Esophageal variceal sclerotherapy using 3% phenol in glycerine.
Endoscopic variceal sclerotherapy (EST) has been proved to be effective in the control of acute bleeding and prevention of recurrent bleeding from esophageal varices in patients with portal hypertension. A variety of sclerosants are available for the purpose. However, these are expensive and not freely available in many developing countries. Phenol in glycerine (active agent phenol) has been used effectively as a sclerosant for peripheral varicose veins. Hansen used phenol in arachis oil for EST and reported good results. We are presenting our results of EST using 3% phenol in glycerine for the control of variceal bleeding.
Endoscopic variceal sclerotherapy (EST) was carried out using 3% phenol in glycerine in 50 patients (37 males, 13 females, aged 10-60 years) with portal hypertension of varying etiology, who were referred to the endoscopy section of the Departments of Surgery and Gastro enterology between September 1983 and August 1984. Seventeen patients underwent emergency EST during primary acute variceal bleeding and an additional 7 cases during a rebleed. (These 7 patients originally belonged to elective EST group). Elective EST was done in 25 patients who were diagnosed to have variceal bleeding in the past; 8 patients underwent prophylactic EST [Table 1]. All patients with active hematemesis, were rescucitated before emergency diagnostic endoscopy. If found to be bleeding from esophageal varices, immediate EST was carried out to control the bleeding. Before elective EST, a thorough clinical and biochemical evaluation and liver biopsy were done. A splenoportogram was also done before EST and repeated afterwards where ver possible to demonstrate obliteration of esophageal varices.
No sedation was given. Topical pharyngeal anaesthesia with 4% xylocain liquid was used. Endoscopy was performed using a fibreoptic forward viewing gastroscope (Olympus GIF D3) and the extent of esophageal varices was recorded as grades 1 to 4. About 3-7 ml of sterile 3% phenol in glycerine (prepared in the hospital pharmacy) was injected intravariceally into each column just about the cardio-esophageal junction and about 5 cm. proximally if required. A total of 15 to 40 ml of sclerosant was injected at each session. No accessories to compress the varices were used during or after EST.
Being viscid, phenol in glycerine solution cannot flow through standard variceal injectors. Hence an indigenously designed injector, consisting of a polythene tube fitted with an 18 gauge needle protruding about 6 mm beyond its distal end was used [Fig. 1].
EST was repeated at intervals of 1 and 3 weeks. Follow up endoscopy was carried out at intervals of 1 and 3 months and disappearance of old varices or appearance of new varices was recorded. Further sclerotherapy was done if there was either apparance of new varix or a history of rebleeding.
An autopsy was performed on patients who died following EST and histopathological changes in injected varices and surrounding esophageal walls were studied.
Control of Active Bleeding
Emergency EST was successful in controlling primary active variceal bleeding in 15 of 17 patients (88.2%) and rebleed in 6 of 7 patients (85.7%), giving an overall success rate of 87.4%. All the patients with unsuccessful EST were cirrhotics in Childs C class.
Rebleeding was observed in 12 of 45 patients (26.6%) who were followed up [Table 2]. Seven of these required emergency EST to control the bleeding. Three patients had self limiting rebleed, not requiring hospitalization. Two patients had rebleed following an attack of hepatic coma, and were treated conservatively. One patient died due to documented uncontrolled variceal rebleed and two others due to rebleed with hepatocellular failure. Over 90% of the patients had rebleeding before completion of the initial 3 EST sessions [Table 2].
Pre and Post EST Splenoportogram (SPG)
In 3 patients in whom SPG was done before and after EST, the esophageal varices were found to have been obliterated.
Two patients developed self limiting mild fever ranging from 37.4-380C and retrosternal pain lasting from 24-48 hours. Clinical examination and skiagram of the chest did not reveal any abnormality. One other patient developed pulmonary embolism followed by right pleural effusion. A barium swallow did not reveal esophageal perforation. He recovered with conservative therapy. No patient was found to have esophageal mucosal ulcerations at the time of follow up.
Five out of 17 patients in emergency group died during hospital admission, giving an overall in-hospital mortality of 29.4%. The cause of death was persistent variceal haemorrhage in 3 and progressive hepatocellular failure in 2 patients. There was no early mortality in elective and prophylactic group.
