15 Aug 2016
Nigel Dougherty presents a case of GDV in a New Zealand sheepdog and discusses management of the stomach’s ischaemic gastric fundus and greater curvature.
Nigel Dougherty
Job Title
Figure 1. A grey-green area of necrosis on greater curvature of the stomach, avulsion of short gastric arteries and heterogenous discolouration of much of the surrounding greater curvature.
Having euthanised a gastric dilatation-volvulus (GDV) case intraoperatively, the author has since held reservations about his decision, thinking retrospectively perhaps the case was quite salvageable.
This article is a case write-up illustrating some of the quandaries faced with managing GDV cases, given the subjectivity of clinical methods used to assess the viability of the stomach wall. It looks at the strength of evidence to “go” with such clinical methods of assessment and puts sensitivity and specificity of such clinical methods into context, comparing them with some of the technologies considered to help identify tissue that is not viable.
A seven-year-old entire male huntaway dog (New Zealand sheepdog) with suspected acute gastric dilatation-volvulus (GDV) was presented within an hour of its owner having noticed and relieved abdominal distention by percutaneous needle gastrocentesis.
Marked tympanic cranial abdominal distention was apparent on admission. The dog was ambulatory, with mentation slightly depressed. Major body system assessment indicated moderate systemic hypoperfusion likely consistent with impending obstructive shock. Point of care blood lactate would have provided initial and ongoing insight into whole body perfusion.
Suspicion of GDV justified surgery without radiographic confirmation. Preoperative stabilisation was brief, involving a second percutaneous needle gastrocentesis and provision of lactated ringers at 90ml/kg/hour per cephalic vein. Preoperative morphine (10mg, slowly IV) preceded prompt induction with alfaxalone (2mg/kg, slowly IV) for rapid intubation.
Anaesthesia was maintained with isoflurane. Electrocardiographic monitoring was not available. In retrospect, a continuous rate infusion of morphine-lidocaine-ketamine would likely have conferred more balanced analgesia.
Per exploratory celiotomy, the stomach was draped ventrally by greater omental leaf, indicating clockwise rotation of the pylorus relative to the gastric mesenteric axis (Tivers and Brockman, 2009). The locality of the pylorus in the left cranial abdominal quadrant suggested volvulus of at least 180°. The junction between the fundus and gastric body along the stomach’s greater curvature was compromised with an L-shaped grey-green colour covering about 5cm × 2cm (Figure 1), the area being consistent with that considered most commonly liable to necrosis (Tivers and Brockman, 2009).
This is coincident with proximity to the short gastric arteries, itself considered to be at highest risk of necrosis (Monnet et al, 2006) because these are often avulsed in the torsion (Fossum et al, 2007).
Much of the remainder of the stomach’s greater curvature showed heterogeneous discolouration – small blotchy to streaked purple and black areas of less than 1cm with deep red-purple peripheries.
Derotation was performed per the technique described by Fossum et al (2007) for clockwise torsions and, although the gastrosplenic ligament was verified not to have torsed after derotation (Fossum et al, 2007), the spleen remained congested and extensive thrombosis of the collateral supply to it was suspected. Gastric wall discolouration improved, but not significantly after 10 minutes of repositioning.
Together with the presence of apparent gastric wall necrosis and potential for thrombosis, the prognosis was considered ominous enough to justify intraoperative euthanasia.
Considering the systemic stability of the patient, the confined locality of gastric necrosis evident and ambiguity over interpretation of the heterogeneous discolouration of the remainder of the stomach wall, the decision not to undertake partial gastrectomy or invagination was questioned, retrospectively.
This was particularly given the collateral circulation offered to the wall of the greater curvature by short gastric arteries and the left gastroepiploic artery, and through anastomoses between left and gastroepiploic arteries, respectively (Bezuidenhout et al, 1996; Budras et al, 1994). This prompted a review of the literature to assess the strength of evidence to euthanise this patient.
One challenge facing management of GDV cases lies with the correct identification of all the non-viable, ischaemic gastric wall, to be able to combine judgements about the extent and distribution of ischaemia/necrosis with overall assessments of whole body perfusion to institute appropriate medical and surgical correction or to elect euthanasia under appropriate circumstances.
