Comparison of Nociceptive Effects of Buprenorphine, Firocoxib, and Meloxicam in a Plantar Incision Model in Sprague–Dawley Rats

Abstract

Due to their reduced frequency of dosing and ease of availability, NSAIDs are generally preferred over opioids for rodent analgesia. We evaluated the efficacy of the highly COX2-selective NSAID firocoxib as compared with meloxicam and buprenorphine for reducing allodynia and hyperalgesia in rats in a plantar incision model of surgical pain. After a preliminary pharmacokinetic study using firocoxib, Sprague–Dawley rats (n = 12 per group, 6 of each sex) were divided into 6 groups: no surgery (anesthesia only), saline (surgery but no analgesia), buprenorphine (0.05 mg/kg SC every 8 h), meloxicam (2 mg/kg SC every 24 h), and 2 dosages of firocoxib (10 and 20 mg/kg SC every 24 h). The nociception assays were performed by using von Frey and Hargreaves methodology to test mechanical allodynia and thermal hyperalgesia. These assays were performed at 24 h before and at 20, 28, 44, and 52 h after start of surgery. None of the analgesics used in this study produced significantly different responses in allodynia or hyperalgesia from those of saline-treated rats. In the Hargreaves assay, female saline-treated rats experienced significantly greater hyperalgesia than did males. These findings add to a growing body of literature suggesting that commonly used dosages of analgesics may not provide sufficient analgesia in rats experiencing incisional pain.

Institutions performing studies using animals have both an ethical responsibility and a legal requirement to administer appropriate analgesia to research subjects.²⁵ Controlling postoperative pain can hasten return to normal function, minimize healing time, and prevent the development of chronic pain. In humans, postoperative pain is prevalent, with 80% of patients reporting acute pain, ranging from moderate to extreme, after surgery.⁴ Postoperative pain is not restricted to human medicine but also extends to veterinary analgesia. Therefore, we must continue to scrutinize analgesic protocols and strive to implement novel therapies.

During a surgical incision, peripheral tissue injury leads to central sensitization. This process results in hyperalgesia, an increased response to a painful stimulus, and allodynia, a painful response to a normally benign stimulus.⁶⁰ Several models that are used to test nociception in rodents have demonstrated repeatability, including plantar incision, antigen-induced inflammation, and capsaicin injection.^7,8 Among these, the plantar incision model is minimally invasive and allows for objective measurements of acute postoperative pain.^11,12 In addition, this model may be particularly advantageous, because surgical incision pain in humans appears to be better controlled in terms of persistency, duration, and pain management than are other causes of nerve injury and inflammation.⁶⁰ In the plantar incision model, allodynia is assessed by using the von Frey assay of mechanical stimuli,⁴⁷ whereas hyperalgesia is assessed by using the Hargreaves assay of thermal stimuli.²³

When making decisions regarding analgesia, choosing an appropriate medication that maximizes pain control and minimizes side effects is important. The 2 most commonly used classes of analgesics in rodents are NSAID and opioids. In human and veterinary medicine, the use of cyclooxygenase (COX) 2-selective NSAID is preferred because of lower potential for the adverse gastrointestinal and kidney side effects that result from COX1 inhibition.²⁹ However, the most commonly used NSAID in laboratory rodents are ketoprofen (COX nonselective), meloxicam (minimally COX2-seletive), and carprofen (variably COX selectivity by species).^31,37,43 Although newer NSAID such as firocoxib have an extremely high selectivity for the COX2 pathway (approximately 400-fold more selective for COX2)³⁷ and are being used more frequently in companion and large animals, little is known about the efficacy of firocoxib in laboratory rodents. A recent study has assessed the use of this medication in mice,⁴⁹ but no published dose determination or efficacy studies are available for rats.

Buprenorphine, widely considered the ‘gold standard’ opioid analgesic for rodents,^5,18,52 is a Schedule III controlled substance under the Controlled Substances Act.⁴² Many researchers do not hold a DEA license, thus creating a barrier to use of this medication. Slow-release formulations of buprenorphine cannot be obtained in New York due to state restrictions against compounded controlled drugs for research use in animals. Standard-formulation buprenorphine must be dosed fairly frequently (every 6 to 8 h in mice and rats) to maintain efficacy, thus increasing the need for handling and injections.^5,6 In addition, potentially detrimental side effects of buprenorphine have been reported in rodents and humans, including increased activity, reduced food intake, respiratory depression, and pica.^35,46,56 These considerations must be weighed against the potential analgesic relief that the medication is intended to provide.

