Evaluation of the analgesic efficacy of grapiprant compared with robenacoxib in cats undergoing elective ovariohysterectomy in a prospective, randomized, masked, non-inferiority clinical trial

Abstract

Objectives

The main objective of this study was to compare the postoperative analgesic effects of grapiprant with those of robenacoxib in cats undergoing ovariohysterectomy (OVH).

Methods

In total, 37 female cats (age range 4 months–10 years, weighing ⩾2.5 kg) were enrolled in a prospective, randomized, masked, non-inferiority (NI) clinical trial. Cats received oral robenacoxib (1 mg/kg) or grapiprant (2 mg/kg) 2 h before OVH. Analgesia was assessed via the Feline Grimace Scale (FGS), the Glasgow Composite Measure Pain Scale-Feline (CMPS-F), von Frey monofilaments (vFFs) and pressure algometry (ALG) 2 h before treatment administration, at extubation, and 2, 4, 6, 8, 18 and 24 hours after extubation. Hydromorphone (<8 h postoperatively) or buprenorphine (>18 h postoperatively) were administered to cats with scores of ⩾5/20 on CMPS-F and/or ⩾4/10 on FGS. NI margins for CMPS-F and vFFs were set at 3 and −0.2, respectively. A mixed-effect ANOVA was used for FGS scores (P <0.05). Data are reported as mean ± SEM.

Results

The data from 33 cats were analyzed. The upper limit of the 95% confidence interval (CI) (0.35) was less than the NI margin of 3 for CMPS-F, and the lower limit of the 95% CI (0.055) was greater than the NI margin of −0.2 for vFFs, indicating NI of grapiprant. The FGS scores were greater than baseline at extubation for both treatments (1.65 ± 0.63; P = 0.001); however, there was no difference between treatments. There was no difference between treatments, nor treatment by time interaction, for vFFs (P <0.001). The CMPS-F scores for both treatments were higher at extubation but returned to baseline after 4 h (P <0.001). For ALG, there was no difference in treatment or treatment by time interaction. The robenacoxib group had lower pressure readings at extubation and 6 h compared with baseline.

Conclusions and relevance

These results indicate that grapiprant was non-inferior to robenacoxib for mitigating postsurgical pain in cats after OVH performed via ventral celiotomy. The impact of grapiprant for analgesia in OVH via the flank is unknown.

Introduction

Ovariohysterectomy (OVH) is a common surgical procedure performed to control feline populations and prevent undesirable behaviors and reproductive pathology. The administration of pre-emptive analgesics is essential for improving postoperative recovery and animal welfare.^1,2 Non-steroidal anti-inflammatory drugs (NSAIDs) are commonly used in multimodal analgesic protocols and are efficacious in treating surgically induced inflammation and pain.^{1 –4}

The isoenzymes cyclooxygenase (COX)-1 and -2 produce prostaglandins (PGs) that are upregulated during inflammation.^5,6 These isoenzymes are also involved in renal and gastrointestinal (GI) homeostasis and play an important role in central and peripheral nociceptive pathways. ⁷ NSAIDs target COX-1 and COX-2 and decrease the production of PGs, and thus reduce inflammation and pain. ⁵ Both non-selective COX inhibitors and COX-2-selective inhibitors have been reported to be effective in providing postoperative analgesia in cats after orthopedic and soft tissue surgeries.^{1 –3,5,8 –12} In other studies, robenacoxib demonstrated superior analgesia when compared with placebo in cats undergoing onychectomy, ⁸ and provided effective analgesia similar to that provided by meloxicam for orthopedic surgery ⁹ and OVH. ³

Although NSAIDs are efficacious in providing analgesia for postoperative pain, there is concern about GI and renal adverse effects, which have been reported in cats. ¹³ These adverse effects are due to the inhibition of the constitutive functions of COX enzymes and the subsequent alteration of production of PGs involved in the maintenance of tissues. ⁵ Despite their superior safety profile compared with non-selective COX inhibitors, COX-2-selective NSAIDs still present a potential concern for adverse effects.^5,6,10,14 Currently, robenacoxib and meloxicam are the only NSAIDs labeled for use in cats in the USA. Since cats are purportedly sensitive to the adverse effects of NSAIDs, further studies to identify safe and effective NSAIDs for perioperative use are warranted. ⁶

