Pharmacological therapy has the advantages of being generally applicable and capable of being initiated as soon as a diagnosis of variceal hemorrhage is suspected, even prior to diagnostic EGD. A recent meta-analysis of 15 trials comparing emergency sclerotherapy and pharmacologic treatment (vasopressin alone or in combination with nitroglycerin, terlipressin, somatostatin, or octreotide) suggests that pharmacological therapy should be considered the first-line treatment of variceal bleeding.(43) However, the most effective, safe, and widely accepted drugs are somatostatin and terlipressin, neither of which is available in the United States.

Vasopressin and the somatostatin analogue, octreotide, are available in the United States (and are on the VA formulary). The use of vasopressin is limited by the presence of side effects. Its efficacy and safety are significantly improved by the addition of nitrates.(44) Nevertheless, side effects of combination therapy are still higher than those associated with terlipressin or somatostatin.(8) Vasopressin is administered at a continuous infusion of 0.2--0.4 units/minute, and can be increased to a maximum of 0.8 units/minute. It should always be accompanied by intravenous nitroglycerin at a starting dose of 40 mg/minute that can be increased to a maximum of 400 mg/minute, adjusted to maintain a systolic blood pressure >90 millimeters/hectogram (mm/Hg). Continuous infusion of vasopressin/nitroglycerin cannot be recommended for more than 24 hours because of an increased incidence of adverse effects.

Results of trials of octreotide have been controversial, with two recent meta-analyses showing contradictory results. One of them suggests that octreotide has little or no effect when used alone,(8) while the other shows that octreotide improves control of variceal hemorrhage compared with all alternative therapies.(45) Its side-effect profile was similar to placebo or no therapy; however, it did not show a survival benefit. There are two significant flaws of this meta-analysis.

The first is comparing octreotide to other therapies that are not comparable among themselves (no treatment, vasopressin, glipressin, balloon tamponade, and sclerotherapy). The second flaw is excluding the only double-blind placebo-controlled study of octreotide, which showed that octreotide had no effect.(46) Therefore, the efficacy of octreotide in acute variceal hemorrhage remains unclear. Octreotide is probably not useful as a single first-line therapy of acute variceal hemorrhage, but it may be of use as an adjunct to endoscopic therapy.

Endoscopic therapy is highly effective in controlling active hemorrhage and in preventing early rebleeding and has become the gold standard in the management of acute variceal hemorrhage.(27) However, as mentioned above, a recent meta-analysis of 15 trials comparing sclerotherapy with vasoactive drugs (vasopressin, terlipressin, somatostatin, and octreotide) showed no differences in failure to control bleeding, rebleeding, mortality, or transfused blood units.(43) Another meta-analysis (47) compared sclerotherapy and EVL and concluded that both endoscopic therapies appear equally effective in an emergency. One study that specifically addressed the issue of endoscopic therapy in the control of acute variceal hemorrhage showed in fact that EVL was associated with greater efficacy and fewer complications than sclerotherapy.(48)

Combination of pharmacological therapy and endoscopic therapy appears to be the most promising approach in the treatment of acute variceal hemorrhage. The use of pharmacological agents with few side effects allows prolonging therapy to 5 days, the period during which the risk of rebleeding is the highest. In this way, rather than controlling the acute episode (which is achieved by endoscopic therapy), the goal of pharmacological therapy is the prevention of early rebleeding. A recent meta-analysis of eight trials involving 939 patients showed that combined treatment (vasoactive drugs plus sclerotherapy or EVL) improved the initial control of bleeding and 5-day hemostasis (RR, 1.28; 95 percent CI, 1.18-1.39) without differences in mortality or severe adverse events.(49)

Vasopressin is a potent vasoconstrictor with significant side effects related to its vasoconstrictive effect. These include cardiac and peripheral ischemia, arrhythmias, hypertension, and bowel ischemia. As mentioned above, the addition of nitrates reduces the rate of adverse events; however, the combination can only be used continuously for a maximum of 24 to 48 hours to minimize the development of side effects. Octreotide and other somatostatin analogues are safe and can be used continuously for many days (5 days in most trials).

