Gastro-colic fistulae are described as presenting with the clinical triad of diarrhea, nausea/vomiting and weight loss . However, all three features are said to occur in only 30% of patients. Other symptoms include malnutrition with cachexia, anemia, abdominal pain and fecal halitosis that is present in over 50% of patients [1, 2].
Malignant gastro-colic fistulae were first described in 1755 by Haller . Gastro-colic fistulae due to benign peptic ulcer disease were described by Firth in 1920 . Gastrointestinal malignant disease is the predominant cause today: colonic adenocarcinoma in the Western world, gastric carcinoma predominating in Japan [2, 5]. Other malignant causes include gastric lymphoma, carcinoid tumors of the colon and locally invasive malignant tumors of the biliary tree, pancreas and duodenum . Benign causes described include peptic ulcer, gastric tuberculosis, trauma, syphilis, retroperitoneal sarcoma, Crohn's disease and pancreatitis [2, 3].
The overall incidence of gastro-colic fistula has decreased since the advent of effective medical management of gastric ulcer disease. Post-surgical-resection-associated fistulae and fistulae related to the use of non-steroidal anti-inflammatory medications were the most reported causes of benign gastro-colic fistulae [2, 4, 6]. In a single case series from 1955, prior to the advent of H2 antagonists and proton pump inhibitors, it was reported that up to 10% of patients post-gastrectomy for benign gastric ulcer subsequently developed a gastro-colic fistula . Fistulae in gastric ulcer disease in the setting of proton pump inhibitor use are exceedingly rare and to the best of our knowledge this is the first documented case.
A barium enema is the radiological modality of choice for diagnosis of gastro-colic fistulae, with specificity of 90-100% compared with a barium meal that has a false negative rate of 30-70% [1, 3]. Endoscopic investigations are recommended to exclude malignant disease. Computed tomography (CT) has not been evaluated for sensitivity and specificity but has been reported in one case series as a useful adjunct in diagnosis and staging.
The treatment of choice for a gastro-colic fistula is en bloc surgical resection of the fistula tract with a margin of adjacent tissue [1, 3, 4, 8]. This allows disease free margins in malignant disease and decreases the recurrence rate in benign disease, which has been reported to be up to 12%. The recurrence rate is higher if simple excision of the fistula tract is used for initial management .
Several cases of medical or minimally invasive management of gastro-colic fistulae have been described and may be suitable where malignant disease has been excluded and/or surgical intervention is not appropriate. Endoscopic injection of the fistula tract with fibrin has shown to be effective in several case reports .
Prognosis for gastro-colic fistula has been thought to be quite poor. Between 1963 and 1994, the longest recorded survival post-resection for gastro-colic fistula due to malignant disease was nine to ten years [1, 5]. Post-operative mortality has been reported to be as high as 25%, presumably due to co-morbidity and de-conditioning of the patients .
One case series of six patients reported one post-operative death due to underlying co-morbid conditions. The remaining cases were followed for a mean of 66 months, with one further death due to an unrelated underlying co-morbid condition . However, there have been very few recent studies and advances in surgical techniques and post-operative care as well as nutritional optimization suggest empirically that prognosis may have improved.