Long‐term postoperative pain evaluation in dogs with thoracolumbar intervertebral disk herniation after hemilaminectomy

Abstract Background Chronic neuropathic pain is a common complication in people with spinal cord injury (SCI) but has not been investigated in dogs. Objective To determine the reliability of measuring spinal mechanical sensory thresholds (MSTs) in dogs and to compare MSTs of healthy dogs and dogs with SCI caused by acute thoracolumbar intervertebral disk extrusion after hemilaminectomy over a 1‐year period. Study Design Prospective study. Animals Thirty‐two healthy and 40 SCI dogs. Methods Dogs were divided into group 1 (healthy Dachshunds), group 2 (healthy dogs including several breeds), and SCI group. The MSTs were measured using algometry at an incision (thoracolumbar) and control site. Dogs in group 1 were tested once; those in group 2 were tested for 5 consecutive days; and SCI dogs were tested on days 7, 14, 28, 42, 180, and 365 postoperatively. The MSTs were compared among days in healthy and SCI dogs and between SCI and healthy dogs using mixed effect models. P < .05 was considered significant. Results At the incision site of SCI dogs, MST was significantly lower than in healthy dogs for 42 days postoperatively, but not subsequently. However, 4/27 dogs had control site MST below the reference range 1 year after surgery. Conclusions and Clinical Importance Mechanical sensory thresholds normalize by 6 months after surgery in most dogs with SCI. Approximately 15% of SCI dogs may develop chronic neuropathic pain. Improving long‐term pain assessment of SCI dogs is important for offering treatment options and advising owners.


| INTRODUCTION
Pain assessment is essential for effective patient care in veterinary medicine. According to the International Association for the Study of Pain (IASP), "pain is an unpleasant sensory and emotional experience, associated with actual or potential tissue damage," and will impact the patient's quality of life (QOL). 1 It can be classified as nociceptive or neuropathic. Nociceptive pain is caused by a noxious stimulus that is processed by a normally functioning somatosensory system as a physiologic response and has been studied extensively in veterinary patients. [2][3][4][5][6][7][8] Neuropathic pain is an inappropriate response caused by damage or dysfunction of the somatosensory system, and human patients report burning, stabbing, numbness, and tingling sensations. 1,3,5,6,[8][9][10][11][12][13][14] It is likely to be more difficult to identify neuropathic pain in veterinary patients, but dogs with Chiari malformation and syringomyelia are thought to exhibit signs consistent with this type of pain. [15][16][17] Approximately 50% to 80% of people with spinal cord injury (SCI) develop chronic neuropathic pain, and it is possible that the same is true in dogs after SCI. 4,8,18 Intervertebral disk herniation is the most common cause of SCI in dogs, and successful outcome is based on neurological recovery. The possibility that affected dogs suffer chronic neuropathic pain that impacts the quality of their recovery has not been investigated because of the challenges associated with recognizing and quantifying pain, and in particular, neuropathic pain in pet dogs.  19,20 More objective methods such as quantitative sensory testing using a variety of different stimuli (mechanical, thermal, vibration) also have been used in dogs. 17,[21][22][23][24][25][26][27][28][29][30][31][32][33] Pressure algometry is a technique used to establish mechanical sensory thresholds (MSTs). Gradually increasing pressure is applied to the area of interest until a response is obtained that indicates a noxious threshold has been reached, and studies have demonstrated its reliability and feasibility in dogs. 21,24,29,33 Our aims were to establish the range and reliability of mea-

| Data collection
Physical and neurologic examinations were performed on all dogs by the same clinician (Natalia Zidan). Minimum data collected from all dogs included body weight, breed, sex, and age. During the neurologic examination, the cutaneous trunci reflex (CTR) also was recorded and categorized as normal or hyperesthetic. A hyperesthetic CTR was characterized by a strong reflex contraction of the cutaneous trunci muscle in response to light touch to the skin (Video S1). In addition, for the SCI dogs, severity of neurological signs was graded using the modified Frankel score (MFS) 39 and the modified open field score (OFS) (ranging from 0 to 12) at presentation and at days 42 and 365 postoperatively. 40 At each evaluation of SCI dogs, postoperative pain was assessed by the same clinician (Natalia Zidan) using a surgical site manipulation score (SSMS; Supporting Information Data S1), 41 before measurement of MST.
Assessments from days 7, 14, 28, 42, 180, and 365 postoperatively were collected for analysis. In addition, at follow-up evaluations, owners were asked if dogs were showing any signs of pain at home.

