View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

MANAGEMENT OF ACUTE ESOPHAGEAL VARICEAL HEMORRHAGE

Gin-Ho Lo Department of Medical Education, Digestive Center, E-Da Hospital, I-Shou University, Kaohsiung, Taiwan.

Acute esophageal variceal hemorrhage (AEVH) is a severe complication of . Its management has rapidly evolved in recent years. Traditional methods included vasoconstric- tor and balloon tamponade. Vasoconstrictors were shown to control approximately 80% of the bleeding episodes and are generally used as a first-line therapy. Following the use of vasocon- strictors, endoscopic therapy is often used to arrest the bleeding varices and prevent early rebleeding. A meta-analysis showed that the combination of vasoconstrictor and endoscopic therapy is superior to endoscopic therapy alone for controlling AEVH. Balloon tamponade may be used to achieve temporary control of the hemorrhage in case of severe bleeding. A transjugu- lar intrahepatic portosystemic stent shunt may be needed in patients with refractory acute variceal hemorrhage. Surgical intervention is now widely contraindicated during acute variceal hemorrhage, except for patients with good liver reserve. Conversely, apart from the control of acute variceal hemorrhage, prophylactic antibiotics were shown to be helpful in the prevention of bacterial infection and to prevent early variceal rebleeding. With the introduction of new treat- ment modalities and the measures taken to manage patients with AEVH, the mortality due to AEVH has significantly decreased in recent years.

Key Words: acute variceal hemorrhage, banding ligation, sclerotherapy, vasoconstrictors (Kaohsiung J Med Sci 2010;26:55–67)

Acute esophageal variceal hemorrhage (AEVH) is a is noted to be a requisite in patients with variceal devastating complication of portal hypertension. Eval- rupture [2]. Poor hepatic reserve and the appearance uation of the natural history of AEVH showed that it numerous red color signs on the large varices are is associated with a mortality rate of 40%, and a high important factors predicting AEVH [3]. In Taiwan, incidence of early rebleeding of 30–50% among the for unknown reasons, gastroenterologists have noted survivors [1]. The factors responsible for AEVH are that the incidence of AEVH is higher in the winter not well known. Portal pressure, variceal pressure, season. and the tension of the esophageal variceal wall have There are a handful of treatment modalities avail- been demonstrated to be associated with AEVH. able for the management of AEVH. Vasoconstrictors Hepatic venous pressure gradient (HVPG) >12 mmHg and endoscopic therapy are the most frequently used tools to control AEVH. Appropriate use of these mo- dalities should improve the survival of patients with Received: Oct 22, 2009 Accepted: Oct 23, 2009 AEVH. This article reviewed the relevant literature and Address correspondence and reprint requests to: Dr Gin-Ho Lo, Department of Medical Education, summarized the limitations and advantages of the Digestive Center, E-Da Hospital, 1 Yi-Da Road, various treatment modalities used for AEVH; and to Kaohsiung 824, Taiwan. provide practical guidance for clinicians to manage this E-mail: [email protected] severe medical emergency.

Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 55 © 2010 Elsevier. All rights reserved. G.H. Lo

FIRST AID FOR AEVH Endotracheal tube insertion to prevent aspiration and the use of sedatives may be required in some Upper gastrointestinal hemorrhage originating from critical cases, particularly alcoholic patients with gastroesophageal varices generally bleeds more pro- or withdrawal symptoms. fusely than other non-variceal lesions. Episodes of However, the endotracheal tube should be removed AEVH may present with either and/or once the endoscopic therapy has been performed and tarry stool, . The presence of bloody the vital signs have stabilized. vomitus may represent either more profuse bleeding Because of the critical situations and poor out- or bleeding from a source not far from the mouth, comes associated with AEVH, experts usu- such as the . Before endoscopic confirmation ally recommend that patients with AEVH are managed of the source of bleeding, any history of portal hyper- in intensive care units. However, in our own experi- tension or the use of nonsteroidal anti-inflammatory ence, only a small proportion of AEVH patients in drugs should be evaluated and signs of portal hyper- critical situations such as torrential hemorrhage, im- tension should be ascertained. Tracing the history of pending hepatic failure or actually require care portal hypertension can be helpful to guide decision- at intensive care units. making for first aid because endoscopists are not The outcomes of AEVH have been demonstrated to always available. depend on the severity of , as reflected by the Nasogastric tube insertion and water irrigation is Child-Pugh class, active bleeding during necessary to remove fine material, blood clots, or fresh [6], HVPG > 20 mmHg [7], the presence of bacterial blood for patients with acute upper gastrointestinal infection and impaired renal function [8]. bleeding. The use of a nasogastric tube does not nor- mally increase the risk of further esophageal variceal Vasoconstrictors rupture. Nasogastric tube irrigation can be used to Vasoconstrictors are often used as the first-line therapy assess the severity of hemorrhage, the amount of blood for AEVH because they are usually safe and easy to ap- loss, and to aspirate bloody materials to prevent the ply. Vasopressin was the first vasoconstrictor approved occurrence of hepatic encephalopathy in patients with approximately 20 years ago and had been widely portal hypertension. Moreover, application of naso- used to arrest AEVH. The mechanisms by which vaso- gastric irrigation will provide a clear visual field dur- pressin modulates hemostasis involve direct stimula- ing endoscopy, which is helpful for endoscopists to tion of vasoconstriction of splanchnic vessels with a accurately identify the bleeding . reduced portal pressure, and on the esophageal wall A clear visual field is also required to differentiate blee- and its vessels, to reduce blood flow. The efficacy of va- ding originating from or esophageal sopressin in arresting AEVH varies widely, and was varices in patients with concomitant gastroesophageal reported to range from 9% to 93% [9–14]. Pooled esti- varices. Blood transfusion is usually required in most mates of the trials have yield overall values of about patients with AEVH. The amount of blood required 60–75% [15]. Before the era of the newer vasoconstric- should be carefully evaluated. Most experts suggest tors, vasopressin was generally shown to be benefi- that hemoglobin should be kept at around 8 g/dL cial in controlling AEVH. However, vasopressin is and hematocrit at around 30% [4]. This is necessary associated with a high frequency of complications to avoid over-transfusion, which may accentuate the such as hypertension, hyponatremia, severe arrhyth- disturbances in portal hypertension that lead to con- mia, abdominal cramp, coronary artery spasm or myo- tinual hemorrhage [5]. Overzealous transfusion may cardial infarction, and skin, intestinal or scrotal also induce pulmonary edema and further worsen- ischemia, and has limited its usage and offset its ing of hemodynamic instability. In contrast, patients effectiveness [9–14,16]. The addition of nitroglycerin with AEVH are more likely to show impaired renal to vasopressin was shown to enhance the efficacy of function compared with patients with non-variceal hemostasis and reduce the incidence of associated bleeding, partly due to the frequent association of complications [11,12]. Moreover, hyposensitivity to AEVH with status. Thus, careful monitoring vasopressin during acute variceal bleeding, particu- and evaluation of the necessity and rapidity of blood larly during shock, was noted among cirrhotic patients transfusion in patients with AEVH is mandatory. [17]. After the discovery of newer vasoconstrictors,

