CURRENT CONCEPTS IN THE TREATMENT OF BLEEDING ESOPHAGEAL VARÍCES: CONTROVERSIES IN STRATEGY TECHNIQUE AND LONG-TERM OUTCOME

*M. Jung, *A. Zipf.

*Johann Wolfgang Goethe University Medical Hospital Frankfurt. Department of Medicine II. Gastroenterology and Hepatology, Pneumologyand Allergology (Head: Prof. W.F. Caspary, M.D.) Theodor-Stem-Kai 7. D-60590 Frankfurt am Main. Germany.


Introduction.-

For over 20 years now, endoscopic variceal sclerosis has been the first line treatment of bleeding esophageal varices. This technique is judged to be relatively problem-free and effective and is associated with a high success rate as far as hemostasis is concerned. A plethora of controlled and non-controlled studies has meanwhile established the relevance of sclerotherapy, with success rate of 90 to 95% being reported for primary hemostasis [1, 20,26,36,45]. Opinions are divided with regard to a long-term control of bleeding, where sclerosis is regarded with greater skepticism [3,23,45].

Although endoscopic sclerosis is used thoughout the world, there are major differences of opinion, sometimes even controversies with regard to the means of application and the course of treatment (Tab. 1).

Controversies.-

METHODS AND TECHNIQUES:

The discussion about methods and techniques has been raging for a long time. Should either paravariceal or intravariceal injection be practiced or should a combination of both forms be used? Hardly a single study is available which shows consistent comparability of the sclerotic techniques reported [6, 29,37].

Theoretically, injection therapy follows two goals. The proponents of the paravariceal approach emphasize the fact that initial hemostasis is achieved by the lateral placement or edema intended to blanket the bleeding varix. Continued application of the sclerosing agent creates a local inflammation of the esophageal wall which should ultimately cover the varix with fibrous tissue thus detaching it from the surface. Scarring tissue is produced over the deep-seated venous cord which is ideally maintained to withstand continued blood flow [7].

Advocates of the intravascular technique favor the thrombotic obliteration of the vessel since, they claim, this is the only way to achieve hemostasis effectively and sustainably [37]. Kitano and co-authors even demand that the entire esophageal mucosa be eliminated by replacing it with a fibrous coating, proclaming this to be the only feasible method of complete variceal eradication [17].

Pathological-anatomical studies on autopsy specimens have encouraged doubts about these selective approaches. Even when the procedure is performed skillfully, for or intravascular injection consistently results in undesirable thrombotic displacement of the vessel or inflammatory changes in the region of the esophageal wall. Hence, it can be stated that none of the respective techniques exclusively produce the desired outcome, but always involve a local reaction or venous obliteration [13]. The disparity of opinions about the type: of technique has meanwhile y¡elded to a pragmatic approach to sclerosing techniques, with the current preference being a combination of the intra and paravascular application.

SCLEROSING AGENTS:

There is a wide spectrum of injected agents (Tab. 2) [32]. The aim of injection treatment is to induce a local inflammatory response in the esophagus which heals partially by means of necrotic changes, sometimes via an ulceration stage, and leads to fibrous tissue formation in the varices-bearing esophageal section or to local vascular thrombus.

The list of sclerosing agents ranges from fatty acid derivatives like ethanol amine oleate and sodium morrhuate via synthetic agents like sodium tetradecyl sulfate to polidocanol (Aethoxyserolâ) [5,17,26,27,33,42,47].

Aethoxisklerolâ contains a mixture of alcohol and polidocanol which is applied in a concentration of 0.5-2%. The preparation is chiefly used in Western Europe as a 1% injection solution administered intra and paravascularly. In Asian countries, pure alcohol is the preferred sclerosing agent [31]. Bucrylate (Histoacrylâ), a rapidly polymerizing substance, causes local hemostasis after strictly intravascular application and, in addition to its indication in gastric varices, can therefore also be used with caution in the esophagus [38]. A 3% aqueous phenol solution is reported to be associated with a lower rate of toxic side effects than the conventional sclerosing agents [24]. Fibrin glues (Tissucolâ, Immuno, Heidelberg, Beriplastâ, Behring, Heidelberg) have the advantage of producing a local thrombus without triggering more extensive damage to the surrounding tissue [9]. However, the product has not been tested in large-scale randomized studies.

