 |
       |
|
|
ENDOVASCULAR TREATMENT OF GASTROINTESTINAL HAEMORRHAGE
SHREYAS MASRANI, HEMANT DESHMUKH
Division of Interventional Radiology, GS Medical College and KEM Hospital, Parel, Mumbai 12.
INTRODUCTION
Percutaneous intervention in gastrointestinal haemorrhage has emerged today as a potent, quick and versatile treatment modality for the treatment of gastrointestinal haemorrhage. It has tremendous logistic advantages for it does away with all the attendant problems associated with anaesthesia and laparotomy and generates a situation where diagnosis of the underlying pathology, localization of lesions and finally endovascular treatment can be done in one sequential session through a single femoral artery access taken under local anaesthesia. It is the only minimally invasive technique in controlling haemorrhage in the small bowel; an area where endoscopes cannot reach and in vascular lesions that are 'out-of-sight' for the endoscopist such as pseudoaneurysms of the splenic and hepatic arteries.
History
Though Rastelli in 1959 and Margulis in 1960 showed active extravasations of contrast medium in the lumen of the bowel in active gastrointestinal haemorrhage,[1,2] it was the landmark work of Rosch and Baum in 1971[3] which showed that gastrointestinal haemorrhage could be controlled with the selective intra-arterial infusion of vasoconstrictor drugs. In 1972 Rosch, Doter and Brown[4] controlled acute gastrointestinal haemorrhage in a patient by embolisation of the gastro-epiploic artery with autologous clot. Since then the advent of a variety of catheters and micro-catheters with co-axial systems as well as a legion of embolic materials has placed embolotherapy at the forefront of emergency treatment of gastrointestinal haemorrhage. The vascular intervention radiology department at our institution, for instance, with all its logistic limitations has performed 28 embolisations for gastrointestinal haemorrhage from Apr 2000 until date with only 3 failures, giving a high success rate of 89.2%.
GENERAL CONSIDERATIONS
The factor of paramount importance for successful embolotherapy is the availability of an experienced interventionist for the following reasons:
1.The diagnosis of a positive angiographic result needs subtle interpretation as the parenchymal phase of the angiogram can appear as extravasation or the tracking of extravasated contrast in the lumen can create a pseudovein appearance. These can be falsely interpreted by the novice. Also it has been shown that bleeding rates as low as 0.5 ml/minute should be detectable arteriographically[5] but at such a low rate, the extravasated contrast medium can appear as dense as the mucosa itself.
2.Manoeuvring the co-axial micro-catheter systems requires exquisite skill on the part of the interventionist whereby optimal utilization of the trackability and torquability of that hardware is achieved for selective and if necessary super-selective catheterisation. All this needs familiarity with the hardware and judicious use of new hardware in the market; this comes only with experience.
3.For deciding against embolisation in equivocal angiographic findings.
4. For identifying anomalous origin of vessels which are particularly common in the coeliac and superior mesenteric arterial axes.
With the routine pre-procedure work-up being a haematocrit and coagulation profile, femoral access is achieved by Seldinger's technique under local anaesthesia. In those cases where frank haematochezia is the main manifestation, the first artery to be catheterised [usually by an SIM I (sidewinder) catheter] is the inferior mesenteric artery. This also prevents overlap of the territory of inferior mesenteric artery by excreted contrast in the bladder from higher vessels. When the site of haemorrhage is not clear, the superior mesenteric artery is the first vessel to be catheterised as it supplies territories proximal as well as distal to the ligament of Treitz. Identification of the vascular lesion by CT angiography or the site of bleeding by endoscopy can limit the number of vessels to be studied to just one (as in single vessel coeliac angiograms for pseudoaneurysms of the splenic artery).
The rate of contrast injection is important as the extravasation is to be picked up during the arterial phase of the study. For the coeliac and superior mesenteric angiograms 15 ml at the rate of 15 ml/sec at a pressure of 300 psi is used. For the inferior mesenteric angiogram, a brisk hand injection of 6-7 ml suffices.
