Using Cageside Measures to Evaluate Analgesic Efficacy in Mice (Mus musculus) after Surgery

Abstract

Recent studies have revealed some of the most frequently used analgesics in mice are not effectively treating postoperative pain. Our laboratory sought to compare and assess the validity and reliability of 2 cageside pain assessments that we recently developed for use in mice—nesting consolidation and grooming transfer tests. We then applied these tests to compare the efficacy of commonly used analgesics—buprenorphine (0.1 mg/kg SC every 12 h for 48 h) and carprofen (30 mg/kg in drinking water for 72 h)—alone and in multimodal combination as a refinement for treating postoperative pain in mice. Briefly, C57BL/6 and CD1 male and female mice underwent assessment under conditions of baseline, anesthesia–analgesia, and laparotomy. Results showed that multimodal analgesia displayed the greatest analgesic coverage over the postoperative period, whereas buprenorphine showed slightly less coverage, and carprofen and saline groups displayed signs of pain at most postoperative time points. After anesthesia–analgesia, buprenorphine and multimodal mice lost significant body weight in the absence of a painful stimulus and displayed other significant drug-related changes. Animals treated with carprofen showed few drug-related changes after anesthesia–analgesia but also demonstrated minimal benefit from postsurgical analgesia. Overall, multimodal analgesia was more effective for treating postsurgical pain in mice than the single-analgesic protocols we tested; however, effects on weight loss need to be considered during analgesic selection. Nesting consolidation and grooming transfer tests were valid and highly reliable over time, in inbred and outbred mice, in male and female mice, under different housing conditions. In addition, the nesting consolidation test had excellent reliability between observers. These findings can be used in refining the detection and treatment of postoperative pain in mice.

Mice are one of the most frequently used species for pain research, where they serve as models for inflammatory, neuropathic, visceral, cancer and postoperative pain.³⁶ Despite the extensive use of mice in pain research, the field of laboratory animal medicine still needs a sensitive, objective, cageside pain assessment method as well as evidence-based recommendations for efficacious analgesia in this species. The goal for analgesic therapy ranges from providing total elimination of pain to achieving a level that is tolerable without compromising wellbeing.³⁶ Historically, evoked measures, which capture reflex responses such as mechanical hypersensitivity by using von Frey filaments, have been the primary approach for assessing changes in nociception through evidence of hyperalgesia or allodynia. These methods focus on nonspecific differences in sensitivity and thus fail to capture the change in affective state or valence, associated with spontaneous pain. Recently, there has been a shift to using spontaneous measures of pain to more accurately understand the degree of pain and assess analgesic efficacy. Spontaneous measures focus on the effect of pain on an animal's performance of natural behaviors or activities, which in mice include burrowing,^22,23 nesting,^12,24,37,41 various locomotor parameters,^1,9,16 and facial grimacing.²⁹ Using these clinically relevant spontaneous measures has increasingly identified deficiencies in some of our commonly used analgesics, previously deemed sufficient according to evoked measures. This disconnect has led to the need to further characterize the efficacy of our most relied-on analgesics and ensure that mouse welfare and the scientific validity of the models they provide are not compromised.

A common category of mouse pain that requires further attention is postoperative pain. Inadequate pain management can lead to distress and negative physiologic consequences, such as immune dysfunction^2,3 and behavioral changes,²⁶ which ultimately not only affect animal welfare but the validity of scientific data.³⁶ In 2017, the International Association for Studying Pain focused their efforts on pain after surgery. One of their recommendations is the adoption of a multimodal approach that combines various types of analgesics and local anesthetics, to reduce dependence on a single mode of treatment.²⁰ According to the National Research Council Committee on Recognition and Alleviation of Pain in Laboratory Animals, multimodal analgesia should be considered for moderate pain and is recommended for severe pain procedures.²² However, few publications report the use of a multimodal approach to analgesia⁶ or evaluate its efficacy in mice.^1,4,10,34,35

Because postoperative pain arises from the activation of various pain pathways, mechanisms, and transmitter systems,¹¹ patients could benefit from the use of a multimodal approach. For example, combining an opioid, which minimizes central hypersensitivity by centrally decreasing nociceptive input into the CNS, with an NSAID, which acts centrally to reduce changes associated with nociceptive input as well as minimizes inflammation at the surgical site,¹¹ potentially will result in more effective pain management. Instead of acting on a single pathway, this practice embodies a multifaceted approach. Current evaluations of single-class analgesics commonly used to control postoperative pain in mice, such as buprenorphine and carprofen, have indicated significant deficiencies related to the analgesics’ durations or dosing. Numerous studies have found that a single injection of standard buprenorphine hydrochloride at a dose within or near our currently used range of 0.05 to 0.1 mg/kg¹¹ provides only partial^22,43,45 to no analgesic coverage^1,25 after various surgical procedures in mice. Similarly, the commonly used NSAID carprofen has shown incomplete coverage for analgesia in postoperative mice.^1,22,32

Given the deficiencies seen with common single-class analgesics, we sought to compare the efficacy of buprenorphine and carprofen, alone and in combination with postoperative analgesics in mice. Because the evaluation of pain in mice is often challenging due to their stoic ‘prey behavior,’³³ we used a combination of assessment measures to ensure that we would capture multiple aspects of pain. We chose to include 2 pain-scoring systems we recently developed—nesting consolidation and grooming transfer tests—in the interest of promoting the growth of practical cageside measures that can be used in real time and with minimal resources. Therefore, the goals of the current study were (1) to determine the best-practice postoperative analgesia protocol for commonly used analgesics, buprenorphine and carprofen, by using multiple postoperative assessments and (2) to establish the validity and reliability of 2 practical, cageside measures for detecting postoperative pain in mice. We hypothesized (1) that a multimodal analgesic protocol would provide superior postoperative analgesia coverage than single-analgesic protocols and (2) that nesting consolidation and grooming transfer test would be valid and reliable measures for postoperative pain in mice. In accomplishing these aims, we seek to develop a refined and practical approach to detecting and treating postoperative pain in mice.

Materials and Methods

This study was approved in advance by the IACUC of the University of Michigan (Ann Arbor, MI) and was conducted in an AAALAC-accredited facility in compliance with the Guide for the Care and Use of Laboratory Animals.¹⁹

Animals.

Main study.

Adult C57BL/6 (n = 47; 22 female, 25 male; age, 6 to 12 wk) and CD1 (n = 52; 22 female, 30 male; age, 7 to 12 wk) mice (Mus musculus) were obtained from Charles River Laboratories (Raleigh, NC). Mice were negative for fur mites, pinworms, and the following viruses: Sendai virus, pneumonia virus of mice, mouse hepatitis virus, minute virus of mice, mouse parvovirus, mouse norovirus, Theiler murine encephalomyelitis virus, reovirus, enzootic diarrhea of infant mice virus, lymphocytic choriomeningitis virus, ectromelia virus, adenovirus, mouse cytomegalovirus, mouse pneumonitis virus, polyoma virus, Hantaan virus, mouse thymic virus, and lactate dehydrogenase-elevating virus. Animals were housed singly or in pairs of the same strain or stock and sex in IVC (model MS7115U54MGPSHR ventilated rack, Allentown Caging, Allentown, NJ) with corncob bedding (Anderson's Bed-O'Cobs, Frontier Distributing, Maumee, OH) in a temperature-controlled room (72 ± 2 °F [22.2 ± 1.1 °C]) on a 12:12-h light:dark cycle with unlimited access to food (Laboratory Rodent 5001, PMI LabDiet, St Louis, MO) and reverse-osmosis–purified water through automatic watering or in water bottles for groups receiving oral carprofen. All mice received cotton squares (Ancare, Bellmore, NY) for standard enrichment; mice housed under the ‘single with nest’ scheme also received an Enviropak (WF Fisher and Son, Somerville, NJ). After arrival, mice were acclimated for a minimum of 3 d prior to experimental manipulation. Body weight was recorded daily during weekdays, and water consumption measured daily by weighing the bottles. A control water bottle kept in a separate cage containing no animals in the same housing rack and removed from the rack and cage at all time points when cages would be manipulated (that is, during each assessment activity) was used to calculate any loss of water from the bottle that was not associated with consumption. This control bottle weight was subtracted from the raw water bottle weights to provide a more accurate measurement of water consumption. Blood and urine were collected from a subset of C57BL/6 and CD1 saline control and carprofen-treated mice at euthanasia and used for evaluating liver and renal changes associated with high-dose carprofen water. A total of 6 CD1 male mice, 4 C57BL/6 males, and 1 C57BL/6 female were removed from the study due to findings of peritonitis at necropsy (n = 2), development of preputial abscesses prior to surgery (n = 4), or severe fighting resulting in wounding (n = 5).

