Implications of urine F2-isoprostane metabolite concentration in horses with colic and its potential use as a predictor for surgical intervention




Reasons for performing study: F2-isoprostanes have been used extensively to quantify lipid peroxidation in association with risk factors in various diseases. Horses with colic may have intestinal ischaemia and/or inflammation characterised by oxidative stress and increased production of isoprostanes.

Objectives: To gather preliminary data regarding the feasibility of using urine F2-isoprostanes and isoprostane metabolites as early screening tools for the presence of gastrointestinal disease requiring surgical intervention in horses and ultimately develop a stall-side test capable of identifying these horses as early as possible for timely referral.

Methods: Concentrations of urine isoprostane and isoprostane metabolite were determined by mass spectroscopy and normalised to urine creatinine (Cr) concentrations in urine samples from 42 healthy control horses and 43 horses with gastrointestinal pain or colic.

Results: Horses with colic were treated medically (n = 21) or surgically (n = 22). Mean ± s.d. concentrations of urine isoprostane and isoprostane metabolite were significantly higher in horses with colic (2.94 ± 1.69 and 0.31 ± 0.22 ng/mg Cr, respectively), compared to control horses (1.89 ± 1.39 and 0.22 ± 0.08 ng/mg Cr, respectively). Mean urine isoprostane metabolite concentrations were significantly higher in horses undergoing surgery (0.38 ± 0.28 ng/mg Cr) compared to controls and medical colics (0.26 ± 0.11 ng/mg Cr). Nonsurvivors had significantly higher mean urine isoprostane metabolite concentrations (0.47 ± 0.39 ng/mg Cr) than control or surviving colic horses (0.29 ± 0.24 ng/mg Cr).

Conclusions: Measurement of urine isoprostane metabolite concentration may be a useful prognostic indicator in equine colic.

Potential relevance: Urine isoprostane metabolites may aid in early recognition of surgical colic. Isoprostanes are a potential therapeutic target to prevent further systemic and gastrointestinal tissue injury in horses with colic.


Colic or abdominal pain is one of the most prevalent and challenging diseases encountered by equine practitioners. Approximately 4% of horses experience at least one colic episode per year and a fatal outcome occurs in 11% of cases (Proudman et al. 2002). While the majority of horses with abdominal pain can be treated successfully on the farm, some require rapid referral for intensive medical therapy or surgical intervention. Horse owners are increasingly aware of the various diseases that result in abdominal pain in horses and expect veterinarians to provide information regarding the prognosis prior to referral and surgery.

Parameters proposed as prognostic aids for equine colic include, but are not limited to, heart rate, packed cell volume, peritoneal fluid colour, response to analgesia, and blood or peritoneal fluid lactate concentrations (Moore et al. 1976; Parry et al. 1983; Puotunen-Reinert 1986; Orsini et al. 1988; Morris et al. 1991; Seahorn et al. 1994; Furr et al. 1995; Thoefner et al. 2000; Proudman et al. 2002; van der Linden et al. 2003; Garcia-Seco et al. 2005; Hinchcliff et al. 2005; Mair and Smith 2005a; Proudman et al. 2005, 2006; Niinisto et al. 2010). In general, any one parameter can be useful in a particular type of colic; however, no single predictor seems to provide accurate prognostic information for all types of colic, especially mild nonstrangulating surgical cases affected by displacement or impaction. Increased blood lactate concentration was only accurate in predicting survival for horses with large colon volvulus, not less severe conditions (Latson et al. 2005; Johnston et al. 2007). Increased peritoneal fluid and plasma lactate values correlate with the need for surgery in clinical colic cases, but are not useful as a sole indicator for predicting survival in nonstrangulating lesions (Thoefner et al. 2000). Plasma and peritoneal fluid D-lactate (a specific bacterial metabolite) were significantly elevated in horses with septic peritonitis or gastrointestinal rupture compared to controls, but were not useful in identifying intestinal ischaemia in horses (Yamout et al. 2010). Alkaline phosphatase in peritoneal fluid may help in identifying ischaemic or inflammatory bowel lesions in horses with acute colic but was not useful in predicting survival(Saulez et al. 2004). There remains a need for a reliable predictor for the need for surgery, especially in the milder surgical cases.