Despite urgent sclerotherapy and/or pharmacological therapy, bleeding cannot be controlled or has an early recurrence in about 10 to 20 percent of patients. Shunt therapy, either shunt surgery (in Child A patients) or TIPS, has proven clinical efficacy as salvage therapy for patients that fail to respond to endoscopic or pharmacological therapy.(50; 51) Although it had been suggested that bleeding from gastric varices was more difficult to control with TIPS than bleeding from esophageal varices, a recent study showed equal effectiveness of TIPS in both situations.(52)

Sclerotherapy is not optimal for patients bleeding from gastric fundal varices. A recent randomized study compared EVL to obliteration with butyl cyanoacrylate in patients actively bleeding from gastric varices.(53) Initial control of hemorrhage, rebleeding rate, treatment-induced ulcers, and survival were all significantly better in patients treated with cyanoacrylate obliteration. Unfortunately, cyanoacrylate is not licensed for use in the United States.

Balloon tamponade is very effective in controlling bleeding temporarily. However, its use is associated with potentially lethal complications and should be limited to patients with uncontrollable bleeding for whom a more definitive therapy (e.g., TIPS) is planned.

Results of ongoing large trials of another somatostatin analogue, lanreotide, should be helpful in establishing the value of somatostatin analogues as adjuncts to endoscopic therapy in the management of acute variceal hemorrhage. Therapies aimed at improving hemostasis (e.g., activated factor VII, antifibrinolytic agents) are ongoing, and results of a recent pilot randomized, placebo-controlled trial using activated recombinant factor VII (rFVIIa) showed a benefit in Child B and C cirrhotic patients.(54)

Candidates are patients who have recovered from an episode of acute variceal hemorrhage, have had no evidence of hemorrhage for at least 24 hours, and in whom pharmacological therapy for the control of acute variceal hemorrhage has been discontinued.

Nonselective beta-blockers and sclerotherapy have reduced variceal rebleeding and death in treated patients compared with untreated controls. In these studies, rebleeding rates of 57 to 63 percent are described in untreated controls compared with rates of 42 to 43 percent in treated patients.(8; 27; 47) A meta-analysis of 10 randomized trials comparing propranolol to sclerotherapy in the prevention of variceal rebleeding shows comparable rates of variceal rebleeding and survival for both therapies, with a significantly higher rate of side effects with sclerotherapy.(8)

Therefore, treatment with nonselective beta-blockers is preferable to sclerotherapy in the prevention of rebleeding. beta-blockers are used at the same doses recommended for prevention of first variceal hemorrhage (see above). As mentioned previously, the combination of a nonselective beta-blocker and ISMN has a synergistic portal pressure-reducing effect and could theoretically be more effective than beta-blockers alone. Only one study has performed a direct comparison between the combination of propranolol plus ISMN and propranolol alone.(55)

This study showed a benefit of combination therapy (33 percent versus 41 percent rebleeding rate), but it was not statistically significant. However, data collected from different randomized clinical trials show lower median rebleeding rates (~33 percent) in patients treated with combined pharmacological therapy compared with rebleeding rates in patients treated with nonselective beta-blockers alone (~50 percent).(8) Therefore, the pharmacological therapy of choice in the prevention of variceal rebleeding is probably the combination of a nonselective beta-blocker and a nitrate.

EVL, compared with sclerotherapy, reduces the rebleeding rate, side effects, and, importantly, mortality.(56) Therefore, EVL is considered the endoscopic treatment of choice in the prevention of variceal rebleeding. EVL sessions are repeated at 7- to 14-day intervals until variceal obliteration, which usually requires two to four sessions. Once eradicated, EGD to evaluate recurrence of varices is usually repeated at 6- to 12-month intervals.

