Clinical presentation and outcome of gastric impactions with or without concurrent intestinal lesions in horses

Abstract Background Gastric impactions (GI) have been identified as primary lesions (lone GI; LGI) or associated with other intestinal lesions (concurrent GI; CGI). Anecdotally, CGI resolve more rapidly with a better prognosis than LGI. Objectives To determine clinical, laboratory, and ultrasonographic findings, and short‐ and long‐term survival in horses with GI. We hypothesized that LGI carries a worse prognosis than CGI. Animals. Seventy‐one horses from 2 referral hospitals (2007‐2022). Methods Retrospective cohort study. Gastric impactions were defined as feed extending to the margo plicatus after ≥24 hours of fasting. Clinical, diagnostic and outcome findings were compared between LGI and CGI. Long‐term survival was determined by a questionnaire. Results Twenty‐seven horses had LGI, 44 had CGI. Large intestinal lesions (32/44) were more common than small intestinal lesions (12/44). Concurrent gastric impactions resolved more slowly than LGI (LGI median 2 days, range 0‐8; CGI median 4 days, range 1‐10; P = .003). Short‐ (LGI 63%, 17/27; CGI 59%, 26/44; P = .75) and long‐term survival (LGI 3.5 ± 1.9 years; CGI 2.3 ± 2.3 years; P = .42) were not significantly different. However, Lone gastric impactions were more likely to experience gastric rupture (LGI 29.6%, 8/27; CGI 11.4%, 5/44; P = .05). Lone gastric impactions were 8.7 times more likely to require dietary changes (LGI 72.7%, 8/11; CGI 25%, 4/16; 95% confidence interval [CI], 1.53‐49.22; P = .01). Gastric impactions recurred in 21.7% (LGI, 6/20; CGI, 4/26; P = .23) of affected horses. Conclusions and Clinical Importance Lone gastric impactions and CGI present similarly with a comparable prognosis, but LGI are more likely to rupture. Long‐term dietary changes are often necessary for horses with LGI.


| INTRODUCTION
Gastric impactions (GI) are a well-recognized but poorly understood cause of colic in horses. They have been defined as persistent, progressive accumulations of dehydrated ingesta that remain within the stomach after prolonged fasting, often defined as ≥24 hours. 1 Classification systems used to describe GI include acute vs chronic or primary vs secondary, if an inciting cause can be identified. 1 However, GI often progresses insidiously initially with few clinical signs, and chronicity can be difficult to establish. Feed accumulation can occur after ingestion of expansive, poorly digestible, or excessive feed and has been associated with hepatic disease, particularly pyrrolizidine alkaloid toxicosis. [2][3][4] However, GI also can occur without any apparent inciting cause 5 and has been observed in the presence of other gastrointestinal lesions, such as large colon displacements or volvuli. 6 Anecdotally, GI occurring concurrently with other intestinal lesions (concurrent gastric impactions [CGI]) appear to resolve more rapidly and carry a better prognosis compared with cases where GI are the only abnormality identified (lone gastric impactions [LGI]). Diagnosis of either condition can be challenging because clinical signs are often vague, ranging from weight loss to poor performance or colic. 5 The primary aim of our retrospective cohort study was to describe and compare clinical and diagnostic findings, response to treatment and short-and long-term outcomes in horses with LGI and CGI. We tested the hypothesis that treatment of horses with LGI took longer and carried a worse short-and long-term prognosis than did treatment of those with CGI.