Aortic reconstructive surgery has a 1% incidence of developing aorto-enteric fistula, the clinical setting is suggestive of the complication which has a 70% mortality rate. A lateral projection of an abdominal aortogram is the only image needed for arriving at the diagnosis.
COMMON CONDITIONS RESPONDING TO EMBOLOTHERAPY
Proximal to the ligament of Treitz, peptic ulcer disease including post-op bleeds after surgery for duodenal perforations forms a significant chunk of the patients presenting as massive upper gastrointestinal bleeds either in the form of frank haematemesis or as frank blood in the RT aspirate. A recent study of a series of cases from KEM hospital over the last 2-1/2 years showed splenic artery pseudoaneurysms forming almost half the number of cases presenting as gastrointestinal bleed; either as 'haemosuccus pancreaticus' (bleeding into the pancreatic duct and then into the duodenum manifesting as either haematemesis, melaena or both) or as bleeding into the pigtail catheter placed in the pseudocyst. Percutaneous biliary drainage procedures presenting with haemobilia as a complication can also be successfully embolised.
Distal to the ligament of Treitz, small bowel neoplasms (spindle cell tumours) form a major chunk with rarities such as Meckel's diverticular bleeds encountered once in a while particularly in paediatric patients.
Distal to the ileo-caecal valve diverticular disease is the commonest cause of GI haemorrhage6 followed by neoplasms and angiodysplasias of the right colon. However finding an area of angiodysplasia does not mean that it is the source of bleeding and other causes such as small bowel carcinoma, leiomyoma or Meckel's diverticulum can be the cause of bleeding as angiodysplasias often exist in elderly non-bleeding patients.[7,8]
Dieulafoy's disease is a rarity where bleeding occurs due to a submucosal arterial protruding a mucosal defect. This can be a cause of severe haematochezia especially when the colonoscopy has failed to identify any mucosal based lesion.
Vascular involvement constitutes an important aspect of GITB. During the ulcerative phase of the disease, superior mesenteric angiography shows hypervascularity in the affected region.[9] This explains GI haemorrhage in TB. Selective embolisation or vasopressin infusion is an important therapeutic modality in patients with intestinal tuberculosis presenting with massive GI bleed.[10]
Role of super-selective catheterization in embolotherapy
Superselective catheterization and embolisation has become possible with the advent of co-axial microcatheter systems such as Magic 3F/2F Dasher-Tracker and Mass transit systems. It minimises the spill over embolisation of non-target vessels that can lead to pancreatitis, gallbladder infarction, hepatic and splenic infarction and splenic abscess. It minimises the infarction of the embolised tissue. This is uncommon in the upper GI tract where organs have a verdant collateral blood supply, it is of grave concern in the small bowel or colon. In a large study of 38 patients observed over a 2 year period by AA Nicholson in 1998.[11] The platinum coil embolisation technique was used to control colonic haemorrhage. Initial inferior and superior mesenteric angiography was performed via a femoral approach using 5-French gauge femorovisceral catheters. Once the bleeding point was identified a 2.2-French gauge coaxial Tracker-18 catheter (Terumo, Piscataway New Jersey, USA) was passed over a torquable seeker guide wire, down the femorovisceral catheter in relation to site. Superselective angiography was used to confirm the position of the catheter in relation to the bleeding point. Appropriate size platinum microcoils were then advanced down the catheter into the bleeding vessel using a pusher wire. Final super-selective angiography confirmed coil position and cessation of bleeding.
Guidelines for transcatheter embolisation in GI haemorrhage
1.The embolisation should be super-selective and only the bleeding artery with its feeder should be occluded. If more than one artery is to be embolised the number should be kept minimum. During catheterisation care should be taken to minimise spasm and avoid initimal dissection of the parent artery. This is ensured by looking for free backflow of blood from the catheter once it is positioned in the ostium of the parent vessel. On the first instance blood is allowed to just enter the syringe to check for free backflow, subsequently a free backflow of infusion fluid in the catheter is considered sufficient.
