Abstract

M

B

, A

A

, M

Z

, S

B

,

W

D

Esophageal Varices

Part I – Pathophysiology, Diagnostics, Conservative

Treatment and Prevention of Bleeding

Żylaki przełyku

Część I – Patofizjologia, diagnostyka, leczenie zachowawcze

i zapobieganie krwawieniom

1_{Department of Otolaryngology, Faculty of Dentistry, Silesian Piasts University of Medicine in Wrocław, Poland} 2_{Department of Gastrology and Hepatology, Silesian Piasts University of Medicine in Wrocław, Poland}

3_{Hospital No. 1 of Silesian Piasts University of Medicine in Wrocław, Poland} 4_{Dental Practice, Świdnica, Poland}

5_{Department of Pharmacology, Silesian Piasts University of Medicine in Wrocław, Poland}

Adv Clin Exp Med 2008, 17, 3, 351–357 ISSN 1230−025X

REVIEWS

© Copyright by Silesian Piasts University of Medicine in Wrocław

Abstract

Prevention of the development of esophageal varices is controversial and not fully documented. At the same time, variceal bleeding is the most dangerous complication of portal hypertension. It appears suddenly and presents as rapid, massive hematemesis and leads to hemovolemic shock if not properly treated. The direct cause is often hard to find. Varix rupture may be caused by local trauma or a sudden increase in portal pressure. Therefore it is criti− cal to introduce proper treatment immediately. Cessation of bleeding and restoration of adequate blood volume is life−saving and prepares the patient for upper digestive tract endoscopy. Endoscopy performed immediately (or after controlling possible hypovolemic shock) allows location of the site of bleeding and rapid introduction of causative treatment (Adv Clin Exp Med 2008, 17, 3, 351–357).

Key words:portal hypertension, complication, treatment, prophylaxis.

Streszczenie

Zapobieganie powstawaniu żylaków przełyku jest nadal dyskusyjne i nie do końca udokumentowane. Krwawienie z nich to najniebezpieczniejsze powikłanie nadciśnienia wrotnego. Pojawia się zwykle nagle i przebiega jako gwałtowne, obfite wymioty krwawe, prowadzące, przy braku odpowiedniego postępowania, do wstrząsu. Bezpośrednia przyczyna jest często trudna do ustalenia. Pęknięcie żylaka może spowodować zarówno miejscowy uraz, jak i nagły wzrost ciśnienia wrotnego. Wdrożenie leczenia musi być wówczas jak najszybsze. Powstrzymanie krwawienia i uzupełnienie ilości krwi krążącej ratuje i przygotowuje pacjenta do wziernikowania górnego odcin− ka przewodu pokarmowego. Wykonana natychmiast (ew. po wyprowadzeniu chorego ze wstrząsu) endoskopia umożliwia ustalenie miejsca krwawienia i jak najszybsze wdrożenie leczenia przyczynowego (Adv Clin Exp Med 2008, 17, 3, 351–357).

Słowa kluczowe:nadciśnienie wrotne, powikłanie, leczenie, profilaktyka.

Hematemesis and melena are often the first symptoms of bleeding from the upper digestive tract, i.e. the esophagus, stomach, or duodenum. Despite popular beliefs, these are not rare symp− toms. Reports from intensive care units inform that they are the primary cause of admission in 10% of

bleedings do not cause significant medical conse− quences, but some may be life−threatening. Hematemesis and melena may or may not be pre− sent at the same time and do not provide enough information about the location or extent of the bleeding. Hematemesis can be observed even in duodenal bleeding, while the presence of blood in stool samples depends on the amount of blood reaching the small intestine (minimum: 50 ml). If intestinal passage is accelerated, it is even possible to detect fresh blood in the stool. Digestive tract bleeding may be simulated by bleeding from the upper respiratory tract (blood swallowed from the nasopharynx, nasal cavity, or pharynx). Distressed patients usually cannot provide useful information about the volume of vomit, and the blood almost always reaches the small intestine. Lack of symp− toms of hypovolemic shock does not exclude the possibility of massive hemorrhage (over 30% of blood loss), and the patient’s condition may be quite good at the beginning. Locating the source of bleeding, assessing its intensity, and introducing appropriate treatment have significant impact on the patient’s life [2].