Acute circulatory failure secondary to intestinal ischaemia or infarction has been proposed as the most common cause of death in horses with gastrointestinal disease (Parry 1987; White 1990; Mair and Smith 2005b). To decrease the occurrence of complications and increase patient survival, early and accurate recognition of an ischaemic segment of bowel is essential (Arden and Stick 1988; Mair and Smith 2005b). Intestinal ischaemia includes all conditions in which the blood supply is inadequate to meet metabolic demands (Paterno and Longo 2008). Some large colon displacements and impactions may lead to mild intestinal ischaemia, secondary luminal distension and oedema, and subsequent production of cytotoxic oxygen radicals (Paterno and Longo 2008). There are currently no reports in the literature describing changes in oxygen radicals in these milder nonstrangulating colics.

The discovery of isoprostanes in 1990 uncovered a novel facet of free radical biology (Morrow et al. 1990). They are a group of prostaglandin-like compounds formed in vivo by nonenzymatic free radical-induced peroxidation of arachidonic acid and are reportedly sensitive indicators of oxidative stress and ischaemia in humans and animals (Marnett et al. 1985) and provide an objective assessment of disease progression and response to therapeutics (Morrow et al. 1990). Isoprostane concentrations are increased in various pathophysiological states, including ischaemia-reperfusion injury, atherosclerosis, diabetes, experimentally-induced oxidative stress and inflammation (Bailey et al. 2004; Kaviarasan et al. 2009; Ishii et al. 2010; Loke et al. 2010; Wu et al. 2010) and can be measured in a variety of biological fluids, including plasma and urine. Isoprostanes are highly unstable in plasma and can become falsely elevated by autooxidation of plasma arachidonic acid in improperly processed and stored samples. As it is frequently difficult to achieve proper processing in a clinical setting, plasma isoprostane concentrations are not well-suited for use on clinical cases. Peritoneal fluid is a potentially ideal body fluid for isoprostane quantification generated during hypoxic conditions involving the gastrointestinal tract; however, the presence of lipid and cells inherently present in peritoneal fluid would require treatment similar to plasma (Wittek et al. 2010). Since urine does not have a high lipid content, autooxidation is not a concern (Morrow et al. 1990) and may be useful in clinical cases.

Limited work has been done to investigate isoprostanes in horses. Isoprostanes have been shown to be a useful marker of oxidative stress and airway inflammation in bronchoalveolar lavage fluid (Kirschvink et al. 1999, 2002a,b,c). Isoprostanes are not only markers of oxidative stress, but also evoke biological responses such as vasoconstriction and enhanced neutrophil extravasation (Sametz et al. 1999; Zahler and Becker 1999; Basu and Helmersson 2005; Comporti et al. 2008). In laminar tissue of horses with experimentally induced laminitis, isoprostane concentrations were increased and associated with potent venoconstricting properties (Noschka et al. 2009). As such, isoprostanes may represent viable targets for the development of more effective therapeutic strategies for equine laminitis and other related diseases associated with oxidative stress to alleviate the adverse biological responses associated with disease. Although elevated concentrations of isoprostanes have been detected in tissues and urine from animals and human patients with various conditions associated with oxidative stress (de Zwart et al. 1999), their presence in horses with gastrointestinal disease compared to healthy horses has not been explored.

The metabolism of F2-isoprostanes results in the production of an entire profile of derivatives; however, a single metabolite predominates, 2,3-dinor-5,6-dihydro-15-F2t-isoprostane (referred to hereafter as isoprostane metabolite). The increased stability and longer half-life of this metabolite compared to the parent compound (Roberts and Morrow 2000), suggests that the metabolite may provide a more accurate measurement in determining the need for surgical intervention. Periods of hypovolaemia and dehydration that often occur in seriously ill colic patients result in oxidative stress in renal tissues that may result in the production of additional isoprostanes and lead to falsely elevated plasma or urine isoprostane concentrations. Because isoprostane metabolites are only generated in the liver during the metabolism of plasma isoprostanes, urine isoprostane metabolites are believed to represent a more accurate systemic index of oxidative stress than urine isoprostanes (Roberts and Morrow 2000).

