Comparison of different methods for endoscopic hemostasis of bleeding canine esophageal varices

GASTROENTEROLOGV

LIVER

AND

BILIARY

1983:84:1455-61

TRACT

Comparison of Different Methods for Endoscopic Hemostasis of Bleeding Canine Esophageal Varices DENNIS M. JENSEN, MICHAEL L. SILPA, JORGE I. TAPIA, DANIEL B. BEILIN, and GUSTAV0 A. MACHICADO Veterans Administration Medical and Research Services. Wadsworth for the Health Sciences, and Center for Ulcer Research and Education.

Despite udvances in the therapy of acute esophageal variceal hemorrhage, morbidity and mortality remain high. Continued severe variceal hemorrlmge remains a major clinical problem in poor risk patients who cannot tolerate emergency surgery for hemostasis. Several endoscopic hemostatic methods might be effective for variceal hemostasis. hut they have not been systematically evaluated. Using a reproducible canine model of esophageal varices, several hemostatic modalities were tested and compared to determine which were most effective in stopping variceal bleeding. Methods tested were endoscopic sclerotherapy, argon laser, neodymiumyttrium-aluminum-garnet laser, monopolar electrocoagulation, bipolar electrocoagulation. ferromagnetic tamponade, and endoscopic heater probe. Both neodymium-yttrium-aluminum-garnet laser and endoscopic sclerotherapy provided reliable hemostasis in acuteJJ, bleeding canine varices. Large heater probe controJJed bleeding 50% of the time, and all the other methods stopped bleeding in less than half the trials. Rebleeding after balloon inflation proximal to the coagulated bleeding site did not

Received June 15. 1982. Accepted November 12. 1982. Address requests for rel)rints to: Dennis M. Jensen. M.D., Division of Gastroenterology. Room 44-146 CHS. UCLA Center for the Health Sciences. Los Angeles. California 90024. This study was supported in part by Veterans Administration Research Funds and NIH Grant AM 17328 [to the Center for IJlcer Research and Education). This work was presented at the meeting of the American Gastroenterological Association, May 17. 1982, Chicago, Illinois. The authors thank Anita Boesman and Sandra Reigel for preparing the manuscript; Janet Elashoff for statistical assistance; Cooper Medical Corporation for use of the argon laser; Molectron Medical for the YAG laser; American Cystoscope Makers Inc. (ACMI) for use of the HICAP unit and a therapeutic endoscope: Walker Scientific for loan of the ferromagnetic tamponade unit; and Olympus Corporation for a therapeutic endoscope. (, 1983 t)vthe American Gastroenterological Association 001fi-5085/83'13~1455-07$03.00

Hospital and UCLA Center Los Angeles. California

occur with neodymium-yttrium-aluminuminum-garnet laser or endoscopic sclerotherapy-treated varices but did occur with the other methods. The principal differences betweeh neodymium-yttrium-aluminum-garnet laser and endoscopic sclerotherapy were the ease of application of neodymium-yttriumaluminum-garnet laser, the higher frequency of esophageal ulcers or erosions with neodymium-yttrium-aluminum-garnet laser, und the Jack of variteal obliteration with neodymium-,yttrium-aluminum-garnet laser. Despite many advances in the medical and surgical management of bleeding esophageal varices over the and case fatality rates remain last 40 yr, morbidity high. Continued severe variceal hemorrhage remains a major therapeutic problem, particularly in patients who are poor risks for emergency surgery. Several new hemostatic methods might be effective and safe for variceal hemostasis, but they have not been systematically evaluated or compared under similar conditions. Recently reported clinical techniques for variceal hemostasis include transbepatic variceal embolization (11, esophageal transection and reanastamosis (z), neodymium-yttrium-aluminum-garnet (YAG) laser (31, and endoscopic sclerotherapy (ES) (4-8).

Many endoscopic hemostatic devices have been carefully tested for control of ulcer bleeding but not variceal hemorrhage. Those studies have been facilitated by standardized models, such as canine gastric, duodenal, and esophageal ulcers (9;10) and erosions (11). Similar animal studies for evaluation of esophageal variceal hemostasis have not been possible because of the lack of a reproducible large-animal model. Several endoscopic hemostasis methods useful for control of nonvariceal bleeding might control variceal hemorrhage. These include argon laser photocoagulation (ALP) (12Z16j. monopolar electroco-

1456

JENSEN ET AL.

GASTROENTEROLOGY Vol. 84. No. 6

agulation (MPEC) (17,18), bipolar electrocoagulation (BPEC) (14,16,17,19,20), heater probe (HP] (21-24), and ferromagnetic tamponade (FT) (25,26). We recently developed a reproducible model of canine esophageal varices (27,28), which enabled us to test and compare methods of variceal hemostasis. The purpose of this study was to evaluate and compare endoscopic methods for control of acute variceal hemorrhage under similar conditions in a canine model of bleeding esophageal varices.