Portal Hypertension (PH)

The celiac system consists of: the superior mesenteric vein, which collects blood from the small intestine and the right part of the colon, the splenic vein, supplied by the inferior mesenteric vein that directs blood from the spleen, left part of the colon, and rectum, and the left gastric vein.

These three vessels connect and create the por− tal vein (or portal system), optionally supplied by pancreatic or vesicular vessels. The portal vein has no valves and it drains blood into the liver, where the blood from the portal vein is mixed with that from the hepatic artery and drains further to the inferior vena cava. The system of these vessels usu− ally differs in architecture. Normal portal pressure varies between 4–8 mm Hg. In case of an obstruc− tion, when natural connections needed to sustain normal flow from the celiac system to the inferior vena cava become insufficient, the portal pressure increases greatly. Smith [3] classified PH into: pre− sinusoidal obstruction: a) extrahepatic (usually por− tal vein obstruction) and b) intrahepatic (such as schistosomatosis, bilharziosis); sinusoidal obstruc− tion: Wilson’s disease, hepatic cirrhosis; extrasinu− soidal obstruction: a) intrahepatic (post−alcoholic cirrhosis), b) extrahepatic (Budd−Chiari syndrome), c) caused by cardiac dysfunctions (pericarditis, right ventricular insufficiency), d) caused by increased blood flow (arteriovenous fistulas).

The etiology of PH among adults in Europe and in the USA is dominated by chronic liver dis−

eases, usually hepatic cirrhosis, with 90% of cases are caused by intrahepatic block. Over 50 years ago, Ludington [4] stated that in 92% of patients with Laennec’s cirrhosis, chronic (1–7 years) alco− holism was the primary cause. In children, obstruction of the portal vein is more common (pre−sinusoidal obstruction in ca. 10%), while extrahepatic block, which could cause obstruction of the hepatical veins due to thrombosis or exter− nal pressure (Budd−Chiari syndrome), is less com− mon [5].

PH is not a separate disease entity, but a com− plex hemodynamic disorder. When it is present for a prolonged period of time it leads to the develop− ment of an extensive network of portosystemic collaterals. One of these collateral systems is through the vessels in the esophageal mucosa that collects blood from the left (coronal) gastric vein. The blood is later drained to the azygos and hemi− azygos veins and later to the superior vena cava. The submucosal vessels of the esophagus are sur− rounded by a small amount of perivascular tissue and do not have valves. These factors make them prone to increasing blood pressure, which leads to their dilatation, called esophageal varices. The blood flow in the varices is bidirectional and changes with the breathing cycle due to the exis− tence of perforating veins arising from the parae− sophageal veins [6].

Esophageal varices (EVs) are visible in endo− scopic examination just under the esophageal mucosa and present as columns of various size and shape. They are usually purple−blue and may look like small lumps or long folds. Varices can usually be found in the lower parts of the esophagus, just above the cardia, but may reach to the aortic arch. The advancement of vascular changes can be clas− sified by various scales describing the stages from the dilatation of “flat” vessels to large varices clos− ing the lumen of the esophagus, covered by thinned mucosa with visible dilated submucosal veins (cherry−red spots) [7]. For simplification, the 40−year−old Conn classification is still used in publications concerning EVs [8]:

1. Grade I – visible only during one phase of respiration/performance of the Valsalva maneuver. 2. Grade II – visible during both phases of res− piration.

3. Grade III – 3–6 mm.

4. Grade IV – varices – > 6 mm.

tality risk. Group I describes patients who do not have significant liver dysfunctions and may not present any clinical symptoms. Group II classifies patients with clinically visible but not advanced hepatic insufficiency that usually increases during bleeding, infections, operative stress, or trauma. Group III consists of patients with persistent, sig− nificant, and heavy hepatic insufficiency and short life expectancy. CPS takes into account total level of bilirubin and serum albumins, the presence of hepatic encephalopathy, and ascites, most serious complications of chronic hepatic insufficiency besides bleeding from the EVs.

