Variceal Upper GI Bleeding



Fig. 3.1
Natural history of portal hypertension: annual risks of varices and variceal



Endoscopy plays a critical role in every aspect of the management of patients with varices and variceal hemorrhage including hemostasis of acute hemorrhage, screening, and primary and secondary prophylaxis. While many modalities are under investigation, there are currently no reliable methods of predicting which cirrhotic patients will have esophageal varices without endoscopy [10]. As such, all patients newly diagnosed with cirrhosis should undergo screening endoscopy to assess for varices (Fig. 3.2). The screening endoscopy will determine who to treat, with what to treat, and the how long until the next surveillance endoscopy.

A352733_1_En_3_Fig2_HTML.gif


Fig. 3.2
Large (grade 3) esophageal varices

Options for prophylaxis against recurrent bleeding from esophageal varices include non-selective beta blockers and/or endoscopic esophageal variceal ligation (EVL). Non-selective beta blockers (propranolol and nadolol) reduce portal pressures via beta-1 adrenergic blockage of cardiac output and beta-2 adrenergic blockage of splanchnic vasoconstriction. The combination reduces portal pressures. The dose of NSBBs is adjusted to a target heart rate of 50–55 or a decrease in 25% of baseline heart rate. Esophageal variceal ligation is a local therapy at the level of the varices themselves and has no effect whatsoever on portal pressures. As such, NSBBs are in theory useful for the prevention of the two other forms of decompensation from portal hypertension (ascites, hepatic encephalopathy), while EVL prevents only variceal bleeding.


Primary Prophylaxis (Fig. 3.3)


If no esophageal varices are detected on initial screening endoscopy, the suggested surveillance interval is 2–3 years, depending on whether the offending etiology of cirrhosis is resolved or not. If, for example, the offending etiology is alcohol or HCV, and the patient is abstinent or cured, respectively, then an interval of 3 years is acceptable. Non-selective beta blockers are not recommended as a large multicenter randomized controlled trial showed no differences between placebo and non-selective beta blockers in preventing the development of varices [11]. The goal for these patients is to prevent decompensation by treating the root cause of cirrhosis and any other offending processes—including obesity, alcohol, and viral hepatitis. Recent studies suggest statins may also be of benefit in cirrhosis by decreasing fibrogenesis, improving liver microcirculation, and decreasing portal pressure in cirrhosis [12]. If at any time these patients develop decompensated cirrhosis—ascites, hepatic encephalopathy, or variceal bleeding—they should undergo EGD at that time, and EGD should be repeatedly annually. Simply speaking, once a cirrhotic patient develops ascites, variceal hemorrhage, or hepatic encephalopathy, they should undergo annual variceal screening on a lifelong basis or until they undergo liver transplantation.

A352733_1_En_3_Fig3_HTML.gif


Fig. 3.3
Primary prophylaxis algorithm

If esophageal varices are detected on screening EGD, three specific parameters should be noted and documented, as they determine risk of hemorrhage (high versus low) and thus the choice of prophylaxis:


  1. (1)


    Size (small, <5 mm; large, >5 mm)

     

  2. (2)


    Red wale marks or other high-risk stigmata (cherry red spot, white nipple sign)

     

  3. (3)


    Child-Pugh Class.

     

High-risk patients are defined as patients with:



  • Medium/large varices (>5 mm) regardless of red wale or Child Class



    • or


  • Small varices (<5 mm) with red wale signs/stigmata



    • or


  • Small varices (<5 mm) in a Child C patient.

Low-risk patients are defined as patients with:



  • Small varices (<5 mm) without red wale signs/stigmata in a Child A patient.

Primary prophylaxis in high-risk patients can administered be via either non-selective beta blockers (NSBB) or (not and) endoscopic variceal ligation (EVL). The decision should be a function of resources, expertise, patient preference, and risks. (Table 3.1).


Table 3.1
Pros and cons of NSBB versus EVL



















 
NSBB

EVL

Pros

1. Low cost

2. No repeat EGD risk (never again)

3. Reduces portal HTN (ascites/HE)

1. Done at screening EGD

2. Fewer contraindications

3. Fewer side effects

4. No titration

Cons

1. 15% intolerance

2. 15% contraindication

3. Under dosing (HR > 55)

4. Indefinite therapy

1. Post-EVL ulcer bleeding

2. Repeat EGD risk

3. Cost of endoscopy

4. Does not reduce portal HTN (ascites/HE)

The pros of NSBB are low cost and none of the risks of endoscopy. Most significantly, once NSBB is initiated appropriately, no repeat endoscopy is required. The cons include a relative or absolute contraindication rate of over 15% (bradycardia, hypotension, peripheral vascular disease, asthma, insulin-dependent diabetes, refractory ascites, SBP), while another 15% of patients simply stop due to intolerance from side effects (most commonly dizziness, fatigue, weakness, sexual side effects, or dyspnea). NSBB should likely be avoided in patients with refractory ascites or spontaneous bacterial peritonitis as studies have shown increased mortality in these subsets of patients [13, 14]. In addition, treatment is life long, as discontinuation results in return of bleeding risk. Propranolol has a higher side effect rate (17%) than nadolol (10%) [15]. The other risk is inadequate dose titration to reach a resting heart rate 50–55 or decrease in baseline heart rate by 25%.