The data reported on the amount of sclerosing agent used for local injection vary. No uniform limit has been set. Depending on the site of the depot in the esophagus, 0.5-3 mi injected longitudinally produce a more or less pronounced wall edema and favor the later formation of ulcerations. The result may be inflammatory and ulcerating wall penetration going as far as mediastinitis in the case of uncontrolled application. Sterile pleural effusions have even been observed [1,10,49]. In contrast, the instillation of a fibrin glue solution is reported to accelerate the healing of sclerotic ulcers [34].

Depending on the extent of the varices, around 10-40 ml of sclerosing fluid can apparently be injected per session wihout risk [10,26].

SCOPE OF SCLEROTHERAPY:

Should the varices in the entire esophagus be sclerosed or is partial sclerotherapy sufficient? Due to their subepithelial location in the aboral section of the esophagus, esophageal varices are strikingly prominent and are most prone to rupture here [39]. It is justified to limit sclerotherapy to the distal 5-7 cm because of this morphological idiosyncrasy [46]. On the other hand, bleeding is also diagnosed a marked distance away from the ora serrata. Sclerotherapy of esophageal varices is therefore undertaken in different regions of the esophagus: only in the lower third or along the entire course or the varices up into the mid or even upper third.

ENDOSCOPIC INSTRUMENTS:

Endoscopic sclerotherapy can be performed with either rigid or flexible instruments. Rigid devices have a larger diameter and therefore tend to be more difficult to handle, usually requiring intubation anesthesia. Since fiber optic scopes with a wide suction canal have become available, rigid endoscopes are virtually never used any more in sclerotherapy [26,45].

SEDATION:

Endoscopic sclerotherapy is carried out under intravenous sedation, and if appropiate, on an out?patient basis. General anesthesia is to be preferred in critical situations with intense hemorrhage and poor visibility. Though endotracheal intubation is more complicated, it does protect against aspiration and can facilitate the procedure under adverse local circunstances.

Course of treatment.-

Emergency endoscopy is indicated if variceal bleeding is suspected, its first major aim being to secure any actual rupture of the varix, then to exclude a second, non-variceal bleeding source. Opinions differ once again about the frequency of additional bleeding sites in the presence of esophageal varices. In up to 30%, however, a non-variceal origin in addition to that from the varices themselves or a different source of bleeding altogether are reportedly encountered [25,46].

What does the optimal acute therapy of blecaing esophageal varices consist of? To date, no consensus has been reached about any standardized procedure.

Initial hemostasis by means of balloon tamponade or concomitant drug therapy was thought to be the primary goal in order to then plan for elective sclerotherapy under more favorable conditions once hemostasis had been achieved. Since active variceal obliteration under emergency conditions always represents a more complicated and risky procedure and the outcome hinges on the skills of the examining physician, emergency sclerotherapy was contemplated with reticence [26,46]. Today, due to the fact that enodoscopists have acquired greater routine with this technique, sclerotheraphy has become predominant in the treatment of acute variceal ruptures.

Several controlled and randomized studies have been published on the procedure for acute variceal bleeding which examine the endoscopic technique in emergencies or as an elective endoscopic therapeutic concept in connection with classical forms of variceal management.

Clinical studies.-

In patients with acute variceal bleeding, Paquet and Feussner compared endoscopic sclerotherapy with balloon tamponade alone (Sengstaken-Blakemore tube) (Tab. 3) [27].One controlied study on 43 patients showed the clear advantages of sclerotheraphy over a treatment with ballon tamponade math regard to initial hemostasis and rebleeding.

Definitive control of bleeding by endoscopic sclerotherapy was statistically significantly superior to balloon tamponade on its own. The 30-day mortality also produced significantly better results after endoscopic therapy. These success rates were reconfirmed after 6 months.

Westaby et al, compared emergency sclerotherapy with elective sclerotherapy after previous drug therapy in 50 patients with 64 bleeding episodes (Tab. 4) [47]. Twelve hours after admission to the hospital, patients with bleeding esophageal varices were either sclerosed immediately or an elective intervention was planned after pretreatment with vasopressin and nitroglycerin. Emergency sclerotherapy succeded in controlling bleeding significantly more frequently than the combination of drug and elective measures. Rebleeding ocurred in both groups with equal frequency. Although mortality was lower in the group treated under emergency conditions, there was no significant difference to the other group.