Encephalopathy is a neuropsychiatric group of symptoms of various extent. It may present as mild personality disorders or changes in the sleep/wake cycle to a coma or severe mental/intellectual dis− orders. Encephalopathy develops because some of the blood does not pass through the liver due to the presence of porto−systemic fistulas. This leads to insufficient detoxication of the blood from nitrous substances (ammonia) and its toxic influence on neural tissues [10]. The primary source of ammo− nia is the digestive process (bacterial digestion), occurring mostly in the colon. Many treatments were thus based on diminishing the amount of pro− tein in the diet, use of antibiotics (neomycine) or laxatives, or changing the pH of the digestive sys− tem with lactulose [11].

Ascites, which develops in 50% of patients with portal hypertension in hepatic cirrhosis within 10 years, is caused of the decompensation of liver functions. The clinical condition of patients with ascites is usually poor and only half of them have a life expectancies of over two years [12]. High pressure in the sinusoids and low oncotic pressure caused by the loss of serum albumins allow the flu− ids to enter the perisinusoidal space (Disse’s spaces), between the hepatocyte and a sinusoid. The overproduction of lymph in the liver makes its drainage through the thoracic duct impossible. The excessive fluid leaks through the liver capsule and accumulates in the peritoneal cavity. In most cases, ascites can be treated with diuretics; the drug of choice is usually spironolactonum (Spironol, Aldacton). If pharmacological treatment is not suf− ficient, repeated paracentesis or persistent perito− neovenous drainage with Le Venn’s shunt is applied. As these procedures lead to significant loss of albumins and may create a risk of infection, they are less commonly used [13].

Besides primary disorders and ulceration of the duodenal and gastric mucosa, the most common cause of digestive tract bleeding are EVs. They develop in about 50% of patients with hepatic cir− rhosis, and 30% of them present with bleeding with− in two years from the time of diagnosis. One in

seven incidents of bleeding ends in the patient’s death, and 50% of untreated adults die within one year after the first bleeding episode [14, 15]. It is necessary to add that only 10% of bleedings from EVs occur in patients with a healthy liver [16]. For 70% of the patients the first bleeding is not the last, and it returns in a couple of months. In children, where half of the patients present with normal liver structure and function, the risk of bleeding is associated mostly with PH and varies between 22% (hepatic cirrhosis) to 80% (portal vein occlusion). It also seems that children tolerate bleeding better than adults. Massive, life−threaten− ing bleeding occurs less often, and only about 8% of children die within six weeks from the first incident [17]. Significantly less common than EVs are varices penetrating the stomach. The locations of the esophageal and gastric varices (GVs) in the wall of the digestive tract is the main difference between them. GVs are located more deeply, under thicker mucosa. This may explain why bleeding from GVs occurs less frequently, but when it happens it is usu− ally more dangerous and harder to stop [18].

While the mechanism of EV development is well described, the direct cause of the bleeding is often hard to find. Pressure in the portal vein is not the only factor. Tripathi et al. [19] observed recent− ly that 8% of patients with EV bleeding had low (< 12 mm Hg) pressure in the portal vein. One can risk the statement that if esophageal venous vessels (the most critical factor in the whole pathology of portal hypertension) were located deeply enough, the problem of EV bleeding would not exist. Of the many factors that cause the rupture of EVs, the most often mentioned are trauma to the varices by food particles, inflammation caused by gastro− esophageal reflux, and a sudden increase in pressure in the vessels (coughing, vomiting, defecation, physical exercise). EV bleeding usually occurs sud− denly and presents as sudden hematemesis and leads to hypovolemic shock if untreated [18]. Thus all patients with suspected EV bleeding must be hospitalized, preferably in intensive care units. Treatment should be provided in specialized hospi− tals with access to various treatment methods. Since every EV bleeding event is a life−threatening situa− tion, treatment must be introduced as soon as possi− ble and should cover three basic steps: 1) treatment of hypovolemia, 2) localization and termination of bleeding, and 3) preventing recurrence of bleeding.