| MATERIALS AND METHODS
Clinical records from 2 equine referral centers were searched for horses diagnosed with GI from 2007 to 2022. Gastric impactions were defined as presence of feed material within the stomach, extending to the level of the margo plicatus, after at least 24 hours of fasting. Cases were excluded if the fasting period was <24 hours or if horses had known dental or hepatic disease. Horses <1 year old and donkeys also were excluded. A definitive diagnosis was achieved by gastroscopy, palpation of the stomach during exploratory laparotomy or necropsy. Confirmed cases were divided into CGI or LGI, depending on the presence or absence of concurrent intestinal lesions suspected of causing signs of colic.
Clinical and laboratory information extracted from the case records included patient signalment, history, presenting clinical signs and laboratory findings. In cases of CGI, the nature of the concurrent lesion was noted. This classification was based on the interpretation of diagnostic or surgical findings by the attending clinician. Reports of ultrasonographic examinations, particularly stomach size but also descriptive abnormalities, were collated and compared if available. A gastric silhouette visible over ≥6 intercostal spaces during ultrasonographic examination was considered enlarged. 7,8 Gastroscopy reports were assessed and a subjective measure of impaction size during gastroscopy was recorded: moderate GI were defined as feed material extending to the level of the margo plicatus, whereas large GI were defined as feed extending above the margo plicatus. If feed material extended to the level of the cardia or into the esophagus, GI were defined as very large. If horses underwent exploratory laparotomy, surgical findings including a subjective assessment of stomach size as described in the surgeon's report (normal, moderate, large, or very large) were included. Ultrasonographic size estimates were compared with surgical or gastroscopic size estimates. Sensitivity (Se) and specificity (Sp) of ultrasonographic gastric silhouette enlargement (≥6 intercostal spaces) were calculated to determine whether ultrasonographic size could be used to predict the size of GI as determined at surgery or gastroscopy. If present, gastric ulcers were recorded, both descriptively and scored on a scale of 4, based on recommended guidelines. 9,10 Treatment methods and response to treatment were noted, including volume of IV and enteral fluids administered, use of carbonated soft drinks or prokinetics, time to GI resolution as determined by gastroscopy, time to discharge, and any recurrence during hospitalization. Short-term survival was defined as survival to discharge.
Owners were contacted using an online or telephone questionnaire to evaluate long-term outcome, recurrence of a GI or colic signs, and any management changes implemented. Ethical approval was obtained from the Social Sciences Research Ethical Review Board (SSRERB) of the Royal Veterinary College.

| Statistical analysis
Data were analyzed using commercially available software (IBM SPSS Statistics 28.0.0.0). Normality of the data was assessed using Shapiro-Wilk tests. Continuous data were represented as mean ± SD if normally distributed, or median and range (minimum-maximum) if not normally distributed. Categorical data were represented as numbers and percentages. Chi-squared tests and Fisher's exact tests were used to compare categorical measurements. Comparisons between parametric continuous data were made using a Student's t-test; nonparametric data comparisons were made using Mann-Whitney U tests.
Other reasons included chronic esophageal obstruction, poor performance, pyrexia, anorexia, and weight loss. The median duration of clinical signs was 12 hours (1-1344 hours). The horse that presented after 56 days (1344 hours) was excluded from the analysis involving the duration of clinical signs, because the clinical signs were intermittent in nature and the specific onset of signs before presentation could not be determined in this horse. Removing this horse, the median duration of clinical signs was 10 hours (1-672 hours). The duration of clinical signs before presentation was significantly longer in the LGI group than in the CGI group (LGI, 24 hours; range, 1-672 hours; CGI, 6 hours; range, 1-96 hours; P = .01). Seven percent (5/71) had experienced gastric rupture on arrival, which was confirmed by exploratory laparotomy (n = 3) or necropsy (n = 2). Of these, 60.0% (3/5) were classified as LGI and 40.0% (2/5) as CGI (P = .3).

| Clinical variables
Clinical and laboratory variables at presentation are presented in Table 1.

| Diagnostic tests
Ultrasonographic measurements of the gastric silhouette were available in 26/71 (36.6%) cases. Of these gastric silhouettes, 24% (15/26) were considered enlarged, extending to ≥6 intercostal spaces. Ultrasonographically enlarged gastric silhouettes were not significantly associated with a large or very large size estimate of gastric impaction during surgery (P = .29) or during gastroscopy (P = 1.00). Using presence of large or very large GI during surgery as the gold standard, ultrasonographic findings had poor sensitivity and specificity of 67% and 20%, respectively, when used to diagnose GI. When comparing ultrasonographic enlargement with the presence of large or very large GI on gastroscopy, sensitivity was 55% and specificity was 50%. Rectal examination was performed in 54/71 (76.1%) cases (LGI, 70.3%, 19/27; CGI, 79.5%, 35/44). Rectal examination was performed in all horses with CGI that did not undergo immediate exploratory laparotomy or euthanasia. Of the LGI group, rectal examination was not performed in 4 horses; 3 of these did not present for colic signs. Two of these horses were referred for treatment of GI and 1 was referred for recurrent esophageal obstruction but gastroscopy identified an extensive GI. The fourth horse was a Shetland Pony and rectal examination was not performed because of its small size.