Grapiprant is an antagonist at the PGE₂ prostanoid receptor, EP4.^11,12,15 EP4 receptors are localized in the dorsal root ganglion of the spinal cord and peripheral nociceptors, and are associated with mediation of inflammatory pain.^{16 –19} Due to the selectivity of grapiprant at this receptor, the production of prostanoids and thus homeostatic functions are preserved. Hence, grapiprant may reduce the likelihood of GI and renal adverse effects associated with traditional NSAIDs. ¹² Currently, grapiprant is licensed for and has been shown to be efficacious in attenuating canine osteoarthritis with a published oral dose of 2 mg/kg every 24 h. ¹¹ Ross et al ²⁰ showed that grapiprant was efficacious in attenuating acute soft tissue pain associated with OVH. Therefore, safe alternatives to traditional NSAIDs to broaden the spectrum of treatments for acute feline pain are desired. Safety and pharmacokinetic studies evaluating grapiprant in healthy cats have been performed, with no significant adverse effects or gross histopathologic lesions reported.^12,15 However, data regarding the efficacy of grapiprant in treating acute surgical pain are limited.

The aim of the present study was to compare the postoperative analgesic efficacy of grapiprant with that of robenacoxib in cats. The primary outcome measures of analgesic efficacy were evaluated by a validated feline pain scale, the Glasgow Composite Measure Pain Scale-Feline (CMPS-F), ²¹ and mechanical nociceptive threshold testing (MNT) using von Frey filaments (vFFs). Secondary outcome measures included the evaluation of analgesic efficacy with another validated feline pain scale, the Feline Grimace Scale (FGS),^{22 –24} and MNT using pressure algometry (ALG). The null hypothesis was that grapiprant would be inferior to robenacoxib in providing postoperative analgesia evaluated via validated feline pain scales and MNT in healthy cats undergoing elective OVH.

Materials and methods

Study design

This study was a prospective, randomized, masked, non-inferiority (NI) clinical trial comparing robenacoxib and grapiprant, conducted in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines. An NI design was chosen due to ethical concerns about utilizing a placebo that would be required for a superiority trial. ²⁵ In addition, an NI trial would allow for the comparison of potential adverse effects of the referent drug (robenacoxbi) vs the study drug (grapiprant). ¹² The study protocol was approved by the University of Tennessee, Knoxville, Institutional Animal Care and Use Committee (number 2889-0222).

Animals

Mixed-breed female shelter cats presenting to the University of Tennessee College of Veterinary Medicine for elective OVH were enrolled in this study. Data collection occurred in the area where the cats were housed and acclimated to minimize the stress of transport and acclimation to a new environment. Cats were deemed healthy based on complete physical examination and routine laboratory testing (packed cell volume, total solids, blood glucose concentrations). Cats with evidence of systemic disease, pre-existing OVH, weight <2.5 kg or age ⩽4 months or ⩾10 years, or cats intolerant to handling and oral treatment administration were excluded. Cats were acclimated to the study environment for 12–18 h before commencement of the study. Each cat was individually housed in kennels to allow pain assessments and to facilitate incision healing. Kennels were enriched with climbing perches and blankets, and cats had free access to clean litter boxes. ²⁶ They were fasted for 8–12 h before anesthesia but had ad libitum access to water.

Treatment allocation and administration

Each cat was randomly allocated to one of two treatment groups: robenacoxib (1 mg/kg; Onsior; Elanco) (ROB group) or grapiprant (2 mg/kg; Galliprant; Elanco) (GRP group) administered orally (PO). In order to minimize variability in tissue handling, two experienced surgeons (JW and RD) were randomly assigned to perform the OVH. A two-factor randomization using a split-split plot design was used to allocate the cats to treatments, so that cats aged ⩾6 months and those aged <6 months, as well as the individual surgeons, had equal representation (JMP version 16; SAS Institute). Robenacoxib or grapiprant was given PO, 2 h before induction of anesthesia. Cats received another dose of their treatment medication 24 h after extubation and were discharged.

