Reasons for performing study: There are few objective data on return to use and performance in horses following colic surgery.
Objective: To investigate return to functional use of horses following colic surgery and factors associated with a negative outcome.
Methods: The North Carolina State University Equine Colic Database was reviewed for horses that underwent exploratory celiotomy for colic (2003–2010). Horses were excluded from the study if they survived <6 months, had no intended use preoperatively, or if further data were not available at attempted follow-up. Information retrieved included history, background, use, and selected pre-, intra-, and post operative factors. Telephone interviews were used to obtain follow-up data. Logistic regression was used to investigate associations between clinical data and outcome, reported as odds ratios with a 95% confidence interval and corresponding P value.
Results: Of patients surviving to 6 months, 133/195 (68%) were performing their intended use and 85/156 (54%) were at or above preoperative performance. At one year, 145/190 (76%) horses were performing their intended use and 101/153 (66%) were at or above preoperative performance. Animals were significantly less likely to return to use/performance if they had a previous celiotomy, stall rest for an orthopaedic condition, a nonstrangulating lesion type, incisional hernia, diarrhoea or laminitis.
Conclusions: The overall prognosis for return to use and performance following colic surgery is fair to good. Multiple pre- and post operative factors may affect the likelihood of return to use and performance.
Potential relevance: Targeted owner education regarding preoperative lameness, post operative rehabilitation and treatment for complications, such as incisional hernioplasty, may help inform owners about their horse's potential for return to use and performance following colic surgery.
Return to intended use and performance of horses following colic surgery is of major concern to the equine industry. The majority of horses in the USA (98.8%) have been reported to have a primary use such as pleasure riding, competition/show, breeding or farm/ranch work . Frequently, the potential for return to use and performance must be weighed against the cost of surgery, the possibility of intensive post operative care, and prolonged convalescence. Objective data on successful post operative return to intended use and performance following colic surgery would assist owners and veterinarians in the decision-making process.
Numerous studies have investigated horse survival following colic surgery. Reported variables affecting survival include patient background, duration of colic, physical examination findings, haematology findings, peritoneal fluid analysis, lesion type, surgical procedure performed and duration of surgery [2–12]. Two previous studies reported an 81% and 90.1% return to expected performance following colic surgery, respectively [13,14]. Return to intended use and performance following exploratory celiotomy with respect to patient details, clinical findings, surgical findings or post operative complications is poorly defined.
The purpose of this study was to determine the return to functional use of horses following colic surgery, factors that may predispose to a suboptimal outcome and the owner's perception of performance post operatively.
Materials and methods
The North Carolina State University (NCSU) Equine Colic Database was used to review data from horses that underwent exploratory celiotomy for acute colic from June 2003 to March 2010. Horses were excluded from the study if they survived <6 months, had no intended use preoperatively, or if further data were not available at attempted follow-up. In horses that underwent multiple colic surgeries at NCSU within a 6-month interval (n = 5), the horse was recorded as having had a previous celiotomy, and follow-up performance data were collected following the last surgery. In horses that underwent sequential celiotomies at NCSU at an interval >6 months (n = 1), the horse was recorded as having had a previous celiotomy, but separate follow-up data was collected for each incident.
Information retrieved from the database included age, sex, breed, intended use, previous surgical colic history, history of stall rest for orthopaedic conditions, duration of anaesthesia, surgical data and post operative complications. Physical examination findings, clinical pathology data and presence of nasogastric reflux were all recorded as measured at admission. Age was described as a median and range. The categories for age were 0–14 and ≥15 years. Reference range values provided by the laboratory were used to divide haematological variables into groups (below, within and above the reference range). Packed cell volume was considered abnormal if above 40%. Heart rate was considered elevated at 60 beats/min. Rectal temperature was considered elevated if ≥38.6°C. Nasogastric reflux was defined as greater than 2 l of gastric reflux fluid obtained at initial examination. Surgery and anaesthesia times were considered prolonged if ≥2 h or ≥3 h, respectively. Laminitis was defined as lameness in one or more feet with concurrent clinical signs of increased digital pulses and/or palpable heat at the dorsal hoof wall. Laminitis was recorded if it developed any time between surgery and 6-month follow-up. Laminitis was termed in-hospital if the complication developed during post operative hospitalisation. Catheter-related complications were defined as swelling and evidence of inflammation around the catheter site or vein, with or without occlusion. Diarrhoea was defined as passage of non-formed faeces with a consistency insufficient to remain above the stall bedding. Post operative reflux was defined as >2 l of nasogastric reflux obtained post operatively. Incisional complications were defined as dehiscence or any incisional drainage. Post operative colic was defined as recurring signs of colic prior to discharge.