Diagnostic Methods and

Nonsurgical Treatment

trolytes, and recovery of acid−base homeostasis. Securing the airways often requires intubation. Most patients with EV bleeding also present with hepatic cirrhosis and require intensive treatment of coagulation disorders, substitution of proteins, and hepatic encephalopathy prophylaxis. It is often necessary to provide plasma coagulation factors (V, VIII, IX, X). As the amount of replaced blood or blood−based products increases (greater number of donors) the risk of post−transfusion complica− tions increases. The symptoms often do not corre− late with the extent and/or duration of the bleed− ing. Unmonitored transfusion may lead to increased portal pressure and induce secondary bleedings [20]. Pharmacotherapy provides a chance to reduce the amount of blood and blood− based product transfusions. The optimal moment for endoscopy is when the hypovolemic shock has been brought under control. Diagnostic endoscopy of the esophagus, stomach, and duodenum is mandatory in every single incident of upper diges− tive tract bleeding. It is the fastest diagnostic tool that provides information about the location of the bleeding, it is well tolerated by patients, and pro− vides over 90% of correct diagnoses. The risk of the patients’ life is low (0.59%) [18]. Before endoscopy became a routine evaluation, almost 50% of bleeding sites were not properly identified and almost 30% of patients with PH bleed occa− sionally from other sources. Removing the blood in the stomach not only increases visibility, but also reduces the risk of developing encephalopa− thy in patients with hepatic cirrhosis [21]. During diagnostic endoscopy, endoscopic treatment should also be implemented. If the bleeding is unexpected and extensive, and the patient’s condi− tion is poor, the diagnostic and therapeutic possi− bilities are sometimes limited. Even an experi− enced endoscopist may have difficulty in localiz− ing the bleeding spots or determining its cause.

When assessing the risk of bleeding, two other diagnostic tools play an important role: Doppler ultrasonography and echoendoscopy. Color Doppler ultrasound (CDU) allows one to deter− mine the diameter, speed, and direction of blood flow in the portal vein and also determines the presence of occlusive material (thrombus) in the portal system. CDU can also be used to determine the direction of blood flow in the left gastric vein; when reversed, it is often accompanied by varices of the gastric fundus [22]. In only ca. 5% of all cases can varices of the gastric fundus be found alone, without accompanying EVs (usually caused by portal thrombosis secondary to pancreatic dis− orders). In these cases the flow in the left gastric vein is not reversed and the bleeding risk is low. It has been demonstrated that a speed of reverse

blood flow in the left gastric vein greater than 15 cm/s is associated with large EVs and a higher risk of bleeding. The sensitivity of classic endo− scopic ultrasound sonography (EUS) in diagnos− ing varices of the gastric fundus is twice as high as that of endoscopic evaluation, but only half as sen− sitive in diagnosing EVs. The lower sensitivity of EUS in the detection of EVs is due to two factors. First, the frequency used in classic US varies between 7.5–12 MHz and their focus falls past the varices. Second, filling the balloon with water compresses the varices and results in their disap− pearance or severe diminishment [23]. High−reso− lution ultrasound (20 MHz) introduced thorough the nose or a gastroscope allows the visualization of varices without the necessity of filling the bal− loon. This method proved to be more sensitive and additionally allows one to access the size, diame− ter, and thickness of the varix walls as well as cal− culate their cross−section area [24].

Three methods are available for measuring the pressure inside the varices; their value is a sensitive indicator of bleeding risk. The first method is based on measuring intraesophageal pressure during relaxation and after filling a balloon and compress− ing the varices. The difference usually correlates with intravariceal pressure. The second method is a modification on the first, but the balloon is filled under EUS control until the flow in the varices stops. The third method measures the pressure inside the varices during the formation of the peri− staltic wave caused by swallowing water, and the pressure is measured at the moment of their depres− sion [23]. Another way to assess bleeding risk is to determine the total cross−sectional area of the varices in the distal parts of the esophagus. Miller et al. [25] concluded that an increase in variceal cross−sectional area of 1 cm2_{increases the risk of} bleeding 75 times in one year. An advantage of echoendoscopy is the possibility to visualize periesophageal varices. These are present in almost all patients with PH and are joined with perforating veins. The presence of large periesophageal varices correlates with the risk of recurrent EV bleeding after endoscopic treatment (sclerotherapy and liga− tion). EUS combined with CDU during treatment may be used to control the efficacy of therapy by confirming diminished flow in the varices and per− forating vessels. It may reduce the number of planned and performed procedures.

acceptable comfort, and their placement is techni− cally difficult. Pressure in the balloons should not be higher than 40 mm Hg and should be lowered every 2–4 hours for 10–15 minutes. Using inflated balloons for over 12 hours is considered danger− ous. They may cause aspiration pneumonia, car− dio−respiratory insufficiency, or ulceration or necrosis of the esophageal mucosa. Long−term complications include paraesophageal hiatal her− nia, esophageal constrictions, and adhesion [26].