Main study design.

A fractional factorial design of 4 (control and analgesic treatments) × 2 (sex) × 2 (strain/stock) × 3 (housing schemes) × 2 (replicates), for a total of 96 animals, was used to complete this study, similar to others using this approach.^13,41 This method accounted for unwanted variance, resulting in maximized power while reducing sampling size.¹⁵ The data was statistically assessed to confirm that the smallest subgroups did not have relevant differences from one another and could, therefore, be combined to address larger variables such as strain, sex, analgesic treatment or housing. Briefly, both strain/stock and sexes of mice were randomly assigned into one of 4 treatment groups (saline, buprenorphine, carprofen or multimodal) and one of 3 housing schemes (pair, single or single with nest). Interanimal variability was addressed by having each mouse serve as its own control, thus undergoing multiple pain assessment tests at baseline, anesthesia–analgesia, and surgery conditions across multiple time points. The baseline condition did not involve any experimental manipulation beyond performing the assessments. The anesthesia–analgesia condition occurred 1 wk after the baseline condition and involved the mice undergoing anesthesia and receiving their respective analgesics. This condition was included to capture all the nonpainful but potentially behavior-modifying components of surgery, such as anesthesia induction, surgical preparation, anesthetic recovery, and the influence of the analgesic itself. After the anesthesia–analgesia condition, there was a 1-wk washout period before the surgery condition occurred. The surgery condition included all the components of the anesthesia–analgesia condition with the addition of an exploratory laparotomy. By including these 3 conditions, we were able to compare between anesthesia–analgesia and baseline and between surgery and anesthesia–analgesia to evaluate the effects of the anesthetic–analgesic drugs and pain, respectively. Each condition contained the following time points at which the animals underwent pain assessment: hours 1 to 2 (900 to 1100), 3 to 4 (1100 to 1300), 7 to 8 (1500 to 1700), 11 to 12 (1900 to 2100), 23 to 24 (7:00 to 9:00), 35 to 36 (1900 to 2100), and 47 to 48 (7:00 to 9:00). Two time points, at hours 11 to 12 and 35 to 36, occurred during the dark cycle. Administration of the first analgesic treatment for anesthesia–analgesia and surgery conditions was considered the start of that condition. Any additional analgesic treatment doses required during a condition were administered 10 to 30 min prior to starting an assessment. At each time point, mice underwent evaluation for nesting consolidation, grooming transfer, mechanical threshold, and locomotor activity.

Anesthesia–analgesia.

Mice underwent anesthetic chamber induction with 4% to 5% isoflurane (VetOne, Boise, ID) carried in 100% oxygen from medical air tanks at a flow rate of 1 L/min and then were switched to a nose cone and 1.5% to 3% isoflurane for anesthesia maintenance (approximately 10 min). Waste anesthetic gas was scavenged by using activated charcoal (F/AIR canister, Braintree Scientific, Braintree, MA) placed near an exhaust ventilation snorkel. Animals were kept on an external heat source to help maintain appropriate body temperature and underwent surgical preparation, which involved placing ophthalmic lubricating ointment (Puralube Vet Ointment, Dechra, MWI Animal Health, Boise, ID) on eyes, clipping the caudal abdomen from the xyphoid to the pubis, aseptically cleaning the abdomen, and surgical draping. Analgesics were administered shortly after induction. Mice received 1 of 4 analgesic treatments: saline (0.1 mL SC every 12 h for 48 h; 0.9% Sodium Chloride Injection, USP, Hospira, MWI Animal Health, Boise, ID); buprenorphine hydrochloride (0.1 mg/kg SC every 12 h for 48 h; Buprenex Injectable, Reckitt Bencksier Pharmaceuticals, Richmond, VA); carprofen (30 mg/kg in drinking water provided 24 h prior to anesthesia–analgesia or surgery and 48 h postanesthesia and surgery; Rimadyl, Zoetis, Kalamazoo, MI); or multimodal treatment consisting of buprenorphine hydrochloride (0.1 mg/kg SC every 12 h for 48 h) and carprofen (30 mg/kg in drinking water provided 24 h prior to anesthesia–analgesia or surgery and 48 h post-anesthesia and surgery). We chose the twice-daily dosing scheme for buprenorphine because this dosing interval is the one most commonly recommended in laboratory animal text books and formularies^7,39 and, although some evidence indicates that more frequent dosing of buprenorphine is needed to provide effective pain control, whether this scheme is needed when buprenorphine is paired with a NSAID in a multimodal approach is unclear. We kept the twice-daily dosing scheme for the buprenorphine-only group to provide an accurate scientific control for the multimodal group so that we could perform direct comparisons between analgesia groups more easily. We chose for carprofen to be administered in the drinking water, in light of the practicality of this route and because of recent literature showing carprofen's stability in water, ready consumption by mice, and capacity to achieve steady-state drug blood levels.²¹ We selected the dose of carprofen according to our pilot studies, which indicated that despite reaching steady-state levels above a hypothesized therapeutic threshold³¹ after surgery, a dose of 10 mg/kg provided poor analgesic efficacy (Figures 1 through 4), and in light of other literature indicating similar deficiencies in analgesia for postoperative mice.^21,24,32 Mice recovered on 100% oxygen and were returned to their cage once fully ambulatory. All anesthesia events occurred between 0600 to 1100.

Figure 1.

Plasma carprofen concentration (mean ± SEM) in CD1 female mice (n = 4 per time point; age, 8 to 12 wk). Mice were placed on carprofen-treated water (dose, 10 mg/kg) at 24 h before surgery (time 0 h) and were taken off the treated water at 48 h. The dashed line indicates the hypothesized therapeutic threshold (24 µg/mL), which is based on values from dogs.³¹

Figure 4.

Mechanical threshold (mean ± SEM; g) from 32 CD1 female mice (age, 8–12 wk; single housing) that were tested during baseline and surgery periods. Mice were placed on carprofen-treated water (dose, 10 mg/kg) at 24 h before undergoing laparotomy and remained on the treated water until 48 h after surgery. *, Values differ significantly (P < 0.05).

Figure 2.

Nesting consolidation scores (mean ± SEM) from 32 CD1 female mice (age, 8 to 12 wk; single housing) that were tested under baseline and surgery conditions. Mice were placed on carprofen-treated water (dose, 10 mg/kg) at 24 h before undergoing laparotomy and remained on the treated water until 48 h after surgery. Over a 24-h period, mice drank, on average, 7 mL water at baseline and 12 mL carprofen-treated water at surgery. *, Values differ significantly (P < 0.05).

Figure 3.

Grooming transfer scores (mean ± SEM) from 32 CD1 female mice (age, 8–12 wk; single housing) that were tested during baseline and surgery periods. Mice were placed on carprofen-treated water (dose, 10 mg/kg) at 24 h before undergoing laparotomy and remained on the treated water until 48 h after surgery. *, Values differ significantly (P < 0.05).

Surgery.