2. Except for occluding the nidus of a lesion, embolisation should always be peripheral. More proximal embolisation which can be unsafe may have to be considered in patients with coagulation defects.
3. If intravenous Octreotide, a long acting somatostatin analogue (1/2 life : 8 hours) has been used by the clinicians in the wards prior to referral for embolisation it should be only after a time lag where the Octreotide effect is over, otherwise vasospasm can occur on catheterising the vessel.
4. Choice of proper occlusive material is important for safety of embolisation.
5. Antibiotic coverage should be used due to the easy contamination of gelfoam and the bowel bacteria which can easily migrate across the ischaemic bowel wall.
EMBOLIC MATERIALS
1. Gelfoam : An absorbable sponge derived from Gelatin, its use as an embolic agent was first described by Carrey and Grace in 1974. Where they used it to control haemorrhage from duodenal ulcer.[12] The advantage of gelfoam is that it works well irrespective of underlying coagulopathies. Gelfoam embolisation causes a severe pan-arteritis characterised by leucocytic infiltration in all layers of the vessel wall as well as by disruption of the intima and elastic tissues. Gelfoam promotes clot formation and thrombosis is commonly seen distal to the embolisation site. The duration of vascular occlusion lasts from a few days to a few weeks with the pan-arteritis resolving within 4 months.
2. Embolisation coils : Gianturco and Wallace introduced steel coils for permanent vascular occlusion.[13-15] These coils induce thrombosis, to facilitate which wool strands (tails) are attached to the coils. This induced severe granulomatous arteritis which extend through the adventitia and hence the wool was subsequently replaced by Dacron. Commercially they are available in various lengths, various diameters and are made of stainless steel. Microcoils are those which are used with microcatheters in cases of super-selective embolisation. We, in our institute have used tracker 18 co-axial system with steel and Tungsten microcoils and nowadays the use of the new 'Mass transit' microcatheters (Cordis, USA) is the latest addition to the armamentarium at our disposal. The usual coils for larger artery embolisation have a diameter of 0.035 inch while the microcoils have a diameter of 0.018 of an inch.
Platinum and tungsten coils are also in use for platinum is highly thrombogenic, radio-opaque and biocompatible. Added synthetic fibres play a major role in contributing towards the thrombogenicity of the platinum coils. Commercially they are available in straight, curved and complex (helical forms).
Coils can be deployed using a guide wire on pusher (Newton TSFNB or TSFNC Betson TSFB wire guides are recommended for the same. The detachable coils (Jackson detachable coils or MDS) have a special delivery system for accurate positioning of the coil.
The major advantage of coils over gelfoam is the ability to provide permanent embolisation which is most desirable in treating pseudoaneurysms such as those of the splenic artery and in superselective embolisation with microcoils in areas of tenuous collaterals such as the large bowel, where a wall identified small sub-mucosal bleeding artery can be safely and permanently embolised without risk of necrosis to a large segment of the bowel.
Complications of embolotherapy
The main complication of bowel infarction is avoidable to a great extent by the use of superselective microcatheter systems. Minor complications such as fever and pain are known and easily managed. Failure can occur in post-op cases, in large pseudoaneurysms and in acute conditions such as acute pancreatitis.
Limitations of trans-catheter control of GI haemorrhage
Transcatheter embolisation can fail in the long run if the basic underlying pathology is untreated or uncurable. Also if the patient is not actively bleeding at the time of angiography, which happens in patients treated by Octreotide infusion, the angiography will fail to show extravasation due to diversion of the blood pool away from the splanchnic vasculature. Thus in our experience Octreotide administration could probably explain a high incidence of negative angiographies even in the presence of acute gastrointestinal haemorrhages. Acute pancreatitis was observed in 492 patients by Testart J et al in 2001. Fourteen (2.8%) developed an arterial erosion revealed by a haemorrhage either in the digestive lumen, in the peritoneum or via previously placed drainage. The eroded artery was the splenic artery in six patients, a pancreatico-duodenal artery in five patients. An initial haemostasis was attempted by embolization in four patients: one died, three others had bleeding recurrence; splenocorporeal pancreatectomy in four patients, three had bleeding recurrence; arterial ligature in four patients; three had bleeding recurrence.