Anesthesia and surgical procedure

Two hours after treatment administration, cats received alfaxalone (2 mg/kg; Alfaxan Multi Dose IDX; Jurox) and hydromorphone (0.1 mg/kg; Hydromorphone; Hospira) intramuscularly (IM). After IM premedication, the cats were left undisturbed for 10–15 mins to allow the drugs to take effect. Sedation was considered appropriate once the cats were amenable to intravenous (IV) catheter placement. A 22 G peripheral IV catheter (SurFlash; Terumo) was placed in the cephalic vein using a clean technique. Anesthetic induction was performed with propofol (PropoFlo 28; Zoetis) IV titrated to effect to allow the placement of a feline supraglottic airway device (V-gel; Docsinnovent). Anesthesia was maintained with isoflurane delivered in 100% oxygen and titrated to achieve a surgical plane of anesthesia. Lactated Ringer’s solution (LRS; Hospira) was administered IV at 3 ml/kg/h via a calibrated syringe pump (Medfusion 3500; InfuSystem). The anesthetic time was measured from the time of induction to the time the inhalant was discontinued.

End-tidal partial pressure of carbon dioxide (EMMA capnograph; Masimo), hemoglobin oxygen saturation (Masimo Rad 57; Masimo), heart rate (HR) and rhythm, respiratory rate (RR), rectal temperature (AmerisourceBergin digital thermometer; AmerisourceBergin) and Doppler blood pressure (Doppler flow detector 811-B; Parks Medical Electronics) were monitored and recorded throughout anesthesia. Body temperature was supported via use of a forced warm air blanket (Bair Hugger; 3M).

Cats underwent OVH via ventral celiotomy. The surgical time was measured from the time the surgical incision was made to complete closure of the incision. After completion of the surgery, isoflurane administration was discontinued, and the supraglottic airway device was removed when the swallowing reflex resumed. Rectal temperature, HR and RR were monitored until the animal was ambulatory and achieved normothermia.

Analgesic efficacy testing

The analgesic efficacy of the treatments was first assessed in real time via two validated feline-specific pain scales (FGS and CMPS-F)^{21 –23} and then by MNT (vFFs and ALG). Pain scoring and MNT assessments were performed by the same masked individual (EKP), who has extensive experience with the use of both pain scales. Evaluations with pain scales and MNT were performed upon admission (baseline), just before treatment administration (T₋₂), at extubation (T_ex), and 2 (T₂), 4 (T₄), 6 (T₆), 8 (T₈), 18 (T₁₈) and 24 (T₂₄) h after extubation. Cats receiving a score of ⩾5 on the CMPS-F and/or a score of ⩾4 on the FGS were administered hydromorphone (0.05 mg/kg) IM or buprenorphine (0.15 mg/kg; ²⁷ Simbadol; Zoetis) subcutaneously. Selection of the rescue analgesic was based on the time of the pain score in relation to extubation time. If the results of the pain scales indicated the need for rescue analgesia ⩽8 h after extubation, cats were administered hydromorphone. If the results of the pain scales indicated the need for rescue analgesia ⩾18 h after extubation, cats were given buprenorphine. Cats were reassessed for pain 30 mins after administration of rescue medications and at subsequent time points. Data from cats receiving rescue analgesia were excluded from statistical analysis.

MNT with both devices was applied on the dorsal area of the flank (control site) and three sites alongside the incision (cranial, middle and caudal) at a distance of 2 cm lateral to the incision. ²⁸ The vFF (Touch Test Sensory Evaluators; North Coast Medical) and a pressure algometer (Somedic Algometer, Somedic Production AB) were used for MNT after pain scoring at the time points described above. The vFFs are calibrated filaments used to apply force in the range of 0.008–300 g (0.7845–2941.176 Pa) to the skin. Testing was started with the 0.008 g filament, and filament size was increased until a positive reaction (head turning, growling, attempts to bite) was elicited. MNT using vFFs for each site was performed in triplicate and the average of the values that elicited a positive reaction was considered the final score. A minimum of 30 s between each test was allowed in order to avoid temporal summation. ²⁰ The vFF testing was performed before ALG testing.

A pressure algometer with a 2 cm² probe tip and a slope of 50 kPa/s was used based on a previously described technique in dogs and humans.^20,29 The algometer was used at the same locations as the vFF assessments. Measurements were taken in triplicate and the pressure was slowly increased until a reaction, as described above, was observed. ²⁰ The pressure measurement (kPa) at the time of the reaction was recorded for each time point.