Celiac blood circulation and pressure levels in the portal system can be controlled with many pharmacological agents: alpha− and beta−blockers (clonidine, propranolol), nitroglycerine, calcium channel blockers (nifedipine), somatostatin and its analogue (octreotide), anti−diuretic hormone (ADH) and its analogues (terlipressin), and prostaglandin inhibitors [27, 28]. The efficacy of balloon tamponade (50–70%) increases with application of ADH by secondary, selective reduc− tion of blood flow in the portal system. Lowering the pressure makes it easier to stop the bleeding. Unfortunately, the therapeutic effect of ADH is sustained for only about 1 hour, and every subse− quent dose is less effective than the previous. It also has several side effects, such as ischemia of the extremities, necrosis at the site of drug deliv− ery, and cardiac and CNS ischemia. It is also con− traindicated for elderly patients and people with coronary heart disease. To prevent the side effects, ADH is often followed by nitroglycerine, which increases the hemodynamic effect of ADH, or sodium nitroprusside [29]. The synthetic analogue of ADH (terlipressin) acts faster, has fewer side

effects, and can be administered in a single dose [36]. In adults, somatostatin is also commonly used in substitution for ADH (it also reduces celi− ac blood−flow) [30, 31]. Somatostatin and octreotide reduce pressure in the portal vein [32]. They are safe to use and do not have many side effects. Jenkins et al. [33] compared the efficacy of somatostatin and ADH in controlling EV bleeding (87% and 74% at the beginning, 67% and 68% after 24 hours, and 53% and 58% after 48 hours). It has also been shown that metoclopramide (by causing contraction of the lower esophageal sphincter) may be useful in controlling EV bleed− ing [34]. It is also recommended to administer antibiotics (i.v.) after confirmation of bleeding from EVs, usually cephalosporin or quinolones, since they reduce the number of associated pneu− monia or peritonitis [35].

In 1980, Lebrec showed that in patients with hepatic cirrhosis and advanced EV, non−selective beta−blockers significantly reduce the risk of bleeding [36]. These agents diminish the size of varices, have positive chronotropic and inotropic effect on the heart muscle, and therefore reduce the blood flow in the portal system [37]. Oral dose of 80–160 mg of propranolol after several hours reduces the pressure in the portal vein by 12% and reduces the flow through the collaterals and azy− gos vein as well as total hepatic blood flow. Therefore, beta−blockers may be used in the chronic reduction of PH, and when introduced early enough are more effective in preventing first episode of bleeding than in recurrent bleedings [36, 38]. This method, although it may require life−long treatment, is cost−effective and safe, especially when well−controlled optimal doses of medication are administered. There is a problem with patients who do not tolerate beta−blockers well enough due to their side effects (hypoten− sion, headaches, vertigo) or with alcoholics who refuse to cooperate. There are also not sufficient data regarding the use of propranolol in children with PH and it is therefore not recommended [39]. The precise dosage of propranolol that reduces the heart rate by 25% is very difficult to determine in children, mostly due to their highly variable phys− ical activity during the day. It is also contraindi− cated in children with cystic fibrosis. Molsidomine in a single 4−mg dose can also be used. It dilates celiac vessels and slows blood flow in the portal vein by 25% [40]. Chronic therapy with beta− blockers has a strong position in the treatment and prophylaxis of EV bleeding. Medications influ− encing the hemodynamic of the digestive system are also important factors in the treatment of EV bleeding.

Conservative treatment often saves the

Fig. 1.Numerous bunches of varices visible in the lumen of a rigid esophagoscope

patient’s life, allows for recovery from shock and the localization of the bleeding site, and prepares for invasive treatment. Reduction of bleeding gives enough time to transport the patient to a bet−

ter−equipped specialized hospital. However, com− plete eradication of EVs can only be achieved by surgical treatment.

References

[1] Chaibou M, Tucci M, Dugas MA, Farrell CA, Proulx F, Lacroix J:Clinically significant upper gastrointesti− nal bleeding acquired in a pediatric intensive care unit: a prospective study. Pediatrics 1998, 102, 933–938.

