On initial examination, the patient was found to be orthostatic but in no apparent distress. Her abdomen was benign but rectal examination was notable for black, guaiac-positive stool. There were no stigmata of chronic liver disease. Initial nasogastric lavage revealed “coffee-ground” material that cleared after 500 cc irrigation with normal saline.

Laboratory testing revealed the following: hemoglobin, 8.7 g/dL; hematocrit, 25.3%; mean corpuscular volume, 91.5 fL; platelet count, 155 × 10³/microliters; blood urea nitrogen, 49 g/dL; mildly elevated serum aspartate aminotransferase at 47 U/L; and low albumin of 2.8 g/dL. The remainder of the liver function tests, serum chemistries, and coagulation profile were all within normal limits.

Readers are encouraged to respond to George Lundberg, MD, Editor of MedGenMed, for the editor's eye only or for possible publication via email: glundberg/at/medscape.net

After volume resuscitation and blood transfusion, urgent upper endoscopy was performed. The stomach and duodenum were clear of either fresh or old blood. There was an isolated cluster of gastric varices within the fundus, appreciated only on retroflexion. A red, flat spot was noted on one of the suspected varices. A single hemoclip was placed over this red spot and the varix was then noted to bleed actively (Figure 1), filling the cardia with blood. Soon thereafter, the bleeding ceased spontaneously. Given the findings of gastric varices, a computed tomography (CT) scan was obtained demonstrating patency of the splenic vein, and normal appearing liver, spleen, and pancreas. Gastric varices in the fundus were confirmed with involvement of the left renal vein (Figure 2).

Figure 1 Gastric varix with hemostasis clip applied, retroflexed view.

Figure 2 Abdominal CT demonstrating gastric varices.

The patient subsequently remained hemodynamically stable without further signs of gastrointestinal hemorrhage. After obtaining informed consent, repeat upper endoscopy was performed with the intention of variceal obliteration therapy with glue. On repeat endoscopy, a few columns of grade 1 nonbleeding esophageal varices were appreciated. Once again, multiple, large nonbleeding gastric varices were noted in the fundus, one of which showed blood staining. Injection of 2-octyl cyanoacrylate (Dermabond; ETHICON Products; Somerville, New Jersey; off-label indication) was achieved through use of an endoscopic needle with administration of 1 cc of a 1:1 dilution of cyanoacrylate mixed with the fat-soluble radiographic contrast agent, ethiodol. A 1-cc ethiodol flush was administered immediately following each injection (Figure 3). A total of 5 injections was applied to the cluster of varices. Subsequent manipulation of the gastric variceal bed revealed significant hardening consistent with successful glue obliteration (Figure 4; Video). The patient tolerated the procedure well and there were no immediate complications. A subsequent chest radiograph demonstrated a large collection of ethiodol within the variceal bed in the gastric fundus, but also some very small fragments within the pulmonary vasculature (Figure 5).

Figure 3 Injection of 2-octyl cyanoacrylate.

Figure 4 Successful obliteration with hardening of variceal bed.

Video Injection of 2-octyl cyanoacrylate. Click here to view.

Figure 5 Postprocedure chest radiograph.

The patient did not demonstrate any signs of pulmonary compromise during the remainder of her hospitalization and has not had any further episodes of gastrointestinal bleeding in the months since her procedure.

This patient's presentation with significant gastric variceal hemorrhage was striking given the absence of preexisting liver disease. In reviewing her history, it is possible that she may have developed occult portal hypertension as a result of methotrexate-induced hepatotoxicity, perhaps explaining her hypoalbuminemia. She had not had previous liver biopsy and prior liver function tests were unavailable because she was visiting from another country. Certainly further investigation as to the etiology of this patient's varices will be required, including performance of liver biopsy with possible portal-pressure gradient measurement. However, the greatest concern during her current hospitalization was the initial management of the bleeding gastric varices.