Statistical analysis

A priori power analysis estimated a minimum sample of 17 cats per treatment group required to detect NI with α = 0.05 and β <0.2 based on previous studies.^20,30,31 The NI margins for CMPS-F and vFFs were set at Δ = 3 and Δ = −0.2, respectively.^20,28 These data are presented graphically and display the mean difference between robenacoxib and grapiprant for any given hour. The mean difference for CMPS-F was calculated by subtracting the mean CMPS-F score for robenacoxib from the mean CMPS-F score for grapiprant. NI was shown for CMPS-F if the upper limit of the 95% CI of the mean difference between the ROB and GRP groups was below the NI margin. For vFFs, the opposite applies, since higher values indicate better analgesic efficacy; therefore, NI was demonstrated if the lower limit of the 95% CI of the mean difference between treatments was greater than the NI margin. For CMPS-F, FGS, vFFs and ALG, a mixed-effect, repeated measures ANOVA was used to analyze the effect of treatment, time and treatment by time interaction, with treatment as the between-subject factor and time as the within-subject factor. Normality of data was tested using Q-Q plots and ANOVA residuals. Pairwise multiple comparisons were analyzed using the Tukey post-hoc test. The incidence of rescue analgesia was compared between groups using Fisher’s exact test. Significance was set at P <0.05.

Results

In total, 37 cats were enrolled; however, 33 cats completed the study due to treatment failure in four cases. There was no difference in mean body weight (GRP 3.1 ± 0.5 kg; ROB 3.1 ± 0.5 kg; P = 0.93), age in years (GRP 0.8 ± 1.7; ROB 1.8 ± 0.8; P = 0.75), surgical time (RD 13 ± 4 mins; JW 12 ± 3.5 mins; P = 0.47) or anesthesia time (RD 27 ± 10 mins; JW 28 ± 5.8 mins; P = 0.5). The mean dose of grapiprant administered was 2.4 ± 0.6 mg/kg and the mean dose of robenacoxib administered was 1.4 ± 0.4 mg/kg. In the ROB and GRP groups, 3/18 (16.6%) and 1/19 (5.2%) cats received rescue analgesia, respectively. The incidence of treatment failure did not differ between treatments (P = 0.34). A post-hoc power analysis revealed adequate power to detect a difference between treatments with only 15 cats in the GRP group. In the ROB group, two cats with FGS scores of 4/10 at T₄ received hydromorphone and a third received buprenorphine at T₂₄. One cat in the GRP group had an FGS score of 4/10 at T₁₈ and received buprenorphine. Cats that received rescue analgesia had an FGS score of <4/10 within 30 mins of the administration of analgesic agents.

The results of FGS, CMPS-F, vFFs and ALG are summarized in Table 1. For the FGS, there was no difference between treatments or treatment by time; time was the only significant factor. In both the GRP and ROB groups, the FGS scores were greater at T_ex in comparison with baseline (T_–2) but returned to baseline after T₂ (Figure 1). Since the upper limit of the 95% CI was less than the NI margin (Δ = 3), grapiprant was determined to be non-inferior to robenacoxib based on CMPS-F (Figure 2). Furthermore, there was no difference in CMPS-F scores between treatments or treatment by time. The CMPS-F scores were greater at T_ex and T₂ in comparison with baseline for cats in the ROB group (P <0.001). CMPS-F scores were also greater at T_ex in comparison with baseline in the GRP group (P <0.001). For both groups, CMPS-F scores returned to baseline after T₄ (Figure 3).

Table 1.

Results for CMPS-F, FGS and MNT at the level of the incision at 2, 4, 6, 8, 18 and 24 h after extubation in 33 cats undergoing OVH, randomly assigned to receive either GRP or ROB 2 h before premedication

Variable	GRP (SE)	ROB (SE)	MD	SED	95% CI	∆
CMPS-F	0.157	0.162	−0.27	0.005	0.181–0.343	3
vFF incision (Pa)	0.071	0.072	0.008	0.01	0.011–0.055	−0.2
FGS	0.179	0.216	−0.197	0.037	NA	NA
Algometer incision (kPa)	1.63	1.6	−1.31	−0.03	NA	NA

CI = confidence interval; CMPS-F = Glasgow Composite Measure Pain Scale-Feline; ∆ = NI margin; FGS = Feline Grimace Scale; GRP = grapiprant; MD = mean difference; MNT = mechanical nociceptive threshold testing; NI = non-inferiority; OVH = ovariohysterectomy; ROB = robenacoxib; SE = standard error; vFF = von Frey filament

Figure 1.