In the current study, we first conducted a pharmacokinetic analysis of firocoxib in Sprague–Dawley rat plasma to determine an appropriate dosing frequency. We then evaluated the use of firocoxib for reducing allodynia and hyperalgesia in Sprague–Dawley rats by using a plantar incisional model of postoperative pain, comparing the efficacy of firocoxib with those of meloxicam and buprenorphine. In light of previous studies in our lab using firocoxib in a similar mouse model of plantar incisional pain,⁴⁹ we hypothesized that firocoxib and buprenorphine would produce comparable antinociception in rats.

Materials and Methods

Animal subjects.

Male and female Sprague–Dawley rats (age, 10 to 11 wk) were obtained from Harlan Sprague–Dawley (Indianapolis, IN) and housed in an AAALAC-accredited facility. Rats were free of hantavirus, lymphocytic choriomeningitis virus, mouse adenovirus 1, pneumonia virus of mice, parvoviruses (Toolan H1 virus, Kilham rat virus, rat parvovirus, rat minute virus), sialodacryoadenitis virus, reovirus 3, Sendai virus, Theiler encephalomyelitis virus, Bordetella bronchiseptica, cilia-associated respiratory bacillus, Citrobacter spp., Clostridium piliforme, Corynebacterium kutscheri, Mycoplasma pulmonis, Pseudomonas aeruginosa, Salmonella spp., Streptobacillus moniliformis, Streptococcus pneumoniae, Streptococcus spp., Helicobacter spp., Klebsiella spp., Pasteurella pneumotropica, endoparasites, and ectoparasites. Rats were randomly housed as same-sex pairs in microisolation cages and acclimated for at least 1 wk prior to study initiation. Pairings remained constant throughout the course of the study. Room environmental parameters included controlled temperature (20 to 22 °C), humidity (30% to 70%), and photoperiod (14:10-h light:dark cycle), with lights on at 0600. All cages contained autoclaved white pine wood shavings bedding material (Hancock Lumber, Hancock, ME), and rats had free access to irradiated pelleted food (no. 7012, Harlan Teklad, Madison, WI) and reverse-osmosis–purified chlorinated water. Animal strain and numbers were chosen to maintain consistency with other pharmacokinetic and nociceptive studies in rats. This study occurred at Cornell University (Ithaca, NY) while all authors were affiliated with Cornell. All study work was approved by the Cornell University IACUC. All animals were housed in accordance with the Guide for the Care and Use of Laboratory Animals.²⁵

Drugs.

Published recommended dosages of firocoxib in dogs^30,53 and an allometric dose and interval scaling worksheet⁵⁷ were used to help determine dosing of firocoxib (20 mg/mL; Merial, Duluth, GA) for rats. A single dose of 20 mg/kg administered subcutaneously was selected for pharmacokinetic analysis. For behavioral testing, buprenorphine (0.3 mg/mL; Hospira, Lake Forest, IL) was diluted in sterile 0.9% physiologic saline to a final concentration of 0.1 mg/mL to increase volume for injection. Meloxicam (5 mg/mL; Norbrook Laboratories, Newry, United Kingdom) and firocoxib (20 mg/mL; Merial, Duluth, GA) were not diluted. All drugs were administered subcutaneously on the dorsum. Buprenorphine and meloxicam doses were selected based on recommended usage in rats.^13,19

Firocoxib pharmacokinetic study.

Placement of tail vein catheters.

Sprague–Dawley rats (7 male, 2 female; age, 10 wk) were towel-restrained, and vasodilation of the lateral tail veins was induced via application of commercially available hand warmers (HotHands, Heatmax, Kobayashi Consumer Products, Duluth, GA) for 5 to 10 min. A 22-gauge, 1-in. intravenous catheter (Surflo, Terumo Medical, Elkton, MD) was placed into either the right or left lateral tail vein and capped with an injection plug (Surflo, Terumo Medical). A few rats required smaller (24-gauge) catheters. Catheters were secured in place by using brown medical tape, and white tape was used to cover the cap to maintain catheter integrity. Catheters were flushed with heparinized saline prior to each sample collection and were locked by using the same solution after sampling. Catheters were in place no longer than 32 h.

Blood collection.