Materials

and Methods

Canine Model of Esophageal

Varices

The reproducible model of canine esophageal varices has been reported (27,28). Briefly, a side-to-side portacaval anastomosis was created, the inferior vena cava was ligated cephalad to this shunt, and an ameroid constrictor (Three Point Products, Montreal, Canada) was placed around the portal vein. Over the ensuing 4-6 wk, moderate- to large-sized esophageal varices developed in about half the animals. Animals with small varices had a second laparotomy and complete ligation of the portal vein. After this operation, the esophageal varices enlarged.

Creation and Measurement Randomization

of Hemorrhage-

After sodium pentobarbitol anesthesia and endotracheal intubation, mongrel dogs were anticoagulated with sodium heparin (ZOO U/kg initially, then 100 U/kg . h). Fluid and blood losses were replaced with intravenous normal saline. Twenty dogs with two to four moderate- to large-sized esophageal varices each were used for these studies. During each experiment, three to six variceal bleeds were induced in each animal. Endoscopy was performed with a therapeutic two-channel panendoscope (ACM1 TX-82, Stamford, Conn., or Olympus Corporation TGF-2D, New Hyde Park, N.Y.). To induce bleeding, the varix was punctured with a 19-gauge needle attached to an endoscopic catheter. The bleeding rate was quantified through a second large suction catheter connected to a calibrated vial. Each bleeding varix was randomized to control or treatment with one of the hemostatic modalities. The control punctures were observed for 5 min, then requantitated. A balloon cuff on the shaft of the endoscope cephalad to the varix served to increase intravariceal pressure and distend esophageal varices. It was also used with sclerotherapy to keep sclerosants in the injected segment and increase contact time. After successful hemostasis, the balloon cuff was inflated for 3 min cephalad to the treatment site to distend the varices and increase intravariceal pressure. This was used to test for rebleeding. Following hemostasis, recovery, and complete healing of any esophageal lesions (erosions or ulcerations induced by treatment), animals with moderate or large esophageal varices were used again for these acute hemostasis experiments.

Sclerotherapy

Agents

For hemostasis with sclerotherapy, z ml of the test agent was injected intravariceally using a 23-gauge needle in a flexible endoscopic catheter. We injected the varix intravariceally, directly at the bleeding point. Only one injection was made for each bleeding varix. The balloon cuff was inflated cephalad to the bleeding site before and for 2 min after the injection. Sclerosants tested were 5% sodium morrhuate (Eli Lilly & Co., Indianapolis, Ind.), 5% ethanolamine oleate (Evans Medical Ltd., Liverpool, England), 1.5% sodium tetradecyl sulfate (Elkins-Sinn, Inc., Cherry Hill, N.J.), cefazolin (Eli Lilly & Co.), 95% and 47% ethanol, and a mixture of 0.5% tetradecyl and 50 U/cm” thrombin (Parke-Davis, Detroit, Mich.) in 50% dextrose (TTD). Endoscopic

Lasers

Laser treatment was applied directly on the variteal bleeding point, after washing the area with water. Coaxial COZ aided in visualization of the bleeding point during treatment. Protective filters were placed over the endoscopic eyepiece when either YAG or argon laser treatments were performed. 171 argon laser (Cooper Medical A 20-W Spectrophysics Corp., Mountain View, Calif.) with 400-pm quartz lightguide and coaxial CO* gas jet was used as described (12). The argon lightguide had a 9” full angle of divergence, and treatment distance was l-3 cm. A power setting of 9 W with pulse duration of 1-2 s was used. In preliminary trials with argon laser for canine variceal hemostasis, lower power (6-8 W) was less effective than 9 W. Higher power (11-15 W) caused frequent tissue vaporization and ulceration without increasing efficacy. Also, esophageal damage and worsening hemorrhage frequently resulted after 220 applications of 9 W (or more) on or around the bleeding point. Therefore, a maximum of 20 applications was made. A 120-W YAG laser (Molectron Corp., Sunnyvale, Calif.) with a 600-pm quartz lightguide and coaxial CO2 was used, as described (29). A power setting of 70 W with a 0.5s pulse duration was used. In preliminary trials with YAG laser, higher powers (90-120 W) caused more tissue injury and lower powers (30-60 W) were less effective. The laser lightguide had a 10” full angle of divergence and the treatment distance was 1-4 cm.

