall activity and changes in specific behaviors are documented responses to pain in a variety of species.^{10,16, 17, 21, 26} This association may in part be due to the stress response to surgery and anesthesia, the postoperative inflammatory cytokine cascade, or a direct result of postoperative pain, which may limit movement. In our study, surgery greatly affected 2 behaviors, travel and rearing activity. The degree of reduction in travel and rearing in our study was significant and prolonged. These findings differ from those in another study¹⁷ of postoperative pain in rabbits, in which inactive pain behaviors (for example, twitching, wincing, flinching) were found to be more reliable indicators of pain than were active behaviors (for example, standing, exploring). Our goal was to identify active behaviors that are readily apparent and could be incorporated into a postoperative assessment scheme; therefore inactive measures of pain were not assessed in our study design. Measures of postoperative pain in both human and animal clinical studies often include a dynamic or interactive component, because pain that may be minimal or even absent in the recumbent patient will become apparent during normal motor activities.^5,16,30 Arguably, improved pain control in rabbits should result in a less marked reduction in activity; however, analgesics, especially opioids, can result in sedation, which can contribute to decreased activity.

Although our findings support our hypothesis that activity and behaviors are relevant to the assessment of the postoperative rabbit, the effort required to monitor these variables is intensive compared with those required for body weight, food and water consumption, and fecal output. Behavioral assessments in this study were not scored in real time, thus allowing for immediate assessment and response (adjustment of dose, dosing frequency, and others). This type of study does not benefit the individual animal; it can be useful in formulating a comprehensive pain assessment system to be used in future studies of analgesic efficacy.

One major limitation in this study was the lack of a known effective analgesic regimen. There was no statistically significant difference between rabbits receiving saline (control), buprenorphine, transdermal fentanyl, or ketoprofen. This lack of a significant difference between treated and control groups is partially supported by other studies.^8,17,29 The present study involved the placement of telemeters (to gather other physiologic data not presented here) in addition to ovariohysterectomy; this additional procedure could result in more pain or discomfort than that normally associated with a standard ovariohysterectomy. In addition, the small sample size was likely a factor in the lack of a discernible difference between treatment groups. The time of day and duration of observation may also have affected the outcome of this study. The optimal time of day and length of observation for pain in rabbits is not well characterized; however, another study¹⁷ looking at postoperative rabbits found them to be less active in the morning as compared with the afternoon. Because the data show an approximately 80% to 100% reduction in behaviors between 1 and 2 d after surgery, we feel that the study was not adversely affected by the recording of behaviors during the morning hours. Other explanations for the lack of a difference between treatment groups may include inappropriate doses or dosing frequencies, ineffective choices of analgesics, and the inability of our method of assessment to detect differences. Future studies using larger group sizes and alternative analgesics, dosages, and dosing frequencies are still needed. The lack of a robust response to analgesics has been documented in other studies involving postoperative rabbits.^8,17,27,29 Ideally future studies would be extended in duration, to allow time for all of the measured variables to return to baseline values.

Future steps would involve consideration of a multimodal approach with the use of nonsteroidal antiinflammatory drugs, opioids, and local anesthesia at the surgical site combined. Previous studies have documented that a multimodal approach to analgesia can have numerous benefits including improved efficacy.^7,16,26,33 Anecdotal evidence from our facility for rabbits undergoing orthopedic surgeries (anterior cruciate ligament transection) suggests that combination therapy using fentanyl, tramadol, and meloxicam results in better postoperative food consumption, body weight retention, and better overall ‘attitude’ than a single analgesic (fentanyl) does. In addition, consideration needs to be given to altering drug doses or dosing frequency (or both), recognizing that dosing frequencies exceeding 2 times daily are challenging in most facilities. Furthermore the duration of treatment may need to be prolonged, thereby increasing the potential for adverse clinical effects associated with nonsteroidal antiinflammatory drugs (for example, gastrointestinal ulceration, renal damage, hepatotoxicity) and perhaps interfering with research objectives. However, pain also is associated with physiologic disturbances that can affect a research study, and animals that are not eating, drinking, or behaving normally likely are not going to be good research subjects.

In summary, our data suggest that food and water intake, activity, and fecal output may be useful variables in assessing analgesic efficacy. However, none of the 3 analgesics, when used at clinically recommended doses, were able to ameliorate the significant behavioral changes in rabbits after surgery. These data demonstrate the need for additional research in the areas of pain assessment and analgesic efficacy in rabbits.