Blood collection followed Cornell IACUC-approved guidelines regarding collection volume and frequency. Blood was collected immediately before drug administration and at 1, 4, 8, 12, 18, 24, and 32 h after treatment. Rats were gently restrained, and 0.3 mL heparinized saline was used to flush the catheter in a pulsatile fashion. A 21- or 23-gauge, ¾-in. winged infusion set (Surflo, Terumo Medical) was inserted into the injection plug, with the extension tubing cut; 4 to 6 drops of lock solution were discarded. At each time point, approximately 200 µL of blood was collected into lithium heparinized tubes (Microvette, Sarstedt, Newton, NC). Some rats required gentle stroking of the tail to stimulate blood to flow through the catheter. Samples were immediately centrifuged, and plasma was stored at –80 °C. No replacement fluids were given, because the cumulative volume of blood collected was less than or equal to 10% of the calculated total blood volume.

Plasma sample analysis.

Firocoxib concentrations in rat plasma were measured by using liquid chromatography–dual mass spectroscopy. Development of standards and plasma analysis were performed by Adsorption Systems (Exton, PA). Briefly, stock solutions (1.00 mg/mL of free drug) were prepared in DMSO. Standards were prepared in Sprague–Dawley rat plasma containing heparin as the anticoagulant. Plasma samples were extracted via acetonitrile precipitation in a 96-well plate format on a liquid handling system (Quadra 96 model 320, Tomtec, Hamden, CT). To test for specificity, the lower limit of quantitation in rat plasma matrix (i.e., 5 ng/mL) was determined so that no peaks coincided with the retention time of firocoxib or at the MS/MS transition of firocoxib. To test for sensitivity, the area count response was optimized at the lower limit of quantitation of the assay. In this case, the area count response was close to 500 at 5 ng/mL.

Pharmacokinetic data analysis.

Plasma firocoxib concentrations were plotted against time for individual rats using both linear and semilogarithmic graphs. Data were analyzed using commercially available software (PK Solutions 2.0, Summit Research Services, Montrose, CO). The program assumes the disposition phase of the drug follows first-order processes as indicated by linearity of the distal portion of the semilogarithmic plots. Parameters calculated included the maximal plasma concentration, the time of the maximal plasma concentration, the plasma concentration at 24 h, and the elimination half-life.

Behavioral Testing.

Surgery.

Anesthesia was induced by placing rats in an induction chamber and introducing 3% to 4% isoflurane in 100% oxygen at 1.0 L/min. Once rats were unconscious, a nonrebreathing anesthetic circuit with nose-cone delivery of 1% to 2% isoflurane in pure oxygen was used for anesthetic maintenance. Rats were placed in sternal recumbency on a heated surgical platform. The surgical site located on the plantar surface of the right hind paw was aseptically prepared by using chlorhexidine scrub and solution. The foot was passed through a small cut made in a sterile surgical drape to create a sterile field. A plantar incision was created as previously described.¹⁴ Briefly, after aseptic preparation, a no. 11 blade was used to create a 1-cm longitudinal incision through the skin and fascia, starting 3 to 4 mm from the proximal edge of the heel and extending toward the toes. The underlying plantaris muscle was elevated by using curved forceps, without disturbing the muscle origin or insertion. After hemostasis, the skin incision was apposed with a single cruciate ligature of 6-0 nylon (Ethilon, Ethicon, Cincinnati, OH). After recovery from anesthesia, rats were returned to their home cages. All incisions were checked daily by a veterinarian. Rats in the sham surgery group were anesthetized and prepped for surgery but did not receive a surgical incision. All surgeries were performed between 1200 and 1500 by a veterinarian (TP), who was blind to the treatment groups.

Nociception testing study design.

Rats were randomly assigned to 1 of 6 groups (n = 12 per group, 6 per group of each sex) at the time of surgery. The person conducting nociceptive testing was blind to treatment group allocation throughout the entire study. Treatments consisted of firocoxib at 10 mg/kg SC every 24 h; firocoxib at 20 mg/kg SC every 24 h; meloxicam at 2 mg/kg SC every 24 h, buprenorphine at 0.05 mg/kg SC every 8 h; and subcutaneous saline every 24 h as depicted in Figure 1. The last (6th) group underwent anesthesia and aseptic preparation but did not receive surgery; this group received a subcutaneous injection of saline every 24 h. The first doses of all drugs were administered 30 min before surgery. After surgery, rats received meloxicam, firocoxib, or saline as appropriate once daily at 1200, whereas buprenorphine was given at 0600, 1400, and 2200 daily until 72 h after surgery.

