Results
Serum pancreatic (DGGR) lipase concentrations were measured in 205 equids. These included 42 ponies, 39 cobs/crosses, 34 sports horses, 26 warmbloods/crosses, 24 Thoroughbred/crosses, 15 Shetland ponies/miniature horses, 12 heavy horses and 13 others (including 3 donkeys). There were 117 geldings, 84 females and 4 entire males. The median age was 13 years (interquartile range [IQR] 9.5; range 0.3-36 years); only 3 horses were younger than 12 months of age. The degree of pain at admission was categorised as mild in 158 animals, moderate in 40 and severe in 3; one animal showed no signs of pain at the time of admission (but was painful at the time that referral was organised). The median heart rate at admission was 52 bpm (IQR 16; range 32-100). The major findings of FLASH ultrasound scans were unremarkable/normal in 114 animals; distended small intestines were recorded in 37, changes compatible with colitis in 16, thickened small intestinal walls in 14, excessive peritoneal fluid in 11, possible colonic displacement in 3, abdominal mass/neoplasia in 3, haemoperitoneum in 2, gastric distension in 2, intussusception in one, and liver abnormality in one (n=204). Gastric reflux (>1 litre) was recorded in 15 horses (range 1-25 litres). The final diagnosis categories were medical colic (n=107), surgical colic/euthanased without surgery (n=53), colitis (n=13), peritonitis (n=15) and other (n=6). The median duration of illness prior to hospitalisation was 12 hours (range 2->100 hours). The outcomes of the cases were survived (n=147), euthanised (n=47) and died (n=11).
The median white blood cell count was 7.2 x 109/L (IQR 4.55; range 1.2-35.6 x 109/L) (n=185), neutrophil count 5.3 x 109/L (IQR 4.2; range 0.4-31.8 x 109/L) (n=185), packed cell volume 36.25% (IQR 16.6; range 16.0-86.3) (n=196), serum Amyloid A concentration 13.95 mg/L (IQR 923; range 0-9900 mg/L) (n=186), total plasma protein concentration 60.0 g/L (IQR 10; range 26.0-86.0 g/L) (n=199), plasma albumin concentration 36.5 g/L (IQR 6.25; range 16.0-49.0 g/L) (n=186), plasma lactate concentration 2.0 mmol/l (IQR 19.7; range 0-19.7 mmol/l) (n=144), peritoneal nucleated cell count 1.4 x 109/L (IQR 3.9; range 0.3-387.0 x 109/L) (n=122) and peritoneal lactate concentration 0 mmol/L (IQR 20.7; range 0-20.7 mmol/L) (n=104). There was a statistically significant association between the systemic lactate concentration and severity of pain (p = 0.03). There were statistically significant associations between all of the systemic haematological/biochemical measurements and the type of colic (p < 0.01, except WBC (p = 0.03)). There were statistically significant associations between peritoneal lactate concentration (p = 0.01), blood lactate concentration (p = 0.02), albumin concentration (p < 0.01) and PCV (p = 0.02), and survival of the colic episode.
The median serum pancreatic lipase concentration in all 205 horses was 17 U/L (IQR 14 – 27; range 1 – 3484). Using the methodology described by Johnson et al (2019), the lipase concentration was categorised as normal in 124 animals (60.5%) and elevated in 81 (39.5%) (mildly elevated in 55 (26.8%), moderately elevated in 14 (6.8%) and severely elevated in 12 (5.9%)). Serum lipase concentration was less than twice the upper reference limit (<2x URL) in 178 and >2x URL in 27 horses. There was no correlation between age and pancreatic lipase concentration (Spearman’s rank: rho = 0.03; p = 0.68), and no significant difference between the lipase concentration groups (normal, mildly, moderately, or markedly elevated) and age (Kruskal-Wallis test; p = 0.93), or between horses with lipase concentration < or > 2xURL and age (Kruskal-Wallis tests: p = 0.62). There was no significant difference in breed and lipase concentration (Kruskal-Wallis tests p = 0.33 for lipase categories, and Chi-Squared p=0.39 for < or > 2xURL), or in sex and lipase concentration (Kruskal-Wallis tests p = 0.52 for lipase categories, and Chi-squared p=0.33 for < or > 2xURL). There were no significant associations between any of the haematological or biochemical findings and the lipase concentration groups.