The purpose of this prospective clinical study was to compare urine concentrations of isoprostane and its predominant metabolite in a population of horses with colic to those in healthy, control horses. Our aim was to determine the feasibility of using urine isoprostane and isoprostane metabolite concentrations as an early screening tool for the severity of colic and to determine the need for surgery. We hypothesised that urine concentrations of isoprostane and its metabolite could be used as prognostic indicators for the outcome of clinical cases of equine colic syndrome and as predictors for the need for surgical intervention. The long-term goal of this investigation is to identify the presence of oxidative stress in horses with colic and to reduce the number of deaths due to colic by developing a stall-side test capable of identifying colic horses in need of surgery as early as possible and expediting their timely referral.

Materials and methods


This prospective study included 42 clinically healthy (control) horses and 43 horses with colic (21 medical and 22 surgical) admitted sequentially to the Veterinary Medical Teaching Hospital at the Virginia-Maryland Regional College of Veterinary Medicine and fitting the inclusion criteria of gastrointestinal pain and urine collection within 6 h of admittance of a live horse. Healthy control horses were selected from the teaching herd or from horses attending a local horse show. All samples were collected between September 2008 and August 2009. All procedures were approved by the Institutional Animal Care and Use Committee and the Hospital Board. Informed consent was obtained from the owner for each horse included in the study.

An initial physical examination was performed on all horses presented as colic cases. Individual case data recorded included age, sex, breed, heart and respiratory rates, body temperature, capillary refill time and results of abdominal auscultation, per rectum examination and nasogastric intubation at the time of admission. Venous blood was submitted for complete blood cell count and serum chemistry. The history, duration of colic signs prior to presentation, location and length of intestinal lesion, specific diagnosis, presence of a strangulating obstruction at either surgery or necropsy, ideal treatment recommended at the time of urine collection, necessity for surgical intervention for resolution of the problem (medical vs. surgical), treatment elected by owner to pursue, complications and outcome (discharged from the hospital, died or subjected to euthanasia) were recorded.

Urine sample collection

Urine samples were collected from a total of 85 horses by free catch whenever possible; however, if horses had not urinated within the first 6 h following hospitalisation, urine was collected by sterile catheterisation (n = 8). Samples were aliquoted into sterile tubes and immediately frozen and stored at -80°C for mass spectroscopy and urine creatinine determination.

Assay for isoprostanes

Samples were shipped on dry ice via overnight courier as one batch to the Eicosanoid Core Laboratory at Vanderbilt University Medical Center for quantification of urine isoprostane and isoprostane metabolite and urine creatinine concentrations. Free urine F2-isoprostane (15-F2t-isoprostane or 8-iso-PGF2α) and isoprostane metabolite concentrations were quantified by selected ion monitoring gas chromatography negative ion chemical ionisation/mass spectrometry employing [2H4]8-iso-PGF2a as an internal standard (Agilent GC/MS system)1 using stable isotope dilution methodology. Compounds were analysed as pentafluorobenzyl ester, trimethylsilyl ether derivates monitoring the M-181 ions, m/z 569 for endogenous F2-isoprostanes and m/z 573 for 8-isoPGF2a. The F2-isoprostanes elute as a series of chromatographic peaks during a 20 s interval and quantification is based on the primary peak eluting at the same time as the internal standard (Morrow and Roberts 1999). Data were expressed in ng/ml and the detection limit of the assay was 2 pg. Isoprostane and isoprostane metabolite values were normalised to urine creatinine concentration to adjust for the specific hydration status and intersubject differences in renal function of each horse at the time of collection and results are expressed in ng/mg urine creatinine.