Figure 1.

Timeline of nociceptive testing and analgesic administration. Red arrowheads indicate administration of buprenorphine (0.05 mg/kg). Blue arrowheads indicate the administration of firocoxib (10 and 20 mg/kg), meloxicam (2 mg/kg), and saline. All substances were administered subccutaneously. Green shaded areas indicate periods of allodynia and hyperalgesia behavioral testing. The black star indicates the time of plantar incision surgery.

Nociceptive behavioral studies.

All rats underwent behavioral testing for mechanical allodynia followed by thermal hypersensitivity testing at 24 h before surgery (baseline) and at 20, 28, 44, and 52 h after surgery (Figure 1). Morning testing was performed between 0800 and 1100 to facilitate evaluation of the nadir of analgesic efficacy (close to the 24 h mark after dosing), and evening tests were performed between 1600 and 1900 to evaluate the peak of analgesic efficacy. Acclimation periods of 10 min were provided after rats were moved to the testing chamber and between mechanical and thermal testing. Rats were weighed once after baseline testing (at –24 h) and again after testing at 20 and 44 h after surgery.

Response to mechanical stimuli.

Allodynia was measured by using an electronic von Frey apparatus (model number 390, IITC Life Science, Woodland Hills, CA) as described elsewhere⁴⁷ with a few modifications. Briefly, rats were placed on an elevated perforated metal platform and confined in individual clear plastic boxes (9 × 18 × 12 cm). The sensor tip was applied perpendicularly to the plantar aspect of the right hind paw approximately 3 to 4 mm medial to the surgical incision by using enough force to cause withdrawal of the paw from the sensor tip. The paw withdrawal threshold (PWT) was determined as the force (in grams) at which the rat withdrew its foot. This assay was performed 5 times per rat at each time point, with a 5-min interval between tests. The median of the 5 test values was subtracted from the median of baseline values to obtain the change in PWT (ΔPWT) for each rat at each time point. Allodynia was defined as a positive ΔPWT.

Response to thermal stimuli.

Hyperalgesia was evoked by using a radiant heat stimulus, and responses were measured by using a thermal analgesia meter (model 390, Plantar Analgesia Meter, IITC Life Science).⁶ Briefly, rats were placed on an elevated glass platform in individual clear plastic boxes (9 × 18 × 12 cm). Radiant heat from a light source was applied to the plantar surface of the hind paw from beneath the platform glass. The light beam was focused on the plantar tissue surrounding the incision. The paw withdrawal latency (PWL) was recorded as the duration (in seconds) of light exposure needed to evoke a brisk paw withdrawal. A cut-off time of 25 s was set to prevent thermal injury to the footpad. This assay was performed 3 times per rat at each time point, with a 10-min interval between tests. The median of the 3 test values was subtracted from the median of baseline values to obtain a ΔPWL for each rat at each time point. Hyperalgesia was defined as a positive ΔPWL.

Antinociceptive effects of chosen analgesics in nonsurgical animals.

Due to an apparent lack of antinociception at our chosen analgesic dosages, we conducted additional nociception testing to evaluate the effects of these drugs in animals without surgery. We purchased an additional 8 male Sprague–Dawley rats (age, 10 wk), which were randomly pair-housed on arrival. These rats received no surgery or anesthesia. After baseline von Frey and Hargreaves data were collected according to the same procedure as described earlier, rats randomly received a single dose of firocoxib at 10 mg/kg SC, firocoxib at 20 mg/kg SC, meloxicam at 2 mg/kg SC, or buprenorphine at 0.05 mg/kg SC. At 4 h after drug administration, rats again underwent von Frey and Hargreaves behavioral testing to determine the change relative to their individual baseline values. The same rats were used for a total of 4 wk in a randomized crossover study, with a 6-d washout period between tests, to achieve 8 rats for each treatment group.

Plantar incision histology.