Mean ± SEM Feline Grimace Score for 33 cats receiving either robenacoxib (dashed line) or grapiprant (solid line) 2 h before surgery (T₋₂), at extubation (T_ex) and 2, 4, 6, 8, 18 and 24 h after extubation. *Within the robenacoxib group, the value differs significantly from T₋₂. ^†Within the grapiprant group, the value differs significantly from T₋₂

Figure 2.

Non-inferiority graph demonstrating the mean difference (red dotted line) along with the 95% CI (thick black horizontal line) between grapiprant and robenacoxib. Black dashed lines indicate the non-inferiority margin set at Δ = 3 for CMPS-F. CI = confidence interval; CMPS-F = Glasglow Composite Pain Scale-Feline

Figure 3.

Mean ± SEM Glasgow Composite Pain Scale scores for 33 cats receiving either robenacoxib (dashed line) or grapiprant (solid line) 2 h before surgery (T₋₂), at extubation (T_ex) and 2, 4, 6, 8, 18 and 24 h after extubation. *Within the robenacoxib group, the value differs significantly from T₋₂. ^†Within the grapiprant group, the value differs significantly from T₋₂

Since the lower limit of the 95% CI was greater than the NI margin (Δ = −0.2), grapiprant was determined to be non-inferior to robenacoxib based on vFF results (Figure 4). In the GRP group, at T₂, the pressure exerted at the incision by the vFFs was less than baseline (T₋₂) (P <0.01). For the ROB group, the pressure at the level of the incision was lower at T_ex compared with baseline (T₋₂) (P <0.01) (Figure 5). There was no difference between treatment or treatment by time. For ALG, there was no difference in treatment, time or treatment by time interaction. For the ROB group, pressure readings via algometry were lower at T_ex, T₆ and T₁₈ compared with baseline (Figure 6).

Figure 4.

Non-inferiority graph demonstrating the mean difference (red dotted line) along with the 95% CI (thick black horizontal line) between grapiprant and robenacoxib. Black dashed lines indicate the non-inferiority margin set at Δ = −0.2 for von Frey filaments. CI = confidence interval

Figure 5.

Mean ± SEM von Frey filament pressure (Pa) readings near the surgical incision for 33 cats receiving either robenacoxib (dashed line) or grapiprant (solid line) 2 h before surgery (T₋₂), at extubation (T_ex) and 2, 4, 6, 8, 18 and 24 h after extubation. *Within the robenacoxib group, the value differs significantly from T₋₂. ^†Within the grapiprant group, the value differs significantly from T₋₂

Figure 6.

Mean ± SEM algometer pressure (kPa) readings near the surgical incision for 33 cats receiving either robenacoxib (dashed line) or grapiprant (solid line) 2 h before surgery (T₋₂), at extubation (T_ex) and 2, 4, 6, 8, 18 and 24 h after extubation. *Within the robenacoxib group, the value differs significantly from T₋₂

Discussion

These results indicate that grapiprant was non-inferior to robenacoxib for mitigating postsurgical pain in cats after OVH performed via ventral celiotomy. The impact of grapiprant for analgesia in OVH via the flank is unknown. Therefore, the null hypothesis was rejected. The results of validated pain scales and MNT support the use of grapiprant for treating acute soft tissue pain in cats 24 h postoperatively.

The overall treatment failure rate was 16.7% for the ROB group and 5.3% for the GRP group. King et al ⁹ evaluated the efficacy of either pre- or postoperative robenacoxib when compared with placebo in cats undergoing forelimb onychectomy and OVH or castration. The reported treatment failure was 19.7% for cats that received robenacoxib compared with 41% in the placebo group. Similar studies ^{8 , 32} evaluated the efficacy of oral robenacoxib vs placebo in cats undergoing forelimb onychectomy and OVH with similar results to the present study, where rescue analgesia was required in 16.5% and 43.3% of cats administered robenacoxib and placebo, respectively. A study conducted by Teixeira et al ³³ compared the analgesic efficacy of grapiprant vs carprofen in cats undergoing OVH, which reported no difference between the treatments when the CMPS-F was used. However, contrary to our study, rescue analgesia was required in 67% of the cats receiving grapiprant compared with 18% of cats receiving carprofen according to the UNESP-Botucatu Multidimensional Composite Pain Scale. ³³ This is a greater percentage of required rescue analgesia compared with the present study, in which only 5.2% of cats treated with grapiprant required it. These authors theorized that grapiprant exhibited a late analgesic effect due to insufficient time for EP4 receptor exposure and thus lack of antagonism of this receptor.^33,34 It is also possible that grapiprant is less effective than traditional NSAIDs due to the blockade of only one of four PGE2 receptors, EP4, which leaves the pronociceptive EP2 receptor active. This is consistent with the results obtained in dogs with experimentally induced synovitis, where grapiprant was less effective than carprofen. ³⁵ Results similar to those of the present study were shown in dogs undergoing OVH, where grapiprant was non-inferior to carprofen in attenuating postoperative pain.^20,36 In addition, higher pain scores could be attributable to higher stress and anxiety in shelter cats. The study conducted by Zeiler et al, ³⁷ which assessed behavior scores in hospitalized cats, concluded that the demeanor scoring system was effective in detecting behavior changes but required 2 days of acclimation to see notable differences. In the present study, cats were acclimated for only 12–18 h before surgery, and thus stress and anxiety may have contributed to the misclassification of pain.

