Prospective controlled trial of elective endoscopic sclerotherapy in comparison with percutaneous transhepatic obliteration of esophageal varices in patients with nonalcoholic cirrhosis

GASTROENTEROLOGY

1987;93:1205-9

Prospective Controlled Trial of Elective Endoscopic Sclerotherapy in Comparison With Percutaneous Transhepatic Obliteration of Esophageal Varices in Patients With Nonalcoholic Cirrhosis HIDETAKA TERABAYASHI, KUNIHIKO OHNISHI, TAKAFUMI TSUNODA, HISASHI NAKATA, MASAYUKI HIDE0 TANAKA, SHINJI IIDA, FUMIO NOMURA, and KUNIO OKUDA First

Department

of Medicine,

Chiba

University

The results of a prospective randomized controlled trial of elective endoscopic intravariceal sclerotherapy carried out over a 36-mo period in comparison with elective percutaneous transhepatic obliteration of varices (PTO) are presented. Sixty-six patients with nonalcoholic cirrhosis were randomized after they had stabilized, usually between 7 and 14 days after variceal bleeding had stopped following medical treatment (balloon tamponade and vasopressin infusion). Thirty-three patients were assigned to the sclerotherapy group and the other 33 patients were assigned to the PTO group. The mean follow-up period was similar in both groups. There was no significant difference in demographic, clinical, and laboratory data between the two groups. Six patients (18%) in the sclerotherapy group and 21 (64%) in the PTO group had at least one episode of gastrointestinal bleeding during the follow-up period (p < 0.005). Three patients in the sclerotherapy group and 1 patient in the PTO group bled from lesions other than varices; therefore the incidence of variceal bleeding was 9% in the former and 61% in the latter (p < 0.005). The cumulative variceal bleeding rate was significantly lower in the sclerotherapy group than the PTO group (p < 0.05). Five patients in the sclerotherapy group died during the

Received December 2, 1986. Accepted June 8, 1987. Address requests for reprints to: Kunihiko Ohnishi, M.D., First Department of Medicine, Chiba University School of Medicine, Chiba 280, Japan. This work was supported by a research grant from the Intractable Diseases Division, Public Health Bureau, and Educational Ministry grants 60570311 and 61480186. 0 1987 by the American Gastroenterological Association 0016-5085/87/$3.50

School

of Medicine,

Chiba,

SAITO,

Japan

follow-up period but none died of recurrent variceal bleeding. Nineteen patients in the PTO group died and 10 of them died of bleeding from varices. The cumulative survival rate was significantly better in the sclerotherapy group (p < 0.05). These results indicate that elective endoscopic intravariceal sclerotherapy is superior to elective PTO in the prevention of recurrent variceal hemorrhage and mortality in nonalcoholic cirrhosis. The benefits of endoscopic sclerotherapy (ES) in reducing the frequency of subsequent rebleeding from esophageal varices were demonstrated in four controlled trials in which most patients had alcoholic cirrhosis (l-4). However, there has been no consensus as to whether ES improves long-term survival. No controlled trials of elective ES have been conducted in nonalcoholic cirrhosis. We evaluated the effects of elective ES in comparison with elective percutaneous transhepatic obliteration of esophageal varices (PTO) (4-6) by a prospective randomized controlled study.

Patients and Methods From January 1983 to April 1986, 66 patients with biopsy-proven nonalcoholic cirrhosis (presumably as a sequela to chronic viral infection) entered a randomized controlled trial aimed at investigating the efficacy and safety of ES in comparison with PTO in the elective treatment of hemorrhage from esophageal varices. All patients gave written informed consent. During the same period 5 patients with alcoholic cirrhosis with variceal Abbreviations used apy; PTO, percutaneous

in this paper: transhepatic

ES, endoscopic obliteration.

sclerother-

GASTROENTEROLOGY Vol. 93, No. 6

1206 TERABAYASHI ET AL.

Table

Age W

1.