Rats in the surgical nociception study were euthanized with CO₂ in accordance with the AVMA Guidelines for Euthanasia² at the 96-h (24 rats) or 76-h (48 rats) time points. Initially, rats were euthanized at 96 h to accommodate additional assay time points to evaluate the decline of nociception back to baseline levels. Once we realized that nociception was not decreasing with time, the latter time points were excluded from the final dataset, and the study endpoint was moved to 76 h. The right hindfoot was detached at the tarsal joint and placed in formalin for fixation. Soft tissues were removed from the plantar surface of the feet, and serial sections of the incision site were embedded in wax and cut into slides. Tissues were stained with hematoxylin and eosin for histologic analysis. Histologic grading was performed on the most central section of the incision. Reepithelialization was determined by measuring the width of the surgical incision at the widest point. Three additional parameters—granulation tissue, angiogenesis, and inflammatory infiltrate—were used to grade the incision, as described elsewhere (Figure 2).^22,48 Reepithelialization was measured using the 5× objective, granulation tissue was determined by using the 10× objective, and inflammatory cells and angiogenesis scores were averaged across 10 high-power fields. Slides were digitally scanned (Aperio ScanScope CS2, Leica Biosystems, Nussloch, Germany) and viewed by using ImageScope software (Leica Biosystems).

Figure 2.

Histologic scoring paradigm for grading rat plantar incision sites at study end point.

Statistical analyses.

For behavioral analyses, mean withdrawal responses were analyzed via mixed-model ANOVA, with a random effect of rat and fixed effects of treatment group, time point, sex, and the treatment group × time point interaction. Pairwise comparison between treatment groups were made via least squares means with Tukey correction for comparing a family of 6 estimates. Histology scores (i.e., granulation tissue, inflammatory infiltrate, angiogenesis) were analyzed by using an ordinal logistic test with unbounded variance, with a random effect of rat and fixed effects of sex and treatment group. Reepithelialization was measured by using standard least squares model with a random effect of rat and fixed effects of treatment group, sex, and the treatment group × sex interaction. Data are expressed as mean ± SEM, and the threshold for significance was established at a P value of less than 0.05. Analyses were performed by using JMP Pro 12 (SAS Institute, Cary, NC) and R Stats package 3.3.1 (R Foundation for Statistical Computing, Vienna, Austria). Figures were made by using Prism 8 (GraphPad Software, La Jolla, CA).

Results

Pharmacokinetic assay for subcutaneous administration of firocoxib.

Two male rats were removed from the study due to insufficient blood flow through the catheter. All plasma samples taken before injection (i.e., 0 h) tested below the limit of quantitation (5 ng/mL). The median for the elimination half-life was 2.5 h (range, 1.9 to 3.1). Medians for the maximum plasma concentration, time at maximum plasma concentration, and concentration at 24 h were 5890 ng/mL (4170 to 7730), 4 h (4 to 8), and 43 ng/mL (11 to 177), respectively. After subcutaneous administration, maximum plasma firocoxib concentrations occurred at 4 to 8 h in 3 of the 7 rats. Firocoxib was not detected in the plasma of 5 of the 7 rats at 32 h after subcutaneous administration.

Behavioral assays.

Mechanical allodynia.

Mixed-model analysis found a statistically significant association between ΔPWT and treatment group (F = 2.46, P ` 0.04). Time point, sex, and the time point × treatment group interaction were not significantly associated (P > 0.05). Baseline PWT (24 h prior to surgery) did not differ significantly between groups (F = 0.55, P = 0.74). Saline group ΔPWT values were significantly different from nonsurgical ΔPWT values at 20, 44, and 52 h after surgery but not at 28 h (Figure 3 A). No treatment group was significantly different from the saline group at any timepoint.

Figure 3.

Summary of analgesic treatment effects on nociception after plantar incision surgery in adult Sprague–Dawley rats. Rats (n = 72; 36 male and 36 female) were tested for allodynia and hyperalgesia after plantar incision surgery. Graphs show the median change in (A–C) paw withdrawal threshold (in grams) and (D–F) paw withdrawal latency (in seconds) after treatment with subcutaneous buprenorphine (0.05 mg/kg every 8 h), firocoxib (10 or 20 mg/kg every 24 h), meloxicam (2 mg/kg every 24 h), or saline (every 24 h); 12 rats (6 male, 6 female) did not undergo surgery. Median test values were subtracted from median baseline values for each respective treatment group. Panels A and D show values for all rats; male-only values are depicted in panels B and E, whereas female-only values are depicted in panels C and F. Paw withdrawal threshold (a measure of mechanical allodynia) was evaluated by using an electric von Frey meter. Paw withdrawal latency measuring thermal hyperalgesia was evaluated using a thermal heat source in the Hargreaves assay. Baseline thresholds were obtained 24 h prior to surgery; assays were performed at 20, 28, 44, and 52 h after surgery. Data are expressed as mean change from baseline ± SEM; positive values indicate increased allodynia and hyperalgesia. (A) Changes in paw withdrawal threshold values of the saline group were significantly different from those of nonsurgical animals at 20, 44, and 52 h after surgery but 28 h. (D) Changes in paw withdrawal latency of the saline group were significantly different from those of nonsurgical rats across all time points. No value for any treatment group was significantly different from the saline group at any time point.