Other Thermal

Coagulation

Units

A monopolar electrode, large and small bipolar electrodes, and large and small heater probes were tested. Each type of endoscopic probe and unit allowed target irrigation of the bleeding point during treatment. In preliminary studies of bleeding canine esophageal varices treatment, esophageal ulceration and worsening of hemorrhage frequently resulted after 15-20 applications of 20-J (or more) pulses on or around the bleeding point. Therefore a maximum of 20 applications was made for these thermal methods. In preliminary trials with electrocoagulation and heater probe for canine variceal hemostasis, lower energy of application (
ESOPHAGEAL

June 1983

than 20 J, but higher energy (25-50 J/pulse) caused frequent esophageal ulceration without increasing efficacy. Therefore, pulses of 20 J each were applied directly on the bleeding point of the varix. A monopolar electrode (model CD-3L, Olympus Corporation) was used with a standard Valleylab SSE2-K electrosurgical unit (Boulder, Colo.). Water was infused through the center of the probe for irrigation. Light touch was used for electrocoagulation. An analogue computer quantitated joules of power delivered with each pulse (30) for both MPEC and BPEC. A prototype endoscopic BPEC unit (BICAP Unit; ACMI, Stamford, Conn.) was used with a pulse duration of l-l.5 s and an intensity setting of 7. Both large and small BPEC probes were tested, with tip diameters of 3.25 and 2.25 mm, respectively. A prototype endoscopic HP (Auth, Seattle, Wash.) was tested with large (3.0 mm) and small (2.3 mm) probes.

Ferromagnetic

Tamponade

A 60-Hz, 208-V Walker Scientific FT unit with model UR-1365 power source (Walker Scientific, Worcester, Mass.) was tested. The power setting was 85% of maximum and treatment distance was -6 cm. The tamponade solution consisted of 10 g of 325 mesh iron powder (Alfa Products, Danvers, Mass.), 1 g carboxymethylcellulose [Sigma Chemical Co., St. Louis, MO.), 1 ml of glycerin, 10 ml of water, and 1000 LJ of topical thrombin (ParkeDavis, Detroit, Mich.). Five to seven milliliters of this solution were applied on the bleeding varix with an endoscopic catheter, and the magnet was turned on for 10 min. After 10 min, the electromagnetic power was turned off, and the animal’s position was changed such that the varix was not dependent. The FT solution was gently washed from the varix with water via a catheter. Continued bleeding was assessed visually. Criteria

VARIX HEMOSTASIS

IN CANINES

1457

static efficacy of each method is shown in Figure 1 and Table 1. The mean (-t SE) pretreatment bleeding rate for all varices was 3.4 k 0.6 ml/2 min [Table 1). The control punctures continued to bleed and after 5 min, the mean control bleeding rate was 2.5 2 0.6 ml/2 min. Pretreatment bleeding rates were higher in the YAG, small BPEC, and MPEC groups than other treatment groups or control (Table 1). Endoscopic

Sclerotherapy

The hemostatic efficacy of ES was 98%. Five to ten different varices were treated with each sclerosing agent. The only failure was one of the five ethanolamine trials. The other sclerosants were 100% effective. This difference was not statistically significant. In this acute study, there were some minor differences among the sclerotherapy agents tested in ease of injection and handling. Ethanolamine was more viscous than the other agents and therefore more difficult to inject through the 23gauge needle. The TTD required mixing before use. Forty-seven percent ethanol and 1.5% sodium tetradecyl sulfate required dilution of available shelf concentrations. Direct injection into the variceal bleeding point with ES was technically much more difficult than either laser treatment. Proximal balloon cuff infla-

for Hemostasis

Treatment was applied until bleeding stopped, severe ulceration occurred, or >20 treatment applications per varix were required. If bleeding did not cease by the twentieth treatment application, or if deep ulceration and worsened bleeding occurred with treatment, that trial was considered a failure for hetiostasis. Statistics

IAG

Statistical comparisons were made using the paired t-test, Fisher’s exact test, or the Bonferroni multiple comparison method (31). The statistical significance level of p < 0.05 was used throughout. Unless otherwise stated, differences

were

not significant.

Results A total of 148 bleeding canine esophageal varices were randomized to control or treatment. Ten varices were treated with each method except ES, which was used to treat 43 varices. The hemo-

Figure

ES

HP-I

ALP

FT

wEr2.I

BFEC,I

HP.I

m

CoNlRoc

1 Hemostatic efficacy expressed as a percentage of bleeding varices controlled with each modality. Efficacy = initial hemostasis minus rebleeding after balloon distention. n = 10 for each category except sclerotherapy, n = 43; and n = 25 for controls. YAG is neodymiumYAG laser; ES is endoscopic sclerotherapy; HP is

heater probe, large (HPl) and small (HPs); BPEC is bipolar electrocoagulation, large (BPECl) and small (BPECs); PT is ferromagnetic tamponade; ALP is argon laser photocoagulation; and MPEC is monopolar electrocoagulation. The efficacy of ES and YAG was significantly higher than controls. Endoscopic sclerotherapy and YAG were significantly more hemostatic than any other modality.