Thus it is noteworthy that durable haemostasis could not be obtained neither by embolization nor by ligature in necrotic tissues. This could explain the difference in the results of arterial erosion treatments in chronic and in acute pancreatitis. Therefore, it is suggested that haemostatic procedures should be performed away from necrotic tissues, or eventually done after their removal.
In lesions with diffuse pathologies such as angiodysplasias or Osler-Rendu-Weber disease, it can work only as temporary palliation as the large surface area of the pathology can cause recurrent haemorrhages in the future.
REFERENCES
1. Rastelli GC, Magnani L, Bocchialini C. L 'impiego dell' arteriografia nella deiagnostica dele emorragic del tubodigerente. Minerva Chir 1959;14:1188.
2. Margulis AR. Heinbecker P Bernard HR. Operative mesenteric arterography in the search for the site of the bleeding in unexplained gastrointestinal haemorrhage. Surg 1960; 48 : 534.
3. Nusbaum M, Baum S. Radiographic demonstration of unknown sites of GI bleeding. Surg Forum 1963; 14 : 374.
4. Rosch C, Dotter CT, Brown MJ. Selective arterial embolization. Radiol 1972; 102 : 303-6.
5.Grainger and Allison Diagnostic Radiology; Gastrointestinal angiography chap. 52 (2) : 1098-99.
6. David Sutton. Texbook of Radiology and Imaging; Arteriography and Interventional angiography Textbook of Radiology and Imaging Chap. 25 : 711.
7. Reuter S, Redman H, Cho K. Gastrointestinal angiography Saunders monographs in clinical radiology 3rd edition. 1983; USA.
8.Margulis, Burhenne's. Alimentary tract radiology. 5th edition, Mosby Year Book, Inc. Missouri, 1994; 1 : 994.
9.Shah P, Ravi R. Role of vasculitis in the natural history of abdominal TB-evaluation by mesenteric angiography. Indian J Gastroenterology 1991; 10 : 127-30.
10.Shah P, Ravi R. Control of life threatening hemorrhage in cecal TB by transcatheter embolisation. Indian J Gastro-enterolol 1989; 9 : 149-50.
11.Transcatheter coil embolotherapy : a safe and effective option for major colonic haemorrhage AA Nicholson, DF Ettles, JE Hartley, I Curzon, PWR Lee, GS Duthie, JRT Monson Department of Radiology, Hull Royal Infirmary, East Yorkshire, University of Hull, Academic Surgical Unit, Castle Hill Hospital, Cottingham, East Yorkshire, UK. GUT 1998; 43 : 79-84.
12.Carey LS, Grace DM. The brisk bleed: controlled by arterial catheterization and gelfoam plung. J Can Assoc Radiol 1974; 25 : 113-5.
13.Gianturco C, Andersson JH, Wallace S. Mechanical devices for arterial occlusion. AJR 1975; 124 : 428-35.
14.Wallace S, Gianturco C, Andersson JH, et al. Therapeutic vascular occlusion utilizing steel coil technique: Clinical applications. AJR 1976; 127 : 381-87.
15.Chuang VP, Wallace S, Gianturco C. A new improved coil for tapered tip catheter for arterial occlusion. Radiol 1980; 135 : 507-9.
16.Arterial erosions in acute pancreatitis. Testart J, Boyet L, Perrier G, Clavier E, Peillon C. Clinique Chirurgicale, Service Pr. Watelet, CHU de Rouen, 1 rue de Germont, 76031 Acta Chir Belg 2001; 101 (5) : 232-7; discussion 237-9.
 |