[2] Ziółkowski B:Krwawienia z górnego odcinka przewodu pokarmowego. Terapia 2003, 139, 46–49.

[3] Smith JL, Graham DY:Variceal hemorrhage: a critical evaluation of survival analysis. Gastroenterology 1982, 82, 968–973.

[4] Ludington LG:A study of 158 cases of esophageal varices. Surg Gyn Obst 1958, 106, 519–526.

[5] McCormick PA, Burroughs AK: Hemodynamic evaluation of portal hypertension. Hepatogastroenterology 1990, 37, 546–550.

[6] Hashizume M, Kitano S, Sugimachi K, Sueishi K: Three−dimensional view of the vascular structure of the lower esophagus in clinical portal hypertension. Hepatology 1988, 8, 1482–1487.

[7] Paquet KJ, Bussing V, Kliems G, Lazar A:Current therapeutic study in bleeding esophageal varices in babies and children and long−term results of endoscopic paravariceal sclerotherapy over twenty years. Eur J Pediatr Surg 1994, 4, 165–172.

[8] Conn HO: Ammonia tolerance in the diagnosis of esophageal varices. A comparison of endoscopic, radiologic, and biochemical techniques. J Lab Clin Med 1967, 70, 442–451.

[9] Binmoeller KF, Soehendra N: Nonsurgical treatment of variceal bleeding: new modalities. Am J Gastroenterol 1995, 90,1923–1931.

[10] Albrecht J, Jones EA: Hepatic encephalopathy: molecular mechanisms underlying the clinical syndrome: J Neurol Sci 1999, 170, 138–146.

[11] Nance FC, Kline DG: Eck’s fistula encephalopathy in germfree dogs. Ann Surg 1971, 174, 856–862.

[12] Gines A, Fernandez−Esparrach G, Monescillo A, Vila C, Domenech E, Abecasis R, Angeli P, Ruiz−Del−Arbol L, Planas R, Sola R, Gines P, Terg R, Inglada L, Vaque P, Salerno F, Vargas V, Clemente G, Quer JC, Jimenez W, Arroyo V, Rodes J: Randomized trial comparing albumin, dextran 70, and polygeline in cirrhotic patients with ascites treated by paracentesis. Gastroenterology 1996, 111, 1002–1010.

[13] Leveen HH, Christoudias G, Ip M, Luft R, Falk G, Grosberg S: Peritoneo−venous shunting for ascites. Ann Surg 1974, 180, 580–591.

[14] Dy SM, Cromwell DM, Thuluvath PJ, Bass EB:Hospital experience and outcomes for esophageal variceal bleeding. Int J Qual Health Care 2003, 15,139–146.

[15] Graham DY, Smith JL:The course of patients after variceal hemorrhage. Gastroenterology 1981, 80, 800–809.

[16] Bosch J, Pizcueta P, Feu F, Fernandez M, Garcia−Pagan JC: Pathophysiology of portal hypertension. Gastroenterol Clin North Am 1992, 21, 1–14.

[17] Bretagne JF, Dudicourt JC, Morisot D, Thevenet F, Raoul JL, Gastard J:Is endoscopic variceal sclerothera− py effective for the treatment of gastric varices. Dig Dis Sci 1986, 31, 505.

[18] Ryan BM, Stockbrugger RW, Ryan JM:A pathophysiologic, gastroenterologic, and radiologic approach to the management of gastric varices. Gastroenterology 2004, 126, 1175–1189.

[19] Tripathi D, Therapondos G, Jackson E, Redhead DN, Hayes PC:The role of the transjugular intrahepatic por− tosystemic stent shunt (TIPSS) in the management of bleeding gastric varices: clinical and haemodynamic corre− lations. Gut 2002, 51, 270–274.

[20] Hintze RE, Binmoeller KF, Adler A, Veltzke W, Thonke F, Soehendra N:Improved endoscopic management of severe upper gastrointestinal hemorrhage using a new wide−channel endoscope. Endoscopy 1994, 26, 613–616.

[21] Bohnacker S, Sriram PV, Soehendra N:The role of endoscopic therapy in the treatment of bleeding varices. Baillieres Best Pract Res Clin Gastroenterol 2000, 14, 477–494.