Owners were contacted for follow-up at one week, one month, 6 months and annually thereafter. Data recorded included if the patient was alive and presence of complications (recurrent colic, incisional hernia, laminitis or other). Recurrent colic was defined as more than one episode of colic in the follow-up period. Return to intended use and return to full level of performance were recorded for the intervals of 6 months and one year post operatively. Return to use was defined as horses who resumed their preoperative use or discipline. For young animals or those in training, return to use was defined as the ability to achieve their intended use as expected. Return to full performance was defined as the ability to perform athletically at a level that was equal to or greater than preoperative levels, or expected level of performance for young animals. Owners were asked to rate their overall satisfaction regarding patient usability or performance post operatively as compared with the preoperative or expected use/performance of the animal (excellent, good, fair, poor, no comment). Intended use of the patient was categorised as either athletic or nonathletic (breeding or trail riding) activity. Patients with a nonathletic use were investigated and reported with respect to return to use indices, but only horses with a preoperative athletic use were included in the return to performance analyses.
A series of univariable and multivariable logistic regression models were developed and examined to assess the potential association of all individual variables with return to use and performance. Outcome variables were ‘return to use’ and ‘return to performance’ and these were evaluated at 6 months and one year, resulting in 4 outcome variables. Continuous independent variables evaluated were divided into 2 or 3 categories for statistical analysis. Categories were based on reference range values provided by the NCSU clinical pathology laboratory, or those routinely used by clinicians.
Separate univariable and multivariable analyses were performed for each of the 4 outcome variables using SigmaStat (Version 3.11, 2004)a[15,16]. All variables were added into a multivariable logistic regression model using a forward stepwise approach [17,18]. Variables were chosen for inclusion in the model and potential confounding was assessed by evaluating changes to odds ratios and confidence intervals as variables were added and deleted from the model. The impact of individual variables on the model was assessed by reviewing changes in the estimated standard error values of the maximum likelihood coefficients. Where appropriate, based on the numbers of variables, a Hosmer-Lemeshow test statistic was used as a post-estimation assessment of the final model .
A multivariable model was fitted for 3 of the outcome variables and reported as the final model for the respective outcome variables. Insufficient data were available to assess 6-month return to performance using multivariable logistic regression. Consequently, univariable logistic regression was used and reported to examine the potential association of variables with 6-month return to performance outcome. Odds ratios (OR) were reported with 95% confidence intervals (CI) and corresponding P values.
Data were retrieved and reviewed from the NCSU Equine Colic Database for 372 horses that recovered from exploratory celiotomy at NCSU for acute colic, from June 2003 to March 2010. Thirty-eight horses (10.2%) died after surgery, prior to discharge; 16 horses (4.3%) died after discharge, before 6-month follow-up; 95 (25.5%) were lost to complete follow-up; and 29 (7.8%) had no intended use. Therefore 194 horses (52.1%) were included in the study. One horse had 2 surgeries performed at NCSU greater than 6 months apart, so separate clinical and follow-up data were recorded for each episode. This yielded 195 reported episodes of recovery from colic surgery.
Age of the population ranged from 0.5 to 25 years (median 9 years, 25th percentile 5 years, 75th percentile 14 years). There were 105 geldings, 72 mares and 18 stallions. There were 79 patients of stock horse breeding (Quarter Horse, Appaloosa or Paint Horse), 33 Thoroughbreds, 27 Warmbloods, 20 Arabians, 8 draught breeds and 28 horses of other breeds. Intended use required athleticism in 156 of 195 of patients. Reported athletic uses included jumping (n = 58), showing (n = 35), western performance (n = 23), dressage (n = 20), racing (n = 8), pulling/driving (n = 6) and halter (n = 6). Horses placed in the nonathletic use group (39/195) included animals solely used for breeding (9 mares, one stallion) and trail riding horses (n = 29).