In the present study, despite the CMPS-F scores not indicating the need for rescue analgesia for either group, the FGS scores qualified 16.6% and 5.2% of cats from the ROB and GRP groups, respectively, for rescue analgesia. While a strong correlation between CMPS-F and FGS scores has been reported, divergences in scores between the systems in the present study are likely due to inherent differences between these scoring systems. ²³ The FGS is entirely based on the evaluation of facial expression and head position. ²² Although the CMPS-F includes the evaluation of facial expressions, it also includes behavior and response to palpation of painful areas. ²¹ Therefore, the divergences in scores indicate that it may be beneficial to include multiple pain scales when evaluating analgesic treatments. ³⁸ Nevertheless, both pain scales have been used to detect pain in cats receiving an NSAID after an OVH and have demonstrated a response to analgesic intervention.^23,24,39 Lastly, the difference in pain scores could be attributable to higher stress and anxiety in shelter cats. As discussed previously, the study conducted by Zeiler et al ³⁷ showed that at least 2 days of acclimation were needed to detect behavioral changes, and the shorter acclimation in the present study may have resulted in stress and anxiety being misclassified as pain.

Hyperalgesia and incisional pain were evaluated using pressure algometry. To the authors’ knowledge, pressure algometry has not been validated in cats. However, in the present study, ALG values consistently decreased after T_ex and returned to baseline at T₂₄, although not significantly. These findings supported the presence of postoperative surgical pain but since the decrease in ALG value was not different between treatment groups, it was determined that grapiprant was as effective as robenacoxib for analgesia in cats undergoing OVH. However, this study may have been underpowered to detect differences in ALG. Our findings also correlate with previous studies.^{20,40 –43} Ross et al ²⁰ compared the efficacy of carprofen and grapiprant in attenuating acute postoperative pain in dogs undergoing OVH by using ALG. That study reported similar results, with a decrease in ALG values and a return to baseline 24 h postoperatively. In another study conducted by Nicholls et al, ⁴⁰ the correlation between MNT via algometry and CMPS-F was evaluated in cats undergoing OVH. Although there was no correlation between the CMPS-F and ALG values, the decrease in ALG values postoperatively potentially indicated the presence of pain. The investigators suggested the subjectivity in pain scoring between individuals, the questionable validity of MNT in cats and the inability of the CMPS-F to distinguish fear and anxiety from pain as possible explanations for the lack of correlation between the two metrics. ⁴⁰ In the present study, there was only one evaluator (EKP) performing pain scoring and MNT; therefore, the evaluation of interrater variability was not possible. Nevertheless, ALG scores in the pre- and postoperative periods followed a similar trend to the pain-scoring systems and were not different between groups.