Demographic and Clinical Characteristics Treatment Groups at Entry

Sex (M:F) Child’s grade A (%I B (%) c (%I

of

PTO (n = 33)

Sclerotherapy (n = 33)

55.4 + 11.0 28:5

53.1 -t 8.4 29:4

9 (271

11 (33) 15 (46) 7 (211

16 (49) 6 (24)

Complications at admission Hepatocellular carcinoma (%) 5 (15.2) Ascites (%) 11 (33.3) Encephalopathy (%) 5 (15.2) Blood chemistry 1.9 t 1.6 Total bilirubin (0.2-0.8 mg/dJ)” Albumin (3.5-5.0 gldl)” 3.3 t 0.4 Alanine transaminase (O-40 IUIL)a 86 + 56

7 (21.2) 13 (39.4) 4 (12.1) 1.7 ” 0.9 3.3 ‘- 0.4 84 ? 59

PTO, percutaneous transhepatic obliteration. Values are mean 2 SD unless otherwise specified. ’ Normal values.

bleeding, who were habitual drinkers, were seen but were not included in the study. All patients had had at least one episode of bleeding from ruptured esophageal varices, documented by emergent endoscopy, and had received conventional medical treatment (tamponade by the Sengstaken-Blakemore tube and vasopressin), which achieved hemostasis until the elective treatment. Two patients who continued bleeding in spite of medication and 2 patients who suffered from early rebleeding between the first episode and the time of randomization were not included in this trial; the former z patients were treated by surgery that consisted of esophageal mucosal transection, splenectomy, devascularization of the left and short gastric and paraesophageal veins, and pyloroplasty (7). One patient died soon after; the other is alive. The latter 2 patients were treated by ES and they are alive. Randomization was carried out between 7 and 14 days after cessation of hemorrhage when the patient had become stabilized. They were randomly allocated by sealed envelope to the ES group (n = 33) or the PTO group (n = 33). Clinical and biochemical data of each group at entry are shown in Table 1.Patients were classified according to the modification of the Child’s grading (8) and they were classified as Child’s C patients on the basis of the status at entry, with the designation made according to the single worst criterion (serum albumin 53 g/dl, total bilirubin 23.0 mg/dl, ascites, or hepatic encephalopathy). In the PTO group, portal vein catheterization was carried out without general anesthesia as described previously (6). Portography, splenic venography made at the splenic hilum, and superior mesenteric venography were performed to demonstrate the variceal vessels. The variceal vessels were then catheterized selectively, and they were obliterated by an average of three stainless steel coils (MWCE-38-4-3, 5-5, 5-8; Cook Co., Bloomington, Ind.) followed by 5-10 ml of 99.9% ethyl alcohol. The efficacy of cessation of variceal flow was assessed by postobliteration venography. Obliteration was complete in 32 cases and incomplete in 1 case.

Endoscopic sclerotherapy was performed using an Olympus EF-B3 (Olympus Corporation of Japan, Tokyo, Japan), to which an oral side balloon made of Teflon was attached, a 23-gauge needle (Olympus, MN-SL), and an anal side balloon as described by Takase et al. (9). Under sedation with minor tranquilizer and analgesics, the endoscope was inserted into the esophagus and the oral side balloon on the scope was filled with 15 ml of air to stop the variceal blood flow into the azygos vein. The needle was inserted into the varices and 3-20 ml of 5% ethanolamine oleate in 36% Urographin (30% amidotrizoate) was injected into each varix, which was confirmed under fluoroscopy; a maximum of 40 ml was injected into three to four varices for each treatment session. After removal of the endoscope, the injection site was compressed for 2 h with the anal side balloon inflated with 15 ml of 76% urographin. Endoscopic sclerotherapy was performed at weekly intervals until all varices became barely visible or disappeared (10). On average, 2.4 treatment sessions were needed to eradicate the varices. After PTO or ES, patients were followed at the outpatient clinic and rebleeding was monitored. In the PTO group, if recurrent bleeding occurred and if the patient was a suitable candidate, surgery (7) was performed. Patients belonging to Child’s B and C grades and those who refused surgery continued to receive PTO and/or the standard medical therapy which included volume replacement, blood transfusion, intravenous vasopressin, and Sengstaken-Blakemore tube tamponade. In the ES group, when rebleeding occurred, patients were treated by ES or the standard medical therapy, or both. Otherwise, they were not treated even if newly developed varices were detected by periodic endoscopic examination (3-6-mo interval). The control trial was designed to terminate when the probability value of the difference of the cumulative survival between the sclerotherapy and PTO groups reached cO.05, assuming a significant criterion of 0.05. Statistics Results are given as mean + SD. Comparisons of two groups were made by the Student’s t-test and ,$ test with Yates’ correction. The cumulative variceal rebleeding rate was calculated by the modified life table method (11) from the time of the elective treatment. The cumulative survival rate was calculated by the Kaplan-Meier method (12) from the time of the elective treatment. Statistical analyses were made by the Mantel-Haenszel method (13).