Thermal hyperalgesia.

Mixed-model analysis found statistically significant associations between ΔPWL and treatment group (F = 6.22, P ` 0.0001), sex (F = 10.36, P ` 0.0020), and the time point × treatment group interaction (F = 1.77, P = 0.04). Baseline PWL (24 h prior to surgery) was not significantly different between treatment groups (F = 0.37, P = 0.87) but was different between sexes (female, 15.7 ± 0.8 s; male, 11.4 ± 0.8 s; F = 16.32, P ` 0.0001). ΔPWL values were significantly different between the saline and nonsurgical groups at all time points (Figure 3 D). As was found for mechanical allodynia, ΔPWL values did not differ between the saline group and any other treatment group.

Tissue analysis and histology.

Gross analysis.

A total of 27 rats (saline group, n = 5; buprenorphine group, n = 9; firocoxib 10 mg/kg group, n = 4; firocoxib 20 mg/kg group, n = 5; meloxicam group, n = 4) removed their sutures before the end of the study. The highest number of rats had removed their sutures at the 20-h time point (n = 13).

Histologic analysis.

Incision width at time of necropsy was significantly associated with treatment group (F = 29.52, P ` 0.0001), sex (F = 6.84, P = 0.01), and the treatment group × sex interaction (F = 17.84, P ` 0.0001). Time of euthanasia did not have a significant effect on incision width (F = 0.43, P = 0.52). Pairwise comparisons revealed that buprenorphine-treated rats, particularly buprenorphine-treated female rats, had significantly wider wounds postoperatively than did rats in other groups. Granulation tissue scoring was significantly associated with treatment group (χ²[5] = 27.36, P ` 0.0001), but this difference was due to the lack of granulation seen in nonsurgical animals; no significant difference was noted between treatment groups comprised of rats that had surgery. There was no significant association between treatment group and angiogenesis. Inflammatory cell infiltrate was significantly associated with treatment group (χ<sup>2</sup>[5] = 19.48, P = 0.0016), again due to nonsurgical rats. None of the treatments significantly alleviated inflammation as compared with saline-treated rats. Inflammatory infiltrate was not influenced by sex (χ<sup>2</sup>[1] = 0.68, P ` 0.41) or time of euthanasia (χ²[1] = 0.12, P ` 0.73).

Nonsurgical rats.

Analysis found no overall statistically significant association between mechanical allodynia and treatment group (Figure 4); however pairwise comparisons revealed that the ΔPWT of firocoxib at 10 mg/kg was significantly different than that of buprenorphine (t = 2.45, dF = 24.42, P = 0.02), firocoxib at 20 mg/kg (t = –2.23, dF = 24.4, P = 0.04), and meloxicam (t = –2.25, dF = 26.1, P = 0.03) Similarly, no overall statistically significant association was found between thermal hyperalgesia and treatment group, but pairwise comparisons revealed a significant difference between the ΔPWL of firocoxib at 20 mg/kg and buprenorphine (t = 2.21, dF = 25.3, P = 0.04)

Figure 4.

Changes in paw withdrawal threshold (PWT) and paw withdrawal latency (PWL) after administration of select analgesics in rat without surgery. Eight male rats that did not receive anesthesia or surgery underwent a randomized crossover study. Baseline von Frey (PWT) and Hargreaves (PWL) values were obtained, after which rats received a single subcutaneous dose of buprenorphine (0.05 mg/kg), firocoxib (10 or 20 mg/kg), or meloxicam (2 mg/kg). Behavioral testing was repeated at 4 h after dosing. Values reported here are the change relative to baseline (PWT, g; PWL, s); positive values indicate increased nociception. The same 8 rats were used for a total of 4 wk, with a 6-d washout between tests. *, P ` 0.05.