Incisional hernias were reported in 12 of 195 (6%) horses. Size was recorded as estimated by the owner in 11 cases, ranging from 3 to 8 cm in diameter. Two horses had multiple small diameter hernias. In only one horse was the hernia considered by the owner and referring veterinarian to be of a size and configuration to contraindicate athletic activity. Surgical repair was performed in one patient. All other horses were treated with stall rest and/or abdominal support.
Diarrhoea developed in 58 of 190 (31%) horses in this population post operatively. Only one horse that developed diarrhoea during hospitalisation was reported to have persistent diarrhoea at the 6-month follow-up.
There were 10 horses that developed laminitis between the time of surgery and 6-month follow-up. Laminitis developed during hospitalisation (in-hospital laminitis) in 5 of 195 (3%) horses. Two of these horses had strangulating intestinal lesion types, 3 had nonstrangulating intestinal lesions.
Previous celiotomy was a historical finding in 25 of 195 horses. Of the 25 cases with a history of previous celiotomy, 2 had 2 prior celiotomies for recurring left dorsal displacement of the large colon; the remaining 23 had a single historical celiotomy.
Lesion diagnosis was nonstrangulating small intestinal in 38 horses, strangulating small intestinal in 36, nonstrangulating large intestinal in 99, strangulating large intestinal in 19 horses, and there were no significant findings on exploratory celiotomy in 3 horses. In horses with strangulating lesions, 42 of 55 (76%) returned to their intended use at the time of 6-month follow-up. Horses with simple, nonstrangulating lesions returned to use at the time of 6-month follow-up in 89 of 137 (65%) cases.
Nine horses were undergoing stall rest for an orthopaedic condition/lameness at the time of surgery. Of the athletic horses, only one of 7 returned to full performance at one year. Three of these failures to return to performance were attributed to persistent lameness by the owner.
Numbers of horses returning to use and performance at the intervals of 6 months and one year for all investigated variables are reported in the supplemental information (Items S1 and S2).
Six-month return to use and performance
At 6 months, 133 of 195 (68%) of horses were performing their intended use. In the final multivariable logistic regression model, horses were significantly less likely to return to use at 6 months if they had an incisional hernia, post operative diarrhoea, in-hospital laminitis, history of previous celiotomy, or a nonstrangulating lesion type (Table 1). The Hosmer-Lemeshow statistic was 5.341 (P = 0.721), indicating adequate goodness of fit of the model.
Table 1. Final multivariable and/or univariable logistic regression modelling of variables significantly associated with decreased return to use and performance at the intervals of 6 months and one year
|Hernia|| || || || |
| (95% CI)||(2.4–58.8)||–||(1.1–11.9)||(1.3–36.5)|
| P value||0.003||–||0.040||0.022|
|Diarrhoea|| || || || |
| (95% CI)||(1.1–4.5)||–||–||–|
| P value||0.036||–||–||–|
|In-hospital laminitis|| || ||–|| |
| (95% CI)||(1.0–100.5)||–||–||–|
| P value||0.012||–||–||–|
|Laminitis|| || || || |
| (95% CI)||–||–||(1.4–20.4)||(0.9–82.1)b|
| P value||–||–||0.013||0.060b|
|Prior colic surgery|| || || || |
| (95% CI)||(1.3–7.1)||(1.0–5.7)a||–||–|
| P value||0.012||0.046a||–||–|
|Strangulating lesion type|| || || ||–|
| (95% CI)||(0.2–0.9)||–||–||–|
| P value||0.024||–||–||–|
|Prior stall rest (orthopaedic issue)|
| (95% CI)||–||–||–||(1.3–101.7)|
| P value||–||–||–||0.030|
At 6-month follow-up, 85 of 156 (54%) horses were at or above their preoperative level of performance. When variables were evaluated independently (on univariable analysis), horses were significantly less likely to return to performance at 6 months if they had a history of previous celiotomy (Table 1). The data did not support the production of a multivariable model.
