Pain scoring systems in hospitalized horses with ocular disease

1 INTRODUCTION

Recognizing pain in hospitalized horses is essential for formulating analgesic plans that promote patient comfort and well-being. Despite its clinical importance, recognizing pain in hospitalized horses is challenging. Recently, species-specific pain scoring systems have been proposed as objective, quantifiable methods for identifying pain in horses. Various pain scales have been developed for several medical conditions of horses. Examples include the horse grimace scale (HGS) and behavior pain score (BPS), which evaluate changes in facial expression and deviations from normal behavior of horses, respectively, as indicators of underlying discomfort.^{1, 2}

Chronic ocular discomfort and prolonged medical treatment are thought to contribute to complications in horses with ocular disease, emphasizing the importance of early pain recognition and aggressive medical or surgical management of these patients. Studies have found variable risk of colic in hospitalized ophthalmic patients, with some showing increased risk of colic³ and others indicating similar risk of colic as elective orthopedic cases.⁴ Associations also have been made between post-operative colic and other factors including pre-operative fluconazole administration and retrobulbar nerve block in surgical ophthalmic patients, indicating that this risk is likely multifactorial.⁵

Current cornerstones of ophthalmologic treatment in horses are medical management using both topical and systemic medications, surgical intervention to treat lesions within or surrounding the eye, and enucleation. Although preserving both the globe and vision is always the desired outcome, enucleation might be elected for various reasons including inability to control pain in the presence of severe ophthalmic disease. Our objective was to evaluate the association between HGS and BPS and clinical progression in hospitalized horses with ocular disease. We hypothesized that changes in pain score over time would be associated with the clinical decision to continue medical management or pursue ocular surgery or enucleation. We further hypothesized that an increase in pain score would be associated with the decision to enucleate.

2 MATERIALS AND METHODS

2.4 Clinical progressions

Clinical progressions were categorized as, (a) discharge from hospital after medical treatment, (b) ophthalmic surgery (eg, keratectomy, conjunctival flap, amnion graft), or (c) enucleation.

Two time periods were used for analysis: time period 1 included all study horses except those that underwent surgery within 48 hours of admission. Time period 2 included all horses that underwent non-enucleation surgery and spanned the period from surgery until a second clinical progression (discharge from hospital, second surgery, or enucleation; Figure 1).

3 RESULTS

Sixty-five horses met the criteria for inclusion in the study. The study population consisted of 24 mares and 41 geldings. Median patient age was 13 years (range, 8 months-29 years), and breeds represented included Quarter Horses (14), Paints (12), Thoroughbreds (7), Warmbloods (7), Appaloosas (5), cross breeds (4), Friesians (3), draft breeds (3), Appendix Quarter Horses (3), Tennessee Walking Horses (2), Miniature Horses (2), ponies (2), and Morgan (1). Forty-two horses were presented for keratitis (corneal ulceration, stromal abscess, fungal keratitis, and immune-mediated keratitis), 9 horses for neoplasia (squamous cell carcinoma), 8 horses for primary uveitis (equine recurrent uveitis), and 6 horses for trauma (corneal puncture or laceration).

Of the 65 horses included in the study, 29 (45%) were discharged after exclusively medical management, 28 (43%) underwent ophthalmic surgery, and 8 (12%) underwent enucleation. Two horses (3%) had 2 ophthalmic surgeries performed. Surgical procedures included corneal graft (11), conjunctival graft (11), superficial (6) and deep (6) lamellar keratectomy, tumor resection (5), penetrating keratoplasty (3), strontium radiotherapy (3), corneal laceration repair (3), photodynamic therapy (3), cryotherapy (3), posterior lamellar keratoplasty (1), phacoemulsification (1), and transscleral cyclophotocoagulation (1), with several horses receiving some combination of these procedures at 1 time.

Hospitalization duration for all horses included in the study from admission to discharge, ocular surgery, or enucleation is summarized in Table 1. When considering all animals that were presented for keratitis, the median duration of hospitalization for patients discharged after exclusively medical management was 15 days (range, 4-59 days), for patients undergoing ocular surgery was 2.5 days (range, 1-44 days), and for enucleation was 3 days (range, 2-15 days). Five of 14 horses with keratitis that underwent ocular surgery did so on the first day of hospitalization, but no horses underwent enucleation on the first day of hospitalization.

When considering all animals that were presented for primary uveitis, the median duration of hospitalization for patients discharged after exclusively medical management was 8 days (range, 4-19 days); a single horse with uveitis underwent surgery 3 days after admission. No horses with uveitis underwent enucleation.