The benefit of EVL is that it can be done at the time of screening EGD, and there is a lower adverse event rates than NSBB (4% versus 13%). However, the adverse events are more serious, specifically bleeding from ligation-induced ulcers, which have in rare cases been fatal. In addition, once started, repeat EVL must be done within 2–4 weeks until varices are obliterated, then another EGD in 1–3 months to ensure obliteration, and finally, every 6–12 months to check for recurrence (Fig. 3.4). In general, it is reasonable to start patients without contraindications on NSBB, and if they cannot tolerate them, switch to the EVL option. Among the subpopulation of patients who have small varices with red wales or Child C Class (Group 2 and 3 above), EVL may be technically more difficult given the smaller variceal size so NSBB is preferred over EVL. Among the subpopulation of patients with medium/large varices without red wale signs or Child C Class, NSBB are preferred over EVL, as well.

A352733_1_En_3_Fig4_HTML.gif


Fig. 3.4
EVL follow up

Primary prophylaxis in low-risk patients is considered optional with non-selective beta blocker to reduce the progression to large varices. There is no long-term evidence to back this recommendation currently. There is no role for EVL. For these patients, upper endoscopy every 1–2 years is recommended. If patients have no signs of decompensation, EGD every 2 years is reasonable. If patients have decompensated cirrhosis, surveillance EGD every year is indicated. As with high-risk patients, if any signs of decompensation develop, they are to have endoscopy at the time of decompensation and repeat endoscopy every year.


Acute Esophageal Variceal Hemorrhage


The patient with active variceal hemorrhage should be admitted to the intensive care unit for immediate circulatory resuscitation, intubation in patients with hematemesis or severe encephalopathy or difficulties protecting their airway, and close hemodynamic monitoring. Endoscopy should be done as soon as possible (usually taken to mean <12 h). Crystalloid or albumin infusions should be used to achieve and maintain hemodynamic stability and tissue perfusion. Red blood cell transfusion should aim for a Hgb 7–8 g/dL. Liberal transfusion strategies targeting Hb 9–11 g/dL have been shown in a randomized controlled trial to be associated with increased mortality likely due to increased portal pressures [16]. There are no specific guidelines concerning transfusion targets for INR or platelets in these patients, but their use as adjunctive therapy to achieve hemostasis is reasonable. There is, in general, no coagulopathy or thrombocytopenia so severe as to preclude endoscopic attempts at hemostasis.

Beyond adequate resuscitation, there is perhaps no medical therapy more beneficial than antibiotics in the bleeding cirrhotic patient. Antibiotic prophylaxis decreases sepsis, recurrence of bleeding, and death [17]. The number needed to treat (NNT) is 4 to prevent sepsis and 22 to prevent death [18]. The most commonly used agents are third-generation cephalosporins (Ceftriaxone 1 g IV every 24 h), given high rates of quinolone resistance in the USA and among this patient population who are often on SBP prophylaxis with quinolones.

Along with antibiotics, intravenous vasoactive medications (Terlipressin, Somatostatin, Octreotide) should be started prior to endoscopy. Octreotide is the only available formulation in the USA. It has been shown to control acute hemorrhage and decrease transfusion requirements, but its effect on mortality is less convincing. [19]. Vasoactive medications should be continued for up to 5 days, because this is the peak period of recurrent hemorrhage, though they can be stopped 24 h after the last evidence of hemorrhage. Upon discontinuation of vasoactive medications, the patient should transition seamlessly to NSBBs or TIPS before discharge if appropriate in selected patients. Pre-procedure intravenous erythromycin may help clear the stomach of blood and clots.

There are two endoscopic therapies for esophageal varices—esophageal variceal ligation (EVL) and endoscopic sclerotherapy (EST).


Endoscopic Sclerotherapy (EST)


EST involves injecting a sclerosing agent into the variceal lumen or immediately adjacent to the varix. EST controls acute variceal bleeding in 70% of patients [20]. There are wide range of sclerosants including sodium tetradecyl sulfate, sodium morrhuate, ethanol, polidocanol, and ethanolamine oleate.

Injections begin distally near the cardia, starting below any bleeding site and working upward in a spiral manner. The sclerosant is injected in 1–3 mL volumes at a time per varix via a needle tip catheter through the working channel of the endoscope. The sclerosant induces thrombosis immediately upon entering the vessel, while adjacent injection creates a tamponade effect via edema and inflammation of adjacent tissue. Overtime, these areas develop fibrosis and obliterate the varix. If bleeding occurs immediately upon retraction of the needle, tamponade can be achieved by simply advancing the endoscope into the stomach and using the body of the endoscope to tamponade.

