Risk factors for laminitis and nonsurvival in acute colitis: Retrospective study of 85 hospitalized horses (2011‐2019)

Abstract Background Acute colitis is a serious cause of morbidity and death in horses. Recent studies have compared clinical features of coronavirus and salmonellosis, but no study has compared clinical features of enteric salmonellosis, coronavirus, and neorickettsiosis. Hypothesis/Objectives To identify risk factors for laminitis and nonsurvival to discharge in horses with enteric salmonellosis, coronavirus, or neorickettsiosis. Animals Eighty‐five horses hospitalized for acute colitis from 2011 to 2019. Methods Retrospective case series. Medical record review (2011‐2019) of adult (≥2 years) horses with colitis. Primary outcomes were laminitis and survival to discharge. Multivariable logistic regression was performed to assess association between variables and the development of laminitis. Stepwise Cox regression was performed to assess association between variables and survival. Results Seventeen of 85 (20%) horses developed laminitis during hospitalization. Neorickettsiosis cases (11/26, 42%) were more likely to develop laminitis than coronavirus (0/16, 0%) cases (odds ratio [OR] 24.48; 95% confidence interval [CI]: 1.33‐451.74, P = .03). There was no significant difference in laminitis between salmonellosis and neorickettsiosis cases (OR 0.27; 95% CI: 0.07‐1.07, P = .06). Admission heart rate (OR 1.08; 95% CI: 1.02‐1.15, P = .01), total solids (OR 0.17; 95% CI: 0.06‐0.54, P = .003), band neutrophils (OR 1248.47; 95% CI: 6.62‐235 540, P = .008), and bicarbonate concentration (OR 0.68; 95% CI: 0.5‐0.92, P = .01) were predictive of development of laminitis during hospitalization. Sixty‐three of 85 (74%) horses survived to discharge: 16/16 (100%) coronavirus cases, 17/26 (65%) neorickettsiosis cases, 14/20 (70%) salmonellosis cases, and 16/23 (70%) unknown cases. Packed cell volume (hazard ratio [HR] 1.17; 95% CI: 1.09‐1.26, P < .001), L‐lactate concentration (HR 1.33; 95% CI: 1.14‐1.55, P < .001), and development of laminitis (HR 7.07; 95% CI: 1.67‐29.95, P = .008) were retained in the final multivariable model for prediction of nonsurvival to discharge. Conclusion and Clinical Importance Nonsurvival and laminitis rates were high, likely related to the presence of neorickettsiosis in the region.

Conclusion and Clinical Importance: Nonsurvival and laminitis rates were high, likely related to the presence of neorickettsiosis in the region.

K E Y W O R D S
coronavirus, neorickettsiosis, salmonellosis 1 | INTRODUCTION Acute colitis or typhlocolitis is a serious cause of morbidity and death in horses. 1 Horses with acute colitis often present with diarrhea, fever, colic, and signs of endotoxemia. Common causes of acute infectious colitis in adult horses include Salmonella enterica, Neorickettsia risticii (Potomac Horse Fever), Clostridium difficile and perfringens, and equine coronavirus. 1 Equine coronavirus is a recently emerging enteric pathogen of horses with a wide variety of clinical presentations, from acute severe diarrhea to fever and colic to lymphopenia. [2][3][4][5][6][7] A recent study of enteric coronavirus and salmonellosis in 42 horses reported a survival rate of 98%, and 1/42 (2.4%) horses in that study developed laminitis. 5 This, however, is in stark contrast to the clinical impression for survival and laminitis in regions of the country where neorickettsiosis is a common cause of colitis. For example, a study of enteric neorickettsiosis in 44 horses reported a 73% survival rate and 16/44 horses (36%) developed laminitis. 8 Risk factors associated with nonsurvival in that study are reflective of severity of disease and included hemoconcentration, hypochloremia, hyponatremia, azotemia, and duration of hospitalization. 8 The severity of neorickettsiosis varies widely between geographic regions. 9 Despite the various etiologic agents associated with colitis and the variability of clinical signs associated with enteric coronavirus, horses presenting for acute colitis can have indistinguishable clinical signs and clinicopathologic data. To date, no study has compared the clinical features of enteric salmonellosis, coronavirus, and neorickettsiosis within a single hospitalized population to determine risk factors for laminitis and nonsurvival to discharge. Therefore, our objectives were to (a) compare the clinical features of enteric salmonellosis, enteric coronavirus, and enteric neorickettsiosis and (b) identify risk factors for the development of laminitis as well as nonsurvival to discharge in horses with colitis caused by Salmonella, coronavirus, and N risticii.  for an additional site of infection (pneumonia, thrombophlebitis). Fifty percent of horses with a diagnosis other than neorickettsiosis received antimicrobials.

| MATERIALS AND METHODS
A diagnosis of laminitis was based on clinical examination by the supervising clinician and was defined as an Obel grade of 1, 2, 3, or 4 with subjectively increased digital pulse pressure, clefting of the coronary band, or both, with or without confirmatory radiographic changes. 10,11 Specific treatments for laminitis included nonsteroidal anti-inflammatory drugs, digital cryotherapy, solar support, pentoxifylline, lidocaine continuous rate infusion, opioids (morphine, butorphanol), gabapentin, and in severe cases, coronary band grooving and foot casts. Nonsurvival was defined as death or euthanasia before discharge; horses who survived to discharge were not followed beyond discharge. In horses that were euthanized, reason for euthanasia was poor prognosis associated with their primary condition (colitis) or because of poor prognosis and poor quality of life associated with severe laminitis.