When considering all animals that were presented for neoplasia, the median duration of hospitalization for patients undergoing ocular surgery was 1 day (range, 1-2 days); the single patient that underwent enucleation did so after 4 days. No horses with neoplasia were discharged after exclusively medical management; all patients underwent surgery or enucleation.

All horses presented for trauma that underwent surgery did so on the first day of hospitalization; no horses underwent enucleation as their first clinical progression, and 1 horse was discharged after exclusively medical management 33 days after presentation.

Of the 65 horses, 17 underwent surgery within 48 hours of admission and 2 had insufficient pain score data for analysis, leaving 46 horses for analysis in time period 1. Twenty-eight horses were included in time period 2 (Figure 1). Because of insufficient group sizes, statistical comparisons were not performed for time period 2, thus only descriptive data are provided. Pain scoring data were available for 879 of 1203 days of hospitalization (72.9% of total days of hospitalization); the remaining 326 days were considered missing data. All days that had scores available had both HGS and BPS recorded.

Non-steroidal anti-inflammatory drug (NSAID) administration was common, with only 1 horse that had not received NSAIDs during hospitalization because of persistent azotemia detected on admission. All other horses received flunixin meglumine at variable doses and duration based on clinician preference and assessment of clinical progression. Two horses were switched from flunixin meglumine to firocoxib after being presumptively diagnosed with right dorsal colitis. Other analgesics administered included gabapentin (4 horses) and butorphanol (2 horses); 1 horse that received firocoxib and 2 horses that received gabapentin did not have sufficient pain score data available to assess trends, but data on these analgesics otherwise are summarized in Table S1. Additionally, 58/65 horses received atropine topically (median frequency, q6h; range, 4-24 hours), and 1 horse received trazodone.

Four horses (6.2%) were diagnosed with right dorsal colitis based on acute decreases in plasma protein and albumin concentrations with or without thickening of the right dorsal colon on transabdominal ultrasound examination. Two horses (3%) had evidence of acute kidney injury based on an acute increase in plasma creatinine concentration that resolved with IV fluid therapy and cessation of NSAID administration.

Three horses (4.6%) had signs of colic, all of which were mild and resolved within 24 hours with supportive care. All 3 horses were suspected of having non-strangulating large colon lesions based on diagnostic testing including transrectal palpation and transabdominal ultrasound examination; 1 of the 3 horses was suspected of having right dorsal colitis. Two of these horses had ophthalmic surgery performed under general anesthesia 2 and 45 days before the colic episodes; the third horse did not undergo surgery or general anesthesia. Of those horses developing colic, 1 horse (that underwent surgery and general anesthesia 2 days before) had an increase in pain scores (2 to 4 for HGS and 8 to 19 for BPS) before colic. The second horse (that underwent surgery and general anesthesia 45 days before) did not have pain scores recorded for 10 days leading up to its colic episode, and the third horse did not have a change in pain scores before its colic episode (HGS 0 and BPS 10 for 2 days before).

No horses died or were euthanized during the study period. However, 1 horse was euthanized 13 days after enucleation after an episode of diarrhea and acute laminitis.

3.1 Time period 1: Admission to clinical progression 1

Of the 46 horses included in time period 1, 29 (63%) were discharged after medical management, 9 (20%) underwent surgery after >48 hours of treatment, and 8 (17%) underwent enucleation. Time from admission to clinical progression ranged from 2 to 59 days. Median hospitalization duration for horses managed medically was 14 days (range, 4-59 days). Median time until surgery or enucleation was 3 days (range, 2-44 days) and 3.5 days (range, 2-15 days), respectively. Pain score data were available for 469/621 days of hospitalization (75.5%) for all horses included in time period 1. No differences were noted in HGS admission scores, HGS slope, BPS admission scores, or BPS slope between horses managed medically and those that underwent surgery (Figure 2; Table 2). Horses that underwent enucleation had higher HGS slopes (P = .02), lower BPS admission scores (P = .01), and higher BPS slopes (P = .01) when compared to horses managed medically. The BPS slope was also higher in horses that underwent enucleation compared to horses that underwent surgery (P = .04). The HGS and BPS scores at admission were moderately correlated (r² = 0.73; Figure 3).

TABLE 2. HGS and BPS admission scores (estimated by y-intercept) and trends (slope) during hospitalization for 46 horses included in time period 1.