Mice underwent anesthesia, analgesia, and surgical preparation as described earlier. All surgeries were performed by the same veterinary surgeon (VO) to ensure a consistent stimulus, and all surgeries occurred between 0600 and 1100. A 1.0- to 1.5-cm midline abdominal incision was made through the skin and then extended through the linea alba. To mimic visceral manipulation associated with various surgeries, a sterile cotton-tip applicator was inserted and moved throughout the abdomen for 30 s. The abdominal muscle layer was closed by using 4-0 silk suture in a discontinuous pattern, and the skin was closed with 2 sterile 9-mm skin staples. Except for 2 mice that had peritonitis at necropsy and therefore were removed from the study, no anesthetic or surgical complications occurred, and mice recovered uneventfully.

Pain assessments.

Nesting consolidation test.

At the beginning of each condition, mice housed under the single or pair scheme were presented with half of a cotton square in each corner of a clean cage. Mice housed under the single-with-nest scheme received a clean cage containing 4 half-squares placed in a row at the end containing the automatic watering system and an Enviropak nest at the opposite end of the cage from the cotton half-squares. The mice were given 48 h to consolidate and shred the 4 pieces of cotton square. During each time point, a blinded observer (VO) scored mice from 1 to 5; for the single- or pair-housing schemes, animals received a score of 1 when no cotton pieces had been grouped together; 2 when cotton pieces were paired; 3 when 3 cotton pieces were grouped together; 4 when all 4 cotton pieces were grouped; and 5 when all 4 cotton pieces were grouped and completely shredded (Figure 5). Cotton pieces were considered to be grouped when they were within a 2.5 × 3-in. area. Scores for the single-with-nest housing scheme were defined as: 1, a cotton piece was within 1 in. of the Enviropak; 2, 2 pieces were within 1 in. of the Enviropak; 3, 3 cotton pieces were within a 1 in. of the Enviropak; 4, all 4 cotton pieces were within 1 in. of the Enviropak; and 5, all 4 cotton pieces were within 1 in. of the Enviropak and there was evidence of shredding of the cotton pieces, with incorporation into the crinkle paper of the Enviropak.

Figure 5.

Nesting consolidation test scoring. Mice were pair-housed, individually housed, or individually housed with nests. At the beginning of each scoring period, mice were presented with their respective ‘start’ cage and were scored (range, 1 through 5) according to their progression to consolidate nesting pieces over 48 h. A grouping was defined as cotton square halves that were within a 2.5×3-in. area of each other and was used for the pair-housed and individually housed schemes.

Grooming transfer test.

At the beginning of each condition, 8 μL of nontoxic florescent powder suspended in mineral oil (Glo Germ Oil, Glo Germ, Moab, UT) was applied to the forehead of each mouse by parting the fur between the ears and administering the liquid as close to the skin as possible. The rapidity with which this substance was groomed off was measured at each time point by scoring the mice as follows: 1, a strong fluorescent signal was present at the application site only; 2, the forepaws or hind paws, particularly the nails, fluoresced; 3, the ears fluoresced; 4, the paws and ears no longer fluoresced, and a trace of fluorescence remained at the original application site; and 5, no fluorescence was present (Figure 6). To assess fluorescence, a blinded observer (VO) placed the home cage containing the mice but without its cage top under 3 UV light fixtures (ADJ Products, Los Angeles, CA) situated 15 cm above the top of the cage. All scoring occurred in a procedural space with the lights off and no enhancement of fluorescence.

Figure 6.

Grooming transfer test scoring. At the beginning of the baseline, anesthesia–analgesia, or surgery period, a fluorescent substance was applied to the skin between the ears of a mouse; according to the scoring system outlined above, the fluorescence was assessed under blue light for its removal during grooming over 48 h. The fluorescence in the above pictures was enhanced to facilitate low-light photography. Images were taken under anesthesia and are representative examples for scoring.

Mechanical threshold.

An electronic von Frey probe (IITC Life Science, Woodland Hills, CA) was used to measure mechanical sensitivity at each time point. Mice were placed individually in clear acrylic boxes containing a wire floor through which the probe could pass. Mice were allowed to acclimate to the apparatus for 5 min before measurement. The probe was applied perpendicular to both left and right hindpaws. Each location was assessed 3 times by a blinded observer (VO), with a 2-min washout period between measurements to minimize sensitization. The mean of the left and right measures were compared across mice by using t testing and was determined not to differ. Therefore, the means of the left and right measures were combined (that is, 6 measures averaged for each time point) and then used to calculate each animal's nociceptive mechanical threshold (peak force in grams).

Locomotor activity.

At each time point, mice were placed individually into MouseTrapp (Neurolytical, Ann Arbor, MI) devices containing touch-screen tablets with software to measure general locomotor activity. A blinded observer (VO) recorded the mice during 15-min sessions; the first and last 150 s of each session were excluded from analysis, to minimize changes associated with handling and acclimation. During each session, the total number of touches, total distance traveled, time to step, average speed, percentage of time spent in the inner portion of the tablet, total number of rears, and average rear duration were collected.

Validity and reliability.

To assess the degree to which nesting consolidation and grooming transfer tests measure pain (validity), the change in score was compared between conditions over time. To evaluate reliability, which we defined as the extent to which an assessment is consistent, we recruited 5 observers (3 veterinarians, 1 registered laboratory animal technician, and 1 undergraduate student) to score animals by using the nesting consolidation and grooming transfer tests. The 5 observers received 1 h of didactic training on the 2 tests, followed by multiple practical sessions (2 h total), during which they scored live animals with an expert (VO). Briefly, 10 CD1 (age, 7 to 12 wk) and 10 C57BL/6 (age, 6 to 12 wk) female mice (Charles River Laboratories, Raleigh, NC) with similar health status to the study animals underwent anesthesia–analgesia and surgery conditions as described. These additional animals received carprofen-treated water (as described earlier) during both anesthesia–analgesia and surgery. Observers were blinded to the conditions and scored the mice under 2 settings: (1) cageside scoring, while the cage remained in the IVC rack; and (2) in a biosafety cabinet (that is, flow hood), with the cage top removed. Animals were scored at hours 6 to 8 (1400 to 1600), 22 to 24 (7:00 to 9:00), and 46 to 48 (7:00 to 9:00) for each condition under the 2 settings. To measure fluorescence under these conditions, observers focused a 21-LED UV flashlight (UVL 1006, Glo Germ) on the mouse in the housing room with the room lights on. An additional observer (VO), who was not blinded to the condition, also scored the animals by using the 2 tests at the beginning and end of each time point for each setting to assess whether individual scores drifted over each 2-h scoring period.

Carprofen pilot study.

We obtained 32 CD1 (age, 8 to 12 wk; Charles River Laboratories) adult female mice with similar health status to those described earlier and placed them in single housing. These mice were used prior to the main study to determine the pharmacokinetics and analgesia efficacy of carprofen administered in the drinking water of postoperative mice. Briefly, mice underwent postoperative pain assessments as described earlier under baseline and laparotomy conditions. Mice were placed on 10-mg/kg carprofen-treated water 24 h before surgery and taken off the treated water 48 h after surgery. The mice were divided into 8 groups of 4 animals each, from which blood was collected through terminal cardiocentesis from each group during one of the postsurgery time points: 0 (surgery), 2, 4, 8, 12, 24, 48, and 72 h. The collected blood was placed in heparinized tubes and centrifuged at 1300 × g for 10 min. Plasma was collected aseptically and stored in –80 °C freezer until shipping. Samples were shipped overnight on ice to the UC Cancer Center Pharmacology Shared Resource Laboratory (Colorado State University, Fort Collins, CO) for evaluation of plasma carprofen levels by using liquid chromatography and tandem mass spectrometry of 50 μL of plasma. A hypothesized therapeutic threshold of 24 µg/mL, based on values from dogs,³¹ was used. Pharmacokinetic and analgesia efficacy results from these mice (Figures 1 through 4) were used to determine the dosage of carprofen administered in the drinking water to animals in the main study and are not discussed in detail.

Statistics.