The number of randomized studies comparing endoscopic sclerotherapy and operative interventions is sparse. Cello et al compared endoscopic therapy with a portocaval shunt operation, although the study population only comprised patients with prognostically unfavorable Child class C (Tab. 5) [5]. As expected, the rebleeding rate was markedly lower in the surgery group, whereas no differences were shown with regard to mortality and survival after 18 months.

Meta-analyses.-

Endoscopic sclerotherapy represents the focal point of randomized studies on the therapy of acute variceal bleeding. According to a meta-analysis conducted by Infante-Rivard based on 7 randomized clinical studies, patients wiht bleeding esophageal varices appear to benefit from repeated sclerotherapy sessions [15]. In comparison to patients on purely conventional, non-endoscopic therapy, the long-term outcome was more favorable on continued sclerotherapy.

A meta-analysis by Bretagne et al examined the value of sclerotherapy based on 40 randomized controlled studies on 3 041 patients [3]. The authors conducted a comparative evaluation of endoscopic sclerotherapy used only in an emergency, used electively and for prophylaxis. Clearly significant results were obtained for sclerotherapy used only in an emergency compared to conventional treatment and in the prevention of rebleeding episodes in comparison with drug therapy. The endoscopic approach had no significant effect on mortality.

Based on a meta-analysis study by Working Team Reports (First United Gastroenterology Week Athens 1992) sclerotherapy was revealed to be the only form of treatment which yielded a significant improvement in short-term survival over conventional treatment [4].

Sclerotherapy of esophageal varices in conclusion.-

Despite its international acceptance, endoscopic sclerotherapy, as an emergency or as an elective procedure, only yields good results over the short-term. Elective sclerosis after drug therapy with nitroglycerin and vasopressin did not convincingly improved the results.

The rebleeding and low long-term survival rates remain unsatisfactory. These parameters are not solely connected to bleeding varices per se, but to a much greater extent to the underlying disease-cirrhosis of the liver and portal hypertension.

The technique of endoscopic sclerotherapy leaves virtually no room for further modification. The hardly favorable longterm prognosis, the local problems and complitations associated with sclerotherapy have led to the development of a less invasive method [1,22,41,49]

Endoscopic ligation of esophageal varices.-

After their animal experiments were successful, Stiegman, Goff and co-workers presented a new method for banding varices [40]. The value of this method was initially tested by the authors in their own series on a consecutively treated number of 146 patients and later confirmed by an American multicenter study [12,41,42].

METHOD:

Technically, endoscopic variceal ligation is comparable with the rubber band suction ligature of hemorrhoids (hemorrhoidal banding) [12]. The method uses a fiber-optic endoscope equipped with a cylinder mounted on the distal tip. An inner cylinder bearing an elastic o-ring is inserted inside the attachment. A nylon trip-wire is fixed to the slot on the inside inner cylinder and passed down the biopsy channel of the endoscope.

The tissue containing the varix is aspirated into the adapter at the tip of the endoscope (ligature chamber). The trip-wire is pulled causing the rubber band to be stripped off and close around the tissue strangulating the vein. The endoscope is then removed and reloaded with a new cylinder pre-fitted with an o-ring. Up to 6 to 10 rubber bands can be used in one ligating session.

An overtube was specially designed for this procedure to avoid having to repeatedly withdraw and insert the endoscope though the oropharynx. This 2-cm wide plastic overtube remains in place during the entire session; allowing the endoscope to be inserted and removed and reloaded easily and also protects the oropharynx and the the tracheobronchial system from the aspiration of blood and stomach contents. During active variceal bleeding, this over-tube represents a further protective measure.

Shortly after application, the variceal column ligated by the o-ring becomes fivid, necrotizes and sloughs off after 2-4 days. The rubber band also later dissolves leaving a fiat, circular ulcer which is limited to the mucosa and submucosa [40,43]. In contrast to an ulcer induced by sclerotherapy, this superficial ulcer does not extend to the muscularis propia [25,31,49]. Rebleeding from these ligature ulcers is observed less frequently and they heal more rapidly. Moreover, variceal ligation more effectively cuts off the proximal blood flow. Even in grade IV esophageal varices, just a few ligations in the esophagus lead to a markedy visible collapse of the veins. The working group around Young carried out a detailed comparison of ulcerations after ligature and after sclerosis, finding that sclerosis-induced ulcers extended deeper and took statistically longer to heal [48].

Esophagoscopic variceal ligation is simpler than the description might lead one to believe.