	Clinical outcome
	Medical (n = 29)	Surgery (n = 9)	Enucleation (n = 8)
HGS Admission Score	1.61 (−0.14, 11.04)	3.00 (0.0, 10.00)	0.31 (−2.33, 6.00)
HGS Slope	−0.03a (−1.00, −0.36)	0.00 (−0.30, 0.50)	0.29a (−2.00, 2.50)
BPS Admission Score	8.00b (6.00, 16.13)	11.00 (5.72, 13.00)	6.00b (3.33, 6.80)
BPS Slope	0.00c (−1.00, 2.00)	0.00d (−1.60, 0.07)	0.27c,d (0.0, 2.00)

• Note: Values with matching superscripts (a, b, c, d) are significantly different. Values are reported as median (minimum, maximum).
• Abbreviations: BPS, behavior pain score; HGS, horse grimace scale.

Descriptive comparisons among animals in each disease category during time period 1 are summarized in Table S2. For horses with keratitis that were discharged after exclusively medical management, median HGS on admission was 2 (range, 0-12), median BPS on admission was 8 (range, 6-17), median change in HGS throughout the time period was −0.050 (range, −1 to 0.36), and median change in BPS throughout the time period was −0.029 (range, −1 to 1). For horses with keratitis that underwent ocular surgery, median HGS on admission was 3.5 (range, 0-10), median BPS on admission was 10 (range, 6-13), median change in HGS throughout the time period was 0 (range, −0.30 to 0.50), and median change in BPS throughout the time period was 0 (range, −1.6 to 0.070). For horses with keratitis that underwent enucleation, median HGS on admission was 1 (range, 0-5), median BPS on admission was 6 (range, 6-11), median change in HGS throughout the time period was 0.37 (range, −2 to 2.5), and median change in BPS throughout the time period was 0.49 (range, 0-2).

For horses with primary uveitis that were discharged after exclusively medical management, median HGS on admission was 0 (range, 0-12), median BPS on admission was 6 (range, 6-19), median change in HGS throughout the time period was 0 (range, −0.65 to 0.14), and median change in BPS throughout the time period was 0 (range, −0.70 to 2). A single horse with uveitis underwent ocular surgery; this animal had an admission HGS of 0, admission BPS of 6, and no change (slope = 0) in both HGS and BPS over hospitalization. No horses with uveitis underwent enucleation.

Most horses with ocular neoplasia (7/9 horses) underwent surgery within the first 48 hours of hospitalization and 1 horse had insufficient pain score data; these therefore were not included in time period 1. The single horse with neoplasia that underwent enucleation during time period 1 had an admission HGS of 0, admission BPS of 6, a change in HGS of 0.2, and no change in BPS (slope = 0).

Five of 6 horses with traumatic ocular injuries underwent ocular surgery on the first day of hospitalization and were excluded from time period 1. The only horse that did not undergo surgery was discharged after medical management; this horse had an admission HGS of 0, admission BPS of 6, a change in HGS of −0.0032, and a change in BPS of 0.

4 DISCUSSION

In our study of hospitalized horses with ocular disease, HGS and BPS pain scores at admission and over time differed in horses managed medically or with ophthalmologic surgery compared to those ultimately undergoing enucleation. Notably, horses requiring enucleation had temporal increases in both HGS and BPS, whereas horses managed medically or with ophthalmologic surgery tended to have stable or decreasing HGS and BPS scores during hospitalization. Horses with keratitis comprised the majority of animals included in time period 1 (36/46 horses, 63%), and subjectively, these horses experienced decreases in HGS and BPS in the time before discharge or surgery and increases in HGS and BPS in the time before enucleation. Sample sizes for horses with uveitis, neoplasia, or trauma were insufficient to draw conclusions for these disease categories. The inability to control pain despite intensive medical management is a common clinical indication for surgical or procedural intervention and likely contributed to the recommendation to perform enucleation in our study population. However, factors other than pain might contribute to the decision to perform enucleation including poor prognosis for vision, low chance of response to medical treatment because of type or severity of the disease process, or financial limitations. In general, the HGS and BPS scores were relatively low for many patients in our population. These results could reflect effective pain management, the patient population, or a combination of factors.