Discussion

The purpose of this study was to determine whether firocoxib is a viable alternative to buprenorphine for controlling postoperative allodynia and hyperalgesia in a plantar incision model of pain in rats. We first conducted a pharmacokinetic study to analyze plasma levels of firocoxib to determine the appropriate dosing frequency in this species. Extrapolations from studies of firocoxib in horses^9,34 resulted in a calculated IC₈₀ dose for rats of approximately 110 ng/mL. Our pharmacokinetic data revealed that subcutaneous dosing of firocoxib at 20 mg/kg results in sustained plasma levels above 110 ng/mL starting at the first hour after injection and persisting for almost 24 h. Given these findings together with a calculated clinical dose of approximately 15 mg/kg from allometric dosing calculations using canine values,³⁸ we administered daily doses of firocoxib at 10 and 20 mg/kg for our subsequent behavioral studies. We included the 10-mg/kg dosage because the 20-mg/kg dose produced plasma levels that were well above extrapolated efficacy levels.

In both the von Frey and Hargreaves nociceptive assays, saline-treated rats had significantly different responses from nonsurgical animals except in the von Frey assay at the 28-h time point. These results validate use of the plantar incisional model to induce nociception in the present study. However, no treatment group in either assay at any time point was significantly different from its respective saline group, suggesting that medicated rats experienced allodynia and hyperalgesia that was similar to that of rats that had received surgery without analgesia. Our treatment group sample sizes were based on a review of published literature,^14,21,47,51 but our standard error suggests that this sample size may have been inadequate to capture the inherent behavioral variation between individual animals. In future studies assessing analgesic efficacy using nociceptive assays, sample size calculations should reflect the fact that including both sexes may increase variability.

Another interpretation of the lack of significantly different values between saline-treated and analgesic-treated rats in the present study could be that the incision was too painful. Mechanical stimulation itself can lead to sensitization,³² which may explain why greater allodynia was present throughout the study. Hindpaw incisional hyperalgesia peaks at 24 h after surgery and typically lasts 3 to 4 d.⁴⁰ We obtained additional data at time points beyond 52 h (not reported in the present study), but found no evidence of nociceptive responses tapering off or returning to baseline after 76 h; these findings support the possibility that extrinsic stimulation of the plantar incision model may be more noxious than standard surgical recovery in this study. In addition, buprenorphine- and meloxicam-treated male rats without plantar incision surgery showed a reduction in ΔPWT and ΔPWL (Figure 4), providing further evidence that our paw incisions may have caused excessive pain, even though we used the same methodology as described in previous publications.^12,49

In support of that claim, 45% of rats that received plantar incision surgery removed their sutures prior to the end point. Studies using plantar incision as a nociception model often exclude such animals, but we felt that their inclusion was important to demonstrate a true clinical scenario more accurately. We included histologic comparison of the incisions to justify the inclusion of these rats and to test for an effect of the treatment group on wound healing, particularly between antiinflammatory- and opioid-treated rats. Although incision sites at time of necropsy were significantly wider in buprenorphine-treated female rats, none of the treatment groups showed significant differences in inflammatory infiltrate, granulation, or angiogenesis that might warrant excluding these animals from our dataset. Analysis of circulating levels of prostaglandins or other inflammatory mediators may be useful in future studies.

We chose the 0.05 mg/kg dose of buprenorphine based on industry standards^13,19 and published efficacy data in rat studies using similar plantar incision models.^18,52 The dose of meloxicam (2 mg/kg) was also chosen from published values.^20,59 However, recent literature suggests that higher dosages of analgesics may be needed for comprehensive rodent analgesia. ^{1,10,26,27,29,33,36,41,50–52,54,58,61} Examples of this include a similar nociception study in rats which found that this dose of buprenorphine only increased thermal withdrawal latencies at one hour after injection; antinociceptive effects were not present at 4 h after injection.²⁷ Various behavioral assays have shown that subanalgesic doses of opioids promote hyperalgesia in mice and in rats.^{3,17,24,28,62} Carprofen, ketoprofen, flunixin, and meloxicam have all recently been reported to be ineffective (or effective only at very high doses) at reducing nociception in models of surgical or incisional pain in both mice^{1,26,33,36,41,51,58} and rats.^52,61 In addition, higher doses of COX2-selective analgesics (coxibs) may be needed to reduce behavioral nociception in mice^10,54 and rats.^29,50 Our dose of buprenorphine may have been too low for this pain model, thus resulting in more allodynia and hyperalgesia, or the analgesic doses may have been too low overall, thereby resulting in insufficient analgesia.

Rats that received buprenorphine were injected every 8 h, whereas rats in other groups were injected every 24 h. Greater animal handling may have negatively affected performance in the behavioral assays.