Statistical analysis was completed by using Prism version 7.00 for Windows (GraphPad Software, La Jolla, CA) and MedCalc version 17.4.4 (MedCalc Software, Ostend, Belgium). Two-way repeated-measure ANOVA with Tukey multiple comparison for posthoc testing was performed to compare within-treatment, -strain or -stock, -sex, and -housing schemes effects between anesthesia–analgesia and baseline conditions (drug effect) and between surgery and anesthesia–analgesia conditions (pain effect) across multiple time points for the following measures: body weight loss, water consumption, nesting consolidation test, grooming transfer test, mechanical threshold, and locomotor activities. For pair-housed mice, only one of the animals was included in the analysis for nesting consolidation and water consumption (which was adjusted to reflect that of a single mouse by dividing the value in half), because we assumed that the 2 animals in each pair shared a given result for these measures. The differences between anesthesia–analgesia and baseline (drug effect) and surgery and anesthesia–analgesia (pain effect) was calculated across multiple time points and used to compare between analgesic treatments, strains or stocks, sexes, and housing schemes by using one-way ANOVA or t testing, depending on the number of variables. For these tests, a P value less than 0.05 was considered significant.

Reliability for nesting consolidation and grooming transfer tests was calculated by determining the interrater reliability across the 5 recruited observers and intrarater reliability for the single observer (VO) over the 2-h time course time points by using an intraclass correlation coefficient (ICC) reported as a single and average measure. ICC single measures represented the reliability of ratings for a single observer, whereas the ICC average provided the reliability of different raters grouped together. Interpretation of the ICC was as follows: excellent reliability, 0.75 to 1.0; good, 0.60–0.74; fair, 0.4–0.59; and poor, less than <0.40.⁸ The consistency of cageside and hood settings was measured by using the Cronbach α, where a value greater than 0.7 was considered acceptable.⁵

Results

Analgesic efficacy.

Differences within control and analgesic treatment groups.

Saline.

Mice in the saline control group showed evidence of isoflurane-associated effects over the first 12 h after anesthesia. These effects included significantly (P < 0.05) decreased grooming transfer scores compared with baseline at 8 and 12 h (Figure 7 A), significantly (P < 0.05) decreased mechanical threshold at 2 and 8 h (Figure 8 A), and significant (P < 0.05) decreases at 2 h in the locomotor parameters of total touches, total distance, total rears, and average rear duration (Table 1). There was substantial evidence of pain at all time points after surgery. Pain effects in saline-control animals consisted of significantly (P < 0.05) decreased body weight at 24 and 48 h (Figure 9 A); nesting consolidation scores at 2, 4, 8, and 12 h (Figure 10 A); grooming transfer test scores at 2, 12, 24, 36, and 48 h (Figure 7 A); mechanical threshold at 4 h (Figure 8 A); and average speed at 4 h (Table 1). In addition, a significant (P < 0.05) increase in average speed was detected at 12 h after surgery.

Figure 7.

Grooming transfer test scores (mean ± SEM) under the following conditions: (A) saline, (B) buprenorphine, (C) carprofen, (D) multimodal, (E) C57BL6, (F) CD1, (G) female, (H) male, (I) pair-housed, (J) single-housed, and (K) individually housed with nests during baseline, anesthesia–analgesia, and surgery periods. #, Value differs (P < 0.05) between anesthesia–analgesia and baseline, indicating drug-associated effects; *, value differs (P < 0.05) between surgery and anesthesia–analgesia, indicating pain.

Figure 8.

Mechanical threshold (mean peak force ± SEM; g) under the following conditions: (A) saline, (B) buprenorphine, (C) carprofen, (D) multimodal, (E) C57BL6, (F) CD1, (G) female, (H) male, (I) pair-housed, (J) single-housed, and (K) individually housed with nests during baseline, anesthesia–analgesia, and surgery periods. #, Value differs (P < 0.05) between anesthesia–analgesia and baseline, indicating drug-associated effects; *, value differs (P < 0.05) between surgery and anesthesia–analgesia, indicating pain.

Table 1.

Summary of significant within-group differences in locomotor activity

	Total no. of touches		Total distance (cm)		Time to first step (s)		Average speed (cm/min)		% of time spent in inner 50% of tablet		Total no. of rears		Rear duration (s)
	AA	Sx	AA	Sx	AA	Sx	AA	Sx	AA	Sx	AA	Sx	AA	Sx
Analgesia
Saline	2 h		2 h					4 and 12 h			2 h		2 h
Buprenorphine									2, 4, 12, 24, and 48 h
Carprofen	2 h		2 h								2 h		2 h
Multimodal									4, 8, 12, 24, 36, and 48 h
Strain
C57BL6	12 and 24 h		12 and 24 h					12 h	2, 4, 24, and 48 h		2, 12, and 24 h		12 and 24 h
CD1	2 h		2 h				48 h		12, 24, 36, and 48 h		2 h		2 h
Sex
Female	2 h		2 h						4, 24, 36, and 48 h		2 h		2 h
Male							48 h		4, 12, 24, 36, and 48 h
Housing
Pair	2 h		2 h		12 h				4, 12, 24, 36, and 48 h		2 h		2 h
Single									24 h
Single with nest									24 h

The time points at which significant (P < 0.05) differences in locomotor activity were present within a group between anesthesia–analgesia and baseline (AA) or between surgery and anesthesia–analgesia (Sx) are indicated.

Figure 9.

Body weight loss (mean ± SEM; g) under the following conditions: (A) saline, (B) buprenorphine, (C) carprofen, (D) multimodal, (E) C57BL6, (F) CD1, (G) female, (H) male, (I) pair-housed, (J) single-housed, and (K) individually housed with nests during baseline, anesthesia–analgesia, and surgery periods. #, Value differs (P < 0.05) between anesthesia–analgesia and baseline, indicating drug-associated effects; *, value differs (P < 0.05) between surgery and anesthesia–analgesia, indicating pain.

Figure 10.

Nesting consolidation scores (mean ± SEM) under the following conditions: (A) saline, (B) buprenorphine, (C) carprofen, (D) multimodal, (E) C57BL6, (F) CD1, (G) female, (H) male, (I) pair-housed, (J) single-housed, and (K) individually housed with nests during baseline, anesthesia–analgesia, and surgery periods. #, Value differs (P < 0.05) between anesthesia–analgesia and baseline, indicating drug-associated effects; *, value differs (P < 0.05) between surgery and anesthesia–analgesia, indicating pain.

Buprenorphine.

Animals treated with buprenorphine showed signs of anesthesia–analgesia-associated effects throughout the 48-h time period. Compared with baseline values, mice had significantly (P < 0.05) decreased nesting consolidation scores at 2 h (Figure 10 B); significantly (P < 0.05) decreased grooming transfer scores at 2, 4, 8, and 12 h (Figure 7 B); and a significant (P < 0.05) decrease in the percentage of time in the inner portion of the tablet at 2, 4, 12, 24, and 48 h (Table 1). Pain effects detected after surgery included significant (P < 0.05) decreases in nesting consolidation scores at 2 h (Figure 10 B) and in grooming transfer test scores at 12, 24, and 36 h (Figure 7 B).

Carprofen.

Anesthesia–analgesia effects for carprofen-treated mice were noted within the first 12 h. These effects included significant (P < 0.05) decreases in grooming transfer scores at 8 and 12 h (Figure 7 C); in mechanical threshold at 4 h (Figure 8 C); and in total touches, total distance, total rears, and average rear duration at 2 h (Table 1). Pain effects were present at most time points after surgery and included significant (P < 0.05) decreases in nesting consolidation scores at 4 h (Figure 10 C) and in grooming transfer test scores at 2, 8, 12, 24, and 36 h (Figure 7 C). Mice given the carprofen-treated water showed significantly (P < 0.05) increased water consumption compared with baseline (Table 2), where mean consumption averaged approximately 16 mL within the first 24 h after both anesthesia–analgesia and surgery conditions, compared with an average of approximately 5 mL at baseline. No differences in water consumption between surgery and anesthesia–analgesia conditions were noted for carprofen- or multimodal-treated animals.

Table 2.