Interestingly, horses managed medically had higher admission BPS and HGS scores (signifying a higher level of discomfort) when compared to horses undergoing enucleation, emphasizing the importance of monitoring pain over time as opposed to at a single time point when using patient comfort to support clinical decisions. In our study population, a higher pain score at admission might have prompted more aggressive medical treatment, bolstering clinical response and aiding in preserving the globe. Additionally, horses presented for acute ocular conditions may have displayed severe pain on admission (particularly if no analgesics had been administered before admission) that ultimately was more responsive to medical treatment, whereas horses referred for chronic recalcitrant conditions may have presented with more moderate pain but an overall poor prognosis for successful medical management. However, because many horses underwent surgery within 24 hours of admission and may not have had pain scores documented before surgery, it was not possible to determine if pain score at admission also might have been associated with the decision to pursue globe- and vision-sparing surgical intervention rather than medical management.

No differences were observed in HGS and BPS scores at admission or throughout hospitalization between horses that were discharged after medical treatment and horses that underwent ophthalmic surgery. Although some factors might dissuade both clinicians and owners from electing ophthalmic surgery (eg, complications from general anesthesia, financial constraints), these procedures also may increase the likelihood of ultimately preserving vision and the globe, supporting their utility as an adjunct to medical treatment even in patients with stable or improving pain levels. In our study population, 17 of the 65 horses underwent ophthalmic surgery on the first day of hospitalization, indicating that the choice often was made to pursue surgical intervention even before medical treatment had been initiated and a response to treatment assessed. In these cases, factors other than patient comfort (eg, type of lesion, severity of condition, clinical experience), likely influenced the decision to recommend immediate surgical intervention. For example, treatments for certain conditions such as corneal laceration or neoplasia can be inherently surgical despite patient comfort. The majority of horses presented for neoplasia or trauma underwent surgery on the first day of hospitalization and were excluded from time period 1. All but 1 horse that presented with ocular trauma underwent surgery within 48 hours of admission, indicating that the traumatic nature of the condition rather than clinical progress was likely the determining factor in the decision to pursue surgical intervention. In addition, 8/9 horses presented for neoplasia underwent surgery within 48 hours of hospitalization, indicating that these interventions were likely planned with owners before presentation or were elected immediately after assessment because of the necessity for excising or debulking masses.

Few previous studies have focused on pain scoring in horses with ophthalmologic disease. One investigated head-related (rather than exclusively ophthalmic) pain and found good interobserver reliability and significantly higher scores in afflicted versus control animals, with a significant reduction in score after treatment.⁶ Another investigated the development of a pain scale (the equine ophthalmic pain scale [EOPS]) for use in horses with ophthalmologic disease and included factors such as behavior, ocular manifestations of pain such as blepharospasm and lacrimation, and physiologic findings. This study found good intra- and inter-observer reliability and determined that there was a significant difference in scores between control and affected animals as well as a decrease in score after treatment.⁷ Collectively, these previous studies support the utility of pain scoring in horses with ophthalmic pain.

One commonly recognized sign of ocular pain in horses (blepharospasm) was not explicitly recorded in our study, and therefore a correlation between blepharospasm and HGS, BPS, or other factors could not be assessed. However, as a measure of facial expression, the HGS includes assessment of orbital tightening and tension above the eye. Although blepharospasm is an important indicator of ocular pain, it is not dependable as a sole measure and can be unreliable because of the evolutionary imperative of prey species to avoid exhibiting weakness as well as the variability in individual pain tolerances and displays. Therefore, although blepharospasm should be considered as a factor when assessing pain in horses with ocular disease, other measures such as pain scales including the HGS and BPS may provide additional information.

All horses except 1 in our study were treated with NSAIDs while hospitalized. Most horses were treated with flunixin meglumine, although 2 horses received firocoxib instead after being diagnosed with right dorsal colitis. The dosage and duration of NSAID treatment were dependent on clinician assessment of the horse's pain and progression of the primary pathology, and therefore it was not possible to objectively assess the effect of a standardized analgesia regimen on these patients. However, subjectively, all horses that received butorphanol or gabapentin in addition to flunixin meglumine or firocoxib in place of flunixin meglumine experienced a decrease in HGS and BPS during the course of analgesic administration except for 1 horse treated with butorphanol that experienced a decrease in HGS but increase in BPS during treatment; this horse had an overlapping course of gabapentin and experienced a decrease in HGS and BPS over the duration of this additional treatment. Also, despite a decrease in pain score, a second horse underwent enucleation after 6 days of butorphanol administration, indicating that factors other than pain were considered in making the decision to enucleate the eye. Despite the decrease in pain score seen during treatment with these medications, other factors including concurrent administration of flunixin meglumine during treatment with butorphanol or gabapentin, administration of atropine, or other factors might have contributed to alleviating pain.