Statistical analysis

Normalised urine isoprostane and isoprostane metabolite concentrations from clinically healthy horses were used to establish normal ranges for the horse. Data from control horses were compared to colic horses using one-way ANOVA and Tukey-Kramer post hoc comparison. Categories for comparison to the control horses included treatment (medical or surgical), survival (yes or no), and presence of a strangulating lesion (yes or no). Three different ANOVA models were fitted for each of the categories. Furthermore, for each of the ANOVA models residual plots were inspected to verify that the errors followed a normal distribution with a constant variance. A 2 sample t test was used to compare control horses to all colic horses. Inspection of Studentised residuals revealed 4 outliers that were further investigated and subsequently excluded from the analysis based on the statistical analysis and review of the case records (one control, one medical, and 2 surgical colics). One additional case was omitted from the analysis because it failed to meet the inclusion criteria for the study. The total remaining cases consisted of 41 control and 38 colic horses (20 medical and 18 surgical). Statistical analysis was performed using SAS/STAT2. Significance was set at P<0.05.


Of the 43 horses admitted for colic, 21 horses met the inclusion criteria for surgical colic and 22 for medical colic. Median age of the study horses was 9.5 years (range 0.4–26 years) with no significant difference between groups. Twenty-seven horses were geldings, 12 were mares and 4 were intact males. A wide variety of breeds were represented and consistent with the general population of horses admitted to the hospital (11 Warmbloods, 7 Quarter Horses, 3 Thoroughbreds, 3 American Paint Horses, other miscellaneous breeds [1 or 2 each]: Arabians, Standardbreds, Saddlebreds, Haflinger, Tennessee Walking Horse, Welsh Pony, Paso Fino, Belgian Draught, Arab-Saddlebred mix, mixed breed).

Clinical signs of gastrointestinal pain were first observed 4–168 h (median 8 h) before admission to the hospital. All horses with gastrointestinal disease had been treated at least once with analgesic drugs (e.g. flunixin meglumine, xylazine hydrochloride, detomidine hydrochloride or butorphanol tartrate) or other medications (e.g. laxatives and fluids) prior to urine sample collection. Among the 38 colic horses, 9 had gastrointestinal inflammation (4 enteritis/colitis, one endotoxaemia, one mesenteric abscess, 2 gastric ulceration, one intestinal rupture), 6 had gastrointestinal strangulation (3 pedunculated lipoma/mesenteric rent entrapment, 3 large colon torsion) and 23 had nonstrangulating gastrointestinal lesions (19 impactions, 9 displacements, one spasmodic colic, one mesenteric haematoma, one Potomac horse fever and one unknown origin).

Of 38 colic horses, 32 were discharged (84.2%). Mean ± s.d. concentrations of urine isoprostane and its metabolite metabolite were significantly higher in the overall colic horse population (2.94 ± 1.69 ng/mg creatinine; and 0.31 ± 0.22 ng/mg creatinine) compared with control horses (1.89 ± 1.39 ng/mg creatinine, P = 0.003 and 0.22 ± 0.08 ng/mg creatinine, P = 0.008).

Urine isoprostane concentrations were significantly higher in medically treated colics (3.31 ± 1.97 ng/mg creatinine) compared to controls (1.89 ± 1.39 ng/mg creatinine; P = 0.004); however, there were no significant differences between surgical colics (2.52 ± 1.23 ng/mg creatinine) and controls or medical colics. Mean urine isoprostane metabolite concentrations were significantly higher in surgical colics compared to control horses or medical colics (P = 0.002) (Fig 1a).

Figure 1.

Urine isoprostane metabolite concentrations normalised by urine creatinine (Cr) concentrations from control horses and horses presenting for colic grouped by (a) treatment; (b) survival; or (c) strangulation. + indicates mean value for each group. Superscript letters indicate significant differences between groups.


Exploratory celiotomy was performed in 17 horses, 12 of which survived to hospital discharge (70.6%). All medical colics were discharged alive. Necropsy, including histopathological evaluation, was performed in all horses that were subjected to euthanasia. One horse was subjected to euthanasia without undergoing surgery for economic reasons, and 4 intraoperatively because of a poor prognosis and the high expense of resection and anastomosis (one each: jejunal rupture, large colon impaction/eosinophilic enteritis, pedunculated lipoma and mesenteric abscess). One horse recovered from surgery but died 4 h following surgery for undetermined reasons. Of 18 horses diagnosed with large colon impaction, 12 were treated medically and 6 surgically. One of the impactions treated surgically had a concurrent large colon volvulus and the horse was subjected to euthanasia during surgery while all other horses diagnosed with an impaction survived until discharge. Of the 9 horses suffering from a right dorsal displacement of the large colon, 5 (presumed right dorsal displacements) were treated medically and 4 surgically. All horses treated for a displacement survived until discharge. All 3 horses diagnosed with a large colon volvulus were treated surgically. One of these horses was subjected to euthanasia during surgery due to the associated cost and prognosis. The other 2 cases recovered and were discharged.