1458

JENSEN

Table

ET AL.

1. Bleeding

GASTROENTEROLOGY

Rates and Effectiveness

Pretreatment bleeding

rate

hemostasis

Rebleedingb after balloon

Effectiveness (%)” lo/lo (100) 42143 (98) 5110 (50) 4110 (40) 3110 (30) 3/10 (30)

Initial0

(ml/2 min)

Mode

n

YAG ES HP (large) ALP FT BPEC

10 43 10 10 10 10

4.6 2.4 2.9 2.5 2.5 2.8

0.9 0.2 0.4 0.4 0.5 0.5

10110 42143 6110 6110 3110 4110

0 0 1 2 0 1

(large1 BPEC

10

6.0 + 1.5

3110

1

2110

(20)

(small) HP (small) MPEC Control

10 10 25

2.4 k 0.2 5.5 2 1.8 2.5 -+ 0.6

3110 2110 0125

1 1

2110 1110 0125

(20) (10) (0)

k f k 2 2 2

For mode abbreviations, refer to Figure 1. Initial hemostasis” rate was No. of varices with hemostasisiTota1 No. treated. Rebleeding after balloonb was the number of successfully treated varices rebleeding with balloon distention proximal to the treatment site. Effectiveness” was defined as initial hemostasis rate minus rebleeding rate after balloon.

tion, after hemostasis with ES, caused no rebleeding. The effectiveness of ES was significantly greater than all other treatments except YAG.

Laser Photocoagulation The YAG laser stopped the bleeding in all 10 varices. The easiest method to use for variceal hemostasis was the YAG laser. Treatment without tissue contact facilitated YAG treatment of bleeding varices compared with ES. No rebleeding occurred with proximal balloon cuff distension. Argon laser photocoagulation [ALP) initial hemostatic efficacy was 60% (6 of 10). Argon laser photocoagulation, like other thermal modalities described below, usually formed a coagulum over the bleeding site, but the varix would continue to ooze. The rebleeding with proximal balloon cuff distention was 33%. The mean pretreatment bleeding rates for successfully treated ALP varices (2.4 ml/2 min) and unsuccessfully treated ALP varices (2.6 ml/2 min) was similar.

Other

Hemostatic

Devices

The monopolar coagulation probe stopped hemorrhage in only 2 of 10 bleeding varices. The mean energy per pulse was 23 + 1.9 J. A preliminary trial with more joules per pulse (30-50) led to esophageal perforation in I animal and severe ulceration in another. The mean bleeding rate before MPEC treatment was 5.5 5 1.8ml/2 min. Penetration into a varix occurred during one MPEC treatment.

Vol. 84, No. 6

This caused massive hemorrhage and near exsanguination. Bipolar electrocoagulation with the small (2.25 mm) and large (3.25 mm) probes was hemostatic in 30% (3 of 10) and 40% (4 of 10) of the varices treated, respectively. The pretreatment bleeding rates for small and large BPEC for the varices that were treated successfully were not significantly different from those that continued to bleed (Table 2). The small heater probe (HP) stopped 3 of the 10 bleeding varices. The large HP was initially effective in 60% of the trials. One ulcer (17%) rebled with proximal balloon distension. For the large HP trials, the initial bleeding rate was not significantly different for those varices treated successfully or those that bled (Table 2). Of 10 varices tested, FT stopped the hemorrhage in three. The mean pretreatment bleeding rate for the varices treated successfully was similar to that for the failures (Table 2). During the follow-up endoscopies, no mucosal damage was found with FT. The superficial erosions or ulcerations induced by HP, BPEC, and ALP healed within 5-10 days. However, deep, chronic ulcers often occurred with monopolar electrocoagulation and with YAG laser. These ulcers required ~2 wk to heal. Monopolar electrocoagulation caused one perforation and death during a preliminary trial. Also, one varix penetration (a hole in the side of the varix) from MPEC treatment resulted in a near exsanguinating hemorrhage. No other methods resulted in perforations. Superficial or deep ulcers resulted from approximately one-fifth of the ES injections. During weekly endoscopies, all the treated esophageal varices were still present after treatment with

Table

2. The Relationship Bleeding Rates

of Initial

Hemostasis

and

Pretreatment bleeding rates” lml/Z min t SEI Modalitv YAG ES HP (large] ALP FT BPEC (large BPEC (small] HP (small) MPEC

