Search

A prospective randomised trial of "planned" chronic injection sclerotherapy (at three weekly intervals) versus "demand" injection sclerotherapy (after variceal bleeds only) was performed in 40 patients who had endoscopically proven variceal bleeding, and who were either graded as modified Child's grade C or were more than 65 years of age. 45% of patients survived for one year and 37% for two years. There was no significant difference in survival between the groups at either one or two years (p < 0.5; p < 0.1). In the "planned" sclerotherapy group 24 gastrointestinal bleeds occurred in 11 patients over the first nine months after study entry. In the "demand" group 35 bleeds occurred in 13 patients over the first 18 months after study entry. In this small group of "poor risk" patients, planned chronic injection sclerotherapy was not associated with improved survival.

The prognosis of patients who bleed from esophageal varices is dismal. Prophylactic treatment of the varix or the elevated portal venous pressure offers a possibility of improving the outlook for these patients. However, as only approximately one-third of patients with varices bleed during their lifetime, correct identification of high-risk patients is vital before embarking on prophylaxis. At present, neither European or Japanese selection criteria are perfect in this respect. The documented incidence of initial variceal bleeding varies between 27% and 48%, and most bleeding episodes occur within the first year after varices are diagnosed. Data from six randomized controlled trials comparing prophylactic beta-blockers with placebo demonstrated a decreased incidence of bleeding in propranolol-treated patients, which in large measure may depend on patient compliance and did not significantly affect survival in all but one study. Early randomized studies of prophylactic sclerotherapy have shown significant reductions in both the incidence of bleeding and mortality, but this promise has not been sustained by subsequent trials, and indeed sclerotherapy was detrimental in two studies. The impressive results in highly selected patients treated in Japan by prophylactic surgery are unlikely to be repeated in a Western setting, involving patient populations that consist predominantly of alcoholic cirrhotics. At present prophylaxis with beta-blockade seems to offer the best therapeutic option, but the future may lie in the development of new interventional techniques such as transjugular intrahepatic portosystemic stent shunting (TIPS) or variceal banding, and ultimately with hepatic transplantation.

In a prospective three centre study oesophageal transection and gastric devascularisation have been compared with endoscopic sclerotherapy in the long term management of bleeding oesophageal varices. Cirrhotic patients (Child's A or B grade) with documented bleeding oesophageal varices were treated initially with emergency sclerotherapy, and after five days stability, were allocated to one of the two treatment regimes. The endoscopic sclerotherapy group underwent regular sclerotherapy until variceal obliteration while those undergoing surgery were not endoscoped unless bleeding recurred, when they were treated by sclerotherapy if appropriate. Ninety two patients were eligible for analysis (68% alcoholic cirrhosis; mean age 50.1 years) and follow up was achieved for a mean of 52.5 months (range 17-83). Mortality in the first three months was greater in the oesophageal transection and gastric devascularisation group (20% v 1%) but by two years the survival curves were the same and thereafter there was no difference in mortality. Rebleeding occurred in 13/41 (31%) patients, undergoing oesophageal transection and gastric devascularisation. The costs incurred during the first year of oesophageal transection and gastric devascularisation treatment were significantly greater than with endoscopic sclerotherapy (4369 pounds v 1067 pounds, p < 0.0001) and the high rate of rebleeding in the surgical group meant that no cost savings occurred in subsequent years. It is concluded that oesophageal transection and gastric devascularisation confers no benefit over endoscopic sclerotherapy in terms of long term survival and that it is not cost effective as judged by the current health care costs in the United Kingdom.

Bleeding as a complication of liver disease can occur in the absence of recognised haemostatic defects. It is now possible to measure the concentration of endogenous heparinoid substances in the blood using a competitive binding assay. One such substance, heparan sulphate (normal range < 600 ng/ml) was assayed in the plasma of 49 patients admitted because of oesophageal varices. In 27 patients with recent upper gastrointestinal bleeding the median plasma heparan sulphate value was 1700 ng/ml (interquartile (IQ) range 900-3900) compared with 390 ng/ml (IQ range 256-800) in 22 patients with no recent bleed (p < 0.01). As heparan sulphate is metabolised by the same route as exogenous heparin, an attempt to establish a cause for the raised heparan concentrations was made by measuring the clearance of exogenous heparin in 10 portal hypertensive patients and 10 controls. The median half life of heparin in plasma in the portal hypertensive patients (25.5 minutes; IQ range 22-34) was significantly longer (p < 0.007) than the median half life in the controls (18.7 minutes; IQ range 17-21.5). Thus, there is evidence of raised concentrations of endogenous heparin like substances in portal hypertensive patients after gastrointestinal bleeding. These high concentrations may result from reduced hepatic clearance.