Urine isoprostane concentrations were significantly higher in surviving colic horses (3.00 ± 1.69 ng/mg creatinine) compared with controls (P = 0.012); however, there were no significant differences between nonsurvivors (2.61 ± 1.76 ng/mg creatinine) and controls or survivors. Urine isoprostane metabolite concentrations were significantly higher in nonsurvivors than controls or survivors (P = 0.001) (Fig 1b).


Urine isoprostane concentrations were significantly higher in nonstrangulating (n = 31) colic horses (3.03 ± 1.74 ng/mg creatinine) compared with control horses (P = 0.011). Urine isoprostane concentrations were not significantly different in strangulating (n = 7) colics (2.55 ± 1.46 ng/mg creatinine) compared to either controls or nonstrangulating colics. The overall ANOVA F test showed that mean urine isoprostane metabolite concentrations were significantly different between groups based on whether the lesions were strangulating or not (P = 0.020). However, statistical significance for this outcome disappeared after correcting for multiple comparisons (Fig 1c).


Isoprostane biosynthesis, secretion and excretion are regulated in part by the availability of precursors required for isoprostane synthesis, such as dietary and tissue arachidonic acid, as well as tissue oxygen concentration, and the generation of various free radical species (Basu and Helmersson 2005). Reduced segmental blood supply to the intestinal tract in horses with gastrointestinal disease may be due to functional constriction or mechanical obstruction of blood vessels and can lead to inadequate tissue perfusion and oxygenation, rapid injury and death of the highly energy-dependent mucosal epithelial cells, and increased release of arachidonic acid from injured cells and lipoproteins in blood (Moore et al. 1995). Nonsteroidal anti-inflammatory drugs, the first line of treatment for horses with colic, control pain by reducing arachidonic acid products involved in inflammation and pain (Moore et al. 1981). Release of cell membrane-derived arachidonic acid (Moore et al. 1995) and hypoxaemia-induced oxygen radicals (Shebani et al. 2000) resulting from cell injury in strangulating and nonstrangulating gastrointestinal conditions make the involvement of isoprostanes in the pathogenesis of colic likely. In this study, concentrations of isoprostane and isoprostane metabolite in urine were significantly higher in horses with colic than in healthy horses. Experimental ileal stasis in rats, a model similar to nonstrangulating causes of equine colic syndrome, e.g. impactions of the ileum or large colon, resulted in elevated isoprostane concentrations (Shebani et al. 2000). In these milder cases of nonstrangulating obstructions, locally generated isoprostanes may contribute to vasoconstriction, worsening ischaemia in the bowel secondary to oedema formation. Concentrations of isoprostane metabolite were significantly higher in horses requiring surgical intervention than in horses treated medically, as well as in horses that failed to survive. These preliminary findings suggest that urine concentrations of isoprostane metabolites may be useful as an indicator of prognosis and the need for surgical intervention in horses with colic.

Isoprostane metabolism and excretion and type of colic could account for differences between the pattern of changes of isoprostane and isoprostane metabolite concentrations in urine. Because isoprostane metabolite is more stable and has a longer half-life than the parent compound, it may be more accurate in determining the need for surgical intervention (Roberts and Morrow 2000). Values for isoprostane metabolite in healthy horses were tightly grouped with a small standard deviation, whereas the deviation for the isoprostane concentrations was quite large. This difference in standard deviation substantiates the concept that isoprostane metabolite may more closely reflect systemic lipid peroxidation independent of any influence of hydration status. Because of the increased stability and lack of renal oxidation, the remainder of the discussion will focus solely on isoprostane metabolite.