There was a statistically significant difference in the disease category and pancreatic lipase concentration categories (Kruskal-Wallis test: p < 0.001) as well as a significant difference between the disease category and lipase concentration < or > 2xURL (Chi-squared p<0.01) , with colic cases having higher lipase concentrations than colitis and peritonitis cases: medical colics, n = 107, median lipase = 17 U/L (IQR 14 – 30); surgical colics /euthanased, n = 53, median lipase = 21 U/L (IQR 16 – 32); colitis, n = 24, median lipase = 13 U/L (IQR 11 – 17); peritonitis, n = 15, median lipase = 15 U/L (IQR 13 – 18); others, n = 6 median lipase = 84 U/L (IQR 28 – 1094). There were statistically significant differences in lipase concentrations between these disease categories (Dunn’s post-hoc test) (Figure 1). Significant differences were present between medical colic and colitis cases (p < 0.001), surgical colic and colitis cases (P < 0.001), and surgical colic and peritonitis cases (p = 0.02). There was no significant association between heart rate and lipase concentration group (Kruskal-Wallis: p = 0.80) or between heart rate and lipase concentration < or > 2xURL (Kruskal-Wallis p=0.24), and no correlation between heart rate and lipase value (Spearman’s rank: rho = 0.09; p = 0.21). There was strong evidence of an association between pain severity and lipase values (Kruskal-Wallis: p = 0.01) and between pain severity and lipase concentration < or > 2xURL (Chi-squared p=0.05), with higher lipase concentrations in horses with more severe pain: mild pain, n = 158, median lipase = 17 U/L(IQR 14 – 25); moderate pain, n = 40, median lipase = 19 U/L (IQR 15 – 47); severe pain, n = 4, median lipase = 558 U/L (IQR 22 – 2186). There was no evidence of an association between outcome (classified as survived, euthanasia or died) and lipase concentration (Kruskal-Wallis: p = 0.36), however there was a difference in outcome across lipase concentration groups categorised as < or > 2xURL (Chi-squared p=0.04). There was no difference in duration of colic between lipase concentration groups (Kruskal-Wallis: p = 0.10) or lipase concentration categorised as < or > 2x URL (Chi-squared p = 0.63).).
Differences were identified in pancreatic lipase concentration category (Dunn’s post hoc test) for the following FLASH ultrasound findings: no abnormalities vs colitis, p < 0.001; distended small intestines vs colitis, p < 0.001; colitis vs colonic displacement, p = 0.05. The median pancreatic lipase concentration of horses with different findings on FLASH ultrasonography were as follows: horses with no significant abnormalities on FLASH ultrasonography (n=114) had a median pancreatic lipase concentration of 19 U\L (IQR 14 – 32) (98 horses had lipase < 2xURL and 16 horses > 2xURL); median lipase concentration of horses with colitis (n = 16) was 12 U/L (IQR 10 – 13) (16 horses had lipase concentration < 2xURL and none with lipase concentration > 2xURL); median lipase concentration of horses with distended small intestines (n = 37) was 19 U\L (IQR 15 – 32) (29 horses had lipase concentration < 2xURL and 8 had lipase concentration > 2xURL); median lipase concentration of horses diagnosed with a potential colonic displacement on FLASH ultrasound (n = 3) was 25 U\L (IQR 22 – 36) (all 3 horses had lipase concentration < 2xURL). Of the 12 horses with severely increased pancreatic lipase concentration on admission, 3/12 (25.0%) had spontaneous nasogastric reflux (spontaneous nasogastric reflux was not recorded in any of the other 193 equids), 6/10 (60.0%) had distended and/or thickened small intestine on abdominal ultrasonography (not performed in 2 cases), 7/12 (58.3%) survived to hospital discharge, and 5/12 (41.7%) died or were euthanised (in 2 cases surgery was recommended but the owners opted for euthanasia for financial reasons). Overall, 15/205 equids (7.3%) had reflux (> 1 litre) either spontaneously or on passage of a nasogastric tube recorded at admission. Four of these (26.7%) had normal pancreatic lipase concentrations, 6 (40.0%) had mildly elevated concentrations and 5 (33.3%) had severely elevated concentrations. In total, therefore, 11 of 15 (73%) horses that had gastric reflux identified at admission to the hospital had elevated serum pancreatic lipase concentrations. 21 of 51 horses (41%) where FLASH scanning identified distended small intestines or thickened small intestinal walls had elevated serum pancreatic lipase concentrations (12 mildly elevated, 3 moderately elevated and 6 severely elevated). The final diagnoses in 10 animals with severely increased serum pancreatic lipase concentrations where a diagnosis was reached included non-specific colic (n=4), enteritis (n=2), grain overload (n=1), post-foaling colic and neuropathy (n=1), epiploic foramen entrapment (n=1) and colon torsion (n=1).
Repeated analyses for pancreatic lipase were performed in 15 animals (Table 1). In 11 of these horses that had a serum pancreatic lipase concentration > 21 U/L on admission to the hospital, the concentration reduced in subsequent samples (following surgical or medical treatment and resolution of the abdominal pain).