Initial hemostasis 4.6 2.4 2.9 2.4 3.0 2.3 6.3 1.7 2.2

-t + 2 2 + + t 2 2

0.9 0.2 0.8 0.3 0.9 0.8 2.4 0.3 0.7

No hemostasis

2.9 2.6 2.3 3.3 5.8 2.5 6.4

4.0 + 0.8 2 0.9 -c 0.7 + 0.5 + 2.1 2 0.2 k 2.5

Pb

>0.05 >0.05 >0.05 >0.05 >0.05 co.05 co.05

For mode abbreviations, refer to Figure 1. Pretreatment bleeding rates” of successfully (initial hemostasis) and unsuccessfully (no hemostasis) treated bleeding varices for each method. p”, by Fisher’s exact test, pretreatment bleeding rates for HP (small) and MPEC varices that were treated successfully were significantly different than those that continued to bleed.

ESOPHAGEAL

June 1983

FT and all the thermal modalities. In about one-third of the ES trials, the varix was obliterated after a single injection, but these tended to be the moderatesized rather than larger varices. Discussion Beginning

with splenectomy

over 80 yr ago strategies to control bleeding from esophageal varices. Portosystemic anastomoses, as suggested by Eck, were first performed in humans by Videl in 1903 (32). Therapeutic portacaval shunting has been used extensively in the United States for >35 yr for control of variceal bleeding. The mortality of emergency shunting remains alarmingly high, even with good risk patients. In a large unselected series, Orloff (33) reported a surgical mortality of -40% with emergency portacaval shunt for variceal bleeding. Stabilization of patients with esophageal hemorrhage by nonsurgical techniques and elective shunting, particularly distal or splenorenal types, has been recently advocated (34-38). Medical treatment with tamponade tubes or vasopressin helps control acute variceal hemorrhage in some patients, but control is often temporary. Also, balloon tamponade tubes were found to have a high complication rate (39), reducible by adherence to a rigid protocol for their use (40). Vasopressin has been associated with local gangrene (411, cardiac arrhythmias and myocardial injury (42),coronary vasoconstriction (43), and bowel ischemia (44). We still use these methods clinically but recognize their limitations. Several invasive techniques have been introduced to control acute variceal hemostasis. Esophageal transection and reanastamosis is a promising new surgical technique (2). However, it requires general anesthesia and has a high mortality rate in poor risk patients. Another technique for control of acute variceal hemorrhage, percutaneous transhepatic embolization, was described in 1974 (45). In a recent series, this technique achieved hemostasis in 37 of 52 (73%) actively bleeding patients (1).In the successfully treated patients, 65% rebled. Retreatment with this method is often not possible. Endoscopic injection sclerotherapy was first performed clinically by Crafoord and Frenckner in 1939 (46)but was never used widely in the United States. With the development of flexible fiberoptic endoscopes and catheters, this technique has reemerged. Most sclerotherapy trials have instituted treatment after bleeding has ceased (47-49). However, in some centers, this technique has been used successfully to control acute hemorrhage (4-7). Several thermally active coagulation devices have (32),there have been many therapeutic

VARIX HEMOSTASIS

IN CANINES

1459

provided successful hemostasis of nonvariceal gastrointestinal lesions in animals and patients. With the exception of YAG laser (3),no endoscopic thermal method has been used in a large clinical series of bleeding esophageal varices. Rebleeding rates after YAG laser treatment have been very high. Similarly, in a small series, Brunetaud (50) reported that argon laser controlled bleeding from esophageal varices, but most patients later rebled. Electrocoagulation (bipolar and monopolar) and heater probes that combine heat and pressure might be able to control variceal bleeding by coagulating compressed, large veins. Ferromagnetic tamponade also could affect variceal hemostasis by combining pressure with local thrombin application. Whether or not these techniques could be used effectively for variceal hemostasis has not been previously evaluated. The development of a canine model of esophageal varices (27)enabled us to perform detailed comparisons under similar conditions. The test animals were fully heparinized to prevent spontaneous cessation of variceal hemorrhage and also to mimic patients with variceal bleeding who often have a severe coagulopathy. The canine model differs from human variceal bleeders because animals did not manifest spontaneous bleeding or encephalopathy. However, the bleeding episodes induced by varix puncture closely resemble those diagnosed with emergency panendoscopy in our patients. As seen in Figure 1, only ES and YAG laser treatments achieved consistent variceal hemostasis. Rebleeding did not occur with variceal distention induced with a balloon cuff. The true acute effectiveness (defined as initial hemostasis minus rebleeding with distention, Table 1) of the other methods was no more than 50%. Both ES and YAG were significantly more effective than control models. The thermally active methods that were not reliably effective (ALP, MPEC, HP large and small, and BPEC large and small) did form a surface coagulum at the bleeding site and often slowed bleeding compared with controls. However, this did not always seal the underlying varix. When a balloon was inflated in the proximal esophagus after successful hemostasis with these methods, rebleeding often occurred from the treated site. Balloon inflation resulted in visual distention of the varix and, presumably, increased intravariceal pressure. Rebleeding after balloon inflation did not occur with YAG or sclerotherapy-treated varices. All the thermally active methods caused some esophageal mucosal damage. These lesions tended to be superficial with HP, BPEC, and ALP. Healing within 5-10 days was usual with these methods. With MPEC, deep ulcers and one variceal perforation resulted. Neodymium-yttrium-aluminum-gar-