| Diagnosis of concurrent intestinal lesions
Concurrent intestinal lesions are summarized in  To avoid confounding extragastrointestinal conditions that have been associated with GI, cases with hepatic disease were excluded. 2 Donkeys and foals also were excluded because gastric emptying in donkeys is anecdotally reported to be slower, and foals could suffer from gastric outflow obstruction secondary to duodenal ulceration, which is assumed to be a different disease process. 13 The presence of GI with another intestinal lesion only recently has been described, raising questions of whether GI occur first, poten- and our findings seem to support this hypothesis.

| Exploratory laparotomy
Although some clinical features differed between the LGI and CGI, the overall presentation and response to treatment were remarkably similar. Clinical and laboratory variables were largely comparable.
Although rectal temperature was significantly higher in the LGI group, temperature remained within the normal range, and the difference was not clinically relevant. As might be expected, LGI cases had a significantly longer history of clinical signs before presentation, whereas clinical signs with CGI were usually more acute in onset. A previous study similarly reported a tendency of LGI to present subacutely or chronically, with a median duration of clinical signs of 3 days. 5 It is therefore likely that development of a concurrent intestinal lesion triggered an earlier presentation to a referral hospital than would have been the case if the GI was the only lesion. Theoretically, it is also possible that some GI developed secondary to the intestinal lesion Analysis of other ultrasonographic features such as fluid filling have been used to evaluate the equine stomach, 16  Against expectation, short and long-term outcomes were not significantly different between the 2 groups. Gastric rupture was more common in LGI, which could be associated with the longer duration of clinical signs and later diagnosis and treatment of the condition in this group. Although short-term recurrence of GI was rare, with only 1 horse developing a second GI during hospitalization, 22% of horses experienced at least 1 more GI after discharge, with no differences noted between LGI and CGI cases. This finding is higher than the 11% recurrence previously reported. 5 With longer follow-up, particularly for recently hospitalized cases, a higher recurrence rate might be found.
Considering that an underlying motility issue is suspected to at least contribute to the development of GI, the high rate of recurrence is not surprising. It also corroborates the assumption that the underlying etiology is similar or the same in both groups. In both groups, many horses required long-term management changes and only half returned to their previous diet and exercise regimen. However, horses with LGI were significantly more likely to require substantial long-term dietary changes.
The longer duration of the impaction could have led to chronic and possibly irreversible stretching of the stomach wall, or these horses could have suffered from a more severe motility disturbance. However, overall numbers are small, and findings should be confirmed in a larger number of horses to avoid overinterpretation. Feeding a partial-or completepelleted diet can be a logistical and financial challenge for horse owners, and can lead to alterations in horses' behavior and gastrointestinal pH. 23,24 Clients should be well informed of both short-and long-term implications when horses are diagnosed with GI.
Most limitations of our study are associated with its retrospective design. Reporting details differed among clinicians, hospitals, and over time, and information was not always complete. The exact time of diagnostic procedures such as ultrasonography was not always noted in the records. According to hospital protocols, most examinations were carried out within 1 to 2 hours of surgery or euthanasia, but in some cases more time could have elapsed and findings might have been affected. In addition, despite all efforts to rule out other intestinal lesions in the LGI group, some horses might have had unidentified reasons for their colic signs. Treatment details such as the exact volumes of enteral fluids and carbonated soft drinks often could not be obtained from the clinical records. Therefore, comparing the effects of certain treatments was not possible. Our study only included horses located in the south of England and results therefore might not be reflective of a broader international population. However, GI have been reported globally, arguing against distinct regional differences. 5 The questionnaire required accurate client recall, which may have been difficult, particularly for horses presented early in the 15-year study period. Several horses were dead, and the reasons for euthanasia were often unclear in the absence of necropsy. Some horses had clinical signs suspicious of GI recurrence but these findings were not confirmed by veterinary examination and gastroscopy.
In conclusion, GI can occur either alone or concurrently with other intestinal lesions and could be underdiagnosed in the presence of other intestinal diseases. Although their clinical presentations and outcomes are similar, gastric rupture and long-term complications are more likely in LGI. The recurrence of GI might be higher than previously reported. 5

ACKNOWLEDGMENT
No funding was received for this study. The authors thank the staff at Donnington Grove Equine Hospital for their collaboration.