56 Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 Management of acute esophageal variceal hemorrhage vasopressin has almost completely been replaced by of terlipressin are similar to those of vasopressin, its other drugs in the treatment of AEVH. incidence and severity are significantly reduced. The Terlipressin is an analog of vasopressin, but is most common adverse events associated with terli- slowly metabolized to vasopressin and is released into pressin include hypertension, bradycardia and abdom- the circulation upon administration. Thus, terlipressin inal pain, and range from 10% to 20% [25,26]. has advantages such as a longer half-life (1.5 hours) Somatostatin is naturally synthesized by the body and a lower frequency of complications. Terlipressin is and can be commercially produced for clinical use. generally administered intravenously with a 2-mg It has a very short half-life, only 3 minutes, compared bolus loading dose, followed by 1–2 mg every 4 or 6 with other vasoconstrictors. Thus, somatostatin should hours. Nitroglycerin can also be administered in com- be administered by continuous infusion, usually 3 mg bination with terlipressin to reduce the occurrence every 12 hours. However, the mechanisms of hemo- of complications. The hemostatic rate for terlipressin stasis by somatostatin are not easily understood. The was reported to range from 19% to 80% [18–22]. The postulated mechanisms include the inhibition of hemostatic effect of terlipressin appeared to be lower glucagon, a weak direct action on esophageal varices, in Chinese patients than in Caucasian patients. How- and reduction of postprandial, or post-hemorrhage ever, because these studies were not compared head- splanchnic hyperemia [27]. The efficacy of somato- to-head, we do not know whether ethnicity is statin for AEVH arrest was around 84%, with early responsible for the differences in efficacy. Terlipressin rebleeding rates of around 20–30%, similar to terli- was usually administered at higher doses (i.e. 2 mg pressin [25,28–30]. The reduction in portal pressure rather than 1 mg) and at a higher frequency (i.e. at was more pronounced when somatostatin was ad- intervals of 4 hours rather than 6 hours) in the trials ministered at a dose of 500 μg/hr than 250 μg/hr [31]. carried out in Western countries. This may account for A controlled study conducted in Spain showed that a the difference between studies. However, Chang et al high dose of somatostatin (500 μg/hr) was superior to showed that the hemostatic rates were similar between a low dose of somatostatin (250 μg/hr) for the control patients receiving high-dose (2 mg) and low-dose of active variceal hemorrhage, whereas both doses (1 mg) terlipressin (53% vs. 48%, respectively) [22]. It had similar hemostatic rates in patients without active has been suggested that terlipressin can be adminis- hemorrhage during endoscopy [32]. The complica- tered at a dose of 1 mg every 4 or 6 hours once the acute tions of somatostatin are not common and are gener- bleeding has been controlled [23]. Until now, terli- ally modest. The most common adverse effects include pressin was the only vasoconstrictor shown to improve hyperglycemia, /vomiting, bradycardia, hyper- the survival of patients with AEVH [20]. However, be- tension and [25,26]. Hyperglycemic hyper- cause that study was industry sponsored and terli- osmolarity nonketotic coma was also encountered in pressin was the only vasoconstrictor that has been our study, although this was rare [26]. tested in an ambulance setting, it is unknown whether Octreotide is an analog of somatostatin with a half- terlipressin remains superior when other vasoconstric- life of approximately 2 hours. Octreotide could be tors such as somatostatin are compared under similar administered either intravenously or subcutaneously. circumstances. A meta-analysis showed that, com- The recommended dose of octreotide is 25–50 μg/hr. pared with placebo, terlipressin reduced rates of mor- Octreotide was shown to prevent postprandial hyper- tality and failure to achieve hemostasis [21]. Since emia and reduced portal pressure [33], but the hemo- terlipressin is a powerful vasoconstrictor, it could be a dynamic effect of octreotide on the reduction of portal first choice for patients with AEVH presenting with pressure is controversial [34]. Although the half-life hypotension. Moreover, terlipressin was reported to be of octreotide is prolonged, rapid desensitization was beneficial in the treatment of noted during hemodynamic studies, and the lowering [24]. Consequently, terlipressin could be considered for of portal pressure was found to last for just a few min- patients presenting with AEVH and renal impairment, utes [35,36]. Theoretically, the transient hemodynamic and potentially those with hepatorenal syndrome. effect may hinder its efficacy; however, the majority However, terlipressin is contraindicated for patients of studies showed that octreotide is as effective as with coronary artery disease, peripheral vascular dis- other vasoactive drugs for AEVH hemostasis [37–39]. eases or profound shock. Although the complications The complications of octreotide are similar to those

Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 57 G.H. Lo of somatostatin. Although a meta-analysis has shown complications (i.e. nasal alar necrosis, esophageal that octreotide compares favorably with other va- ulcers, , airway obstruction and soconstrictors [40], European scholars still regard aspiration pneumonia) with balloon tamponade is ap- octreotide as a drug that requires more studies with proximately 15%, and 6% of complications were fatal positive results [41]. Conversely, octreotide is widely [46]. Because of the potential risks, the balloon tam- adopted in the United States of America to control ponade should only be inserted in patients with en- AEVH. doscopic confirmation of the AEVH and some experts It is now widely believed that vasoconstrictors suggest that it should be only performed in intensive should be administered on the suspicion of an AEVH care unit. Once the balloon tamponade is removed, episode. However, the issue of how long the vaso- endoscopic therapy should be performed to check for constrictors should be administered is less clear [29]. and prevent early rebleeding as soon as possible in Nowadays, vasoconstrictors are used to arrest acute patients who have not yet been treated with endo- hemorrhage and to prevent early variceal bleeding scopic therapy. [23,26]. Some hepatologists have suggested using vasoactive drugs for 5 days during AEVH to reduce Timing of endoscopy episodes of early rebleeding [42]. More studies are It is generally agreed that cirrhotic patients suffering needed to evaluate cost-effectiveness and the optimal from acute upper gastrointestinal hemorrhage should duration of administration. receive endoscopy to confirm the bleeding source. However, the most appropriate timing for performing Balloon tamponade the endoscopic examinations is less clear. For patients Balloon tamponade arrests AEVH by a direct tampon- with AEVH, a wide range of critical situations may ade effect. Two types of balloon have been adopted, be encountered. If patients present with stable vital namely the Sengstaken-Blakemore (SB) tube and the signs and nasogastric irrigation reveals fine material Linton-Nachlas (LN) tube. The gastric balloon in the rather than fresh blood, endoscopic examinations may SB tube can only be inflated with 100–200 mL of air or not be urgently required. These patients can be man- water, whereas the LN tube may be inflated with up aged with vasoconstrictors while at the emergency to 600 mL of water. The pressure of the esophageal room [29] and receive endoscopic examination by balloon may be increased to 30–40 mmHg. The use of endoscopists at a convenient time. However, some a higher pressure may achieve a greater hemostatic patients may bleed constantly during this time. For effect; however, this may increase the risk of compli- patients with rapid bleeding on presentation, it is un- cations [43]. The optimal duration of balloon tam- known whether emergency endoscopy and endo- ponade is believed to be around 12–24 hours. It seems scopic therapy within a very short period of time likely that shorter duration of balloon tamponade is after presentation can improve the outcome. In such better if hemostasis is quickly achieved. The effec- critical situations, it is usually very difficult to perform tiveness of stopping hemorrhage by balloon tam- emergency endoscopy. The latest Baveno Consensus ponade is around 80–94% [44–46]. When bleeding of portal hypertension suggested that endoscopy occurs from esophageal varices, permanent hemo- should be performed within 12 hours for patients stasis was obtained more frequently with the SB tube with AEVH [45]. than with the LN tube (30%). In bleeding gastric Some studies have compared emergency endo- varices, the SB tube generally failed, whereas perma- scopic therapy and vasoconstrictor therapy in AEVH nent hemostasis may be obtained with the LN tube and found that vasoconstrictor was as effective as in 50% of cases [44]. However, the use of balloon endoscopic injection sclerotherapy (EIS) for control- tamponade is very unpleasant for the patients and ling bleeding, but with fewer complications [46–50]. the rebleeding rate is about 50% after the balloon Thus, it was claimed that vasoconstrictor can replace is deflated. Thus, balloon tamponade should be re- EIS as the first-line therapy for AEVH [51]. However, served to patients with massive variceal hemorrhage, a study by our colleagues showed that emergency which cannot be arrested by vasoconstrictors or en- endoscopic variceal ligation (EVL) is superior to doscopic therapy, or when endoscopists are unavail- somatostatin infusion for 48 hours to control AEVH able for emergency endoscopic therapy. The rate of without increased risk of complications [30]. Our

58 Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 Management of acute esophageal variceal hemorrhage study suggested that EVL should be performed within Endoscopic treatment of the bleeding sites of varices 24 hours of AEVH.30 can usually achieve satisfactory results [26].

Incidence of active esophageal variceal hemorrhage during endoscopy ENDOSCOPIC THERAPY: TO SCLEROSE OR Active esophageal variceal hemorrhage during endo- TO LIGATE? scopy is a significant challenge for endoscopists. Unstable vital signs may be encountered and the pro- EIS has been widely used to arrest AEVH for more cedure may be hazardous if the blood transfusion does than 30 years [60–63]. The mechanisms of EIS in the not correct shock status and is associated with pro- control of AEVH are mainly achieved by inducing found hypoxemia. Profuse blood in the esophagus thrombosis of the bleeding varices with a sclerosant. may hinder the identification of the sites of bleeding The sclerosants used in this acute clinical setting and increases the likelihood of endoscopic hemostasis include ethamoline, sotradecol, polidocanol, and alco- failure. Fortunately, the incidence of active variceal hol. EIS is easy to administer and offers a highly effec- hemorrhage during endoscopy is not excessively tive hemostatic rate. Before the advent of EVL, EIS high. It is generally believed that approximately one- was the most widely used endoscopic therapy for the third of cases with portal hypertensive bleeding pre- management of AEVH. The hemostatic rate achieved sented as active variceal bleeding during endoscopy, by emergency EIS ranged from 60% to 100%. Emer- one-third had varices that had already stopped bleed- gency EIS has been compared with vasopressin, ing during endoscopy and the other one-third had terlipressin, octreotide, somatostatin, and balloon bleeding from other lesions [54]. In fact, the incidence tamponade [64]. A meta-analysis of six trials compar- of active variceal bleeding during endoscopy varied ing EIS with vasopressin or terlipressin showed a sig- greatly between studies and was reported to range nificant advantage of EIS in controlling bleeding. from 10% to 100% [6,50,55–59]. One of several reasons However, the difference in controlling of bleeding for this disparity is that some studies considered was not significant between EIS and somatostatin acute and active esophageal variceal hemorrhage inter- or octreotide [64]. Of note, EIS was associated with changeably. Variation in disease severity, the interval more complications than the vasoconstrictors. An- from time zero to the performance of endoscopic exam- other meta-analysis comparing EIS with vaso- ination and prior use of a balloon tamponade or vaso- constrictors in the control of esophageal variceal constrictors may influence the incidence of active hemorrhage also confirmed this trend [53]. Thus, esophageal variceal hemorrhage during endoscopy. In it has been suggested that vasoconstrictors could our own experience, active esophageal variceal hem- replace EIS as the first-line treatment for AEVH [53]. orrhage during endoscopy accounts for about 10% of Injection of the tissue adhesive agent histoacryl, all cases AEVH in recent years [6,26]. It should be which is widely used to arrest gastric variceal bleed- noted that patients with unstable vital signs should ing [65], has also been tested in AEVH with some not receive endoscopic examination. If patients with success [66]. However, the use of tissue adhesive to signs of active upper gastrointestinal hemorrhage treat AEVH is not common, possibly because of the require endoscopic examination, the endoscopists difficult and time-consuming techniques, particu- should be very experienced and blood transfusion larly if all esophageal variceal channels are treated in and oximetry monitoring during the procedure are one session. mandatory. In contrast, EVL has been widely used to treat Conversely, endoscopic diagnosis of AEVH may AEVH since 1986. A small cylinder is attached to the also pose a challenge for inexperienced endoscopists distal end of the endoscope to hold the varices under when the varices are not actively bleeding. Inex- suction before ligating the varices with rubber bands. perienced endoscopists should be familiar with the The mechanism by which EVL arrests AEVH is through criteria of acute esophageal variceal hemorrhage. mechanical strangulation of the bleeding varices. Stigmata of recent variceal hemorrhage such as white Active bleeding during endoscopy is a great chal- nipple signs or hematocystic spots should be sought lenge for endoscopists. Furthermore, some experts and treated as quickly as possible in cases of AEVH. have claimed that the cylinder attached to endoscope

Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 59 G.H. Lo may hinder the detection of the site of bleeding when management of AEVH [47]. EIS could be used in cases using EVL during active bleeding. with rapid bleeding that limits the endoscopist’s Several studies have shown equivalent hemosta- visual field. Only one study has compared EVL and tic rates between EIS and EVL in the control of active somatostatin for the control of AEVH and showed bleeding varices (Table) [67–70]. A meta-analysis of that EVL was superior to somatostatin for controlling these studies showed that EVL is as effective as EIS hemorrhage without increased risk of adverse events for controlling of active bleeding varices [71]. How- (Figure 1) [30]. ever, only a few patients with active bleeding varices were included in these studies that focused on the prevention of variceal rebleeding. We have per- PREVENTION OF EARLY REBLEEDING formed a study specifically aimed at comparing EIS and EVL for the management of active bleeding Early rebleeding is generally defined as an episode of varices. Our trial showed that the hemostatic rate rebleeding that occurs within 6 weeks of the initial at 48 hours achieved by EIS and EVL was 76% and bleed. It is estimated that 40% of episodes of early re- 97%, respectively [56]. Complications were more fre- bleeding occur within 5 days of the initial bleeding, quently encountered in patients treated with EIS than which leads to a high mortality. Consequently, the with EVL. Based on these observations, the 2005 latest Baveno Consensus Conference of Portal Hyper- Baveno Consensus has suggested that EVL should tension defined the time frame of 48–120 hours after be the first choice of endoscopic treatment in the the initial bleeding as very early rebleeding [72]. It is generally believed that the administration of a vasoconstrictor such as terlipressin or somatostatin for Table. Ligation versus sclerotherapy in hemostasis of active variceal hemorrhage* 5 days can be used to prevent very early rebleeding. Previous studies showed that prolonged use of either Study Ligation Sclerotherapy terlipressin or somatostatin can significantly reduce the Stiegmann et al [67] 12/14 (86) 10/13 (77) early rebleeding rate [23,49,51]. However, the require- Laine et al [68] 8/9 (89) 8/9 (89) ment for vasoconstrictors is doubtful if AEVH is suc- Gimson et al [69] 19/21 (91) 21/23 (92) Hou et al [70] 20/20 (100) 14/16 (88) cessfully controlled by EVL. In fact, our study showed Lo et al [56] 36/37 (97) 26/34 (76) that, even in cases with active bleeding, the use of Total 95/101 (95) 79/95 (83) emergency EVL without vasoconstrictors could also *Data presented as number of positive cases/total cases (%). achieve a low incidence of very early rebleeding [56].

100 EVL 90 80 70 SMT 60 50 40 30 p = 0.0001 20 free of treatment failure (%) failure of treatment free

Proportion of patients remaining Proportion 10 0 0122436 48 Hours Patients 62 61 60 60 59 EVL at risk 63 52 48 46 43 SMT Figure 1. Comparison of ligation and somatostatin for the control of acute esophageal variceal hemorrhage. EVL=endoscopic variceal ligation; SMT =somatostatin (Adapted with permission from [30]).

60 Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 Management of acute esophageal variceal hemorrhage

Combination of vasoconstrictor and of AEVH is more valuable for patients with active endoscopic therapy hemorrhage during endoscopy and for endoscopists The mechanisms of hemostasis achieved by vasocon- not experienced at endoscopic therapy. Nowadays, strictor and endoscopic therapy are different; thus, it the standard of therapy for AEVH should be a vaso- is logical to combine vasoconstrictor and endoscopic constrictor in combination with EVL. The use of a therapy in the management of AEVH. Most controlled vasoconstrictor can reduce the incidence of very early studies have demonstrated that the combination of a rebleeding. However, the optimal duration of vaso- vasoconstrictor and EIS is more effective than either a constrictor therapy remains to be elucidated. vasoconstrictor or EIS alone [57,59,73]. A meta-analysis revealed that the combination of a vasoconstrictor and endoscopic therapy achieved 2-day and 5-day DO WE NEED ENDOSCOPIC THERAPY? hemostasis rates of 88% and 77%, respectively, which were significantly higher than that achieved by endo- Previous studies showed that EIS may be associated scopic therapy alone (76% and 58%, respectively) [74]. with substantial complications. Because vasoconstric- However, the proportion of cases with successful tors alone showed good efficacy for hemostasis, it hemostasis achieved by endoscopic therapy alone was suggested that endoscopic therapy may not be was significantly lower than that reported by most needed in patients with no active variceal bleeding experts experienced in endoscopic therapy. Most of during endoscopic examination [2,53,47]. We have these trials were carried out during the era of EIS, performed a controlled study aimed at evaluating except for one trial that used EVL as the endoscopic the safety and efficacy of EVL for the management therapy. A recent study conducted in Spain showed of inactive variceal bleeding during endoscopic ther- that somatostatin in combination with EVL instead of apy. Our study showed that patients with inactive EIS significantly improved the efficacy and safety bleeding at endoscopy and who received a combina- of treating AEVH [75]. However, the survival was not tion of terlipressin infusion for 2 days and EVL had significantly improved in patients treated with com- lower rates of very early rebleeding and treatment bination therapy. Thus, the combination of a vasocon- failure than patients treated with terlipressin alone for strictor and endoscopic therapy for the management 5 days. Moreover, the rate of complications was not

Hematemesis or

Resuscitation Suspected variceal bleed Vasoconstrictor, lactulose, antibiotics

Endoscopy Confirmed esophageal variceal bleed EVL/EIS

Success Failure

Continue vasoconstrictors for 2–5 d S-B tube

Rebleeding No rebleeding Success Failure

Repeat EVL or EIS Prevent rebleeding TIPS or shunt operation

Figure 2. Algorithm of treatment for acute esophageal variceal hemorrhage. EVL = endoscopic variceal ligation; EIS = endoscopic injection sclerotherapy; S-B = Sengstaken-Blakemore; TIPS = transjugular intrahepatic portosystemic stent.

Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 61 G.H. Lo significantly increased in patients who received com- should be reserved for patients after unsuccessful bination therapy. Thus, it can be concluded that a endoscopic treatment. vasoconstrictor in combination with EVL remains the treatment of choice for patients presenting with an inactive variceal hemorrhage at endoscopy [26]. PREVENTION OF COMPLICATIONS

The occurrence of AEVH may result in serious com- RESCUE THERAPY FOR PATIENTS WITH plications other than shock per se. As described UNSUCCESSFUL INITIAL THERAPY above, cirrhotic patients presenting with upper gas- trointestinal hemorrhage are predisposed to develop If the combination of a vasoconstrictor and endoscopic renal function impairments, particularly patients in therapy is unsuccessful, the SB tube tamponade offers shock status. On the other hand, bacterial infections are a temporary measure to arrest bleeding. Because the frequently associated with cirrhotic patients present- incidence of rebleeding after removal of the SB tube is ing with upper gastrointestinal bleeding. The use of high, a more definitive therapy is usually required. An prophylactic antibiotics in cirrhotic patients with alternative endoscopic therapy may be considered. For upper gastrointestinal bleeding was reported to reduce example, if initial EVL was unsuccessful, the endo- bacterial infections by 30% and mortality by 9% [79]. scopist could perform EIS using a sclerosant or per- A study by Hou et al demonstrated that the use of form endoscopic obturation with cyanoacrylate. If prophylactic antibiotics in cirrhotic patients with AEVH severe esophageal ulceration with bleeding induced by could reduce incidence of rebleeding [80]. A controlled the endoscopic therapy is noted, a further session of study suggested that the third-generation cephalo- endoscopic therapy would be very difficult. Shunt sporin is more effective than quinolones in the pre- therapy may be considered in patients with good liver vention of bacterial infections in cirrhotic patients reserve after unsuccessful medical therapy. On the presenting with upper gastrointestinal bleeding [81]. other hand, patients with poor liver reserve and unsuc- However, cephalosporin is expensive. A study by our cessful medical therapy are not suitable for shunt sur- colleagues suggested that first-generation cephalo- gery. Instead, a transjugular intrahepatic portosystemic sporin was as effective as quinolones in the prevention stent shunt (TIPS) could be applied in these patients. of bacterial infections in cirrhotic patients presenting However, the prognosis is generally dismal. A recent with upper gastrointestinal bleeding, but with fewer study suggested that TIPS should be performed early complications [82]. in patients with AEVH whose HVPG is > 20 mmHg because these patients generally do not respond well to standard therapies [76]. However, TIPS can only CARE OF PATIENTS AFTER HEMOSTASIS be performed by a few medical centers in Taiwan. Achieving hemostasis is the most important step to TIPS improve the survival of patients with AEVH. Non- Before the introduction of TIPS, devascularization or etheless, the associated complications, other than infec- shunt operation were the only options available for tion, should also be prevented as much as possible. patients with AEVH that was not amenable to med- One of the precipitating factors for hepatic enceph- ical therapies available at the time. Although surgical alopathy is gastrointestinal hemorrhage. Thus, the procedures are generally effective for hemostasis, prophylactic use of lactulose to clear blood products the associated mortality and morbidity rates limit its present in the is mandatory. Dur- clinical utility. TIPS was developed to treat AEVH ing resuscitation, a large volume of fluid including that was unsuccessfully treated using standard med- normal saline, blood and frozen plasma may be trans- ical therapies. Approximately 90% of uncontrollable fused. The excessive fluid may result in aggravation AEVH could be arrested by TIPS [77,78]. The mortal- of ascites in patients with significant blood loss dur- ity rates range from 27% to 55% in patients receiving ing AEVH. Thus, intravenous fluids should be mini- emergency TIPS. TIPS is an invasive procedure with mized once hemostasis and stable vital signs have high risk of hepatic encephalopathy; hence, TIPS been achieved. Diuretics may be required in patients