Regarding EVL versus combination pharmacological therapy (beta-blockers plus nitrates), there are three studies showing different results. One study showed a benefit of combination pharmacological therapy,(57) another showed a benefit of EVL,(58) and a third showed no difference among treatment groups, despite a clear tendency in favor of pharmacological therapy.(59) These differences probably reflect the dosage of medications used, patient population, and, ultimately, center expertise.(60) Both therapies would appear to be equivalent. Therefore, the choice between pharmacological therapy and endoscopic therapy will depend on such factors as local expertise, compliance, tolerability, and patient preference.

The side effects of pharmacological therapy are reportedly more frequent with the combination therapy (beta-blockers plus nitrates) than with beta-blockers alone, mostly in terms of headache and weakness.(8) In fact, in the trial that compared both therapies head-to-head, drug discontinuation because of side effects was significantly greater in the combination therapy group (15 percent versus 2 percent).(55)

In patients that are intolerant to combination therapy, nitrates should be discontinued and beta-blockers alone should be continued. Complications of EVL occur in about 14 percent of cases but are usually minor. The most common complication is transient dysphagia and chest discomfort. Shallow ulcers at the site of each ligation are the rule, but they rarely bleed.

Shunt surgery is very effective in preventing rebleeding; however, it markedly increases the risk of hepatic encephalopathy, without an effect on survival.(27; 61) Not surprisingly, recent meta-analyses of 11 trials that compared TIPS to endoscopic therapy show similar results.(62; 63) That is, even though rebleeding is significantly less frequent with TIPS, post-treatment encephalopathy occurs significantly more often after TIPS, without differences in mortality. Additionally, shunt dysfunction occurs quite frequently, with 77 percent of patients requiring balloon angioplasty or re-stenting in the first year.(64) Unfortunately, Duplex sonography is not a sensitive test in predicting the presence of a hemodynamically significant TIPS stenosis. When occlusion is suspected, shunt status should be assessed by venography and direct portal pressure measurements.(65)

Furthermore, a recent trial showed that, even though pharmacological (propranolol plus nitrates) therapy was less effective than TIPS in preventing rebleeding, it caused less encephalopathy, identical survival, and more frequent improvement in Child-Pugh class with lower costs than TIPS.(66) Therefore, TIPS should not be used as a first-line treatment, but as a rescue therapy for patients who have failed pharmacologic plus endoscopic treatment.

Combining endoscopic therapy with pharmacological therapy is rational because non-selective beta-blockers theoretically will protect against rebleeding prior to variceal obliteration and would prevent variceal recurrence. A recent randomized trial demonstrates that the combination of EVL plus nadolol plus sucralfate is more effective in preventing variceal rebleeding than EVL alone,(67) with rebleeding rates of 23 percent and 47 percent, respectively. Currently, it seems reasonable to combine non-selective beta-blockers with ECL in cases where pharmacological therapy or endoscopic therapy have failed.

As mentioned above, EVL has been shown to be superior to sclerotherapy, and has been associated with lower rebleeding rates, a lower frequency of esophageal strictures, and the need for fewer sessions to achieve variceal obliteration.(47; 56) Therefore, sclerotherapy should no longer be considered an adequate therapy to prevent variceal rebleeding.

Once an ascites PMN count of >250/mm³ is detected, and before obtaining the results of ascites or blood cultures, antibiotic therapy needs to be started. The antibiotic that has been most widely used in the treatment of SBP is intravenous cefotaxime with which SBP resolves in around 90 percent of treated patients.(71-73)

Other third-generation cephalosporins, such as ceftriaxone, have been shown to be as effective as cefotaxime in uncontrolled studies.(74; 75) In a controlled randomized trial, the combination of amoxicillin and clavulanic acid administered intravenously was shown to be as effective and safe as cefotaxime in the treatment of SBP.(76) Patients who develop SBP on prophylactic quinolones (see below) have responded as well to cefotaxime as patients not on prophylaxis.(77)

Cefotaxime is on the VA National Formulary but may be restricted at the facility or Veterans Integrated Service Network (VISN) level. However, other third-generation cephalosporins with a similar spectrum of activity, such as ceftriaxone, are available and should be equally effective. The intravenous preparation of amoxicillin and clavulanic acid is not available in the United States but another beta-blocam/beta-blocamase inhibitor combination, such as ampicillin/sulbactam, would have a similar spectrum of activity. The susceptibility patterns of individual practice settings should be taken into consideration when selecting the antibiotic for SBP.