Water consumption (mL; mean ± SEM) of mice treated with carprofen in their drinking water

		Baseline			Anesthesia–analgesia			Surgery
		24 h	48 h		24 h	48 h		24 h	48 h
Analgesia	Carprofen (n = 17)	5.1 ± 0.9	6.4 ± 0.8		16.0 ± 1.7	7.6 ± 0.5		16.1 ± 1.1	6.6 ± 0.4
Multimodal (n = 17)	4.8 ± 0.7	5.6 ± 0.5		15.0 ± 1.2	6.1 ± 0.3		14.8 ± 1.5	5.3 ± 0.6
Strain	C57BL6 (n = 16)	3.0 ± 0.5	4.0 ± 0.2		12.9 ± 1.2	6.1 ± 0.4		13.4 ± 1.3	5.0 ± 0.5
CD1 (n = 18)	6.7 + 0.9	7.7 ± 0.7		17.0 ± 1.4	7.3 ± 0.5		17.0 ± 1.2	6.6 ± 0.6
Sex	Female (n = 16)	5.2 ± 0.9	5.5 ± 0.6		14.8 ± 1.3	6.4 ± 0.4		15.3 ± 1.0	5.6 ± 0.4
Male (n = 18)	4.7 ± 0.8	6.6 ± 0.8		16.0 ± 1.6	7.2 ± 0.5		15.4 + 1.6	6.2 ± 0.7
Housing	Pair (n = 8)	5.3 ± 0.8	5.5 ± 0.8		14.1 ± 1.5	6.5 ± 0.5		13.5 + 1.1	5.1 ± 0.4
	Single (n = 13)	4.6 ± 0.8	6.4 ± 0.5		16.3 ± 1.4	7.1 ± 0.4		17.1 ± 1.3	6.7 ± 0.6
Single with nest (n = 13)	5.1 ± 0.3	6.0 ± 0.2		16.0 ± 0.5	7.2 ± 0.2		16.3 ± 0.4	6.2 ± 0.1

Carprofen-treated water was administered during anesthesia–analgesia and surgery periods. Water consumption measured from pair-housed mice was divided in half, and only one of the pair was used in statistical analysis

Multimodal analgesia.

Anesthesia–analgesia effects for multimodal-treated mice occurred at all time points. These effects consisted of significant (P < 0.05) decreases in body weights at 24 and 48 h (Figure 9 D); in grooming transfer scores at 2, 8, 12, 24, and 36h (Figure 7 D); and in the percentage of time animals spent in the inner portion of the tablet at 4, 8, 12, 24, 36, and 48h (Table 1). Only the grooming transfer test detected changes associated with pain, identifying significant (P < 0.05) score decreases at 12, 24, and 36 h after surgery (Figure 7 D). No differences in weight loss occurred between mice in the surgery and anesthesia–analgesia groups.

Differences between control and analgesic-treatment groups.

Differences between control and analgesic treatment groups were analyzed across the various assessments by comparing between anesthesia–analgesia and baseline values (drug effect) and between surgery and anesthesia–analgesia data (pain effect). This analysis was performed to compare the relative effects and efficacy of the analgesics. However, comparison between treatment groups should be interpreted carefully, because interanimal variability is no longer controlled, thus decreasing sensitivity.

When we compared the drug effects between control and analgesic treatment groups, mice in the buprenorphine and multimodal treatment groups had lost significantly (P < 0.05) more body weight at 24 h after anesthesia–analgesia than carprofen and saline groups, with animals in the multimodal group losing more than 4 times the weight of the other 3 treatment groups. In addition, the buprenorphine and multimodal groups had significantly (P < 0.05) more total touches and greater total distance at 2 h after anesthesia–analgesia than did the carprofen and saline. Carprofen-treated mice had significantly (P < 0.05) fewer total rears at 2 h after anesthesia–analgesia than the buprenorphine and multimodal groups. Mice given saline or carprofen had significantly (P < 0.05) shorter rear duration at 2 h after anesthesia–analgesia than those in the buprenorphine and multimodal groups. Comparison of differences associated with pain effects revealed that saline-treated mice displayed a significantly (P < 0.05) greater decrease in nesting consolidation scores at 8 h after surgery than animals given multimodal treatment. Compared with multimodal analgesia, carprofen led to a significant (P < 0.05) decrease in total distance traveled at 2 h after surgery. Saline-treated mice had a significant (P < 0.05) decrease in average speed at 12 h after surgery compared with those given buprenorphine. No significant differences of conditions were noted between analgesic groups for mechanical threshold, grooming transfer test, percentage of time spent in the inner portion of the tablet, or time to first step.

Strain- and stock-associated effects.

Differences within strain and stock.

Both outbred CD1 and inbred C57BL/6 mice showed evidence of anesthesia–analgesic effects throughout the 48-h time period. These drug effects included significant (P < 0.05) decreases in nesting consolidation scores during early time points (Figure 10 E and F) and decreased mechanical threshold at 2 h (Figure 8 E and F) for both CD1 and C57BL/6 after anesthesia–analgesia. In addition, C57BL/6 mice had significantly (P < 0.05) decreased grooming transfer scores at 2, 4, 8, and 12 h compared with baseline, whereas CD1 showed significant (P < 0.05) decreases at 2, 8, and 12 h (Figure 7 E and F) after anesthesia–analgesia. Drug effects detected in the locomotor activity of C57BL/6 mice included significant (P < 0.05) increases at 12 and 24 h for total touches and total distance; significant (P < 0.05) decreases in total rears and rear duration at 12 and 24 h; a significant (P < 0.05) decrease in the percentage of time spent in the inner portion of the tablet at 2, 4, 24, and 48 h; and a significant (P < 0.05) decrease in total rears at 2 h (Table 1). For CD1 mice, drug effects on locomotor activity included significant (P < 0.05) decreases in total touches, total distance, total rears, and rear duration at 2 h; a significant (P < 0.05) decrease in average speed at 48 h; and significant (P < 0.05) decreases in the percentage of time spent in the inner portion of the tablet at 12, 24, 36, and 48 h.

Similar to drug effects, pain effects were present in both strain and stock throughout most of the 48-h time period. C57BL/6 mice had significant (P < 0.05) decreases in body weight at 24 and 48 h after surgery compared with anesthesia-analgesia, whereas CD1 mice did not display any significant weight loss (Figure 9 E and F). Other pain effects noted included decreases in nesting consolidation scores at 2, 4, and 8 h for C57BL/6 mice and at 2 and 4 h for CD1 mice (Figure 10 E and F); significant (P < 0.05) decreases in grooming transfer test scores at 2, 8, 12, 24, 36, and 48 h for C57BL/6 mice and at 12, 24, 36, and 48 h for CD1 animals (Figure 7 E and F); a significant (P < 0.05) decrease in mechanical threshold for C57BL/6 mice at 4 h (Figure 8 E); and a significant (P < 0.05) increase in average speed at 12 h for C57BL/6 mice (Table 1).

Interstrain–stock differences.

When we compared the differences associated with drug effects across strain and stock, CD1 animals had 1/3 fewer (P < 0.05) total touches at 2 h and no more than half the total distance (P < 0.05) at 2, 36, and 48 h than C57BL/6 animals. Regarding differences associated with pain, C57BL/6 mice had significantly (P < 0.05) lower grooming transfer test scores at 2 h than CD1 mice. No significant differences were noted between the strain and stock in regard to body weight loss, nesting consolidation, mechanical threshold, average speed, percentage of time spent in the inner portion of the tablet, time to first step, total rears, or rear duration.

Sex-associated effects.

Within-sex differences.