62 Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 Management of acute esophageal variceal hemorrhage who develop ascites. It is unknown whether albumin 4. Burroughs AK, Alexandrino P, Cales P, et al. Sore infusion is helpful in these patients. The role of plate- points. In: de Franchis R, ed. Portal hypertension II. let and frozen plasma transfusion in the management Proceedings of the Second Baveno International Consensus Workshop on Definitions, Methodology and Therapeutic of AEVH is also unknown. The use of activated re- Strategies. Oxford: Blackwell Science, 1996;10–7. combinant factor VII, which corrects prothrombin 5. Castaneda B, Debernardi-Venon W, Bandi JC, et al. time in cirrhotic patients, was found to be ineffective The role of portal pressure in the severity of bleeding in decreasing the risk of 5-day hemostatic failure [83]. in portal hypertensive rats. Hepatology 2000;31:581–6. Twelve hours after EVL has achieved hemostasis, the 6. Lo GH, Chen WC, Chen MH, et al. The characteristics patient can commence a liquid diet. This avoids the and the prognosis for patients presenting with actively bleeding esophageal varices at endoscopy. Gastrointest need for intravenous fluid and provides nutritional Endosc 2004;60:714–20. support. 7. Moitinho E, Escorsell A, Bandi JC, et al. Prognostic Conversely, because the frequency of rebleeding value of early measurements of portal pressure in in patients achieving hemostasis is high, the patients acute variceal bleeding. 1999;117: that survive the AEVH should be advised to receive 626–31. preventive measures [84]. Patients with poor liver 8. Goulis J, Armonis A, Patch D, et al. Bacterial infection is independently associated with failure to control bleed- reserve but survive AEVH may be placed on a wait- ing in cirrhotic patients with gastrointestinal hemor- ing list for liver transplantation. rhage. Hepatology 1998;27:1207–12. 9. Fogel RM, Knauer CM, Andress LL. Continuous intra- venous vasopressin in active upper gastrointestinal CONCLUSION bleeding: a placebo controlled trial. Ann Inter Med 1982;96:565–9. 10. Mallory A, Schaefer JW, Cohen JR, et al. Selective AEVH is a medical emergency. Appropriate resusci- intraarterial vasopressin infusion for upper gastroin- tation, early institution of vasoconstrictors and EVL testinal tract hemorrhage: a controlled trial. Arch Surg and concomitant antibiotic prophylaxis, all are ele- 1980;115:30–2. ments to improve survival. For patients with unsuc- 11. Tsai YT, Lay CS, Lai KH, et al. Controlled trial of vaso- cessful vasoconstrictor and EVL therapy could be pressin plus nitroglycerin vs. vasopressin alone in the rescued by EIS, balloon tamponade and TIPS. The treatment of bleeding esophageal varices. Hepatology 1986;6:406–9. choice of the aforementioned modalities for patients 12. Gimson AES, Westaby D, Hegarty J, et al. A randomized with treatment failure should be individualized and trial of vasopressin and vasopressin plus nitroglycerin based on the individual hospital’s available facilities. in the control of acute variceal hemorrhage. Hepatology The management of AEVH should include a team 1986;6:410–3. with well-trained hepatologists, endoscopists, radiol- 13. Kravetz D, Bosch J, Teres J, et al. Comparisons of intra- ogists, surgeons and nurses. Cases with intractable, venous somatostatin and vasopressin infusions in the treatment of acute variceal hemorrhage. Hepatology difficult-to-manage and very severe variceal hemor- 1984;4:442–6. rhage should be transferred to a medical center with 14. Chojkier M, Groszmann R, Atterbury C, et al. A con- access to a greater wealth of treatment options to trolled comparison of continuous intra-arterial and intra- better treat these patients. The treatment algorithm of venous infusions of vasopressin in hemorrhage from AEVH is shown in Figure 2. esophageal varices. Gastroenterology 1979;77:540–6. 15. D’Amico Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-analytic review. Hepatology 1995; REFERENCES 22:332–54. 16. Gogel HK, Sherman RW, Becker LE. Scrotal and abdom- inal skin necrosis complicating intravenous vasopressin 1. Garcia-Tsao G, Sanyal AJ, Grace ND, et al. Prevention and therapy for bleeding esophageal varices. Dig Dis Sci management of gastroesophageal varices and variceal 1985;30:460–4. hemorrhage in cirrhosis. Hepatology 2007;46:922–38. 17. Tsai YT, Lee FY, Lin HC, et al. Hyposensitivity to vaso- 2. Bosch J, Abraldes JG, Groszmann R. Current manage- pressin in patients with B-related cirrhosis dur- ment of portal hypertension. J Hepatol 2003;38:s54–68. ing acute variceal hemorrhage. Hepatology 1991;13: 3. Beppu K, Inokuchi K, Koyanagi N, et al. Prediction 407–12. of variceal hemorrhage by esophageal endoscopy. 18. Lee FY, Tsai YT, Lai KH, et al. A randomized controlled Gastrointest Endosc 1981;27:213–8. study of triglycyl-vasopressin and vasopressin plus

Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 63 G.H. Lo

nitroglycerin in the control of acute esophageal variceal 34. Lin HC, Tsai YT, Lee FY, et al. Hemodynamic evalua- hemorrhage. Chin J Gastroenterol 1988;5:131–8. tion of octreotide in patients with hepatitis B-related 19. Chiu KW, Sheen IS, Liaw YF. A controlled study of gly- cirrhosis. Gastroenterology 1992;103:229–34. pressin versus vasopressin in the control of bleeding 35. Escorsell À, Bandi JC, Andreu V, et al. Desensitization from esophageal varices. J Gastroenterol Hepatol 1990;5: to the effects of intravenous octreotide in cirrhotic pa- 549–53. tients with portal hypertension. Gastroenterology 2001; 20. Levacher S, Letoumelin P, Pateron D, et al. Early 120:161–9. administration of terlipressin plus glyceryl trinitrate 36. Baik SK, Jeong PH, Ji SW, et al. Acute hemodynamic to control active upper gastrointestinal bleeding in effects of octreotide and terlipressin in patients with cirrhotic patients. Lancet 1995;346:865–8. cirrhosis: a randomized comparison. Am J Gastroenterol 21. Ioannou GN, Doust J, Rockey DC. Systematic review: 2005;100:631–5. terlipressin in acute oesophageal variceal haemor- 37. Hwang SJ, Lin HC, Chang CF, et al. A randomized con- rhage. Aliment Pharmacol 2003;17:53–64. trolled trial comparing octreotide and vasopressin 22. Chang TT, Lee FY, Tsai YT, et al. A randomized con- in the control of acute esophageal variceal bleeding. trolled study of low-dose and high-dose terlipressin in J Hepatol 1992;16:320–5. the control of acute oesophageal variceal haemorrhage. 38. Huang CC, Sheen IS, Chu CM, et al. A prospective ran- J Gastroenterol Hepatol 1991;6:481–4. domized controlled trial of sandostatin and vasopressin 23. Escorsell À, Bandi JC, Moitinho E, et al. Time profile of in the management of acute bleeding esophageal varices. the haemodynamic effects of terlipressin in portal hyper- Chang Gung Med J 1992;15:78–83. tension. J Hepatol 1997;26:621–7. 39. Silvain C, Carpentier S, Sautereau D, et al. Terlipressin 24. Hadengue A, Gadano A, Moreau R, et al. Beneficial plus transdermal nitroglycerin vs. octreotide in the con- effects of the 2-day administration of terlipressin in trol of acute bleeding from esophageal varices: a multi- patients with cirrhosis and hepatorenal syndrome. center randomized trial. Hepatology 1993;18:61–5. J Hepatol 1998;29:565–70. 40. Corley DA, Cello JP, Adkisson W, et al. Octreotide for 25. Feu F, Arbol LRD, Banares R, et al. Double-blind acute esophageal variceal bleeding: a meta-analysis. randomized controlled trial comparing terlipressin Gastroenterology 2001;120:946–54. and somatostatin for acute variceal hemorrhage. 41. de Franchis R. Updating consensus in portal hyperten- Gastroenterology 1996;111:1291–9. sion: report of the Baveno III Consensus Workshop on 26. Lo GH, Chen WC, Wang HM, et al. Low-dose definitions, methodology and therapeutic strategies in terlipressin plus banding ligation vs. low-dose terli- portal hypertension. J Hepatol 2000;33:846–52. pressin alone in the prevention of very early rebleeding 42. de Franchis R. Longer treatment with vasoactive drugs of esophageal varices. Gut 2009;58:1275–80. to prevent early variceal rebleeding in cirrhosis. Eur J 27. Abraldes JG, Bosch J. Somatostatin and analogues in Gastroenterol Hepatol 1998;10:1041–4. portal hypertension. Hepatology 2002;35:1305–11. 43. Conn HO. Hazards attending the use of esophageal 28. Burroughs AK, McCormick PA, Hughes MD. Random- tamponade. N Engl J Med 1958;259:701–7. ized, double-blind, placebo-controlled trial of somato- 44. Teres J, Cecilia A, Bordas JM, et al. Esophageal tampon- statin for variceal bleeding. Emergency control and ade for bleeding esophageal varices. Gastroenterology prevention of early variceal rebleeding. Gastroenterology 1978;75:566–9. 1990;99:1388–95. 45. Vlavianos P, Gimson AES, Westaby D, et al. Balloon 29. Calés P, Masliah C, Bernard B, et al. Early administration tamponade in variceal bleeding: use and misuse. BMJ of vapreotide for variceal bleeding in patients with 1989;298:1158. cirrhosis. N Engl J Med 2001;344:23–8. 46. Haddock G, Garden OJ, Mckee RF, et al. Esophageal 30. Chen WC, Lo GH, Tsai WL, et al. Emergency en- tamponade in the management of acute variceal hem- doscopic variceal ligation versus somatostatin for orrhage. Dig Dis Sci 1989;34:913–8. esophageal variceal bleeding. J Chin Med Assoc 2006; 47. Laine L, Planas R, Nevens F, et al. Treatment of 69:60–7. an acute bleeding episode. In: de Franchis R, ed. 31. Villanueva C, Planella M, Aracil C, et al. Hemodynamic Portal Hypertension IV: Proceedings of the Fourth Baveno effects of terlipressin and high somatostatin dose during International Consensus Workshop. Oxford: Blackwell acute variceal bleeding in nonresponders to the usual Publishing, 2006:217–42. somatostatin dose. Am J Gastroenterol 2005;100:624–30. 48. Fiaccadori F, Pedretti G, Biraghi M, et al. Terlipressin and 32. Moitinho E, Planas R, Bañares R, et al. Multicenter ran- endoscopic sclerotherapy control variceal bleeding and domized controlled trial comparing different sched- prevent early rebleeding in cirrhotic patients. Curr Ther ules of somatostatin in the treatment of acute variceal Res 1993;54:519–28. bleeding. J Hepatol 2001;35:712–8. 49. Escorsell À, Ruiz-del-Arbol L, Planas R, et al. Multicenter 33. Albillos A, Rossi I, Iborra J, et al. Octreotide prevents randomized controlled trial of terlipressin versus scle- postprandial splanchnic hyperemia in patients with rotherapy in the treatment of acute variceal bleeding: the portal hypertension. J Hepatol 1994;21:88–94. test study. Hepatology 2000;32:471–6.

64 Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 Management of acute esophageal variceal hemorrhage