The technique is easy to leam. Problems are initially caused by the limited visibility provided through the loaded adapter, with restrictions to the methods arising during acute bleeding. However, the overtube is even more the subject of discussion than the limited field of vision [2,16,30]. The tiny gap between the diameter of the endoscope and the tube can scrape or even perforate the mucosa when the instrument is advanced through the esophageal channel. Introduction of the overtube via a Savary (Ch 14) or Malony bougie is designed to prevent serious injury during the procedure.

Betside the original set developed by Stiegman and Goff (Bard International Inc. Tewsbury, Ma) with a specially fitted overtube, other modified versions of variceal ligature instruments are currently available (Pauldrach, Garbsen).

RANDOMIZED STUDIES:

In a nation-wide American multicenter study, efficacy and safety of sclerotherapy versus variceal ligature were investigated in 129 patients with bleeding esophageal varices [42]. The major portion of these patients were treated efectively, i.e. after an acute bleeding episode. Significance was only proven with regard to the complication rate in a comparison of the two methods (Tab. 6). Complications in the group of ligated patients were markedly lower, vAth the number of cases of pneumonia associated with therapy proving particularly low. Endoscopy-induced esophageal strictures were significantly higher in the sclerosed patients than in the control group.

No remarkable differences between sclerotherapy and ligature were shown with regard to the long-term prognosis of aft the patients. However, if the prognostically unfavorable Child class C patients were excluded from analysis, the survival of Child A and B patients was significantly extended. These new findings on patients with prognostically favorable stages have indeed motivated further comparative studies, although to date only, few randomized studies are available on these therapeutic methods.

Observations made by Loren Laine and Gimson revealed differences to the American multicenter study, with Laine finding no significant differences in rebleeding rate or mortality, but fewer esophageal strictures after variceal ligature (Tab. 7) [11,18].

Gimson proved that variceal ligation had a favorable effect on rebleeding rate and described rapid variceal obliteration [18]. Complete eradication, however, was only possible in either group to a limited extent (55% EVS vs. 59% EVL) (Tab. 8). Both methods had the same efficacy in acute bleeding and there were no statistical differences with regard to the complication rate of the two techniques. The low number of iatrogenic esophageal strictures was attributed to the quality of the sclerosing agent (ethanol amine oleate) and medical co-therapy with sucralfate.

Hashizume modified this therapeutic strategy in a comparison of ligation versus sclerosing at the start and pure sclerotherapy in the follow-up sessions [14]. Here as well, a reduction in the complication rate was observed in ligature patients. Moreover, the use of rubber band ligature in the follow-up sessions led to a reduction in the number of injections and the amount of sclerosing fluid used.

META-ANALYSIS:

In a meta-analysis that has only been published as an abstract so far, Loren Laine compared sclerotherapy and ligation based on the results of 6 randomized studies [19]. The only significance revealed was the reduction in rebleeding after endoscopic variceal ligation and a trend towards reduced mortality. Curiously, these mostly American studies observed no differences in the respective number of complications.

COMPLICATIONS OF VARICEAL LIGATION:

Despite the obvious reduction in general complication after variceal ligation, specific problems associated with this technique should not be overlooked. The tightness and pain in the chest resulting from placement of more than 5 ligation bands due to transient esophageal constriction is harmless [12]. The symptoms subside after 2?3 days and require a liquid or slurry diet.

The case reports are accumulating in which the broad overtube is regarded as a complication-related instrument [2,16,30]. Repeated incidences of injury to the cervical esophagus accompanied by hemorrhage and perforation have been described, even isolated fatalities [35]. The broad diameter of the tube can also lead to compression of the larynx and trachea and thus to acute oxygen saturation deficiency. The advantage is sealing-off the tracheobronchial system, there by diminishing the consequences of aspiration contrasts, with local damage to the esophageal wall. In our endoscopic department, we refrain from tube implantation during variceal ligation whenever possible and accept that the endoscope has to be inserted and withdrawn repeatedly. Intravenous sedation of the patients is recommended during this procedure.

Like sclerotherapy, endoscopic variceal ligation also has its contraindications. lf the patient's immunity is impaired, the esophageal ulcers cannot be expected to heal in good time. Endoscopic variceal ligation can equally lead to local necrosis with perforation and death if the healing mechanisms fail due to long-term immunosuppression and a reduction in coagulation factors (lowered Quick's time, thrombocytopenia) [35]. In such cases, it is recommended to start with one or two rubber bands in the region where the bleeding is strongest (esophagocardiac junction) and to evaluate the local results after 2-3 days. lf the wounds are healing normally, variceal ligation can be continued with caution.