Both sexes showed evidence of drug effects at all time points after anesthesia–analgesia. These drug effects did not include significant decreases in weight compared with baseline (Figure 9 G and H) but did include significant (P < 0.05) decreases in nesting consolidation scores at 2 h in males (Figure 10 H) and significant (P < 0.05) decreases in grooming transfer scores at 2, 4, 8, and 12 h for females and at 8, 12, and 24 h for male mice (Figure 7 G and H). In addition, other drugs effects such as decreases in mechanical threshold at 2 and 8 h for females and 2 h in males (Figure 8 G and H), as well as various locomotor changes including significant (P < 0.05) decreases in total distance, total touches, total rears, and rear duration at 2 h for females; significant (P < 0.05) decreases in percentage of time spent in the inner portion of the tablet at 4, 24, 36, and 48 h for females and 4, 12, 24, 36, and 48 h for males; and significant (P < 0.05) decreases in average speed at 48 h in males (Table 1). Changes associated with pain effects were noted in both sexes at most time points. Male mice showed a significant (P < 0.05) decrease in body weight after surgery at 48 h, compared with anesthesia–analgesia (Figure 9 H). Both sexes showed significant (P < 0.05) decreases in nesting consolidation scores at 2, 4, and 8 h (Figure 10 G and H) and significant (P < 0.05) postsurgery decreases in grooming transfer scores at 2, 8, 12, 24, and 36 h for females and 12, 24, 36, and 48 h for males (Figure 7 G and H). In addition, the mechanical threshold was significantly (P < 0.05) lower in females at 4 h after surgery, compared with anesthesia–analgesia (Figure 8 G).

Between-sex differences.

Comparing the differences associated with drug effects between the 2 sexes revealed that male mice had a significantly (P < 0.05) lower average speed than females at 48 h. None of the other assessments showed significant differences between sexes.

Housing-associated effects.

Differences within housing condition.

Changes associated with drug effects were commonly noted within the first 24 h after anesthesia–analgesia. Drug effects of paired mice included significantly(P < 0.05) decreased nesting consolidation scores at 2 and 4 h compared with baseline (Figure 10 I); significantly (P < 0.05) decreased grooming scores at 4, 8, and 12 h (Figure 7 I); significant (P < 0.05) decreases in total touches, total distance, total rears, and rear duration at 2 h; significant (P < 0.05) decreases in time to step at 12 h; as well as significant (P < 0.05) decreases at 4, 12, 24, 36, and 48 h in the percentage of time spent in the inner portion of the tablet (Table 1). Single-housed animals demonstrated drug effects including a significant (P < 0.05) decrease in grooming transfer scores at 8 and 12 h (Figure 7 J) and a decrease in the percentage of time spent in the inner portion of the tablet at 24 h (Table 1). Animals that were housed in the single-with-nest scheme showed drug effects such as a significant (P < 0.05) decrease in grooming transfer scores at 8 and 12 h (Figure 7 K); significant (P < 0.05) reduction in mechanical threshold at 2 h (Figure 8 K); as well as significantly (P < 0.05) decreased percentage of time spent in the inner portion of the tablet at 24 h (Table 1). Pain effects were noted throughout the 48-h period for all housing schemes. Pain effects in paired animals consisted of significant (P < 0.05) decreases in nesting consolidation scores at 2 h (Figure 10 I) and significant (P < 0.05) decreases in grooming scores at 12, 24, 36, and 48 h (Figure 7 I). Single-housed animals without nests had significantly (P < 0.05) decreased nesting consolidation scores at 2, 4, 8, and 12h (Figure 10 J); significantly (P < 0.05) decreased grooming transfer scores at 2, 12, 24, 36, and 48 h (Figure 7 J); and significantly (P < 0.05) decreased mechanical threshold at 4 h (Figure 8 J). Mice singly housed with nests displayed signs indicative of pain including significant (P < 0.05) decreases in nesting consolidation scores at 2, 4, 8 and 12h (Figure 10 K); significant (P < 0.05) decreases in grooming transfer scores at 2, 8, 12, 24, and 36 h (Figure 7 K), as well as significant (P < 0.05) decreases in mechanical threshold at 12 h (Figure 8 K).

Differences between housing conditions.

There were no significant differences noted between the housing schemes in any of the assessments.

Validation and reliability of nesting consolidation and grooming transfer tests.

We assessed the validity of nesting consolidation and grooming transfer tests by using construct validity and evaluating the extent to which the scores differed between conditions. For the nesting consolidation test, scores were significantly (P < 0.05) different between anesthesia–analgesia and baseline (drug effect) at 2 h as well as between the anesthesia–analgesia and surgery conditions (pain effect) throughout the first 12 h (Figure 11). The grooming transfer test scores demonstrated significant (P < 0.05) drug effects throughout the first 24 h as well as pain effects at 2, 8, 12, 24, 36, and 48 h (Figure 12).

Figure 11.

Validity of nesting consolidation test by comparison of changes in scores (mean ± SEM) of all mice (n = 88) over time during baseline, anesthesia–analgesia, and surgery periods. #, Value differs (P < 0.05) between anesthesia–analgesia and baseline, indicating drug-associated effects; *, value differs (P < 0.05) between surgery and anesthesia–analgesia, indicating pain.

Figure 12.

Validity of grooming transfer test by comparison of changes in scores (mean ± SEM) of all mice (n = 88) over time during baseline, anesthesia–analgesia and surgery periods. #, Value differs (P < 0.05) between anesthesia–analgesia and baseline, indicating drug-associated effects; *, value differs (P < 0.05) between surgery and anesthesia–analgesia, indicating pain.

The consistency and reliability of these tests were assessed by measuring the Cronbach α between cageside and inflow hood observations and determining the ICC between multiple observers as well as for one expert rater over time. Internal consistency was acceptable (0.96) between cageside and hood observations for the nesting consolidation test. Interrater reliability for nesting consolidation was excellent under both cageside and inflow hood observations, where single–measure ICC was 0.81 (range, 0.73–0.87) and 0.86 (0.79–0.90) and the average-measure ICC was 0.96 (0.93–0.97) and 0.97 (0.95–0.98), respectively (Table 3). In addition, the scores assigned by the expert rater at the beginning and end of each time point displayed excellent intrarater reliability with single-measure ICC of 0.83 (0.76–0.88) and 0.89 (0.85–0.93) and average-measure ICC of 0.91 (0.87–0.93) and 0.94 (0.92–0.96) for cageside and inflow hood observations, respectively. The grooming transfer test showed acceptable internal consistency (0.86) between cageside and inflow hood observations. Interrater reliability was fair to poor, with single-measure ICC of 0.48 (0.37–0.59) and 0.39 (0.28–0.50) for cageside and flow-hood observations, respectively (Table 3). The average-measure ICC for the grooming transfer test were considered excellent at 0.82 (0.75–0.88) for cageside and 0.76 (0.66–0.84) for the inflow hood observations. The grooming transfer test scores assigned by the expert rater at the beginning and end of time points had excellent intrarater reliability single measures of 0.80 (0.73–0.86) and 0.91 (0.88–0.94) and average measures of 0.89 (0.84–0.92) and 0.95 (0.93–0.97) for cageside and inflow hood observations, respectively.

Table 3.

Inter- and intrarater reliability of nesting consolidation and grooming transfer tests under cageside and hood conditions

		No. of observers	ICC single (95% CI)	ICC average (95% CI)
Nesting consolidation test	Cageside	5	0.81 (0.73–0.87)	0.96 (0.93–0.97)
	1	0.83 (076–0.88)	0.91 (0.87–0.93)
	Hood	5	0.86 (0.79–0.90)	0.97 (0.95–0.98)
	1	0.89 (0.85–0.93)	0.94 (0.92–0.96)
Grooming transfer test	Cageside	5	0.48 (0.37–0.59)	0.82 (0.75–0.88)
	1	0.80 (0.73–0.86)	0.89 (0.84–0.92)
	Hood	5	0.39 (0.28–0.50)	0.82 (0.75–0.88)
	1	0.91 (0.88–0.94)	0.95 (0.93–0.97)

Intraclass correlation coefficient (ICC) calculated for single and multiple (average) raters.

Toxicology analysis of carprofen-treated water.