Doses of cefotaxime used in clinical trials have ranged between 2 grams i.v. every 4 hours and 2 grams i.v. every 12 hours. One randomized study compared two different dose schedules of cefotaxime (2 grams every 6 hours versus 2 grams every 12 hours) and showed similar rates of SBP resolution and patient survival with both schedules.(73) Ceftriaxone has been used at a dose of 1 to 2 grams i.v. every 24 hours and ceftazidime at a dose of 1 gram i.v. every 12 to 24 hours. The only study assessing the combination of amoxicillin and clavulanic acid used a dose of 1 gram/0.2 gram i.v. every 8 hours.(76)

Antibiotic treatment can be safely discontinued after the ascites PMN count decreases to below 250/mm³, which was shown to occur in a period of 5 days.(78) Another study shows that 5-day therapy with cefotaxime is as effective as 10-day therapy.(72) Therefore, duration of antibiotic therapy should be for a minimum of 5 days, but given that the median time to SBP resolution in controlled trials is 8 days, this latter duration is probably preferable. A study showed that i.v. ciprofloxacin could be safely switched to oral antibiotics after 2 days of therapy and once a response to therapy is demonstrated by a decrease in ascites PMN.(79)

The antibiotics recommended above have been associated with very few side effects and no renal toxicity. Cirrhotic patients have an increased propensity to develop aminoglycoside-induced nephrotoxicity and, therefore, aminoglycosides should be considered as a last resort in the therapy of infections in cirrhotic patients.(80)

In patients with community-acquired, uncomplicated SBP (i.e., no renal dysfunction, no encephalopathy), a randomized controlled trial showed that oral ofloxacin (or another fully absorbed quinolone) is a good alternative.(81) Ofloxacin is not on the VA National Formulary, but other quinolones, such as levofloxacin, although not investigated in clinical trials, could be used. These other quinolones would have a theoretical benefit, given a broader coverage of gram-positive organisms, which are increasingly the cause of bacterial infections in cirrhosis.(82; 83) However, the use of quinolones for treatment of SBP will depend on the local prevalence of quinolone-resistant organisms.

Cirrhotic patients are particularly prone to develop nephrotoxicity from aminoglycosides and, therefore, their use should be avoided.(80) Because large volume paracentesis (LVP) can be associated with vasodilatation (84) and theoretically can contribute to precipitating renal dysfunction in patients with SBP (who are already predisposed because of the presence of a bacterial infection), the performance of LVP should be delayed until after the resolution of SBP. Likewise, medications that can potentially decrease effective intravascular volume, such as diuretics, should be avoided during acute infection.

Renal impairment, a main cause of death in SBP patients, occurs as a result of a further decrease in effective arterial blood volume that, in turn, probably results from a cytokine-mediated aggravation of vasodilatation. With the objective of determining whether plasma volume expansion can prevent renal impairment, a randomized study comparing cefotaxime and albumin to cefotaxime alone was performed in patients with SBP.(85)

While the rate of infection resolution was the same in both groups, patients who received albumin had significantly lower rates of renal dysfunction (10 percent versus 33 percent), in-hospital mortality (10 percent versus 29 percent), and 3-month mortality (22 percent versus 41 percent) compared with patients who did not receive albumin. The inpatient mortality rate of 10 percent is the lowest described so far for SBP.