In this study, vFFs were also utilized to identify the presence of allodynia or hyperalgesia, which may result from central sensitization formation in the dorsal horn of the spinal cord. ⁴⁴ These plastic changes in the central nervous system from tissue and nerve injury will manifest as abnormal responses to pain (allodynia or hyperalgesia), and thus allow continued firing of damaged afferent neurons. ⁴⁴ Both pain-scoring systems utilized in the present study are subjective and do not distinguish physiologic incisional pain from central sensitization. Similar to other previously conducted studies where hyperalgesia was demonstrated after OVH,^45,46 the use of MNT provides the opportunity to further differentiate the presence of incisional pain, central sensitization and the efficacy of the analgesics being studied. ⁴⁵ Several studies have been conducted utilizing vFFs to measure sensory thresholds in non-painful cats, ⁴¹ cats with osteoarthritis ⁴¹ and dogs undergoing OVH.^20,28,41 Although Adami et al ⁴¹ found that cooperation of cats decreased if repeated testing was performed within 1 h, this was not seen in the present study. This could be due to the strict inclusion criteria, which required cats to be amenable to repeated handling and sensory testing. Similar tolerance to MNT was reported by Addison and Clements, ⁴⁷ where the authors were able to utilize vFFs to differentiate osteoarthritic and healthy limbs in cats. Utilizing the flank (an area away from the incision) as a control aided in differentiating allodynia from incisional pain. ²⁸ Similar to the ALG results in the present study, the vFF values decreased after T_ex but then returned to baseline at T₂₄ independent of treatment. This was demonstrated by the lack of difference in response to vFF over the flank vs the points along the incision. Allodynia and hyperalgesia associated with incisional pain in dogs have also been evaluated with vFFs.^20,28 Lascelles et al ⁴⁵ demonstrated that hyperalgesia due to central sensitization was lessened with the pre- and postoperative administration of carprofen. When comparing carprofen and grapiprant in dogs undergoing OVH, Ross et al ²⁰ demonstrated a lack of hyperalgesia or allodynia via ALG and vFFs. The cats in the present study did not demonstrate allodynia, evidenced by the lack of significant decrease in vFF values of the flank or the incision in either group.

Robenacoxib was selected as the positive control because it is efficacious in attenuating musculoskeletal^3,8,32,48 and soft tissue^1,8,9 pain, and it has comparable analgesic efficacy to tolfenamic acid ² and superior efficacy to meloxicam^9,49 in cats undergoing OVH. In addition, safety and pharmacokinetic^50,51 profiles for robenacoxib in cats have been established, ⁵¹ including use in cats with stable, low grade, chronic kidney disease.^52,53 Lastly, meloxicam and robenacoxib are the only NSAIDs labeled for use in cats in the USA. ⁶ In contrast, there is minimal efficacy data for grapiprant in cats. Pharmacokinetics ¹⁵ and safety profiles ¹² have, however, been established.

Although uncommon, adverse GI effects have been reported in cats receiving NSAIDs.^6,50 In our study, the only detectable GI effects included mild self-limiting soft stool in six cats. Of these cats, five had soft stool before preoperative administration of their respective NSAID. This is similar to previously reported adverse effects of grapiprant ¹² and robenacoxib. ⁵⁰ However, the soft stool could have also been secondary to stress, diet change or GI parasitism. One cat from the GRP group was anorexic before receiving NSAIDs and exhibited anxiety throughout the study yet did not require rescue analgesia. Therefore, pain was an unlikely cause and anorexia was attributed to stress. Cats receiving supratherapeutic doses of either robenacoxib ⁵⁰ or grapriprant ¹⁶ had minimal adverse effects, and our study supports this finding. Therefore, both of these NSAIDs offer safe yet effective analgesia in healthy cats undergoing OVH.

The present study has some limitations. The use of a superiority trial instead of an NI trial may have provided a better assessment of the efficacy of grapiprant in cats. However, due to ethical concerns, a placebo group was not included. A longer acclimation period may decrease stress from environmental changes and alter pain score results in cats that require rescue analgesia. Another limitation of this study was the rounding up of the dose of grapiprant administered to the cats, which may have led to a lower treatment failure rate. While the use of hydromorphone may result in analgesia, the NI analgesic efficacy of grapiprant existed beyond the 2 h duration reported for this dose in cats.^54,55 Lastly, the utilization of highly skilled surgeons who impart minimal soft tissue trauma limits the extrapolation of the results of this study to patients undergoing procedures with significant tissue trauma. Future studies are needed to determine the efficacy of grapiprant in attenuating pain in cats other than that associated with mild to moderate acute pain.

Conclusions

The results of this study indicated that pre- and postoperative administration of oral grapiprant was non-inferior to robenacoxib in attenuating acute soft tissue pain for up to 24 h postoperatively, which was confirmed by the lack of difference in pain scores and MNT between the groups. Therefore, grapiprant may be a suitable analgesic alternative to robenacoxib in cats undergoing OVH.