The subset of animals that received carprofen-treated water and that were analyzed for toxicities showed no abnormalities at necropsy. C57BL/6 female mice on carprofen-treated water had mild elevations in creatinine and BUN (Table 4), and CD1 male mice on carprofen-treated water showed a mild increase in BUN. In both lines of female mice, total bilirubin was lower in those given carprofen-treated water than in the saline-control group. Urinalysis of carprofen-treated female C57BL/6 and CD1 mice did not reveal any abnormalities (Table 5).

Table 4.

Mean blood chemistry results (mean ± SEM) from saline- and carprofen-treated mice

	Sex	Treatment	Creatinine (mg/dL)	BUN (mg/dL)	AST (U/L)	Albumin (g/dL)	ALT (U/L)	Total bilirubin (mg/dL)	ALP (U/L)
C57BL6	Female (n = 2)	Saline	0.3 ± 0.08	27.5 ± 1.5	322.5 ± 22.5	3.3	59.5 ± 22.5	0.2 ± 0.05	91.5 ± 4.5
	Male (n = 3)	Saline	0.3 ± 0.07	22.7 ± 0.3	224.3 ± 78.4	3.3 ± 0.1	58.7 ± 5.8	0.2 ± 0.07	70.3 ± 5.4
	Female (n = 11)	Carprofen	0.5 ± 0.04	31.3 ± 1.7	409.1 ± 35.8	3.2 ± 0.08	70 ± 3.3	0.04 ± 0.02	100.2 ± 6.0
Male (n = 3)	Carprofen	0.3 ± 0.06	29.7 ± 1.5	385.7 ± 87.9	3.3 ± 0.6	81 ± 21.8	0.2 ± 0.05	67 ± 2.5
CD1	Female (n = 1)	Saline	0.15	22	230	3.7	53	0.2	71
	Male (n = 2)	Saline	0.2 ± 0.02	29 ± 3	353 ± 108	3.3 ± 0.05	65.5 ± 9.5	0.3 ± 0.07	80 ± 22
	Female (n = 17)	Carprofen	0.4 ± 0.04	26.1 ± 1.2	324.3 ± 43.1	3.3 ± 0.06	64.2 ± 4.2	0.06 ± 0.02	88.9 ± 5.8
Male (n = 2)	Carprofen	0.3 ± 0.02	35 ± 10	341 ± 140	3.25 ± 0.05	56 ± 9	0.2 ± 0.005	76.5 ± 1.5

Mice in the carprofen group received 30 mg/kg carprofen in their drinking water for 72 h during the anesthesia–analgesia and surgery conditions.

Table 5.

Summary of urinalysis results for female mice given carprofen-treated water

	Color	Clarity	Specific gravity	Glucose	Protein
C57BL6 (n = 7)	straw yellow	Clear	1.025	Negative	1+ to 2+
CD1 (n = 6)	straw yellow	Clear	1.033	Negative	1+ to 2+

Pooled samples from 3 or 4 mice were analyzed by using urine test strips.

Discussion

According to our postoperative pain assessments, analysis within the saline-control and analgesia-treatment groups revealed that mice treated with a multimodal regimen have the most analgesic coverage over the 48 h after surgery, whereas buprenorphine as a single agent showed slightly less coverage, and the carprofen and saline-control groups indicated pain throughout most, if not all, of the postoperative time points. These findings are in agreement with our hypothesis. Despite being the superior analgesic of the treatments tested, multimodally treated animals still showed evidence of pain in the grooming transfer test from 12 through 36 h after surgery. A previous study using buprenorphine and carprofen after mammary fat-pad removal similarly found incomplete analgesic coverage, according to decreased postoperative wheel running activity and visual assessment scoring.¹ In addition, our study found multimodal analgesia to have significant drug effects throughout the 48-h time period after anesthesia–analgesia. Of most concern is its association with a significant decrease in body weight—approximately 1.5 g—at 24 and 48 h after anesthesia–analgesia, independent of a pain stimulus. Another study using a combination of buprenorphine and NSAID for postoperative pain management similarly found reductions in body weight in the absence of an incisional wound or pain.⁴ Although this weight loss after multimodal analgesia is an important consideration, it is equally important to note that it was no greater after surgery nor in comparison to the saline control or any other treatment groups after surgery. In addition, although repeated anesthetic events might have compounded postsurgical responses, all groups underwent the same number and sequence of anesthetic exposures and postsurgical changes. In addition, postanesthetic changes were minimized with the appropriate use of analgesia, indicating that pain was the main driver of postsurgical perturbations in the metrics assessed. Single-agent buprenorphine showed similar efficacy as multimodal treatment, with detection of pain at 12 to 36 h after surgery according to grooming transfer testing, as well as at the 2-h time point, where pain was detected by decreased nesting consolidation scores. We suspect that, in light of the time points involved, the increased pain associated with buprenorphine compared with multimodal treatment may coincide with the 12-h dosing interval, which has previously been suggested to be too long.^23,25 Perhaps this outcome could be improved through more frequent dosing, using a sustained-release formulation as recommended by recent studies,^23,27 or the addition of carprofen for multimodal coverage. Similar to the multimodal group, mice treated with buprenorphine showed drug effects at most of the time points after anesthesia–analgesia. Interestingly, weight loss did not occur after anesthesia–analgesia in the buprenorphine group, as was noted with multimodal analgesia. Reports of buprenorphine-induced postoperative weight loss vary. One study reported no significant changes in postoperative weight loss in mice administered 0.1 mg/kg buprenorphine after surgery,⁴³ whereas others have noted decreases postoperatively after 0.1-mg/kg^25,27 and 2-mg/kg¹⁴ dosages. In our case, perhaps the additive effects of buprenorphine and carprofen result in additional drug-effect changes to what is seen in the mice that received buprenorphine only, potentially leading to the increased weight loss.

Carprofen-treated and saline-control mice showed the greatest evidence of pain throughout most, if not all, of the postoperative time points. Pain was detected at 2 through 36 h in carprofen-treated mice, whereas saline-control animals showed indications of pain throughout the 48h. In addition, saline-control animals lost approximately 1g of body weight after surgery, which was double their body weight lost after anesthesia–analgesia alone. Carprofen-treated animals did not display any significant changes in body weight throughout the 3 conditions, but body weight in the carprofen group showed greater variability compared with the other treatment groups. The reason for higher variability in body weights for this analgesic treatment group is unknown. As expected, saline did not result in many drug effects after anesthesia alone, with significant changes noted only within the first 12 h. Similarly, carprofen exerted drug effects during the first 12 h after anesthesia–analgesia. Mice in the carprofen group readily drunk the carprofen-treated water, especially within the first 24 h of access. Similar to another study that used carprofen-treated water in mice,²¹ our animals consumed more of the carprofen-treated water than untreated water. We suspect this increased consumption is due to the palatability of the drug, but further testing is needed to confirm this hypothesis. Back-calculation of the drug dose according to the mean daily water consumption of 16 mL after 24 h and 6.9 mL after 48 h reveals that a 25-g mouse would have received 3.2 mg and 1.4 mg of carprofen, respectively. This dose is approximately 2 to 4 times higher than the 0.75 mg expected from a 30-mg/kg dose. This finding suggests, despite using a much higher drug dose than the typical 5 mg/kg and exceeding the hypothesized therapeutic threshold plasma levels (Figure 1), single-agent carprofen may not be a suitable analgesic for the management of pain after laparotomy.³¹ Other studies have noted analgesic deficiencies of carprofen in mice at dosages above 5 mg/kg. One study using the mouse grimace scale found that carprofen produced significant reductions in spontaneous pain after laparotomy only when dosed at 20 or 25 mg/kg.³² Nesting complexity measures in mice that underwent sham embryo transfer surgery showed no significant difference between animals treated with 50 mg/kg carprofen and the saline-control group.²⁴ Despite the increased dose of carprofen that we used, no toxic changes were seen on gross necropsy, liver chemistries, or urinalysis of either strain or stock. Mild changes in BUN and creatinine levels occurred within both lines of mice, and further investigation including renal histology should be performed to determine whether these changes are biologically significant.