Observation of colour/return of colour, visualisation and palpation of wall thickness, palpation and observation of pulses in local vessels, observation of peristalsis and thrombi, and assessment of cut-surface bleeding by partial thickness (seromuscular) incision(s) into the enteric wall form the foundation for practical clinical assessment of its overall viability (Matthiesen, 1983; Tivers and Brockman, 2009; Wheaton et al, 1986; Green et al, 2011). Table 1 shows example hypotheses for interpretation of gastric texture and discolouration of the gastric serosa-muscularis.
Given the subjectivity of clinical assessment methods, research has compared them with technologies that have the potential to improve sensitivity and specificity – particularly for correctly identifying non-viable tissue of varying enteric localities.
High sensitivity for identifying tissue as non-viable is more important than high specificity (correctly identifying and retaining all viable tissue), provided critical amounts or localities of enteric tissue are not unnecessarily excised.
At the same time, partial gastrectomy may be associated with elevated risk of mortality, although this was not to be the case in one study (Beck et al, 2006). Furthermore, while causality was not established, Pirker et al (2015) suggested possible over-interpretation of the presence of gastric necrosis, supported by comparatively low prevalence of partial gastrectomies among a large number of GDV survivors in their study in relation to others.
The sensitivity of subjective clinical methods relative to such technologies may vary, depending on whether the aim is to definitively rule tissue as viable or rule out as non-viable.
Their relative accuracy may also depend on the predominant type of vascular compromise involved, and the enteric locality of concern.
Furthermore, given the differing nature of blood supply in various parts of the enteric system and in different species, extrapolations of findings to other species or other parts of the enteric system need to be made with caution (Mann et al, 1982). Variation may also be explained by differences in underlying study design and in how conservative cut-off points used to judge viability in different studies are.
Studies comparing the accuracy of clinical and augmentory methods of detecting non-viable tissue against a histopathological gold standard serve to illustrate such variance. In studies of experimentally compromised cat and dog small intestine, clinical evaluations were found to be unreliable at detecting non-
viable tissue (Mann et al, 1982; Wheaton et al, 1983), as the eye could not detect small areas of ischaemia that later became necrotic.
Such sensitivity at detecting ischaemic injury improved with venous occlusions, rather than with arteriovenous occlusions – probably because ischaemic injury following venous occlusions tends to be more generalised (Wheaton
et al, 1983). By contrast,
fluorescein-fluorescence based methods were deemed particularly accurate at identifying ischaemic small intestinal mucosal necrosis in cases of arteriovenous and venous occlusion, with necrosis developing within a few millimetres beyond fluorescence.
Pulse oximetric methods were insensitive for detection of non-viability, irrespective of type of occlusion (Wheaton et al, 1983; Mann et al, 1982).
The relative merits of clinical versus fluorescein-fluorescence methods for detecting non-viable gastric tissue give opposing results to similar studies involving small intestine as summarised beforehand.
In one study of GDV induced under experimental conditions (Wheaton et al, 1986), where any degree of fluorescence was considered a priori as viable, clinical bases for evaluation were found to be 85% sensitive for correctly identifying devitalised tissue (later confirmed as such by histopathology) – a sensitivity 27% higher than with fluorescein-based assessments.
Furthermore, mucosal-based assessments were considered to be poorer markers of non-viability than musculoserosal assessments of gastric tissue. The gastric mucosal layer was considered by histopathological evaluation to be less hypoxic or more refractory to the effects of hypoxia than the seromuscularis layers and, according to the authors, several mechanisms are suggested to augment gastric mucosal blood flow and provide mucosal cytoprotection – particularly at the transmural pressures manifest during GDV.
Interestingly, and in potential contrast, Monnet et al (2006) found the mucosa-submucosa experienced the most severe reductions in blood flow as measured by laser-Doppler flowmetry (LDF) and coloured microsphere spectrophotometry, although it is unclear whether this was a finding under experimental induction of portal hypertension and/or arterial ligation of all vessels supplying the greater curvature of the stomach. GDV was not induced in this part of the study.
Explanations in the literature for the variation in findings place emphasis on understanding the underlying pathophysiology of compromise. The Wheaton et al (1986) studies were considered by the authors to corroborate findings from their 1983 study and the Mann et al (1982) small intestinal enteric studies to show fluorescein studies may be less accurate as indicators of viability (as opposed to non-viability) when splanchnic venous outflow obstruction is the main determinant of hypoxia.