The dose of albumin used was arbitrary--1.5 grams per kilogram of body weight (g/kg) during the first 6 hours, followed by 1 g/kg on day 3. The group of patients that appeared to be more likely to benefit from the addition of albumin had a serum bilirubin >4 milligrams/deciliter (mg/dL) and evidence of renal impairment at baseline (blood urea nitrogen [BUN] >30 mg/dL and/or creatinine >1.0 mg/dL). Of note, patients assigned to cefotaxime alone had, at baseline, more renal failure and more liver synthetic dysfunction than those randomized to cefotaxime plus albumin.

Although it is not statistically significant, this suggests that patients in the control group were sicker. Perhaps more important, patients randomized to the control group did not receive albumin even if there was evidence of renal dysfunction at baseline or during followup.

Recommending the use of albumin in every patient with SBP depends on the results of confirmatory studies. These studies will better identify subgroups of patients who will benefit from this adjunctive therapy.

In a double-blind, placebo-controlled study, continuous oral norfloxacin was shown to significantly decrease the 1-year probability of developing recurrent SBP from 68 percent (in the placebo group) to 20 percent (in the norfloxacin group).(86) This was even more obvious for the probability of developing SBP caused by gram-negative organisms, which was reduced from 60 percent to 3 percent.

Prophylactic therapy was discontinued after 6 months of therapy and therefore the effect on survival was not determinable. The median survival of patients who develop SBP is around 9 months (87) and, therefore, antibiotic prophylaxis in this setting does not imply an inordinately prolonged period of administration.

The dose of norfloxacin used in the abovementioned study was 400 mg by mouth (PO) QD. Prophylaxis should be continued until liver transplantation or the disappearance of ascites (likely to occur in alcoholics who stop alcohol ingestion).

The development of infections by quinolone-resistant organisms is the main complication of long-term norfloxacin prophylaxis. A recent study, which was performed in a large number of cirrhotic patients hospitalized with an infection, demonstrated that gram-negative bacteria isolated from patients on long-term quinolone prophylaxis were significantly more likely to be not only quinolone-resistant but also trimethoprim/sulfamethoxazole-resistant compared with those of patients not on prophylaxis.(82)

Norfloxacin is not on the VA National Formulary; however other quinolones with a similar spectrum, such as ciprofloxacin, could be used instead at a suggested dose of 250 mg QD. Another quinolone, levofloxacin, is also an acceptable substitute with the added advantage of gram-positive coverage. A trial of weekly ciprofloxacin has shown efficacy in the prevention of SBP.(88)

However, the study has methodological problems and, additionally, the use of intermittent ciprofloxacin has been related to a higher occurrence of quinolone-resistant organisms in feces.(89) This higher occurrence was confirmed in a more recent study using weekly rufloxacin in which all 12 patients tested had Escherichia coli resistant to quinolones in their feces by the end of the study.(90) Additionally, this study showed that daily norfloxacin was more effective than weekly rufloxacin in preventing recurrent SBP due to Enterobacteriaceae species.

Therefore, quinolones administered weekly cannot be recommended. Another trial using oral trimethoprim/sulfamethoxazole (one double-strength tablet daily, 5 days per week) (91) also showed efficacy in the prevention of SBP. However, this trial included patients who had had an episode of SBP and patients who had never experienced an episode of SBP, hindering the interpretation of these results. Nevertheless, in patients who are unable to take quinolones, this alternative is reasonable.

Candidates are cirrhotic patients with ascites not associated with infection or renal dysfunction.(97) Recommendations for uncomplicated cirrhotic ascites apply to patients with uncomplicated hepatic hydrothorax.

Sodium restriction is recommended for all cirrhotic patients with ascites. Although dietary sodium should be restricted to levels lower than urinary sodium excretion, sodium restriction to 2 g/day (i.e., 88 milliequivalents per day [mEq/day]) is a realistic goal, particularly in an outpatient setting. Patients with a baseline urinary sodium excretion >50 mEq/day may respond to salt restriction alone. Most patients will require the addition of diuretics.