Comparing the differences between conditions among the analgesic treatment groups revealed that the mice in the buprenorphine and multimodal groups lost more weight after anesthesia–analgesia and scored higher regarding locomotor activities after anesthesia–analgesia and surgery than did the carprofen and saline-control animals. This increase in locomotor activity can likely be attributed to buprenorphine's ability to cause hyperactivity in mice, which has been noted in previous studies.^17,18,40 Unlike for buprenorphine and multimodal treatment, the drug effects of carprofen did not differ from those of the saline control group, suggesting that carprofen had little behavioral effect that might alter pain assessments. Compared with saline controls, mice in the multimodal treatment group had fewer changes in nesting consolidation scores between surgery and anesthesia–analgesia at early postoperative times, which likely is attributable to its early analgesic coverage.

Analysis of the differences within and between the strain and stock of mice revealed only minor changes. Within strain and stock, both C57BL/6 and CD1 mice showed drug effects throughout the 48 h after anesthesia–analgesia. Direct comparison of condition differences between strain and stock showed that CD1 mice had greater decreases in the locomotor parameters total touches and total distance after anesthesia–analgesia, whereas C57BL/6 mice had lower grooming scores immediately after surgery. C57BL/6 mice were painful throughout the 48-h postoperative period, whereas CD1 mice were painful at all time points except 8 h. Other spontaneous pain assessment tests such as the time-to-integrate-to-nest test⁴¹ and mouse grimace score³³ at baseline and the locomotor activity⁴² and behavioral ethogram scales⁴⁵ after surgery have reported differences in scoring between strains and stocks. The minor differences between C57BL/6 and CD1 mice used in our study are still important to take into account when evaluating for drug and pain effects within a given assessment. However, they did not appear to change the overall trend in drug or pain effects seen between strain and stock. These subtle differences highlight the importance of performing baseline measures to gain an appreciation of how the animals score in the absence of factors that can alter their performance, such as anesthesia, analgesic drugs, and surgical pain.

Only minor differences in changes within and between sexes were noted. Within groups of female and male mice, drug effects were noted throughout the 48-h period after anesthesia–analgesia. A direct comparison of the changes of conditions between the 2 sexes revealed only that males had a greater decrease in average speed at 48 h after anesthesia–analgesia. Subtle differences between sexes have been noted in some spontaneous pain assessments, such as the mouse grimace scale, with C3H and CD1 mice but not C57BL/6 showing differences,³³ and no between-sex differences for multiple strains in the time-to-integrate-to nest test.⁴¹ Again, confirmation of baseline measures between sexes appears to be important for establishing valid detection of drug and pain effects.

Housing similarly showed minimal differences within and between groups. Within housing groups, paired mice showed evidence of drug effects throughout the 48 h, whereas single-housed and single-with-nest animals showed drug effects only within the first 24 h after anesthesia–analgesia. Looking at pain effects within the 3 housing schemes, mice showed evidence of pain throughout most, if not all, postoperative time points. Although comparing between housing groups did not reveal any significant differences for any of the assessments in our study, other studies have noted a more profound effect of social housing on assessments such as locomotor activity³⁸ and time-to-integrate-to-nest test scores.⁴¹ Therefore, similar to strain–stock- and sex-associated differences, we emphasize the importance of validating these assessments by using different housing schemes.

This study used 2 cageside postoperative pain assessments, nesting consolidation and grooming transfer tests. We wanted to validate their use for detecting postoperative pain in mice of different strains and stocks, sexes, and housing conditions. According to construct validity, both assessments follow the predicted time course of showing a significant decrease in scores after anesthesia–analgesia and surgery conditions, with the surgery condition showing the greatest decrease. The time course of pain detection for the nesting consolidation test returned to baseline levels by 24 h, whereas differences in scores after surgery for the grooming transfer test were detectable throughout the 48-h postoperative period. This finding is likely due to the 2 assessments measuring different levels or aspects of pain, where the task of consolidating nesting material potentially requires fewer movements and thus provoke less pain after abdominal surgery than the more vigorous movements associated with assuming different body postures for complete grooming. Alternatively, animals might have a greater drive to build a nest for survival purposes rather than groom themselves thoroughly. Further work with these 2 assessments needs to be performed using different surgeries or models of surgical pain to learn more about the differences of pain detection within and between these scoring systems.

Another goal of this study was to assess the reliability of the nesting consolidation and grooming transfer tests. The nesting consolidation test showed excellent reliability between different observers, conditions, and time. In contrast, the grooming transfer test had excellent reliability between conditions and time, indicating that pain and its alleviation could be detected easily, but poor reliability between raters, suggesting that the test works best for postoperative assessment when the same person scores a set of animals over time rather than when several different raters score the same animals at different times. This outcome is likely a reflection of the greater subjectivity of this assessment, and additional user training might help observers to produce similar scores. Compared with the nesting consolidation test, the grooming transfer scores can be influenced by the technical skill of shining a flashlight on a conscious, moving mouse at the correct angle to obtain fluorescence. For example, we found that it was more technically challenging to detect traces of the fluorescent signal on the skin of the head of C57BL/6 mice, which have opaque black fur, than on CD1 mice, which have white translucent fur and for which eliciting a visible fluorescent signal was easier. In addition, unlike the nest consolidation scores, the grooming transfer scores change along a continuous rather than discrete scale, thus also potentially contributing to variability between observers. These factors ultimately suggest that additional training and clearer distinctions between scores likely would be beneficial for improving interrater reliability in the grooming transfer test.

The performance of these 2 cageside pain assessments appeared to be superior to the other assessments used. Both the nest consolidation and grooming transfer tests demonstrated greater sensitivity in identifying pain, with the nest consolidation test identifying pain throughout 12 h after surgery and the grooming transfer test detecting changes throughout the 48-h postoperative period. Interestingly, mechanical threshold, a common method for assessing pain in rodents, appeared to be a poor indicator of pain. Mechanical threshold detected pain only in saline-control animals at 4 h. Given that the expected time course of pain after laparotomy in mice is expected to last approximately 48 h,³² mechanical threshold grossly underestimates postoperative pain, because it measures changes in nociception, and therefore should not be used for drawing conclusions regarding the clinical detection of pain. Similar to other studies,^14,25,27 body weight was not a reliable indicator of pain in our current study, because changes are nonspecific and can be associated with other variables, such as drugs, as occurred in our multimodal group. The evaluation of locomotor activity was not particularly sensitive in detecting changes associated with pain in our study. Only 1 of the 7 measured locomotor activity parameters, average speed, had an effect in the saline-control animals at a single time point after surgery. Most locomotor activity parameters, however, did detect changes associated with drug effects after the anesthesia–analgesia condition.

A limitation to this work was the brief acclimation period of 3 d, which might have contributed to some of the variation in our results. Most publications in which stress after shipping was assessed recommend a minimum of 24 to 48 h after transportation before research use, given that changes in immune function and body weight may require as long as 2 d to resolve.^28,30 We, therefore, chose a minimum of 72 h to provide an even greater buffer from transportation stress, but behavioral studies have been known to use acclimation periods of several weeks.⁴⁴

In summary, our study shows that a multimodal analgesic approach provides the most efficacious postoperative analgesic regimen in mice under the conditions tested. However, analgesic coverage through multimodal methods still warrants improvement, and the tendency to cause significant weight loss independent of a painful surgical stimulus is important to consider. Oral carprofen does not appear to be appropriate as a single agent to treat pain after laparotomy in mice, because higher-than-recommended dosages still provide little analgesic benefit. Future work on postoperative analgesia in mice would benefit from exploring the use of extended- or sustained-release formulations of buprenorphine and determining whether substituting these formulations might improve pain alleviation from both multimodal and buprenorphine-only regimens. The nesting consolidation and grooming transfer tests were found to be valid and practical cageside assessments that provide robust pain measures independent of mouse strain or stock, sex, or housing variables. Reliability can be improved, particularly for the grooming transfer test, with additional attention to user training. We recommend further validation through future application of these cageside pain assessments to different models of postoperative pain.