Urine isoprostane metabolite concentrations were significantly higher in horses requiring surgical intervention than in control horses or those treated medically. The group of horses requiring surgery included both strangulating and nonstrangulating obstructions. Most surgical conditions including more severe impactions and displacements of the large bowel are potentially characterised by ischaemia-induced oxidative stress due to decreased gastrointestinal perfusion, unregulated lipid peroxidation and tissue damage. Increased urine isoprostane metabolite concentrations may reflect the duration and severity of the condition. Values for the medically treated colic horses and healthy control horses were nearly indistinguishable, suggesting that the horses requiring medical therapy were not as severely compromised in their gastrointestinal perfusion status as the horses requiring surgery.

The results of this study suggest that there may indeed be ischaemia-induced oxidative stress occurring in nonstrangulating lesions. Nonstrangulating conditions have not typically been thought of as related to oxidative stress and ischaemia in the past. However, it appears that surgically corrected impactions or displacements of the large colon induced sufficient isoprostane metabolite concentrations in order to statistically distinguish them from the less severe medically treated colics, suggesting that isoprostane metabolite may be helpful in distinguishing medical from surgical cases of nonstrangulating obstructions. A study in rats found increased urine isoprostane concentrations in animals suffering from intestinal stasis caused by a surgically created U-pouch of the terminal ileum, signifying an association with oxidative stress (Shebani et al. 2000). Further evidence for the role of free radical species in the formation of isoprostanes is that dietary antioxidant intervention with vitamin E, a potent free radical scavenger, and allopurinol, an inhibitor of xanthine oxidase, reduced myeloperoxidase activity and isoprostane production and ameliorated the early inflammatory events associated with intestinal stasis in rats (Shebani et al. 2000). The effectiveness of antioxidants also points out the potential use of isoprostanes as a therapeutic target. The effects of NSAIDs on concentrations of F2-isoprostane and its metabolite in equine colic patients are unknown and would be difficult to predict based on the limited literature available and the altered metabolic status of most colic patients. NSAID use may result in lower breast cancer risk in women via reduced levels of F2-isoprostane (Kato et al. 2006); however, NSAID use in endurance runners resulted in increased oxidative stress as measured by plasma and urine F2-isoprostanes (McAnulty et al. 2007). Nearly all of the colic patients in this study received NSAIDs prior to sample retrieval, making correlation between isoprostane concentrations and NSAID therapy impossible.

The results of our study identify the important role that isoprostane and its metabolite may have in the pathophysiology of colic. Their vasoactive properties may mediate oedema formation, ischaemia and oxidative stress within the bowel wall. Isoprostane may be a viable target for treatment with antioxidants that inhibit the formation of free radicals, even in milder forms of surgical colic cases. Isoprostane may exacerbate oxidative stress in the inflamed colon by inducing mucosal ischaemia secondary to vasoconstriction (Stucchi et al. 2000). These findings support the use of antioxidant or direct cell-protecting treatments such as dimethylsulphoxide, lidocaine, tocopherol, ascorbic acid and allopurinol as adjunctive treatments in surgical colics (Schoenberg and Beger 1993; de Moffarts et al. 2005; Guschlbauer et al. 2010). The routine use of a lidocaine continuous rate infusion started intraoperatively, even in milder colic cases, may diminish post operative oxidative stress and ileus, and enhance the speed of recovery. Measurement of isoprostane and isoprostane metabolite concentrations could not only be useful preoperatively, but as a serial assay post operatively to monitor disease progression and/or resolution similar to the ultrasonographic measurement of large colon wall thickness following correction of large colon volvulus (Sheats et al. 2010).

Although there were relatively few nonsurvivors in the colic horse population in this study, urine concentrations of isoprostane metabolite were significantly higher in these horses when compared to either survivors or healthy control horses. Further studies are indicated to strengthen these findings and further assess the sensitivity and specificity of predicting survival in the aim of identifying a cut-off value that would help determine prognosis or the need for surgery.

Urine isoprostane metabolite concentrations in horses with strangulating obstructions were higher than those in horses with nonstrangulating obstructions or in healthy control horses; however, this difference did not reach statistical significance. This could, in part, be due to the low number of horses with strangulating lesions or due to impaired urine production as a consequence of shock and hypovolaemia. The relatively large standard deviation within this group may make it difficult to find a cut-off value; however, strangulating lesions are easier to diagnose preoperatively than some milder forms of colic requiring surgery such as large colon displacements and impactions making the need for an objective value less urgent. Delineation of the need for surgery in nonstrangulating lesions may provide a more useful application for measurement of urine isoprostane metabolite concentrations than in horses with strangulating lesions.