Endoscopic banding is dependent on the ability to aspirate the variceal channel into the ligation chamber. This procedure can only be performed to a limited degree if the esophagus has already undergone sclerotherapy and exhibits scarring or low-grade varices of class I. Ligation is not possible in the rigid esophagus, leaving thorough sclerotherapy as the only alternative in such cases.

De novo varices.-

lf portal hypertension persists, the occurrence of new varices can be expected in the esophagus, even after sclerotherapy. Esophageal varices are also just as likely to recur after through eradication by ligation.

The de-novo development of hypertensive gastropathy has been reported after variceal eradication by sclerotherapy [8]. Hypertensive lesions in the gastric mucosa can also occur after variceal ligation therapy. Pressure measurements by splenoportography have demonstrated an increase in portal venous pressure after ligation which can trigger the formation of further varices.

Endoscopic and ultrasound studies have meanwhile confirmed the extent of gastric varices in the gastric cardia and fundus and the development of hypertensive gastropathy.

The findings on whether the actual risk of bleeding rises with the development of these de-novo bleeding sources have not been sufficiently evaluated.

Ligation of cardiac varices.-

Endoscopic ligation of gastric varices is generally possible in the regions of the gastric: cardia, fundus and upper corpus. In the case of high-caliber venous channels, ligation is usually inadequate and can result in banding bleeding. Histoacryl can be regarded as first-line treatment in gastric varices since the tissue glue injection promptly and reliably stops the bleeding [38].

To date, the literature contains too few reports (mostly in abstract form) on ligation in the treatment of cardiac varices.

Personal results.-

From July 1993 to February 1995, we treated 29 patients with bleeding esophageal varices using the ligation method. Follow-up data from a period of 7 to 19 months (mean: 15 months) are available on 27 patients (Tab. 9). These data derive from 20 men and 7 women ranging in age from 21 to 75 years (mean=50 years) with Child class A (n=13), Child class B (n=10) and Child class C (n=4). With regard to pre-existing gastric diseases, two patients had hypertensive gastropathy and 10 patients predominantly had varices of the fundus and cardia (Tab. 10). Endoscopic variceal ligation was performed electively in 10 (37%) and during acute bleeding in 17 (62%). Most of the cases involved grade III varices (n=18). Treatment was concluded after a mean of two sessions [1-6] with a mean number of 6 ligations [2-24]. The gastric varices of two patients were treated with Histoacryl injections. Complete eradication was only successful in 33% (n=9) of the patients.

The complications comprised: bleeding, local necrosis, sclerosis-induced ulcers, one case of esophageal perforation with mediastinitis and resultant death. During follow-up, 2 of the patients died from the sequelae of metastatic malignancies 3 and 9 months after the variceal ligation. No rebleeding occurred in these patients, but was otherwise around 22%. Even at follow-up examinations and in cases of rebleeding, variceal ligation was adhered to, proving generally more comfortable than sclerotherapy (Tab. 11). Two patients underwent surgery, one receiving a liver transplant, the other a splenorenal shunt.

De-novo gastropathy developed in 5 cases, but did not lead to hemorrhage (Tab. 12). Our department refrained from conducting any further comparative studies of sclerotherapy versus ligation, esophagoscopic variceal ligation proving to be the more practical and technically cleaner method. Sclerotherapy was only used if aspiration of the variceal channels was not successful in the pretreated esophagus.

TAB. 12: Gastric involvement before and after EVL.

Conclusion.-

Endoscopic variceal ligation is an effective method for treatment of bleeding esophageal varices. The procedure is easy to learn, its handling is simple and ¡t causes fewer complications than conventional sclerotherapy. Some disadvantages are the limited visibility through the loaded adapter and the fact that the endoscope has to be repeatedly inserted and withdrawn to load new rubber bands. The experience gained with regard to the complications caused by the overtube indicate that its use should be limited.

Endoscopic ligation produces rapid eradication of varices and can probably reduce the risk of rebleeding. The divergent results of randomized studies do not yet permit a final statement on the efficacy and safety of this method, but only indicate a trend towards ligation becoming the preferred treatment (Tab. 13).

lt remains to be seen whether the initially optimistic long-term results from the American study will also be confirmed by more recent reports. Not until this is the case will it be possible to speak of a replacement for conventional sclerotherapy.


Ver Bibliografía