Previous studies have produced conflicting information on the prevalence of colic in horses with ocular disease, with 1 study finding no significant difference in horses hospitalized for ophthalmic (8%) or orthopedic (5%) disease⁴ and another finding a high prevalence of colic (21.4%) in horses hospitalized for ocular disease.³ Notably, in the latter study, 40.3% of horses with colic had moderate to severe signs, and 13.9% of colic cases were attributable to cecal impactions.³ In our study, only 3 horses (4.6%) developed colic, and all cases were considered mild. The colic signs experienced by 1 horse were attributed to right dorsal colitis, and in another colic were attributed to a non-specific large intestinal non-strangulating obstruction. The third horse had a suspected large colon impaction 2 days after ophthalmic surgery. Increases in HGS from 2 to 5 and in BPS from 8 to 19 also were noted in the days leading up to the colic episode. Postanesthetic colic incidence rates have been reported to range from 3.6% to 15.5%,^{5, 8-10} indicating that anesthesia may play a role in predisposing hospitalized equine patients to colic. Of the 8.7% of horses that developed delayed fecal passage or colic after undergoing general anesthesia for elective procedures in 1 study, 16.7% of these animals developed a large or small colon impaction, and 11.1% developed cecal impaction or dysfunction.⁸ In another study, 3.7% of horses undergoing anesthesia for elective procedures developed signs of colic, and 30% of these cases were attributed to large intestinal impactions.¹⁰ Therefore, colic and large intestinal impactions are rare but documented adverse effects of general anesthesia, which may make separating the effects of anesthesia and pain in ophthalmic surgical cases more complicated.

Although IV atropine can be associated with colic and decreased gastrointestinal motility,¹¹ results regarding topical ophthalmic atropine are more variable and likely associated with dosage and frequency of administration. One study found that hourly topical administration of atropine led to decreased gastrointestinal motility on auscultation in all horses included in the study, and 4 of the 6 horses had signs of colic.¹² This frequency of atropine administration was selected for experimental purposes and does not mimic frequencies commonly used clinically. Notably, the median atropine dosing interval for our study was 6 hours. Administration every 3 hours has been associated with decreased gastrointestinal motility whereas administration every 6 hours was not.^{13, 14} Therefore, atropine is less likely to increase risk of colic at the frequencies commonly used in clinical treatment of horses with ocular disease such as the frequencies (every 4 to 24 hours) administered to patients in our study.

Although still relatively infrequent, right dorsal colitis was the most prevalent complication observed in our study (4 horses, 6.2%). In addition, 2 horses had evidence of acute kidney injury. Both right dorsal colitis and renal papillary necrosis are well-recognized adverse effects of NSAID administration in horses, with horses that are off feed more predisposed to right dorsal colitis and horses that are not drinking adequately more predisposed to both conditions.^{15, 16} Several studies have investigated differences in the proportion of time spent engaging in behaviors such as eating and drinking in horses subjected to painful surgical procedures such as arthroscopy¹⁷ and castration¹⁸ and have suggested than horses in pain spend less time engaging in both behaviors, thus potentially predisposing to the adverse effects of NSAIDs if pain is not adequately managed.

Our study had some limitations, including small sample size, missing pain scores, variability in initial diagnoses and treatments, and a relatively large number of individuals (5 veterinarians over 2 years) who determined pain scores. Pain scores were not recorded for 27.1% of days of hospitalization for the entire study period, potentially affecting assessment of trends. Despite the relatively large number of veterinarians performing scoring, the HGS has been shown to have high inter-observer reliability,¹ mitigating the effect of multiple observers. Similar studies of inter-observer reliability have not been performed for the BPS, but the EOPS, which incorporates the BPS as part of its grading, was found to have good inter-observer reliability overall, with components of the BPS having fair to excellent reliability.⁸ Finally, because of the small sample size, statistical analysis comparing disease categories could not be performed and assessment was limited to descriptive summary. Further research is warranted to develop an understanding of how disease type contributes to pain score and clinical decision-making in horses with ocular disease.

In summary, horses ultimately undergoing enucleation because of ophthalmologic disease showed increases in HGS and BPS over time when compared to horses discharged after medical treatment or ocular surgery. Conversely, higher HGS and BPS at admission were associated with discharge from hospital when compared to admission scores for horses undergoing enucleation. The HGS and BPS represent potentially useful tools for assessing pain and assisting in clinical decision-making in hospitalized horses with ocular disease.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION

Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Authors declare no IACUC or other approval was needed.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.