The use of isoprostane metabolite concentrations as an additional screening tool to complement measurement of lactate may not only detect oxidative stress associated with strangulating lesions but also milder ischaemic conditions accounting for the severity of nonstrangulating lesions. Lactate seems to be more useful in predicting survival in strangulating lesions, but is less accurate in nonstrangulating lesions. Coupling the use of lactate and urine isoprostane metabolites may provide accurate determination of prognosis or the need for surgery in a wider range of diagnoses. A larger multi-centre study involving more horses in each category would help further define the potential value of isoprostane metabolite concentrations in distinguishing horses with strangulating obstructions from those with nonstrangulating obstructions and to provide an accurate cut-off value.

Isoprostane plasma concentrations can provide a useful index of total endogenous production of isoprostane because concentrations in plasma presumably are derived from all tissues in the body. A potential contribution of local formation of isoprostane in the kidney may confound interpretation of urine concentrations of unmetabolised isoprostane (Roberts and Morrow 1997; Taber et al. 1997). Despite the high specificity and sensitivity of isoprostane for oxidative stress, their quantification in the circulation will only offer information regarding a discrete point in time, since it is cleared rapidly from the circulation (within approximately 16 min in rats) (Morrow et al. 1990). Consequently, if large alterations occur in the production of lipid peroxides over a longer period (several hours to days in colic cases), measurement of isoprostane concentrations in a single plasma sample will not provide an accurate integrated assessment of oxidative stress. In dynamic situations in which there is an oxidant insult for only a relatively short period of time (e.g. ischaemia-reperfusion injury), multiple sequential samplings of blood would be necessary to assess the full magnitude of the increase in isoprostane generation during rapidly changing rates of production (Richelle et al. 1999). Measurement of the more stable metabolised isoprostanes in urine as an index of total endogenous isoprostane production may circumvent the potential contribution of local isoprostane production in the kidney during states of dehydration and provide a more accurate measure of the systemic oxidative stress of a patient over time. Considering all of these aspects, it appears more practical to collect urine samples than plasma samples for the evaluation of isoprostane metabolite from clinical colic cases.

The usefulness of isoprostane metabolite concentrations as a clinically useful predictive indicator for colics is dependent in part on the ability to develop a stall-side assay. A recent study compared enzyme immunoassays with gas chromatography/mass spectrometry in domestic animal species for the measurement of urine isoprostanes as markers of in vivo lipid peroxidation and demonstrated that enzyme immunoassays are not reliable for the determination of isoprostane concentrations in plasma or urine of horses (Soffler et al. 2010). The more stable isoprostane metabolite, however, may be more suitable for an enzyme immunoassay that could be developed as a stall-side assay, similar to the SNAP-test used for quantification of serum IgG concentrations in equine neonates. The ability to reliably measure urine isoprostane metabolite concentrations using a simple stall-side test would be an excellent clinical application as an aid in determining the need for surgical intervention in horses with colic.

The results of this study provide justification for a large multi-centre study on horses with colic to further develop the predictive value of isoprostane metabolites in the need for surgery and to prognosticate during the post surgical period. We were able to detect oxidative stress as an important aspect of the pathogenesis in horse suffering from colic, which provides evidence for the implementation of antioxidant therapies to improve the overall outcome. This contribution to the knowledge of the formation of isoprostanes and their metabolites in horses suffering from colic may potentially aid in the reduction of isoprostane generation through early surgical intervention based on a stall-side test or the prevention of further systemic and gastrointestinal tissue injury in horses with colic by timely referral and surgery.

Conflicts of interest

No conflicts of interest have been declared.

Sources of funding

Supported by funding from the Office of the Vice President for Research, Virginia Tech.

Manufacturers' addresses

1 Agilent, Palo Alto, California, USA.

2 SAS Institute Inc. Cary, North Carolina, USA.