Results Both

the ES and PTO

groups

were

similar

in

age, sex ratio, and type of underlying cirrhosis. The severity of liver disease and the frequency of complications such as hepatocellular carcinoma history of ascites and hepatic encephalopathy

and of were

also equally distributed among both groups (Table 1). The mean follow-up period was similar in both groups [ES, 18.3 * 9.7 mo; PTO, 13.8 t 10.2 mo).

December

1987

ENDOSCOPIC

SCLEROTHERAPY

VERSUS

PTO

1207

In the ES group, 3 patients had recurrent variceal hemorrhage and 3 had nonvariceal bleeding in the gastrointestinal tract (1 due to hemorrhagic gastritis, 2 at an undetermined site). Variceal rebleeding was treated successfully by ES in all 3 patients. Nonvariceal bleeding was controlled conservatively in all 3 patients, but 1 patient bled to death.

0

6

12

16

Follow- up p(lriod

Figure

24

30

Mortality 36

(months)

1 Efficacy of endoscopic sclerotherapy for the prevention of variceal rebleeding in comparison with transhepatic obliteration of esophageal varices. The cumulative rebleeding rate was significantly less in the former (p < 0.05).

Recurrent

Bleeding

Twenty-one (64%) of the 33 patients treated by PTO and 6 (18%) of the 33 patients treated by ES had suffered from at least one episode of gastrointestinal bleeding during the follow-up period (p < 0.005). One patient in the PTO group and 3 patients in the ES group bled from lesions other than varices, and thus the incidence of variceal rebleeding was 20 of 33 (61%) in the former group and 3 of 33 (9%) in the latter group (p < 0.005). The cumulative variceal bleeding rates for the PTO group at 1, 6, 12, 24, and 36 mo were 25%, 62% 62%, 77%, and 77%, respectively; corresponding rates were 0% 0% lOok, 21% and 21% for the ES group (p < 0.05) (Figure 1). Procedures

for Recurrent

Hemorrhage

In the PTO group, 20 patients had a recurrent variceal hemorrhage and 1 patient bled from an undetermined site in the gastrointestinal tract. Variceal rebleeding was controlled by conventional treatment in 10 patients, by a second PTO in 9, and by emergency surgery (7) in 1. Gastrointestinal bleeding was controlled conservatively in 1 patient. Six patients died from bleeding despite the conventional treatment, and 2 of 4 patients in whom recurrent variceal hemorrhage was stopped successfully by the conventional treatment underwent surgery (7). One of 9 patients in whom a second PTO was performed for recurrent variceal hemorrhage succumbed to continuing bleeding. Five of 8 patients in whom recurrent variceal hemorrhage was treated successfully by a second PTO bled again; 3 bled from varices and 2 bled from an undetermined site in the gastrointestinal tract. These 5 patients did not respond to the conventional treatment and died. One of 3 patients who did not bleed again after a second PTO underwent surgery (7).

In the PTO group, 19 patients died during the follow-up period: 7 of Child’s B and 3 of Child’s C grade died from variceal hemorrhage, 2 of Child’s C grade died from gastrointestinal bleeding of undetermined origin, and 4 of Child’s B and 3 of Child’s C grade died from hepatic failure. Of 7 patients who died of hepatic failure, 5 did not rebleed after the first PTO; 3 were Child’s B grade (2 had complicating hepatocellular carcinoma) and 2 were Child’s C grade at the time of entry. Fourteen patients survived during the follow-up period: 9 were Child’s A and 5 were Child’s B grade. Five of the Child’s A patients underwent surgery (7) electively; 2 patients were operated on after the first PTO, 3 after recurrent variceal hemorrhage had been stopped successfully (by a second PTO in 1 and by conservative treatment in 2). These 5 patients survived during the follow-up period. In contrast, one Child’s B patient who underwent emergency surgery (7) for recurrent variceal hemorrhage died of hepatic failure after 4 mo of hospitalization despite cessation of bleeding. In the ES group, 5 patients died during the follow-up period; the cause of death was hepatic failure with hepatocellular carcinoma in a Child’s A case, gastrointestinal bleeding of undetermined origin in another Child’s A case, and hepatic failure in 2 Child’s B and 1 Child’s C cases. Of 4 patients who died without recurrent gastrointestinal bleeding dur-

I 0

6

12 Follow-up

Figure

1

I

16

24

period

I 30

36

(month.)