EST is advantageous for its low cost and no need for second oral intubation like EVL. However, it is less effective and, most importantly, has a much higher complication rate than EVL, specifically more early re-bleeding and esophageal strictures. Other adverse events of EST include fever, retrosternal discomfort/pain, dysphagia, esophageal perforation, mediastinitis, pleural effusion, bronchoesophageal fistula, acute respiratory distress syndrome, and infection. Bacteremia is not uncommon following EST. Endoscopic sclerotherapy is reserved for situations in which EVL is not technically feasible or has failed, and is no longer recommended for secondary prophylaxis of variceal hemorrhage.


Esophageal Variceal Ligation (EVL)


Esophageal variceal ligation (banding) is the endoscopic therapy of choice for controlling esophageal variceal hemorrhage (see Video 3.1 and Figs. 3.5 and 3.6). EVL is superior to endoscopic sclerotherapy in regards to recurrent bleeding, local adverse events including ulceration and stricture formation, time to variceal obliteration, and survival [21]. The band ligator consists a friction-fitting transparent cylindrical cap preloaded with elastic bands, which attaches over the endoscope tip; a spool that fits onto the biopsy port via a Velcro strap; and a trip wire connecting the two, which is passed up through the endoscopic channel.

A352733_1_En_3_Fig5_HTML.gif


Fig. 3.5
Active bleeding from an esophageal varix


A352733_1_En_3_Fig6_HTML.gif


Fig. 3.6
View of a varix post-banding

The procedure involves two intubations. First, a thorough diagnostic upper GI evaluation is done to identify all sources of bleeding and specifically to identify the size, number of esophageal varices, and any high-risk stigmata (wales, white nipple, etc.). The endoscope is withdrawn, and the band ligator is attached onto the endoscope. After re-intubation, banding begins distally from the GE junction upwards, often in a spiral pattern. Starting distally allows for complete visualization and avoids the potential risk of dislodging a band during advancement of the endoscope past a previously captured varix. Once the bulging varix is visualized, the tip is pointed toward it and continuous suction is applied, prolapsing the varix into the cap. Once the screen “reds out,” the spool is twisted and the trip wire releases one of the loaded bands over the varix. Active bleeding or high risk stigmata for bleeding (white nipple/fibrin cap or red wale signs) are the ideal targets for banding. During variceal-band ligation, transient bleeding can occur because of rupture of the varix, but this is usually self limited. The procedure is repeated for each column of varices, moving upwards in a spiral fashion until all columns of varices are flattened, which is ideally achieved before reaching the mid-esophagus. Patient can be placed on liquid diet after the procedure for the first 12–24 h, and then advanced as tolerated. Placement of a naso-gastric tube is typically avoided lest the bands are dislodged, though there is no evidence to our knowledge to support this.

After 3–5 days, ligated sites inevitably slough and produce consistent shallow ulcerations from 3 to 7 days after application (Figs. 3.7 and 3.8). Early re-bleeding occurs in about 10–20% of patients and the majority occurs within the first 5 days. The bleeding is typically due to post-EVL induced ulcers in the setting of high portal pressure, which appears to occur more commonly in patients undergoing EVL after an episode of acute bleeding, with reports as high as 14% [22]. Ligation-induced ulcers heal at a mean of 14 days compared with 21 days for those resulting from sclerotherapy [23]. Proton pump inhibitors may decrease post-EVL ulcer size but have not been demonstrated to significantly reduce bleeding risk [24].

A352733_1_En_3_Fig7_HTML.gif


Fig. 3.7
Post-banding ulcers seen two weeks after banding


A352733_1_En_3_Fig8_HTML.gif


Fig. 3.8
In this photograph, a post-banding ulcer is seen at 6 o’clock while numerous red areas called “wale signs” are seen

Endoscopy should be repeated every 2–4 weeks until varices have been eradicated, which typically requires 2–4 sessions [25]. Then a follow-up EGD should be performed in 1–3 months after obliteration and then every 6–12 months thereafter [26].


Salvage Therapy


For patients in whom all the above attempts at hemostasis have failed or who re-bleed (10–20%), balloon tamponade and covered self-expanding metals stents (SEMS) can be temporizing measures until TIPS can be placed (<24 h). Balloon tamponade is effective in temporarily achieving hemostasis in over 80% of cases but carries an alarming 20% mortality rate from aspiration, asphyxiation due to upward migration into the airway, or esophageal perforation. In addition, bleeding recurs after deflation of the balloon the majority of the time [27]. Intubation for airway protection and sedation is a must when using balloon tamponade. Self-expanding metal stents are feasible, effective, and likely safer, but require technical experience and risk migration [2830]. Patients with early re-bleeding after initial EVL should typically have another endoscopic attempt at hemostasis. TIPS is the final salvage pathway for all patients with refractory esophageal hemorrhage regardless of history of hepatic encephalopathy [31]. Success rates are variable and highly based on local expertise.

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Jul 25, 2017 | Posted by in GASTROENTEROLOGY | Comments Off on Variceal Upper GI Bleeding

Full access? Get Clinical Tree

Get Clinical Tree app for offline access