Successful pharmacological arrest of haemorrhage might avoid the risk of aspiration associated with tamponade and early studies have suggested that the vasoactive agent somatostatin may be as effective and perhaps safer than tamponade in controlling variceal haemorrhage. In our view, vasopressin has not established a role in management but we retain an open mind regarding the potential use of terlipressin in combination with nitroglycerin. It is unlikely that any of these agents can improve significantly our ability to control variceal haemorrhage when compared to balloon tamponade but they may reduce the incidence of pulmonary complications and thereby reduce subsequent mortality. Tamponade has proved successful in controlling acute haemorrhage from oesophageal varices in our hands. Late complications continue to give cause for concern but until effective safe alternatives to tamponade are developed, we continue to advocate its use for emergency control of acute variceal haemorrhage. Our own studies have shown that the high mortality seen in this patient population may reflect the severity of the underlying liver disease rather than failure of a management policy employing oesophageal tamponade for the initial control of acute variceal haemorrhage.

Forty endoscopically proven active variceal bleeds were entered in a prospective trial comparing oesophageal tamponade with SMS 201-995 infusion. Oesophageal tamponade controlled 19 of 20 bleeds over the first four hours and 14 of 18 bleeds over 48 hours. SMS 201-995 infusion controlled 18 of 20 bleeds over the first four hours and 10 of 20 bleeds over 48 hours (p = 0.15). No significant differences between the groups were seen in time to control of bleeding, amount of blood transfused or number of patients crossed over to the opposite treatment. Complications in the oesophageal tamponade group were discomfort due to the tube (17 patients) and chest infection (10 patients), while in the SMS 201-995 group 7 chest infections and one episode of hyperglycaemia occurred, with no symptomatic complaints. The patient survived the admission in 15 of the oesophageal tamponade bleeds and all of the SMS 201-995 bleeds (p = 0.047). An intravenous infusion of SMS 201-995 appears to have comparable efficacy to oesophageal tamponade in variceal bleeding.

Risk factors for complications after injection sclerotherapy were studied in 163 patients undergoing 667 treatments for bleeding oesophageal varices. The overall mortality rate was 7 per cent per injection sclerotherapy session; 16 per cent per acute session and 2.4 per cent per elective session. Acute variceal bleeding was controlled by injection sclerotherapy in 91 per cent of patients. Complications occurred after 16 per cent of injection sclerotherapy sessions. Deaths and complications were significantly associated with poor modified Child's grading (P less than 0.001), the first variceal bleed (P less than 0.001), acute injection sclerotherapy (P less than 0.001) and the use of the rigid oesophagoscope (P less than 0.001).

From March 1979 to April 1988 nine patients with hepatic cirrhosis have presented with acute variceal haemorrhage from gastric varices. Of six patients who underwent emergency laparotomy those with modified Child's grade C (n = 3) died within 30 days of surgery. Six patients have been followed for 2 years or longer. In the three patients who underwent under-running of gastric varices alone, two patients developed oesophageal varices at 3 and 14 months, respectively, and in the third patient gastric varices recurred 3 years after surgery. In the remaining three patients, additional left gastric vein ligation was not associated with recurrence of gastric varices or the development of oesophageal varices.