2. Survival rate from the start of the elective treatment in the endoscopic sclerotherapy group and in the group treated by transhepatic obliteration of varices. The cumulative survival rate was significantly better in the former (p < 0.05).

1208

TERABAYASHI

ET AL.

ing the follow-up period, 3 had complicating hepatocellular carcinoma at entry. The cumulative survival rate in the PTO group at 1, 6,12, 24,and 36 mo was 82%, 66%, 54%, 22%, and ZZ%, respectively, and 97%, 92%, 92%, 77%, and 77% in the ES group (p < 0.05) (Figure 2).

Complications Transient fever higher than 38°C and precordial pain occurred in 10% and 8% respectively, of 42 PTO procedures. Esophageal ulcer developed in 58% of 75 injection courses in the ES group. Disturbances in swallowing continuing for >l wk occurred in 27% of injection courses. Esophageal dilatation using a Celestin dilator was required in 2 of 4 patients who developed dysphagia due to stricture. Microscopic hematuria, hemoglobinuria, pleural effusion, and transient fever higher than 38°C occurred in 13% 4%, 3%, and 21% of injection courses, respectively.

Discussion In Japan, esophageal varices had been treated by surgery (7) or by a medical treatment that included PTO, depending on the condition of the patient and the stage of underlying cirrhosis, up until 1982 when Takase et al. (9) reported their results of endoscopic intravariceal sclerotherapy using a flexible fiberscope, although ES had already become popular in some parts of Europe (14-17)and South Africa (18).In January 1983, we began a prospective randomized controlled trial of ES versus PTO as an elective treatment of esophageal varices in patients with nonalcoholic cirrhosis and with recent variceal hemorrhage. The reason for using PTO as the control was the following: PTO was the only nonsurgical technique available to us and we had already demonstrated that PTO was superior to conventional medical treatments (Sengstaken-Blakemore tube tamponade and vasopressin infusion) in preventing recurrent variceal bleeding (6). We randomized patients when they were stabilized after cessation of variceal bleeding by the conventional medical treatment rather than when they were actively bleeding because only a small proportion of the patients with acute bleeding would be sent to our unit. Most of them were sent to our affiliated hospitals, where only emergency endoscopy and conventional medical treatment could be done, and some time later elective ES or PTO could be performed by our team in these hospitals (6 patients) or in our own unit when beds became available (60 patients). No patient was lost during the follow-up period. In the present study, it was clearly demonstrated

CASTROENTEKOLOGY

Vol. 93, No. 6

that ES significantly reduced the frequency of variceal rebleeding and improved survival rates compared with PTO. During the follow-up period, 19 patients who had undergone PTO died, 10 of recurrent variceal bleeding. In contrast, 5 patients who had had ES died, none of recurrent variceal bleeding. These results indicate that ES reduced mortality related to rebleeding. This result may be related to the different therapeutic endpoints for esophageal varices between the ES and PTO groups. With ES, the weekly injection schedule was carried out with the therapeutic endpoint of obliteration of varices. In contrast, PTO was performed as a single treatment with the therapeutic endpoint of cessation of variceal flow from left or short gastric veins, or both, because we had demonstrated previously that enlarged, tortuous veins arising from thin parallel vessels at the esophagogastric junctions corresponded to varices on endoscopy in both size and course (19). However, varices did not disappear endoscopically in any of the patients after PTO (unpublished observation), suggesting that the obliteration of varices was not complete with the single PTO and that, with time, increased collateral flow led to rebleeding in the remaining or newly developed varices. The survival rate was also related to the Child’s classification of the patient on entry. In the PTO group, all 9 Child’s A patients, 5 of whom B patients had surgery (7), and 5 of 16 Child’s survived, but all Child’s C patients died. In the ES group, in contrast, 9 of 11 Child’s A, 13 of 15 Child’s B, and 6 of 7 Child’s C patients survived. These results indicate that ES prolonged survival in Child’s B and C patients but not in Child’s A. In advanced cirrhosis such as Child’s grade B or C, sinusoidal flow begins to diminish (20) and gastrointestinal bleeding worsens hepatic function significantly by further reducing sinusoidal flow. It remains to be determined whether surgery improves survival of Child’s A patients treated by PTO. Our results in the ES group were very similar to those reported by Takase et al. (9) from Japan. However, variceal rebleeding seemed to be less frequent and survival better in the present study compared with other controlled trials from western countries (2-5) and South Africa (1).These differences are likely due to the difference in the time of entry of patients to the study, condition of patients (our patients did not drink during the follow-up period), and etiology of cirrhosis. As a complication of ES, esophageal ulcer was the most common, as has been reported by others (1,2,4,9,17); gastrointestinal bleeding other than varices occurred in 3 patients in the ES group, similar in incidence to variceal rebleeding. Thus, prevention of esophageal ulcer and gastrointestinal