1460

JENSEN

ET AL.

net laser treatments were all associated with ulcers that required >lO days to heal. All varices were still present on the weekly follow-up endoscopies after treatment with FT and all the thermal methods. What were the possible sources of bias in our study? These might include varix response to treatment, interanimal variation in intrinsic coagulation ability, variceal size, and variceal bleeding rates. Intrinsic coagulation was blocked by intravenous heparin so that spontaneous coagulation was eliminated. Bleeding rates of controls 5 min after randomization were not significantly different than before randomization (2.6 + 0.2 vs. 2.5 ? 0.6 ml/2 min). Only moderate- to large-sized varices (which protruded into the esophageal lumen) were used in this study. Differences among dogs in size or location of varices were not enough to account for the differences in device efficacy. Uniformly good results were found in all animals for ES and YAG compared with all other methods. Varices responded differently to treatment with different methods. For example, edema and acute ulceration were common with YAG laser; acute blanching and change in varix size were common with sclerosis. These differences appeared to vary with the hemostasis method being tested rather than the different animals. Similar results were noted in different animals. We concluded that differences of intrinsic coagulation, varix response to treatment, and varix size among animals did not have a major effect upon hemostatic efficacy results. Overall, our canine model allowed comparison of new or promising hemostatic therapies under similar conditions. Could the initial bleeding rate (IBR) be a source of bias? Not for the effective methods. For example, the YAG group, which was the most successful, had the third highest mean IBR (Table 1). Endoscopic sclerotherapy, with 98% efficacy, had a bleeding rate not significantly different from five less successful methods. In the ALP, large BPEC, and large HP groups, the mean IBR was similar for those varices that continued to bleed and those successfully treated. Although IBR may have affected results with the small HP and MPEC, it did not affect any other hemostatic modalities. For those varices also treated with small BPEC, the mean IBR was higher for those that stopped than for those that did not. Successfully treated FT varices had a higher IBR than the unsuccessful group. In this study, YAG and ES were the only methods that provided consistent hemostasis. Three major differences were apparent for YAG and ES. One difference was the ease of application of YAG laser compared with ES for bleeding varix treatment. Treatment from a distance without catheter-tissue contact contributed to this difference. Second, most YAG-treated varices had acute and chronic esopha-

GASTROENTEROLOGY

Vol. 84, No. 6

geal ulcers compared with less than half those treated with sclerotherapy. The third major difference between them appears to be their chronic effects on the varix. With the power used in this study, YAG laser sealed the bleeding site acutely, but did not obliterate the varix. Endoscopic sclerotherapy can obliterate a varix, as we noted in about one-third of the varices treated with a single sclerotherapy injection These tended to be the moderate-sized rather than the large varices. Usually, variceal obliteration requires a series of treatments (49), particularly for multiple or large esophageal varices. While all the sclerosing agents tested in this study seemed equally effective for hemostasis, they may not be equally effective in obliterating varices. Further studies are warranted to evaluate possible differences among various sclerosing agents in obliteration rates and ulceration frequency in this canine model. Based upon the lack of efficacy of MPEC, BPEC, HP, ALP, and FT in this model of bleeding canine esophageal varices, we would predict similar poor results with these devices clinically. Additionally, we do not consider MPEC safe enough to use in the esophagus because of the danger of deep ulcers, and perforation of the esophagus or varix. More extensive studies are warranted to evaluate the possible acute and chronic differences among sclerotherapy agents and different YAG laser powers. Based upon their excellent hemostatic efficacy, we also plan further studies of YAG laser and sclerotherapy in this canine variceal model to determine ways to reduce the tissue injury while increasing varix obliteration Ultimately, a comparative clinical trial of ES and YAG for acute variceal hemostasis is anticipated. References 1. Smith-Laing

2.

3.

4.

5. 6.

7. 8.