Management of gastric varices often depends on the nature of the underlying etiology. Pathogenesis can be related either to portal hypertension or isolated left-sided (sinistral) portal hypertension from splenic vein pathology. The former can be related to any condition leading to chronic liver disease whereas splenic vein thrombosis or stenosis most often can be attributed to underlying pancreatic disease. Irrespective of etiology, the risk of hemorrhage from a gastric varix is strongly associated with anatomic location, the highest risk being reported with gastric fundal varices.[1] In addition, long-term data have indicated that bleeding from gastric varices tends to be more severe than bleeding from esophageal varices.[1] For immediate control of massive hemorrhage, the 600-cc Linton-Nachlas tube has been shown to have greater efficacy than the 200 cc Sengstaken-Blakemore tube in achieving immediate hemostasis for gastric varices.[2,3] Intended only as a temporizing measure, this intervention should be regarded as a bridge to more definitive therapy. In the case of splenic vein thrombosis, this would include either splenectomy or splenic artery embolization. In the case of underlying portal hypertension, therapeutic modalities include transjugular intrahepatic portosystemic shunting, balloon-occluded retrograde transvenous obliteration, or endoscopic therapy. In recent years, endoscopic therapy, primarily cyanoacrylate glue-injection, has gained increasing international acceptance as the primary intervention for acute gastric variceal hemorrhage.

The first reported experience with the use of cyanoacrylate compounds for the control of bleeding from gastric varices was described in 1975 with an experimental dog model.[4] Good results were achieved using a percutaneous transjugular approach. Subsequent human studies, again using percutaneous catheter obliteration, demonstrated a favorable efficacy profile for cyanoacrylate compounds when compared with absorbable gelatin, thrombin, and monethanol aminoleate.[5] This favorable profile was further confirmed in experimental dog models treated with endoscopic injection therapy.[6] Experience with endoscopic sclerotherapy proved to be less promising, with unacceptably high rates of rebleeding.[7,8] Although band ligation therapy has shown benefit in treating esophageal varices, similar success has not been demonstrated in the management of gastric varices. In a randomized head-to-head trial involving cirrhotic patients with bleeding gastric varices, endoscopic obturation using cyanoacrylate was found to have better rates of initial hemostasis, a lower incidence of rebleeding, and an overall mortality benefit when compared with band ligation therapy.[9]

Subsequent long-term data have further confirmed both the safety and efficacy of cyanoacrylate injection therapy for management of bleeding gastric varices.[10] Nearly all of the previous investigations have worked with N-butyl cyanoacrylate (Histoacryl; B. Braun; Melsungen, Germany); however, the compound 2-octyl cyanoacrylate (Dermabond; ETHICON Products; Somerville, New Jersey) has recently received FDA approval for skin-closure in the United States. The cyanoacrylate compounds are adhesives formed from the reaction of formaldehyde and alkyl cyanoacetate, that produce a polymer. This polymer is then heated to form a monomer that upon exposure to water or base, undergoes an exothermic hydroxylation reaction, leading to polymerization. The 2-octyl cyanoacrylate derivative with its 8-carbon alkyl side-chain leads to slower degradation than seen with the original N-butyl cyanoacrylate compound. This affords greater tensile strength and results in fewer cyanoacrylate break-down products in neighboring tissues, thus reducing inflammation. A recent pilot study examining treatment of gastric fundal varices with 2-octyl cyanoacrylate demonstrated initial hemostasis in 100% of patients and rebleeding in only 4% after mean follow-up of 11 months.[11] The approach used in our patient is similar to that outlined by Seewald and colleagues,[12] with careful attention paid to dilution of the cyanoacrylate with contrast material to prevent polymerization within the catheter. The administration of a single cc per injection helps to minimize the risk of embolization.[13] Use of a steel needle is required to avoid polymerization of the glue within the needle itself. Rapid flush after each injection also serves to prevent the needle from becoming lodged within the varix.

Although rare, reported complications of cyanoacrylate injection therapy have included fever, sepsis, pulmonary embolus, retroperitoneal abscess, splenic infarction, as well as portal and splenic vein thrombosis.[14–17] Some experts currently recommend repeat endoscopy every 3 to 6 months to assess for long-term prevention of rebleeding.[12] Our patient, who resides overseas, is now undergoing follow-up evaluation for rebleeding risk as well as investigation regarding the nature of her newly diagnosed portal hypertension.