One-year return to use and performance
At one year, 145 of 190 (76%) horses were performing their intended use. The final multivariable model demonstrated that horses were significantly less likely to return to use at one year follow-up if they had suffered from an incisional hernia or laminitis (Table 1). The Hosmer-Lemeshow statistic was 0.234 (P = 0.890), indicating adequate goodness of fit of the model.
At one year, 101 of 153 (66%) horses were at or above their preoperative level of performance. The final multivariable logistic regression model demonstrated that horses were significantly less likely to return to performance at one year if they had suffered from hernia formation or were on stall rest for an orthopaedic condition at the time of surgery (Table 1). Laminitis was found to be significant on univariable analysis (OR 10.6, 95% CI 1.2–93.6, P = 0.033), but insignificant during multivariable analysis. It was included in the multivariable model because it was considered to have potential biological importance and improved the overall fit of the model. The Hosmer-Lemeshow statistic was 0.119 (P = 1.000), indicating adequate goodness of fit of the model.
Overall owner satisfaction regarding post operative return to use or performance in the entire group was excellent in 146 of 188 (78%) horses, good in 25 (13%), fair in 7 (4%) and poor in 10 (5%). Seven respondents were undecided or had no comment. For the athletic horse population only, satisfaction regarding performance ability was excellent in 115 of 149 (77%), good in 20 (13%), fair in 5 (3%) and poor in 9 (8%).
The majority of horses returned to use and performance at 6 months. At one year, 101 of 153 (approximately two-thirds) of horses were performing at or above preoperative performance. Owner satisfaction with the post operative performance ability of individuals with an athletic-type use was excellent in 115 of 149 (77%) responses. Only 9 of 149 (6%) reported a poor satisfaction score. Two previous studies reported a higher percentage of animals (81% and 90.1%) returning to expected performance following colic surgery [13,14]. These studies collected follow-up data at a single time period; one study at one year and the other between 9 months and 6 years. Previous studies did not evaluate early return to function, use of the horses, interval for return to performance, or reasons for failure to perform. Consequently, a direct comparison of our results to these previous studies is difficult.
Horses that had a hernia were significantly less likely to return to use at 6 months and one year, as well as performance at one year. Incisional herniation has a reported incidence of 8–16% following exploratory celiotomy for colic [8,19]. Twelve of the 195 (6%) animals in this study population developed an incisional hernia following exploratory celiotomy for colic. Failure for early return to use may be attributed to restricted activity or hernia belt application commonly recommended at our hospital for treatment of early abdominal wall defects. Surgical hernioplasty is often delayed for 4–6 months to allow for fibrous remodelling of the hernia margins. Horses undergoing surgical repair would have inadequate time for rehabilitation at 6-month follow-up. Surprisingly, horses that had an incisional hernia remained less likely to return to performance at one year. By one year, prolonged convalescence as a direct consequence of the incisional hernia or repair thereof should be minimal. The authors speculate that the observed failure to return to use and performance in horses with hernias may be secondary to poor cosmesis (especially in animals used for showing and halter events), restricted owner finances not allowing surgical hernia repair, and owner opinion that a benign abdominal wall defect renders the animal unsound for athletic activity. Proper veterinary management should ensure that hernias are repaired if necessary and that horses are not unnecessarily retired for abdominal wall defects.
Horses were 2 times less likely to return to use at 6 months if they had post operative diarrhoea. Previous studies document an incidence of post operative diarrhoea ranging from 3.8–53% [11,20]. Similarly, diarrhoea developed in 58 of 190 (31%) horses in this population post operatively. One horse that developed diarrhoea post operatively was reported to have diarrhoea at follow-up. Although body condition or weight was not documented in this population, the authors speculate that malabsorption associated with diarrhoea may result in a decreased body condition and prolonged convalescence. These horses may also have prolonged recovery due to latent infection with an organism such as Salmonella sp.; however, this is unlikely as most episodes of diarrhoea were transient and mild. Results of Salmonella sp. cultures were not investigated in this study.