Under such conditions, clinical assessments are thought in a relative context to improve in sensitivity for the detection of viable tissue. However, while caudal vena cava compression and compression and torsion of the portal vein are likely to be critical determinants of hypoxia (Wheaten et al, 1983; Lantz et al, 1984), other causes of ischaemia may also play an important role in its development besides venous outflow obstruction – especially in naturally occurring cases of GDV.
Capillary compression, thrombosis, microembolisation and rupture, particularly of the short gastric arteries, may also contribute significantly to gastric infarction and necrosis (Fossum et al, 2007; Monnet et al, 2006).
Most of the technologies employed to augment clinical assessments provide measures of perfusion or oxygenation (and indirectly of ischaemia) rather than necrosis – for example, coloured microspheres (Hakkinen et al, 1995), fluorescein (Wheaton et al, 1986), LDF (Monnet et al, 2006), pulse oximetry (Wheaton et al, 1983; Mann et al, 1982, also referred to in Fossum et al, 2007) and yet no reliable guide remains to judge whether ischaemia will become necrotic, as pointed out by Mann et al (1982). One of the difficulties, according to these authors, is to be able to predict extensions of thrombotic processes.
Recognising these predictive limitations, clinically-based decisions to perform gastrectomy, or otherwise, in naturally occurring GDV cases were broadly correlated with measures of blood flow in the stomach wall using LDF and microsphere spectrophotometry (Monnet et al, 2006).
Perhaps more significantly, in this study, no non-viable cases were missed using the combined subjective clinical criteria of colour, palpation and peristalsis (implying 100% sensitivity of detection of non-viability), although the specificity of decisions to resect could not be ascertained from the study because histopathology was not undertaken as a gold standard.
Besides the detection of clearly non-viable tissue demanding resection or invagination, a decision, in this case, needed to be made about how conservative margins of excision would need to be beyond necrotic tissue, given the risks of dehiscence following incision into compromised tissue.
The heterogeneity of discolouration does make such delineation more difficult to gauge and, although LDF has not been validated to assess gastric viability under conditions of GDV (particularly with the capillary compression and oedema that manifest), it has proven 100% sensitive in assessing viability of human intestine (Smits et al, 1986).
Particularly under circumstances of greater arterial than venous compromise (Monnet et al, 2006), LDF may have the potential to “map” discrete interfaces where viability is questionable, given the small volume of tissue whose blood flow is measured by the probe (Michelson et al, 1996; Shepherd et al, 1987).
Another decision requirement for this GDV case pertained to the total amount and specific localities of gastric wall that could be resected without unduly elevating the mortality risk. The gastric cardia and distal oesophagus were, perhaps, not as fully assessed, as recommended by Tivers and Brockman (2009).
The junction between fundus and gastric body along the greater curvature of the stomach is, however, the most common site for necrosis following GDV (Tivers and Brockman, 2009) and, as such, will most likely determine the extent of partial gastrectomy required.
Descriptions of the procedure suggest a substantial amount may be removed (Matthiesen, 1983; Fossum et al, 2007), with one dog reported to have survived at least 12 months after resection of the entire greater curvature (Matthiesen, 1983). Ischaemic necrosis of the gastric cardia and oesophagus and extensive damage and necrosis of both the lesser and greater curvatures of the stomach, respectively, carry a grave prognosis and latterly are not considered amenable to resection (Tivers and Brockman 2009; Matthiesen, 1983).
The duration of GDV in this case likely exceeded the six hours considered to significantly raise the likelihood of need for partial gastrectomy (Beck et al, 2006), notwithstanding the time taken for necrosis to develop is highly individual and dependent on a number of factors at play. In retrospect, support for a high sensitivity of clinical methods of viability assessment is further reason to consider if this case was salvageable – especially given the extent of gastrectomy feasible and the stable systemic status of this particular patient during surgery.
Furthermore, there was no deep black or strong demarcation suggestive of full thickness necrosis and, although the serosa appeared to be very compromised and of streaked discolouration, given the proportionately greater volumes of blood flow directed to the mucosa/submucosa, this tissue may have remained viable.
Perhaps, most importantly, given the relative systemic stability of this patient, any concerns about the viability of tissue conserved in the first surgery could have been assessed without unreasonable extra risk to the patient in this case by conduct of a “second look” laparotomy. The need for and timing of this would require intensive postoperative monitoring of the animal’s systemic condition for the most appropriate decisions to be made.
The author would like to thank Jonathan Bray, head of the companion animal group at Massey University, for providing constructive thoughts on the management of the case.