Of 45 surviving patients, 10 died during follow up (1-30 months). The cause of death was variceal bleeding (1 case) variceal bleeding with liver cell failure (2), gastric erosion (1), hepatocellular failure (2) and elective post shunt surgery (4). All six patients who died due to causes other than surgery were cirrhotics in Child's B and C category.
Autopsy was performed in two patients who died within 1-2 weeks of EST. Histopathological examination of the injected varices revealed the presence of fresh thrombus with intimal damage [Fig. 2].
Long Term Follow Up
Two patients are lost to follow up. Thirty-three (94.3%) out of 35 patients are alive for a period ranging from 12-30 months. Of these, 24 patients are cirrhotics (7 Child's A, 4 Child's B and 13 Child's C), 7 patients of extra-hepatic obstruction and 2 of non-cirrhotic portal fibrosis [Table 3]. All the patients are on 3 monthly follow up with endoscopic examination and EST if necessary.
Three per cent phenol in glycerine was found to be an effective sclerosant in controlling acute variceal bleeding in 88.3% of patients. These figures are comparable with the reported results of 70-100% control using other sclerosing agents.,,
The recurrent bleeding rate of 26.6% in our series is comparable with the figures reported by others.,,,,,,,, A majority of rebleeds (91.66%) were observed during the initial treatment period in patients who received less than 3 injection courses. Similar high rates of rebleeding (42 to 52%) during initial treatment period have been reported by several other workers as well.,,, These rebleeding episodes could easily be controlled with further EST and some of them were self limiting.
There were no fatal complications related to the procedure or sclerosant. Only one patient developed a major complication in the form of pulmonary embolism but recovered completely. None of our patients was found to have mucosal ulceration on follow up endoscopy at 1 and 3 weeks' interval. Hansen et al have also reported absence of deep esophageal mucosal ulceration on submucosal injection of 5% phenol in arachis oil. Goodale et al reported a 32% incidence of ulceration with sodium morrhuate while Westaby et al noticed mucosal ulceration in 33% of patients with ethanolamine oleate. Absolute alcohol was found to cause superficial mucosal ulceration in 61% and deep ulcer with necrotic base in 77% of patients. Though mucosal ulcerations are considered to be harmless by some, others have reported severe bleeding from these ulcers.,
The overall early mortality of 29.4% in patients with acute variceal bleeding is comparable with the 30% mortality reported by Soehendra et al and superior to the over 50% mortality reported by others.,, All 5 patients, who died were cirrhotics in Child's C category. The absence of early mortality in the elective and prophylactic EST group indicates that EST with 3% phenol in glycerine is safe.
The late mortality of 20.45% in our series compares well with other reported series.,,,,,,,, Only 2.2% of deaths were due to recurrent variceal bleeding.
Opinions differ regarding the use of tamponade during or after injection, and various technique have been advocated.,, The purpose of tamponade is to produce stasis and thereby prolong the contact of sclerosant with the vessel wall. However, Borsoum et al by mixing radiographic contrast with the sclerosant, have shown that the injected material is cleared so rapidly into the external esophageal veins via the perforating veins that the value of compression was considered doubtful. Sivak et al have therefore advocated a slow steady injection over a period of 10 to 13 seconds. Being viscid, 3% phenol in glycerine has to be necessarily injected slowly and does not get washed away easily into the paraesophageal veins thereby producing maximum local sclerosang effect. We have further confirmed this by injecting 3% phenol in glycerine and contrast mixture under radiological control (unpublished data).
We recommend direct intravariceal injections as our earlier experimental study in dogs revealed that paravenous injection does not produce obliteration of the veins.
Regarding the mechanism of action of the sclerosant, the present autopsy findings are well supported by our earlier experimental findings. The histopathological examination of injected varices revealed presence of intra-luminal thrombus with intimal damage and sclerosis of media. These findings are similar to those reported by Johnson.
Role of prophylactic EST remains controversial. Significantly lower bleeding episodes and death have been reported in patients who were subjected to variceal injection prophylactically than in patients who did not undergo EST.,,
In conclusion, we would like to recommend use of 3% phenol in glycerine as a sclerosant for EST, as it is effective, safe and economical. One session costs less than a rupee and can be prepared in local hospital pharmacy.