Horses were significantly less likely to return to use at 6 months, and to return to use and performance at one year, if they had laminitis. Since laminitis often results in permanent debilitation and persistent lameness, this finding was not surprising. Laminitis is generally reported as a rare post operative complication, occurring in approximately 1% of horses undergoing celiotomy [7,8]. The incidence of horses developing laminitis during hospitalisation was similar in this population at 5 of 195 (3%). Five of 10 reported episodes of laminitis developed after discharge from the hospital, but before 6-month follow-up, suggesting that horses discharged from the hospital following colic surgery may be at an increased risk for laminitis when compared with a normal population of horses. This may be an incidental finding, and additional study with a larger population of animals is warranted.
Horses were significantly less likely to return to use and performance at 6 months if they had a history of previous celiotomy. Owners who will tolerate multiple periods of convalescence may be less likely to return these animals to their intended use. Due to the multiple convalescence periods, these horses may also return after an ideal time or age necessary for their specific intended use.
Horses were significantly less likely to return to full performance at one year if they had a history of recent stall rest for an orthopaedic condition prior to surgery. There was no significant difference at the 6-month follow-up interval for return to use or performance. The delay in return to full performance at one year only is most likely a result of recurrent lameness, but these data were not collected.
Horses with strangulating lesions are often assumed to be at an increased risk of limited return to use because of potential complications associated with ischaemic bowel and endotoxaemia: laminitis, post operative nutrient malabsorption, adhesions, requirement for altered dietary regimens and recurrent colic [3,21–23]. Horses that returned to use at 6 months were more likely to have a strangulating lesion type. These patients often require prolonged hospitalisation, intensive monitoring and care of secondary complications, which may increase the return to use. For example, there were no abdominal hernias in this group of horses. In addition, owners who were willing to invest in a more expensive surgery and aftercare associated with resection and anastomosis may be more likely to value the use/performance of the animal and be more proactive in the return thereof.
In the population studied, the number of horses returning to use was influenced by owner decision to prolong recovery and rehabilitation. Athletic performance was also subjectively determined by owners. Additional studies with a uniform group of athletic horses and quantifiable performance indices may be warranted to provide a more objective assessment of performance post operatively.
This study design was not case-controlled with normal or nonsurgical colic horses as controls and may not account for variables such as natural attrition in a population of performance horses. Additionally, the odds ratios for the risk factors of hernia, laminitis and stall rest were associated with wide confidence intervals. This was probably due to a low incidence of these complications in the study population.
Several parameters in the study were categorised for statistical analysis, as described previously. The size of the study population precluded narrower categorisation of some clinical variables. Categorisation of continuous data may result in an inability to identify all clinically relevant associations of a variable with the outcome as a result of the grouping of a wide range of data. Additional study with a larger population of animals would add further clarity to these relationships.
As the data for the outcome of 6-month return to performance did not support the production of a multivariable model, univariable statistics were reported for this outcome. Univariable results are not adjusted to reflect the potential effects of other factors and there may be statistical error in the screening of large groups of clinical variables. Additional research is needed before an accurate conclusion can be made about the relationship between this variable and the return to performance at 6 months.
Among the 4 outcomes evaluated, certain variables obtained significance for one or 2 of the outcomes only. This may be related to a temporal effect of the variable (i.e. temporary or chronic conditions) or the intensity of performance (i.e. stall rest for lameness was significantly associated with return to performance, but not return to use). Alternatively, the lack of consistent significance for multiple outcome variables may be due to low statistical power and an insufficient sample size to effectively screen for all potentially important variables.
In conclusion, the overall return to general use and performance in horses surviving 6 months following colic surgery was fair to good. A history of previous celiotomy, stall rest for an orthopaedic condition, incisional hernia formation, post operative diarrhoea and laminitis are significant risk factors for failure to return to use and expected performance following exploratory celiotomy for acute gastrointestinal lesions. Targeted owner education regarding preoperative lameness, post operative rehabilitation and treatment for complications, such as incisional hernioplasty, may help inform owners about their horse's potential for return to use and performance following colic surgery.
The authors thank additional contributing surgeons (K. Bowman; Y. Elce; L. Frederico; L. Gonzalez; R. Labens; D. Little; A. Morton; J. Nutt; R. Redding; M. Schramme; J. Tate; J. Tomlinson) and the NCSU Equine Internal Medicine Service.
a Systat Software, Inc., San Jose, California, USA.