Spironolactone is the diuretic of choice. Spironolactone can be started alone or in combination with furosemide. It has been shown that spironolactone alone is as effective as combination therapy (spironolactone and furosemide). However, dose adjustments are needed more frequently in the combination group because of the development of increases in blood urea nitrogen and/or decreases in serum sodium.(98; 99) Therefore, it is preferable to initiate therapy with spironolactone alone.

Diuretics can lead to a reduction in intravascular volume and to renal dysfunction and should not be initiated in patients with a rising creatinine level. Additionally, diuretics should not be initiated in patients with concomitant complications of cirrhosis known to be associated with decreased effective arterial blood volume, such as variceal hemorrhage and SBP.

In patients who develop renal dysfunction (elevation in creatinine >50 percent to a creatinine >1.5 g/dL) diuretics should be temporarily discontinued and restarted at a lower dose after creatinine returns to baseline. Patients who develop hyponatremia (serum sodium <130 mEq/L) while on diuretics should be managed with fluid restriction and a decrease in the dose of diuretics.

The preferred diuretic schedule is to initiate therapy with spironolactone alone at a single daily dose of 100 mg and to increase it in a stepwise fashion to a maximum of 400 mg/day. Because the effect of spironolactone takes several days, it can be administered in a single daily dose and the dose should be adjusted only every 3 to 4 days. If weight loss is not optimal or if hyperkalemia develops, furosemide is then added at an initial single daily dose of 40 mg, increased in a stepwise fashion to a maximum of 160 mg/day. To minimize complications, weight loss in patients without edema should be maintained at a maximum of 1 pound (lb)/day (0.5 kg/day), while a weight loss of 2 lb/day (1 kg/day) is allowable in patients with edema.

The more commonly described complications of diuretic therapy are renal impairment due to intravascular volume depletion (25 percent), hyponatremia (28 percent), and hepatic encephalopathy (26 percent).(100; 101; 102) Spironolactone is often associated with adverse events related to its antiandrogenic activity, mainly painful gynecomastia.

Potassium canrenoate, one of the major metabolites of spironolactone, has a comparable diuretic effect and a lower antiandrogenic activity and could be used in cases in which gynecomastia and mastalgia are side effects of spironolactone therapy. However, this drug is not available in the United States.

Amiloride, another potassium-sparing diuretic, does not produce gynecomastia and is recommended in patients with intolerable painful gynecomastia, but it has a significantly lower natriuretic effect than spironolactone.(103) Amiloride is used at an initial dose of 20 mg/day and can be increased to 60 mg/day. For patients whose natriuretic response on amiloride is suboptimal it may be worthwhile to attempt retreatment with spironolactone.

LVP, plus intravenous albumin, has been shown to be as effective as standard therapy with diuretics but with a significantly faster resolution and the same or a lower rate of complications.(100; 101; 104) Because this therapy is significantly more expensive and requires more resources than the administration of diuretics, it is reserved for patients not responding to diuretics (see below).

However, in hospitalized patients with moderate to tense ascites in whom other complications have been resolved, it is reasonable to initiate therapy with total paracentesis with concomitant albumin infusion followed by the administration of diuretics. This therapy will accelerate the patient's discharge from the hospital.

Diuretics. Two randomized trials have shown significantly lower efficacy of the loop diuretic furosemide used alone compared with spironolactone alone (98; 105) or with the combination spironolactone/furosemide.(98) When furosemide is used alone, sodium that is not reabsorbed in the loop of Henle is taken up at the distal and collecting tubules because of the hyperaldosteronism present in most cirrhotic patients with ascites. Therefore, furosemide should not be used as the sole agent in the treatment of cirrhotic ascites.

Nonsteroidal anti-inflammatory drugs or aspirin. These drugs blunt the natriuretic effect of diuretics and therefore should not be used in cirrhotic patients with ascites.(106; 107) Cyclooxygenase-2 inhibitors may also be detrimental and their use should be avoided until additional clinical data are available.