[22] Komatsuda T, Ishida H, Konno K, Hamashima Y, Ohnami Y, Naganuma H, Asanuma Y, Masamune O:

Color Doppler findings of gastrointestinal varices. Abdom Imaging 1998, 23, 45–50.

[23] Larry S, Miller MD:Endoscopic Ultrasound in the Evaluation of Portal Hypertension. Gastrointest Endosc Clin North Am 1999, 9, 271–285.

[24] Zaman A, Hapke R, Flora K, Rosen HR, Benner K: Factors predicting the presence of esophageal or gastric varices in patients with advanced liver disease. Am J Gastroenterol 1999, 94, 3292–3296.

[25] Miller L, Banson FL, Bazir K, Korimilli A, Liu JB, Dewan R, Wolfson M, Panganamamula KV, Carrasquillo J, Schwartz J, Chaker AE, Black M:Risk of esophageal variceal bleeding based on endoscopic ultrasound evaluation of the sum of esophageal variceal cross−sectional surface area. Am J Gastroenterol 2003, 98, 454–459.

[26] Sauerbruch T, Fischer G:Conservative treatment of upper gastrointestinal bleeding in portal hypertension. Hepatogastroenterology 1991, 38, 350–354.

[28] de Franchis R:Longer treatment with vasoactive drugs to prevent early variceal rebleeding in cirrhosis. Eur J Gastroenterol Hepatol 1998, 10, 1041–1044.

[29] Terblanche J, Krige JEJ, Bornman PC:Endoscopic sclerotherapy. Surg Clin North Am 1990, 70, 341–359.

[30] Jurczyk G:Masywne krwawienie z górnego odcinka przewodu pokarmowego (cz. I). Gastroenterologia 2002, 1, 6–16.

[31] Garcia−Tsao G: Current management of the complications of cirrhosis and portal hypertension: variceal hemor− rhage, ascites, and spontanneous bacterialperitonitis. Gastroenterology 2001, 120, 726–748.

[32] Cello JP, Chan MF: Octreotide therapy for variceal hemorrhage. Digestion 1993, 54, Suppl 1, 20–26.

[33] Jenkins SA, Baxter JN, Corbett WA, Devitt P, Ware J, Shields R:Somatostatin and variceal haemorrhage. Lancet 1984, 8378, 680.

[34] Chojkier M, Conn HO:Esophageal tamponade in the treatment of bleeding varices. A decadal progress report. Dig Dis Sci 1980, 25, 267–272.

[35] Blaise M, Pateron D, Trinchet JC, Levacher S, Beaugrand M, Pourriat JL: Systemic antibiotic therapy pre− vents bacterial infection in cirrhotic patients with gastrointestinal hemorrhage. Hepatology 1994, 20, 34–38.

[36] Lebrec D, Nouel O, Corbic M, Benhamou JP: Propranolol – a medical treatment for portal hypertension? Lancet 1980, 2, 180–182.

[37] Garcia−Pagan JC, Villanueva C, Vila MC, Albillos A, Genesca J, Ruiz−Del−Arbol L, Planas R, Rodriguez M, Calleja JL, Gonzalez A, Sola R, Balanzo J, Bosch J: Isosorbide mononitrate in the prevention of first variceal bleed in patients who cannot receive beta−blockers. Gastroenterology 2001, 121, 908–914.

[38] Hayes PC, Davis JM, Lewis JA, Bouchier IA:Meta−analysis of value of propranolol in prevention of variceal haemorrhage. Lancet 1990, 336, 153–156.

[39] Maksoud JG, Goncalves ME: Treatment of portal hypertension in children. World J Surg 1994, 18, 251–258.

[40] Monnin JL, Vinel JP, Le Quellec A, Pascal JP, Taourel P, Bruel JM, Blanc F, Michel H, Ciurana AJ: Etude par echographie Doppler pulse des effets de la molsidomine et de l’association propranolol−molsidomine sur l’he− modynamique portale, [Evaluation of the effects of molsidomine and propranolol−molsidomine combination on portal hemodynamics by pulsed Doppler ultrasound.] [French]. Gastroenterol Clin Biol 1992, 16, 745–750.

Address for correspondence:

Wojciech Dziewiszek Department of Pharmacology Mikulicza−Radeckiego 2 50−345 Wrocław Poland

Tel.: +48 71 784 14 47

E−mail [email protected]

Conflict of interest: None declared