Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Intravenous albumin in spontaneous bacterial peritonitis

Three Part Question

In [adult patients with spontaneous bacterial peritonitis] does [the use of i.v. albumin] decrease [renal impairment and mortality]?

Clinical Scenario

A 40 year old man presents in the Emergency Department complaining of fever and increased abdominal size for the last two days. He is an alcoholic and the clinical exam shows a distended abdomen with dullness in the flanks. An abdominal ultrasound confirms ascitis. A diagnostic paracentesis is done and reveals a polymorphonuclear (PMN) cell count greater than 250/mm3.

Search Strategy

A computerised literature search of the Cochrane and MEDLINE databases was conducted. The bibliographies of all selected articles found which included information on i.v. albumin in spontaneous bacterial peritonitis were reviewed in an attempt to find other relevant articles. Limits to language (english, spanish, portuguese and french), age (adult) and human patients were applied.
MEDLINE search:
("albumins"[MeSH Terms] OR "albumins"[All Fields] OR "albumin"[All Fields]) AND (spontaneous[All Fields] AND bacterial[All Fields] AND ("peritonitis"[MeSH Terms] OR "peritonitis"[All Fields])) AND ("humans"[MeSH Terms] AND (Clinical Trial[ptyp] OR Meta-Analysis[ptyp] OR Randomized Controlled Trial [ptyp]) AND (English[lang] OR French[lang] OR Spanish[lang] OR Portuguese [lang]) AND "adult"[MeSH Terms])

Cochrane search:
(spontaneous bacterial peritonitis):ti AND (albumin):ti

Search Outcome

MEDLINE Search results:
15 papers found; 3 were considered relevant and critically appraised

Cochrane search results:
8 papers found; 2 were considered relevant and critically appraised

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Sort P. et al
199 patients inclusion criteria: 18-80 years; PMN cell count> 250/mm3; absence of findings suggestive of secondary peritonitis 73 patients excluded 3 dropouts 63 patients randomized to the cefotaxime§ group and 63 to the cefotaxime and albumin* group § dose adjusted to renal function * 1,5g/kg within first six hours of enrolment + 1g/ Kg on the 3rd dayRCTrenal impairmentsignificantly lower among the patients treated with cefotaxime and albumin (10% vs 33%, p=0,002)This is a good quality study, with a good protocol and well stablished criteria and outcomes. The sample size was calculated in 50 patients per group to allow detection of a difference of 25% between the two groups, so this is a study with statistical power. I found no selection bias.
in-hospital mortalitysignificantly lower among patients treated with cefotaxime and albumin (10% vs 29%, p=0,01)
at 3 monthsalso lower in the albumin group (22% vs 41%, p=0,03) odds ratio for death associated with treatment with cefotaxime alone=4.5 (95% confidence interval, 1.0 to 20.9)
Choi C.H. et al
42 patients (tense ascitis and ascitic fluid with PMN cell count>250/mm3) Groups: - Group 1 (21 patients). Large volume paracentesis within 24h after diagnosis + diuretics + i.v. albumin (6-8g per liter of removed ascitic fluid) - Group 2 (21 patients). oral diuretics+ i.v. albumin if serum levels were <3g/dL 2 dropouts in group 2 7 days observationRCTserum creatinine levelsIn group 1 the difference was not significant (p=NS) but in group 2 there was an improvement in creatinine levels after 7 days of treatment (1,4±0,1 vs 1,1±0,1; p=0,012)• ethical approval is not mentioned • in patients who did not respond to cefotaxime, antibiotic was modified either according to the in vitro susceptibility or empirically. The different therapeutic approach can influence the mortality rate. • the number of patients in Child- Pugh class C was slightly higher in group 1
in hospital mortalityno significat differences between the two groups (14,3% Group 1 vs 10,5% Group 2; p=NS)
Fernández J. et al
12 patients (18-80 years; PMN cell count >250/mm3 and absence of findings suggestive of secondary peritonitis) intervention: ceftriaxone 2g iv after diagnosis and then 1g iv/ day + i.v. albumin (1,5g/Kg within 12 hours after diagnosis and 1g/Kg on day 3)observationalrenal impairmentsignificant improvement in serum creatinine levels (1,4 (0,7-2,9) vs 0,8 (0,6-1,5); p=0,004) and in blood urea nitrogen levels (21 (9-69) vs 18 (8-55); p=0,02• the lack of a control group is a limitation of this study, because we cannot establish a cause-effect relationship. • the sample is too little to take any valid conclusions about mortality, for example
Sigal S.H. et al
28 patients= 38 episodes Therapeutical protocol: - low risk for renal failure: 15 patients=18 episodes. No albumin administered - high risk for renal failure (bilirubin>68,4μ mol/L or creatinine>88,4 μmol/L): 21 patients=26 episodes. Albumin administered (1,5g/Kg on day 1 and 1g/ Kg on day 3)observationalrenal impairmentno renal impairment in low risk group vs 19% in high risk (no p value available)• we do not have control groups for each risk group, specially the high risk, where there remains the doubt if the albumin administration made any difference (the authors also measured a huge amount of systemic and splanchnic hemodynamics, endogenous vasoactive systems and IL6 at diagnostic time and after resolution in the high risk group: very few of these items had a p value<0,05) • the p value for renal impairment and mortality measures are not available.
mortalityno deaths in low risk group vs 24% mortality in high risk group (no p value available)
Xue H.P. et al
112 patients (22-70 years; PMN cell count in ascitic fluid >250/mm3 and absence of other infections) Groups: - Group 1 (56 patients): treated with ceftriaxone, adjusted to renal function - Group 2 (56 patients): treated with ceftriaxone + i.v. albumin* *0,5-1g/Kg within 6 hours of enrolment and the same amount on day 3 3 weeks observationRCTrenal impairmentsignificantly lower incidence in group 2 (9% vs 34%; p=0,002)• there is no data about the randomization process nor ethical approval • the statistical analysis is not sufficiently detailed • we do not know about dropouts • size of sample was not made • the administered dose of albumin is different from all the other published studies, which makes it difficult to compare results.
mortalitysignificantly lower in group 2 (9% vs 35%; p=0,01)