Antibiotic prophylaxis. As mentioned above, short-term antibiotic prophylaxis is recommended in cirrhotic patients (with or without ascites) admitted with GI hemorrhage. Long-term prophylactic antibiotics are recommended in patients with ascites who have recovered from an episode of SBP. However, there are insufficient data to support the use of long-term antibiotic prophylaxis in cirrhotic patients with ascites who are not bleeding and who have not had a previous episode of SBP.

In the only placebo-controlled trial of primary prophylaxis of SBP in patients with low-protein ascites, rates of SBP were low and not significantly different between patients treated with norfloxacin (0 percent) and those treated with placebo (9 percent).(108) On the other hand, long-term antibiotic prophylaxis has led to a significant increase in infections due to quinolone-resistant organisms in patients with cirrhosis.(82) Therefore, until future trials identify a higher-risk population in whom SBP can be prevented with antibiotic therapy, long-term antibiotic primary prophylaxis of SBP is not recommended.

A recent double-blind crossover study suggests that the estrogen antagonist tamoxifen at a dose of 20 mg BID may be useful in the management of painful gynecomastia in cirrhotic patients.(109) This needs further investigation.

Candidates are cirrhotic patients with ascites who fail to respond to diuretics (despite adherence to diet and drugs) or who present complications that preclude the administration of adequate doses of these drugs. Recommendations for patients with refractory ascites apply to patients with refractory hepatic hydrothorax, although these patients should undergo in-hospital careful diuretic therapy before the hydrothorax is considered refractory.

Currently, LVP plus albumin is the standard therapy for refractory ascites. In all LVP studies, diuretics are discontinued before and restarted after the procedure. The need to discontinue diuretics prior to LVP has not been well analyzed and, in practice, it is not performed routinely. A trial has shown that the administration of diuretics after LVP is associated with less-frequent recurrence of ascites without any differences in complications. Therefore, sodium restriction and diuretics at the maximal tolerated dose should be used in conjunction with serial LVP.

The need for concomitant administration of intravenous albumin was demonstrated in two trials. The first showed that daily LVP without intravenous albumin is associated with a significantly higher incidence of hyponatremia and renal impairment than LVP with albumin.(111) In another trial, albumin was shown to be associated with a lower incidence of post-paracentesis circulatory dysfunction (PCD) (18 percent) compared with synthetic plasma expanders (38 percent for polygeline and 34 percent for dextran-70).(112) PCD is defined as an increase in plasma renin activity on the sixth day after paracentesis (indicating a decreased effective arterial blood volume). PCD is associated with faster reaccumulation of ascites and a significantly shorter median survival time (10 months versus 17 months).

Because LVP is a local therapy that does not act on any of the mechanisms that lead to the formation of ascites, recurrence of ascites is the rule rather than the exception. The frequency of LVPs is determined by the rate of ascites reaccumulation and, ultimately, by the need to relieve the patient's discomfort. In turn, the rate of ascites reaccumulation depends largely on the patient's compliance with salt restriction and use of diuretics.

Albumin should be administered at a dose of 6-8 g of albumin i.v. per liter of ascites extracted. For paracentesis of less than 5 L, a synthetic plasma expander (Haemaccel® dextran-70) can be used instead of albumin, and it has been suggested that no plasma expansion may be necessary in this setting.(97; 112)

As mentioned above, a complication of LVP, particularly without the concomitant administration of albumin, is PCD, which is characterized by a significant increase in plasma renin activity after paracentesis. PCD appears to be secondary to a worsening in the vasodilatory state.(84) Therefore, LVP should not be performed when there is a worsening in the vasodilatory state of cirrhosis, such as SBP.

