Search

Gastrointestinal bleeding after surgery for cancer of the pancreas contributes significantly to patient morbidity and to patient mortality, if the bypass was performed for palliation. Bleeding after resection of the pancreatic tumor appears to be amenable to therapy, and unaltered by the addition of vagotomy at the time of surgery. Patients undergoing palliative surgery who are expected to live beyond the postoperative period may well benefit from measures to reduce the risk of gastrointestinal bleeding.

Although endoscopic sclerotherapy is effective in controlling bleeding from esophageal varices, the effects of sclerosing agents on the extrahepatic portal and splenic veins have not previously been investigated. This study of 21 men with portal hypertension and variceal bleeding compares the morphology of the portal and splenic veins in 11 who had received endoscopic sclerotherapy versus 10 patients who did not. The mean number of injections per patient was 11 +/- 5, the mean volume of 1.5 percent sodium tetradecyl injected was 23 +/- 15 ml, and the interval between the last injection and surgery was 15 +/- 6.5 days. Among the 11 patients who had endoscopic sclerotherapy, portal vein thrombosis occurred in 4 (36 percent). Two of these patients died from acute liver failure; the other two had shunt procedures. Histologic changes included intimal thickening and medial fibrosis in seven patients, thrombus in four patients, and destruction of the venous architecture in two patients. Of the 10 patients with portal hypertension who did not have endoscopic sclerotherapy, all had medial fibrosis of the portal vein, with thrombus and intimal thickening present in only 1. These findings suggest that endoscopic sclerotherapy for esophageal varices should be used cautiously in patients who may later require a shunt. Moreover, further studies are necessary to evaluate the long-term effects of injecting sclerosing agents into the portal circulation before widespread use of prophylactic sclerotherapy can be recommended.

Forty-two noninfected femoral anastomotic aneurysms that occurred after aortobifemoral bypass were repaired during a 7 year period. Twenty of these were recurrent aneurysms. In an attempt to define the cause of recurrence, a number of factors were studied including infection, suture failure, graft deterioration, defective graft manufacture, graft dilatation, and arterial degeneration. Five grafts in the primary group failed because of inherent defects in the graft structure. Two lightweight Dacron grafts showed generalized graft dilatation, and three unwrapped expanded polytetrafluoroethylene grafts had linear tears at the anastomosis. We believe that such graft degeneration does not represent a current problem because the polytetrafluoroethylene grafts now used are reinforced, and lightweight Dacron grafts are no longer made. Two primary femoral anastomotic aneurysms resulted from broken polypropylene suture. No infections were found in the primary or recurrent groups of aneurysms. Arterial degeneration adjacent to the anastomosis, consisting of a decrease in wall thickness due to loss of smooth muscle with proliferation of elastic fibers and disruption of elastic laminae, was a consistent finding in the remaining 15 primary femoral anastomotic aneurysms and all recurrent femoral anastomotic aneurysms. In many instances of primary femoral anastomotic aneurysm and in all instances of recurrent aneurysms dilatation of the Dacron graft with secondary arterial degeneration was believed to be the cause of aneurysm formation. Graft dilatation without fiber deterioration is an inherent problem in Dacron grafts and reportedly varies from 10 to 50 percent of the original graft diameter. This dilatation plus the use of initial graft diameters larger than the host artery leads to increased tension at the anastomotic site with subsequent arterial degeneration. For the aforementioned reasons, our current practice and recommendation is that repair of all anastomotic aneurysms be with expanded polytetrafluoroethylene grafts sized to approximate the host artery since its dilatation at sites of arterial pressure is minimal.