From all the appraised studies, I think Sort P. et al RCT (1999) is the best one to answer my 3 part question. These authors not only concluded that i.v. albumin decreases renal impairment (RI) and mortality in spontaneous bacterial peritonitis (SBP) but also proved, with a multivariate analyses, that treatment (cefotaxime and albumin or cefotaxime alone) was an independent predictor of RI and in-hospital mortality. Nevertheless, all studies point to the conclusion that albumin administration prevents the deterioration of hemodynamics and renal function in patients with SPB, thus avoiding death. The consensual dose is 1,5g/Kg of body weight soon after the diagnosis and 1g/Kg of body weight on day 3, although Xue H.P. et al (2002) administered a lower dose with the same benefits. Although Sigal S.H. et al trial (2006) is not enough to prove it, it casts doubt as to whether albumin could only be effective in patients at high risk of developing IR. This fact would restrict albumin use and can be of extreme importance, once this is an expensive human derived product. Another remaining question is if albumin is able to improve already existing renal failure.

Clinical Bottom Line

Adult patients with SBP should be treated with i.v. albumin (and antibiotic regimen) to decrease RI and mortality. The consensual regimen seems to be 1.5g/Kg of body weight soon after diagnosis and 1g/Kg of body weigh on the third day.

Level of Evidence

Level 3 - Small numbers of small studies or great heterogeneity or very different population.


  1. Sort P. et al Effect of intravenous albumin on renal impairment and mortality in patients with cirrhosis and spontaneous bacterial peritonitis The New England Journal of Medicine 1999; 341 (6): 403-409
  2. Choi C.H. et al Long-term clinical outcome of large volume paracentesis with intravenous albumin in patients with spontaneous bacterial peritonitis: a randomized prospective study. Journal of Gastroenterology and Hepatology 2005; 20: 1215-1222
  3. Fernández J. et al Effect of intravenous albumin on systemic and hepatic hemodynamics and vasoactive neurohormonal systems in patients with cirrhosis and spontaneous bacterial peritonitis. Journal of Hepatology 2004; 41: 384-390
  4. Sigal S.H. et al Restricted use of albumin for spontaneous bacterial peritonitis. GUT 2007; Apr 56(4): 597-599
  5. Xue H.P. et al Effect of albumin infusion on preventing the deterioration of renal function in patients with spontaneous bacterial peritonitis. Chinese Journal of digestive diseases 2002; 3: 32-34