A double-blind controlled study of long-acting propranolol in the secondary prevention of variceal hemorrhage was conducted in 81 cirrhotic patients. After the index hemorrhage, all patients were treated with injection sclerotherapy on one occasion to secure hemostasis and then randomized within 72 h to propranolol or placebo therapy which was continued for 2 yr. Study endpoints were severe recurrence of variceal hemorrhage or death. Forty-two patients did not fulfill the entry criteria for the study. Thirty-eight patients received propranolol of whom 18 (47%) had further hemorrhage, 14 died, eight had side-effects (2 withdrawals), and 3 did not complete follow-up. Forty-three patients received placebo of whom 33 (77%) had further hemorrhage, 19 died, 5 had side-effects (2 withdrawals), and 5 failed to complete follow-up. The median time from onset of hemorrhage to starting drug therapy was 6 days for both groups. Life table analysis showed an equivalent incidence of further hemorrhage in both groups over the first 60 days, following which the propranolol group did consistently better than the placebo group. There was a significantly lower incidence of rebleeding in modified Child's C patients receiving propranolol (39%) than those on placebo (90%). No statistically significant effect on mortality was seen. In this study, propranolol reduced the incidence of late recurrence of variceal hemorrhage in patients with cirrhosis.

Sixty-two patients presented on 81 occasions with acute-oesophageal variceal haemorrhage. Bleeding required tamponade on or during that admission on 87 occasions and was successful in arresting haemorrhage in 93%. Employing a policy of management in which injection sclerotherapy was the main therapeutic option, control of haemorrhage was achieved on 89% of admissions with an admission mortality of 30%. These results are compared with other reported series.

Over a seven-year period, 138 patients with portal hypertension presented on 223 occasions with endoscopically proven acute variceal hemorrhage. Hemorrhage ceased spontaneously on 92 occasions (41%). On 126 occasions (57%) passage of the four-lumen modification of the Sengstaken-Blakemore tube was required, and hemorrhage was successfully controlled in 98%. Intubation was refused on five occasions (2%). Hemorrhage recurred during these 223 admissions on 47 occasions (21%); on 11 occasions a second rebleed occurred and on two occasions, a third. Tamponade was required during all of these rebleeds and arrest of hemorrhage was achieved in 87%. Hemorrhage in patients with poorer modified Child's grade was less likely to cease with intubation. The overall rate of control in the 186 episodes of hemorrhage requiring tamponade was 94%. There were 28 complications attributed to the use of tamponade in 186 episodes of hemorrhage (15%). On 12 occasions these complications proved fatal (6.4%). In four further patients failure of tamponade to control hemorrhage was fatal.

Forty eight patients who have survived admission with acute variceal haemorrhage have been readmitted and entered into a programme of chronic injection sclerotherapy to prevent recurrent variceal haemorrhage. Two hundred and nineteen elective courses of injections were performed at fibreoptic endoscopy with minor complications occurring in six patients. Twenty three patients rebled on 33 occasions but in only five instances did haemorrhage occur following variceal obliteration. With 67 per cent of patients surviving one year, this treatment option offers a satisfactory means of preventing recurrent haemorrhage when variceal obliteration is achieved.

In order to identify factors predicting survival following acute variceal haemorrhage, data were collected prospectively from 100 admissions in 70 patients managed by a standard policy employing oesophageal tamponade, injection sclerotherapy and, if necessary, oesophageal transection. Of the ten predictive factors identified by univariate analysis, only prothrombin ratio, serum creatinine and the presence of encephalopathy on admission were shown by stepwise logistic regression to have independent significance. The derived regression equation allowed clearer identification than conventional scoring systems of high and low risk groups and successfully predicted outcome in 90 per cent of admissions.

A prospective single blind controlled trial was performed to assess the efficacy of the neodymium yttrium aluminium garnet laser in the management of patients bleeding from peptic ulcers and single vessels. Over 20 months 184 patients were found at endoscopy to be bleeding from peptic ulcers and single vessels; 130 of these failed to meet the criteria for entry into the study, but their bleeding stopped with conservative management. Forty-five patients entered the study: in all 25 who were bleeding from ulcers with spots the bleeding stopped irrespective of the treatment allocated. Twenty patients were bleeding from arteries. Eight were allocated to placebo treatment, and all later underwent emergency surgery for further haemorrhage. Twelve were allocated to receive laser treatment, and only one of eight who actually received it required surgery (p less than 0.002). It is concluded that laser treatment reduces the incidence of further haemorrhage and the need for emergency surgery, but the technique is difficult and not always applicable.