G, Scott J, Long RG, Dick R, Sherlock S. Kale of percutaneous transhepatic obliteration of varices in the management of hemorrhage from gastroesophageal varices. Gastroenterology 1981;80:1031-6. Cooperman M, Fabri PJ, Martin EW, Carey LC. EEA esophageal stapling for control of bleeding esophageal varices. Am J Surg 1980;140:821-4. Kiefhaber P, Nath G, Moritz K. Endoscopic control of massive gastrointestinal hemorrhage by irradiation with a high power neodymium-YAG laser. Prog Surg 1977;15:140-55. Terblanche J, Northover JMA, Bornman P, et al. A prospective evaluation of injection sclerotherapy in the treatment of acute bleeding from esophageal varices. Surgery 1979:85:239-45. Hughes RW. The management of bleeding esophageal varices with sclerosing injections. Mayo Clin Proc 1981;56:580. Fleig WE, Ruttenauer K, Strange EF, Ditschuneit H. A prospective study of sclerotherapy for massive hemorrhage from esophageal varices not responding to balloon tamponade (abstr). Gastrointest Endosc 1981;27:131. Lewis J, Chung RS, Allison J. Sclerotherapy of esophageal varices. Arch Surg 1980;115:476-80. Johnston GW, Rodgers HW. A review of 15 years’ experience

June

9.

10.

11.

12.

ESOPHAGEAL

1983

in the use of sclerotherapy in the control of acute haemorrhage from oesophageal varices. Br J Surg 1973;60:797-800. Protell RL, Silverstein FE, Piercey J, Dennis M, Sprake W, Rubin CE. A reproducible animal model of acute bleeding ulcer-the “ulcer maker.” Gastroenterology 1976;71:961-4. Jensen DM, Machicado GA, Tapia JI, Mautner W. Clotting factors fail to control bleeding from standard ulcers or gastric erosions (abstr). Gastroenterology 1980;78:1187. Jensen DM, Machicado GA, Tapia JI, Franc0 P. A canine model of acute antral erosions and ulcers induced by indomethacin (abstr). Gastroenterology 1980;78:1187. Johnston JH, Jensen DM, Mautner W, Elashoff J. Argon laser treatment of bleeding canine gastric ulcers: limitations and Gastroenterology guidelines for endoscopic use.

1981;80:708-16. 13. Machicado GA, Jensen DM, Tapia JI, Mautner W. Treatment of bleeding canine duodenal and esophageal ulcers with argon laser and bipolar electrocoagulation. Gastroenterology 1981;81:859-65. 14. Jensen DM, Machicado GA, Tapia JI, Mautner W. Bipolar electrocoagulation and argon laser photocoagulation of colonic lesions (abstr). Gastrointest Endosc 1980;26:69. 15. Jensen DM, Machicado GA, Mautner W, Tapia JI, Franc0 P. Endoscopic argon laser photocoagulation of patients with severe gastrointestinal bleeding (abstr). Gastroenterology 1980;78:1188. 16. Johnston JH, Jensen DM, Mautner W. Comparison of laser photocoagulation and electrocoagulation in endoscopic treatment of bleeding canine gastric ulcers. Gastroenterology 1982;82:904-10. 17. Protell RL, Gilbert DA, Silverstein FE, Jensen DM, Hulett FM, Auth DC. Computer-assisted electrocoagulation: bipolar vs. monopolar in treatment of experimental canine gastric ulcer bleeding. Gastroenterology 1981;80:451-5. 18. Papp JP. Endoscopic electrocoagulation of actively bleeding arterial upper gastrointestinal lesions. Am J Gastroenterol 1979;71:516-21. 19. Auth DC, Gilbert DA, Opie EA, Silverstein FE. The multipolar probe-a new endoscopic technique to control gastrointestinal bleeding (abstr). Gastrointest Endosc 1980;26:63. 20. Johnston JH, Jensen DM, Mautner W. A comparison of bipolar electrocoagulation and argon laser photocoagulation with coaxial CO, in the treatment of bleeding canine gastric ulcers (abstr). Gastrointest Endosc 1978;24:200. 21. Protell RL, Rubin CE, Auth DC, et al. The heater probe: a new endoscopic method for stopping massive gastrointestinal bleeding. Gastroenterology 1978;74:257-62. 22. Protell RL, Lawrence DM, Peoples JE, Majoch TR. Auth DC. A new endoscopic thermal cautery which stops experimental bleeding safely (abstr). Gastroenterology 1980;78:1239. 23. Jensen DM, Tapia JI, Machicado GA, Beilin DB, Silpa M. Comparison of electrocoagulation and heater probe for hemostasis in the canine colon (abstr). Gastrointest Endosc 1982:28:151. 24. Jensen DM. Tapia JI, Machicado GA, Beilin DB, Silpa M. Endoscopic heater and multipolar probes for treatment of bleeding canine gastric ulcers (abstr). Gastrointest Endosc 1982;28:151. 25. Smith FE. The use of strong magnetic fields in the control of gastrointestinal bleeding in the dog. Presented at the Western Gut Club, Carmel, Calif., February 1979. 26. Carlson D, Jensen DM, Machicado GA, Franc0 P, Tapia JI. Treatment of experimental bleeding gastric ulcers and erosions with ferromagnetic tamponade (abstr). Gastroenterology 1980;78:1147. 27. Jensen DM, Tapia JI, Machicado GA, Kauffman C, Beilin DB, Franc0 P. A reproducible canine model of esophageal varices. Gastroenterology 1983;84:573-9.