TIPS is considered a second-line therapy for refractory ascites. This recommendation is based mostly on the results of two recent large multicenter studies comparing LVP plus albumin to TIPS.(113; 114) Although, as expected, recurrence of ascites after LVP was significantly greater in patients randomized to LVP plus albumin, there were no differences in mortality. There was, however, a higher rate of severe encephalopathy and a higher cost in the group randomized to TIPS. Therefore, TIPS should be relegated to a secondary position in the treatment of refractory ascites and is mainly indicated in patients who require LVP frequently--that is, three or more LVP/month.(97)

Peritoneo-venous shunting (PVS) is an alternative to LVP plus albumin. In two randomized trials comparing LVP plus albumin to PVS, both procedures were shown to be equally effective, to have a similar rate of complications, and to have a comparable survival rate.(111; 115) Because of its high obstruction rate, PVS required longer admissions for shunt revision or for the management of other more serious complications. The use of PVS has been practically abandoned because LVP plus albumin is a simpler procedure that can be performed in an outpatient setting. Additionally, the placement of a PVS may hinder the future placement of TIPS and may complicate liver transplant surgery given its ability to produce peritoneal adhesions. Therefore, PVS is mostly indicated in patients who require LVP frequently and who are not candidates for TIPS or for transplant.

As mentioned above, long-term prophylaxis with norfloxacin is recommended in patients with ascites who have recovered from an episode of SBP. Currently, there is insufficient data to support the use of long-term antibiotic prophylaxis in cirrhotic patients with ascites who are not bleeding and who have not had a previous episode of SBP. In patients with refractory hepatic hydrothorax, the insertion of a chest tube should be proscribed as this will lead to massive fluid losses, a further depletion of the intravascular effective volume, and to renal dysfunction.

Type 1 HRS is characterized by rapidly progressive renal failure with a doubling of serum creatinine to a level greater than 2.5 mg/dL or a halving of creatinine clearance to less than 20 milliliters/minute (ml/min) in less than 2 weeks. The prognosis of type 1 HRS is extremely poor, with a median survival of about 2 weeks.(96) In type 2 HRS, serum creatinine is greater than 1.5 mg/dL and/or creatinine clearance is less than 40 ml/min, but renal failure progresses more slowly and there is a better prognosis.

The first choice therapy for HRS is liver transplantation. Patients with type 2 HRS have a longer survival rate and this improves their chances to obtain a liver transplant. This is not the case for patients with type 1 HRS whose very short survival rate makes the feasibility of liver transplantation very unlikely, unless survival can be increased by short-term temporizing measures. These are the measures that are under investigation and that have, in uncontrolled studies, been shown to improve renal function and to prolong survival slightly.

In small studies, prolonged (15-day) use of arteriolar vasoconstrictors--such as ornipressin,(117; 118) terlipressin,(119) or the combination of octreotide plus midodrine (120) together with volume expansion with albumin--has shown promise in the treatment of type 1 HRS. However, controlled studies involving larger numbers of patients are required before this treatment can be widely recommended.

More recently, the combination of intravenous noradrenaline (at a dose of 0.5-3 mg/hour) in combination with intravenous albumin given for 10 days was shown to reverse HRS in 10 of 12 patients.(121) It is notable that in many patients this therapy has been discontinued without recurrence of HRS, suggesting that these patients experience either an improvement in liver status (as would occur in alcoholic liver disease with abstinence) or the resolution of a transient decompensating factor such as infection.

Small uncontrolled studies also suggest that TIPS may be useful in the treatment of type 1 HRS and type 2 HRS.(122; 123) More recently, a prospective controlled trial using the molecular adsorbent recirculating system (MARS), a modified dialysis method using an albumin-containing dialysate, was shown to improve 30-day survival in eight patients with HRS compared with five untreated controls.(124) Larger controlled trials are required before the use of TIPS or MARS can be recommended.

Renal venodilators, such as prostaglandins and dopamine (at nonpressor doses), have been used in patients with HRS in an attempt to reduce intrarenal vascular resistance, without an obvious benefit.(125-129) The combination of peripheral vasoconstrictors plus renal vasodilators has also failed to improve renal function in patients with HRS.(130)