December

1987

bleeding other than varices is a problem, to be solved in the future by the use of different sclerosing agents and cytoprotective drugs. It was concluded that elective endoscopic intravariceal sclerotherapy would seem to be superior to elective PTO in reducing variceal rebleeding and mortality in patients with nonalcoholic cirrhosis (Child’s B and C grade).

References 1. Terblanche

2.

3.

4.

5.

6.

7.

J, Bornman P, Kahn D, et al. Failure of repeated injection sclerotherapy to improve long-term survival after oesophageal variceal bleeding. A five-year prospective controlled clinical trial. Lancet 1983;ii:1328-32. The Copenhagen Esophageal Varices Sclerotherapy Project. Sclerotherapy after first variceal hemorrhage in cirrhosis. A randomized multicenter trial. N Engl J Med 1984:311:1594600. Westaby D, Macdougall BRD, Williams R. Improved survival following injection sclerotherapy for esophageal varices: final analysis of a controlled trial, Hepatology 1985;5:827-30. Korula J, Balart LA, Radvan G, et al. A prospective, randomized controlled trial of chronic esophageal variceal sclerotherapy. Hepatology 1985;5:584-9. Benner KG, Keeffe B, Keller FS, Rosch J. Clinical outcome after percutaneous transhepatic obliteration of esophageal varices. Gastroenterology 1983;85:146-53. Ohnishi K, Takayasu K, Takashi M, et al. Transhepatic obliteration of esophageal varices using stainless coils combined with hypertonic glucose and Gelfoam. J Clin Gastroenterol 1985;7:200-7. Hirashima T, Hara T, Takeuchi H, et al. Transabdominal esophageal mucosal transection for the control of esophageal varices. Surg Gynecol Obstet 1981;151:36-40.

ENDOSCOPIC

SCLEROTHERAPY

VERSUS

PTO

1209

Philadel8. Child CC, III. The liver and portal hypertension. phia: WB Saunders, 1964. of 9. Takase Y, Ozaki A, Orii K, et al. Injection sclerotherapy esophageal varices for patients undergoing emergency and elective surgery. Surgery 1982;92:474-9. The gen10. Japanese Research Society for Portal Hypertension. eral rules for recording esophageal varices. Jpn J Surg 1980;10:84-7. 11. Bankroft H. Modified life table in follow-up study. Introduction to biostatistics. New York: Paul B. Hoeber, 1957:19&7. estimation from incom12. Kaplan EL, Meier P. Nonparametric plete observation. J Am Stat Assoc 1958;53:457-87. tests with one degree of freedom: 13. Mantel N. Chi-square extensions of the Mantel-Haenszel procedure. J Am Stat Assoc 1963;58:690-700. 14. Johnson GW, Ridgers HW. A review of 15 years’ experience in the use of sclerotherapy in the control of acute hemorrhage from esophageal varices. Br J Surg 1973;60:797-800. 15. Williams KGD, Dawson JL. Fiberoptic injection of esophageal varices. Br Med J 1979;2:766-7. 16. Paquet K-J, Oberhammer E. Sclerotherapy of bleeding oesophageal varices by means of endoscopy. Endoscopy 1987;10:7-12. 17. Macdougall BRD, Westaby D, Theodossi A, et al. Increased long-term survival in variceal hemorrhage using injection sclerotherapy. Results of a controlled trial. Lancet 1982;i:1247. 18. Terblanche J, Northover JMA, Bornman P, et al. A prospective controlled trial of sclerotherapy in the long-term management of patients after esophageal variceal bleeding. Surgery 1979;148:323-33. 19. Takashi M, Igarashi M, Hino S, et al. Esophageal varices: correlation of left gastric venography and endoscopy in patients with portal hypertension. 20. Ohnishi K, Saito M, Terabayashi its in cirrhosis

Radiology

1985;155:327-31.

H, et al. Portal hemodynam-

in relation to large portal-systemic

Child’s grade (abstr). Hepatology

1986;6:1140.

shunts and