VARIX HEMOSTASIS

IN CANINES

1461

28. Jensen DM, Tapia JI, Machicado GA, Kauffman G, Beilin DB, Franc0 P. A reproducible canine model of esophageal varices (abstr). Gastrointest Endosc 1981;27:131. 29. Johnston JH, Jensen DM, Mautner W, Elashoff J. YAG laser treatment of experimental bleeding canine gastric ulcers. Gastroenterology 1980;79:1252-61. 30. Piercey JRA, Auth DC, Silverstein FE, et al. Electrosurgical treatment of experimental bleeding gastric ulcers: development and testing of a computer control and a better electrode. Gastroenterology 1978;74:527-34. 31. Miller RG. Simultaneous statistical inference. New York: McGraw-Hill Book Co., 1966. 32. Grannis FW. Guido Banti’s hypothesis and its impact on the understanding and treatment of portal hypertension. Mayo Clin Proc 1975;50:41-8. 33. Orloff MJ, Charters AC, Chandler JG, et al. Portocaval shunt as emergency procedure in unselected patients with alcoholic cirrhosis. Surg Gynecol Obstet 1975;141:59-68. 34. Malt RA. Management of bleeding esophageal varices. West J Med 1982;136:143-5. 35. Conn HO, Resnick RH, Grace ND, et al. Distal splenorenal shunt vs. portal-systemic shunt: current status of a controlled trial. Hepatology 1981;1:151-6. 36. Reynolds TB, Donovan AJ, Mikkelsen WP, Redeker AG, Turrill FL, Weiner JM. Results of a 12-year randomized trial of portacaval shunt in patients with alcoholic liver disease and bleeding varices. Gastroenterology 1981;80:1005-11. 37. Rikkers LF, Rudman D, Galambos JT. et al. A randomized, controlled trial of the distal splenorenal shunt. Ann Surg 1978;188:271-82. 38. Zeppa R, Hutson DG, Bergstresser DR, Levi JU, Schiff ER, Fink P. Survival after distal splenorenal shunt. Surg Gynecol Obstet 1977;145:12-6. 39. Conn HO, Simpson JA. Excessive mortality associated with balloon tamponade of bleeding varices. JAMA 1967;202:1359. 40. Pitcher JL. Safety and effectiveness of the modified Sengstaken-Blakemore tube: a prospective study. Gastroenterology 1971;61:291-8. 41. Greenwald RA, Rheingold OJ, Chiprut RO, Rogers AI. Local gangrene: a complication of peripheral pitressin therapy for bleeding esophageal varices. Gastroenterology 1978;74:7446. 42. Beller BM, Trevino A, Urban E. Pitressin induced myocardial injury and depression in a young woman. Am J Med 1971;51:675-82. 43. Drapanas T, Crowe CP, Shim WKT, Schenk WG. The effect of pitressin on cardiac output and coronary, hepatic, and intestinal blood flow. Surg Gynecol Obstet 1961;113:484-9. 44. Conn HO, Ransby GR, Storer EH. Selective intraarterial vasopressin in the treatment of upper gastrointestinal hemorrhage. Gastroenterology 1972;63:634-45. 45.

46. 47.

48. 49. 50.

Lunderquist A, Vang J. Transhepatic catherization and obliteration of the coronary vein in patients with portal hypertension and esophageal varices. N Engl J Med 1974;29:646-9. Crafoord C, Frenckner P. New surgical treatment of varicose veins in the oesophagus. Acta Oto-laryngol 1939;27:422-9. Clark AW, Westaby P, Silk DBA. et al. Prospective controlled trial of injection sclerotherapy in patients with cirrhosis and recent variceal hemorrhage. Lancet 198O;ii:552-4. Brooks WS, Galambos GT. Endoscopic sclerosis of esophageal varices (abstr). Gastrointest Endosc 1981;27. Sivak MV, Stout PJ, Skipper G. Endoscopic injection sclerosis of esophageal varices. Gastrointest Endosc 1981;27:52-7. Brunetaud JM, Enger A, Flament JB, Petit J, Berjot M, Moschetto Y. Utilization d’un laser a argon ionise en endoscopic digestive: Photocoagulation des lesions hemorragiques. Rev Phys Appl 1979;14:385-90.