| Statistical analysis
Statistical analysis was performed using standard statistical software (Stata 15.1MP, StataCorp, State College, Texas). Data were assessed for normality using the Shapiro-Wilk test for normality. Normally distributed variables were reported as mean ± SD. Non-normally distributed variables were reported as median and range. Categorical variables were compared using a Chi-Square test. For continuous variables, normal data were compared across groups using a 1-way ANOVA with post hoc Bonferroni correction, whereas non-normal data were compared across groups using Kruskal-Wallis. Horses were grouped based on final diagnosis, and coronavirus was set as the reference, with the exception of survival analysis, where neorickettsiosis was set as the referent because no coronavirus cases died. Univariate logistic regression was performed to assess association between variables and the development of laminitis, followed by lasso regression to build significant variables into the final predictive multivariable logistic model. Lasso regression is a novel method for selection in multivariable models. In Lasso regression, independent variables are selected and fit to linear, logistic, probit, and Poisson models, and the results of lasso regression are used for multivariable model selection. 12 Univariate Cox regression was performed to confirm the association with the outcome of interest (survival), followed by stepwise Cox regression with Breslow method for ties for determination of the final multivariable Cox regression model with factors that were significantly associated with nonsurvival. Because of complete separation of data, Firth logistic regression was then used. All analysis was conducted with 2-sided tests of hypotheses and P-value of <.05 set as the criterion for statistical significance.

| Animals
Eighty-six horses met the inclusion criteria. A single horse was identified which was both coronavirus and Salmonella PCR positive, and this horse was excluded from further analysis. Therefore, 85 horses were included in the final study. Median age was 8 years (range, 2-26 years). There

| Diagnosis
Twenty-six of 85 (31%) horses were classified as neorickettsiosis, 20/85 (24%) were classified as salmonellosis, 16/85 (19%) were classified as coronavirus, and 23/85 (27%) were classified as unknown. Table 1 summarizes clinical findings at admission stratified by final diagnosis. Coronavirus cases were significantly less likely to have T A B L E 1 Clinical findings of 85 horses hospitalized for acute colitis, stratified by diagnosis. For continuous data, non-normal data are listed as median and range with P value generated by Kruskal-Wallis, and normal data are listed as mean ± SD with P value generated by 1-way ANOVA with post hoc Bonferroni for any significant values. For categorical data, P value is generated by chi-square. % represents within-column frequency .69
This supports the findings of a previous study which determined that the disease features of enteric coronavirus and salmonellosis were similar in horses and supports the difficulty in distinguishing etiologic agents of colitis based on clinical findings. 5 Of note, our record search did identify coinfection with coronavirus and Salmonella in a single horse, although this horse was excluded for the final study. Coinfection with coronavirus and Salmonella has been reported in a horse. 2 In contrast to a previous study of enteric coronavirus and salmonellosis, where a single horse developed laminitis, 20% of horses in our study developed laminitis. 5 The incidence of laminitis associated with colitis in the veterinary literature varies, from a low of 2.4% in the aforementioned study of salmonellosis and coronavirus cases, to 36% in a study of neorickettsiosis cases. 5,8 Almost two-thirds (11/17) of laminitis cases in our study were associated with neorickettsiosis, whereas the remainder were salmonellosis and unknown cases. The prevalence of neorickettsiosis in our geographical region likely explains the difference in laminitis rates as compared to the previous study, as the region of the country in which that study was performed is not affected by enteric neorickettsiosis. 5 No coronavirus cases in our study developed laminitis, and this coincides with both our clinical impression and prior research. 2,4,5 Digital cryotherapy in horses with colitis reduces the incidence of laminitis from 33% in horses not treated with cryotherapy to 10% in horses treated with cryotherapy. 10 Importantly, cryotherapy has been most effective at prevention of laminitis, not necessarily treatment of laminitis. Despite the more widespread use of digital cryotherapy in colitis cases as a preventative method against laminitis, regions of the country continue to see higher incidence of laminitis, most commonly associated with neorickettsiosis. In fact, the study evaluating digital cryotherapy found that 7 horses developed laminitis despite digital cryotherapy, and 6 of those 7 were neorickettsiosis cases. 10 In our study, neorickettsiosis cases were significantly more likely to present with or to develop laminitis than coronavirus and unknown cases, despite the widespread use of digital cryotherapy in colitis cases in our hospital. Recent literature has compared various methods for digital cryotherapy, and while 1 study found that the commercially available sleeve-style ice boot used in our hospital achieved appropriately low skin and lamellar temperatures, 13 another study found that this sleeve-style boot did not maintain hoof surface temperatures below 10 C. 14  suggested that a 2-step approach of fecal and nasal swab testing should be utilized for the diagnosis of coronavirus in horses. 19 This approach was not used in any of the horses in this study because of the retrospective nature of the study, but could be indicated in the future for coronavirus testing in horses. We initially performed statistical analysis both with and without inclusion of the unknown group, and its presence or exclusion did not lead to significant differences in results. Furthermore, survival analysis was complicated by the lack of death in the coronavirus group, resulting in astronomically high HRs or incalculable P values when comparing survival between groups.
Finally, as previously mentioned, the development of severe laminitis might have led to the decision to euthanize a horse and therefore its inclusion in the final model for nonsurvival could be misleading.
In conclusion, our study shows that horses with enteric neorickettsiosis are more likely to develop laminitis than horses with coronavirus. Additionally, we identified risk factors for the development of laminitis and for nonsurvival. This information can be applied in the future to predict the risk of laminitis and nonsurvival in a clinical setting.
Further research with multi-institutional studies is indicated to determine national rates of laminitis and survival in equine colitis cases.

ACKNOWLEDGMENT
No funding was received for this study.

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.