Although our study did not find statistically significant differences between treatment groups, buprenorphine may produce a clinically relevant reversal of nociception (Figure 5).

Figure 5.

Clinical reduction of plantar incision-induced nociception in rats receiving analgesia. Animals (n = 72; 36 male and 36 female) were tested for allodynia and hyperalgesia after plantar incision surgery. Graphs represent the percent clinical reversal in allodynia (A-C) and hyperalgesia (D-F) after treatment with subcutaneous buprenorphine (0.05 mg/kg q8h), firocoxib (10 or 20 mg/kg q24h), meloxicam (2 mg/kg q24h), or saline (q24h). 12 animals (6 male, 6 female) did not have surgery. Median nociceptive test values were divided by median nociceptive baseline values for each respective treatment group. Male-only values are depicted in (b) and (e), while female-only values are depicted in (c) and (f). Paw Withdrawal Threshold measuring mechanical allodynia was evaluated using an electric von Frey meter. Paw Withdrawal Latency measuring thermal hyperalgesia was evaluated using a thermal heat source in the Hargreaves assay. Baseline thresholds were obtained 24 h prior to surgery; assays were performed at 20, 28, 44, and 52 h after surgery.

We considered clinically relevant reversal as percent reversal greater than that of saline treatment. Buprenorphine appears to produce clinical reversal of allodynia in males at the first 3 time points (Figure 5 B) and in females at 20 and 44 h after surgery (Figure 5 C). For hyperalgesia, treatment with buprenorphine appears to produce a clinically relevant reversal at all time points in female rats (Figure 5 F) but not in male rats (Figure 5 E). Therefore, buprenorphine may provide some relief from incisional pain-induced nociception in female Sprague–Dawley rats, and its use may be warranted for this indication; however, it may be ineffective in male rats. In general, the effects of NSAID were inconsistent, except for firocoxib 10 mg/kg, which generally produced less reversal of nociception than saline in both assays and both sexes but with less hyperalgesia in female rats. We do not recommend using these doses of meloxicam or firocoxib for reversal of incisional-induced hyperalgesia. However, in the male rats, saline treatment had the highest percentage reversal of hyperalgesia (Figure 5 E).

The NIH support the inclusion of both sexes in biomedical research,³⁹ which historically has excluded female animals due to a variety of concerns including the effects of reproductive hormones on behavioral assays. Male and female mice^15,44,63 and rats^7,45 are known to react differently to NSAID and opioids, in part due to the effects of reproductive hormones on nociceptive processing. In the present study, we found no difference in allodynia between sexes. However, analyzing the mean ΔPWL values by sex revealed that female rats had greater hyperalgesia in response to thermal stimulation than did male rats, although this association may be due to the significant difference between males and females already present at baseline. This outcome correlates with previous studies in rats, which suggest a more robust hyperalgesic response in females than males.^7,24 Meloxicam is reported to have a higher blood concentration, longer half-life, and slower excretion in female rats compared with males.⁵⁵ Additional studies evaluating sex differences in analgesic efficacy in rats are warranted.

In light of the increase in ΔPWL in male rats treated with firocoxib 10 mg/kg compared with saline (Figures 2 E and 5 E), a crossover study was performed to determine effect of analgesia alone on PWT and PWL in rats without surgery. In this study, firocoxib 10 mg/kg was found to significantly increase allodynia compared with buprenorphine, firocoxib 20 mg/kg, and meloxicam, whereas firocoxib 20 mg/kg significantly increased hyperalgesia compared with buprenorphine (Figure 4). These findings suggest that the postsurgery allodynic and hyperalgesic responses associated with these treatment groups could be due to the injection of the medication alone. Opioids can cause behavioral activation in healthy rodents, but NSAID typically have minimal effect on normal behavior.⁴⁰ Rats that received firocoxib showed no evidence of injection site necrosis or inflammation during necropsy (data not shown), but the increase in allodynia and hyperalgesia in sham-treated animals needs further investigation.

In summary, the dosages of buprenorphine, firocoxib, and meloxicam that we administered did not produce a significant reduction in PWL or PWT as compared with saline-treated rats in a plantar incision model of nociception. Sex differences were found in response to the thermal Hargreaves assay during baseline measurements. Our findings support the continued need for standardized, reproducible models of nociception testing and for further exploration of appropriate, context-specific postsurgical analgesic protocols in rodents.