| Statistical analysis
Summary data were generated on demographics including age, sex, breed, and were compared among groups. Continuous data were evaluated for normality of distribution using the Shapiro-WilkW test.
Because all variables were nonnormally distributed, data were reported as median and range and a nonparametric Wilcoxon rank sum test was used to compare variables among groups.
Baseline paraspinal MSTs were established from the single-day testing summary data on control and thoracolumbar (incision) site results in healthy Dachshund dogs (control group 1). The MSTs of control and thoracolumbar sites were compared using a mixed effect model in which anatomical site was included as a fixed effect and dog as a random effect. To assess interobserver variability, the MST measurements of 2 observers were modeled using analysis of variance in which observers and dog were included as a fixed effect.
The intraclass correlation was reported as the Shrout-Fleiss reliability single score.
Mixed effect models were generated and least squares means estimated and compared to evaluate: test-retest reliability of daily pressure algometry in control group 2 (5 daily measurements), changes in para-

| Mechanical sensory thresholds
Minimal restraint was required to perform MST in control and SCI dogs. One control dog did not want to be held, but tolerated the test well.

| Control group 1: Healthy Dachshunds
In healthy Dachshunds tested once, the mean paraspinal MSTs of the control site and thoracolumbar site were 9.95 ± 0.68 lbs and 8.65 ± 1.41 lbs, respectively, with the thoracolumbar site MST being significantly lower than the control site (P < .001). Agreement between observers was excellent (κ = 0.91; Figure 1).

| Group 2: Test-retest reliability
When testing control group 2 dogs once a day for 5 days, no significant interaction between algometry site and measurement day was identified (P = .81), indicating that no significant change in algometry measurements occurred over time and that the dogs tolerated repeated measurements.

| Cutaneous trunci reflex and spinal palpation
The CTR was evaluated in dogs of all 3 groups. Three dogs (

| Open field score in SCI dogs
To determine whether motor score was associated with MSTs, the OFS also was evaluated in 39/40 dogs with SCI at 6 weeks and in 27/40 dogs that returned for reevaluation at 1 year postoperatively.
No relationship was found between the OFS at 6 weeks (median, 9; range, 2-11) and 1 year (median, 11; range, 5-11) after surgery and the MST at incision and control site in these dogs (P = .9).

| DISCUSSION
We demonstrated that paraspinal MSTs can be measured reliably using pressure algometry, with excellent interobserver reliability and accept- Pressure algometry is considered a reliable tool to evaluate pain in humans [42][43][44] and although it has been used for research in dogs, [45][46][47] relatively few studies have evaluated its feasibility and reliability in dogs. 24,25,29 To evaluate the utility of pressure algometry as an objective means of measuring MST and, by implication, pain, we first evaluated MSTs at 2 different paraspinal sites in 22 healthy Dachshunds.
The interscapular region was selected as a control site because this relatively protected site is less mobile, is an unusual location for IVDE and was considered sufficiently distant from the normal sites of IVDE to serve as a useful control. Indeed, we found that MSTs were higher at this site than at the TL junction in healthy dogs. This observation is consistent with previous studies in human and veterinary medicine that have reported different MSTs at different locations because of differences in nociceptor density, distribution, and types, with some regions being highly innervated and having smaller receptive fields compared to others. 25,48,49 Furthermore, the difference in skin type and amount of overlying soft tissue can contribute to differences in threshold responses. 25,50,51 We then compared interobserver reliability and found that there was high agreement between observers. In general, pressure algometry measurements are more reliable if performed by a single observer, 26 but in our study the observers practiced the application of pressure together and their measurements had high consistency.
One study showed that learning avoidance and anticipation may occur in some dogs over time 25 whereas others found that dogs did not develop local hyperesthesia, learned avoidance, or intolerance. 24,29 We evaluated test-retest reliability in neurologically normal healthy MST threshold analgesia should be provided. Another potential limitation of our study is that the MSTs of sites below the lesion were not evaluated. Spinal cord injury patients may develop signs of neuropathic pain not only at the level of the injury, but also above and below the lesion. 13, 60 We evaluated the level of the injury and areas above the injury. Although investigating MSTs below the level of the injury is important, it was difficult to identify a region in every dog that was not in some way involved with the surgical incision or within a short distance of the site of injury. Future studies on pain below the level of the injury might be indicated after the surgical incision heals, particularly because such neuropathic pain often becomes evident only in the chronic phase of injury. 4,13 Finally, despite efforts to minimize any behavioral influence, assessing MST in animals still can be challenging. Differences in individual temperament may affect the dog's response to pressure algometry, and some studies have shown that hyperactive dogs also appear less willing to cooperate with MST testing, leading to difficulties in data collection. 24 We determined that pressure algometry can be used to establish