50. Jenkins SA, Shields R, Davies M, et al. A multicentre 64. de Franchis R, Primignani M. Endoscopic treatments randomized trial comparing octreotide and injection for portal hypertension. Semin Liver Dis 1999;19: sclerotherapy in the management and outcome of acute 439–55. variceal hemorrhage. Gut 1997;41:526–33. 65. Lo GH, Liang HL, Chen WC, et al. A prospective, ran- 51. Escorsell À, Bordas JM, Ruiz-del-Arbol L, et al. domized controlled trial of transjugular intrahepatic Randomized controlled trial of sclerotherapy versus portosystemic shunt vs. cyanoacrylate injection in the somatostatin infusion in the prevention of early prevention of gastric variceal rebleeding. Endoscopy rebleeding following acute variceal hemorrhage in 2007;39:679–85. patients with cirrhosis. J Hepatol 1998;29:779–88. 66. Sung JJY, Yeo W, Suen R, et al. Injection sclerotherapy 52. Shields R, Jenkins SA, Baxter JN, et al. A prospective for variceal bleeding in patients with hepatocellular randomized controlled trial comparing the efficacy of carcinoma: cyanoacrylate versus sodium tetradecyl somatostatin with injection sclerotherapy in the con- sulfate. Gastrointest Endosc 1998;47:235–9. trol of bleeding oesophageal varices. J Hepatol 1992;16: 67. Stiegmann GV, Goff JS, Michaletz-Onody PA, et al. 128–37. Endoscopic sclerotherapy as compared with endo- 53. D’Amico G, Pietrosi G, Tarantino I, et al. Emergency scopic ligation for bleeding esophageal varices. N Engl sclerotherapy versus vasoactive drugs for variceal J Med 1992;326:1527–32. bleeding in cirrhosis: a Cochrane meta-analysis. 68. Laine L, El-Newihi HM, Migikovsky B, et al. Gastroenterology 2003;124:1277–91. Endoscopic ligation compared with sclerotherapy for 54. Graham DY, Smith JL. The course of patients the treatment of bleeding esophageal varices. Ann Intern after variceal hemorrhage. Gastroenterology 1981;80: Med 1993;119:1–7. 800–9. 69. Gimson AES, Ramage JK, Panos MZ, et al. Randomised 55. Sarin SK, Govil A, Jain AK, et al. Prospective random- trial of variceal banding ligation versus injection scle- ized trial of endoscopic sclerotherapy versus variceal rotherapy for bleeding esophageal varices. Lancet band ligation for esophageal varices: influence on 1993;342:391–4. gastropathy, gastric varices and variceal recurrence. 70. Hui MC, Lin HC, Kuo BIT, et al. Comparsion of endo- J Hepatol 1997;26:826–32. scopic variceal injection sclerotherapy and ligation 56. Lo GH, Lai KH, Cheng JS, et al. Emergency banding for the treatment of esophageal variceal hemorrhage: ligation versus sclerotherapy for the control of active a prospective randomized trial. Hepatology 1995;21: bleeding from esophageal varices. Hepatology 1997;25: 1517–22. 1101–4. 71. Laine L, Cook D. Endoscopic ligation compared with 57. Avgerinos A, Nevens F, Raptis S, et al. Early adminis- sclerotherapy for treatment of esophageal variceal tration of somatostatin and efficacy of sclerotherapy in bleeding. Ann Intern Med 1995;121:280–7. acute esophageal variceal bleeds: the European Acute 72. Burroughs AK, Cales P, Kravetz D, et al. Definition of Bleeding Oesophageal Variceal Episodes (ABOVE) key events—last attempt? In: de Franchis R, ed. Portal randomized trial. Lancet 1997;350:1495–9. Hypertension IV: Proceedings of the Fourth Baveno 58. D’Amico G, Politi F, Morabito A, et al. Octreotide com- International Consensus Workshop. Oxford: Blackwell pared with placebo in a treatment strategy for early Publishing, 2006:11–39. rebleeding in cirrhosis. A double blind, randomized 73. Sung JJY, Chung SG, Yung MY, et al. Prospective, ran- pragmatic trial. Hepatology 1998;28:1206–14. domized study of effect of octreotide on rebleeding from 59. Besson I, Ingrand P, Person B, et al. Sclerotherapy with esophageal varices after endoscopic ligation. Lancet or without octreotide for acute variceal bleeding. 1995;346:1666–9. N Engl J Med 1995;333:555–60. 74. Banares R, Albillos A, Rincon D, et al. Endoscopic treat- 60. Paquet KJ, Feussner. Endoscopic sclerosis and esoph- ment versus endoscopic plus pharmacologic treatment ageal balloon tamponade in acute hemorrhage from of acute variceal bleeding: a meta-analysis. Hepatology esophagogastric varices: a prospective controlled ran- 2002;35:609–15. domized trial. Hepatology 1985;5:580–3. 75. Villanueva C, Piqueras M, Aracil C, et al. A randomized 61. Lo GH, Lai KH, Ng WW, et al. Injection sclerotherapy controlled trial comparing ligation and sclerotherapy as preceded by esophageal tamponade versus immediate emergency endoscopic treatment added to somatostatin sclerotherapy in arresting active variceal bleeding: in acute variceal bleeding. J Hepatol 2006;45:560–7. a prospective randomized trial. Gastrointest Endosc 76. Monescillo A, Martinez-Lagares F, Ruiz-del-Arbol L, 1992;38:421–4. et al. Influence of portal hypertension and its early 62. Westaby D, Hayes PC, Gimson AES, et al. Controlled decompression by TIPS placement on the outcome of clinical trial of injection sclerotherapy for active variceal bleeding. Hepatology 2004;40:793–801. variceal bleeding. Hepatology 1988;9:274–7. 77. Chau TN, Patch D, Chan YW, et al. “Salvage” tran- 63. Sung JJY, Chung SCS, Lai CW, et al. Octreotide infu- sjugular intrahepatic portosystemic shunts: gastric sion or emergency sclerotherapy for variceal hemor- fundal compared with esophageal variceal bleeding. rhage. Lancet 1993;342:637–41. Gastroenterology 1998;114:981–7.

Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 65 G.H. Lo

78. Boyer TD, Haskal ZJ. The role of transjugular intrahep- infections in patients with advanced cirrhosis and atic protosystemic shunt in the management of portal hemorrhage. Gastroenterology 2006;131:1049–56. hypertension. Hepatology 2005;41:386–400. 82. Chen WC, Lo GH, Hsu PI, et al. Quinolones compared 79. Bernard B, Grange JD, Khac EN, et al. Antibiotic pro- with cephalosporin in the prevention of bacterial infec- phylaxis for the prevention of bacterial infection in cir- tion in cirrhotics with upper gastrointestinal bleeding. rhotics with gastrointestinal bleeding: a meta-analysis. Taiwan Gastroenterol 2007;24:82. Hepatology 1999;29:1655–61. 83. Bosch J, Thabut D, Albillos A, et al. Recombinant factor 80. Hou MC, Lin HC, Liu TT, et al. Antibiotic prophylaxis VIIA (RFVIIA) for active variceal bleeding in patients after endoscopic therapy prevents rebleeding in acute with advanced cirrhosis: a multi-centre randomized variceal hemorrhage: a randomized controlled trial. double-blind placebo-controlled trial. J Hepatol 2007; Hepatology 2004;39:746–53. 46(Suppl 1):295A. 81. Fernandez J, Ruiz-del-Arbol L, Gomez C, et al. 84. Lo GH. Prevention of esophageal variceal rebleeding. Norfloxacin vs. ceftriaxone in the prophylaxis of J Chin Med Assoc 2006;69:553–60.

66 Kaohsiung J Med Sci February 2010 • Vol 26 • No 2 急性食道靜脈瘤出血之治療

羅錦河 高雄義大醫院 醫教部

急性食道靜脈出血是可怕的門脈高血壓併發症。其治療近年來已經有長足的進步。傳 統的治療包括血管收縮劑和食道氣球壓迫止血。血管收縮劑大約可以控制 80 ﹪的 出 血，所以通常用來作第一線治療。使用血管收縮劑以後，內視鏡療法常被用來止血及 預防早期再出血。綜合分析顯示，在急性食道靜脈瘤出血的治療，使用血管收縮劑合 併內視鏡的療法比單獨使用內視鏡療法優越。食道氣球壓迫止血則可用來緊急控制洶 湧出血的情況。經頸靜脈肝內血管分流術可用於很難控制的急性靜脈瘤出血的病人。 另一方面，除了控制急性靜脈瘤出血之外，在預防細菌感染以及早期靜脈瘤的再出 血，事先使用抗生素被證明是有用的。隨著新治療方法的出現，近年來急性食道靜脈 瘤出血病人的死亡率已經有了顯著的下降。

關鍵詞：急性靜脈出血，結紮法，靜脈曲張注射，血管收縮劑 （高雄醫誌 2010;26:55–67）

收文日期：98 年 10 月 22 日 接受刊載：98 年 10 月 23 日 通訊作者：羅錦河醫師 高雄義大醫院醫教部 高雄縣燕巢鄉義大醫院醫教部

Kaohsiung J Med Sci February 2010 · Vol 26 · No 2 67