Human albumin infusion for reducing hyponatremia and circulatory dysfunction in liver cirrhosis: A meta-analysis update

Abstract

BACKGROUND

Liver cirrhosis is a progressive disease with high morbidity and mortality requiring effective management strategies to improve patient outcomes. Various therapies including albumin infusion, volume expanders (VEs), and vasoactive agents are used to manage patients with cirrhosis. Despite numerous clinical trials, a comprehensive meta-analysis comparing the effectiveness of albumin infusion against alternative treatments is limited. This study provides the current and comprehensive synthesis of evidence, offering key insights for optimizing therapeutic strategies in patients with liver cirrhosis.

AIM

To systematically update available data on therapies of liver cirrhosis, we performed a meta-analysis to evaluate and compare the clinical efficacy of albumin infusion vs other VEs and vasoactive agents in patients with liver cirrhosis.

METHODS

A literature search from the PubMed and Embase databases (inception till June 2024) focused on hyponatremia (primary outcome) and various outcomes such as gastrointestinal bleeding, hepatic encephalopathy, severe infection, post-paracentesis-induced circulatory dysfunction (PICD), ascites reappearance, spontaneous bacterial peritonitis, hepatorenal syndrome, renal impairment, hospital stay, mortality, and safety was performed. The primary analysis pooled studies that compared albumin infusion with control. In the subgroup analysis, comparisons were made within the stratified treatment categories included in the control group.

RESULTS

Of the 2957 studies retrieved, 31 studies (27 randomized controlled trials and 4 observational studies) comprising 6255 patients were included. Albumin use was significant in reducing odds of hyponatremia [odds ratio (OR) = 0.67; 95% confidence interval (95%CI) = 0.53-0.85] and PICD (OR = 0.38; 95%CI = 0.20-0.71), whereas the reduction in severe infection (OR = 0.55; 95%CI = 0.28-1.07) did not reach statistical significance. In the subgroup analysis, albumin demonstrated a favorable improvement in lowering the incidence of hyponatremia vs inactive/standard medical therapy (OR = 0.54; 95%CI = 0.27-1.09). For PICD, albumin use was significant compared with other VEs (OR = 0.31; 95%CI = 0.11-0.85) but not with vasoconstrictors (OR = 0.63; 95%CI = 0.21-1.91). In the overall subgroup analysis, a significant reduction was observed in hyponatremia (OR = 0.67; 95%CI = 0.53-0.85) and PICD (OR = 0.38; 95%CI = 0.20-0.71).

CONCLUSION

Human albumin has been shown to significantly reduce the incidence of hyponatremia and PICD in patients with liver cirrhosis, whereas its effect on severe infection remains suggestive but not statistically significant.

Key Words: Albumin; Efficacy; Hepatic encephalopathy; Hyponatremia; Liver cirrhosis; Mortality; Paracentesis-induced circulatory dysfunction; Safety

Core Tip: Liver cirrhosis is one of the leading causes of death worldwide, which imposes a substantial public health burden. The efficacy of various therapies, such as albumin infusion, volume expanders (VEs), and vasoactive drugs, used to treat patients with cirrhosis is still not clear. This meta-analysis compares albumin infusion, VEs, and vasoactive agents for the treatment of liver cirrhosis. Albumin infusion significantly decreases hyponatremia and paracentesis induced circulatory dysfunction compared to other treatments. These findings indicate that albumin therapy is more effective against complications of liver cirrhosis compared to other treatments.

INTRODUCTION

Cirrhosis represents the end stage in the evolution of chronic liver diseases, resulting in more than 0.8 million deaths every year worldwide[1,2]. The transition from compensated cirrhosis to decompensated cirrhosis is reflected by the onset of events such as hyponatremia, ascites, hepatic encephalopathy (HE), gastrointestinal (GI) bleeding, and severe jaundice[3,4]. Hyponatremia is a frequent complication of decompensated cirrhosis, and it is the most common electrolyte abnormality observed in hospitalized patients[5]. It is associated with poor prognosis, increased mortality, and neurological complications and has a negative impact on the quality of life of the patients[2,5-8]. Furthermore, ascites in patients with cirrhosis is associated with spontaneous bacterial peritonitis (SBP), hepatorenal syndrome (HRS), and a mortality rate of 50% at 2 years[9,10]. Therapeutic paracentesis is the first line of treatment for patients with cirrhosis and tense ascites, but this invites paracentesis-induced circulatory dysfunction (PICD), which ultimately results in hyponatremia, renal impairment, ascites re-accumulation, and shorter survival[11].

Human albumin is the most abundant protein in the plasma and is an effective intravascular volume expander (VE) owing to its high oncotic activity and prolonged half-life[12]. As albumin is synthesized in the liver, its concentration reduces with hepatic dysfunction[13]. It contributes to 80% of the colloidal osmotic pressure of the plasma, so intravenous administration rapidly increases the circulating blood volume[14]. Evidence suggests that along with the main physiological function of maintaining colloidal osmotic pressure, it also has antioxidant, anti-inflammatory, ligand-binding, and transport functions. In patients with advanced cirrhosis, a decrease in plasmatic albumin because of impaired hepatocellular function with reduced albumin synthesis is observed, which can reach up to a 60% to 80% reduction in advanced cirrhosis[15]. In addition, dilution resulting from the renal retention of sodium and water as well as the increased transcapillary escape rate may contribute to reduced serum albumin[15]. Hypoalbuminemia in turn leads to hypervolemic hyponatremia[5].

Albumin infusions have been extensively used in the past decades in the management of patients with cirrhosis[12]. Current clinical guidelines from the European Association for the Study of the Liver and the American Association for the Study of Liver Diseases recommend the administration of human albumin for specific complications of liver cirrhosis—namely, HRS, SBP, and large-volume paracentesis[16,17], whereas the Japanese and Italian practice guidelines have suggested its use in the management of liver cirrhosis with hyponatremia[18,19].

Effects of albumin have been compared with other means to enhance the effective plasma volume, either fluids or vasoconstrictors, in randomized trials and observational studies. However, the effectiveness of comparison with albumin remains uncertain because of the lack of adequate design and low statistical power of such trials. In addition, there is a paucity of high-quality evidence supporting the use of albumin in other important clinical endpoints of liver cirrhosis such as peripheral edema, overall adverse events, hospital stay days, and hospital mortality. Previous meta-analysis studies have confirmed the benefits of albumin in reducing hyponatremia, post-PICD, GI bleeding, recurrence of ascites, renal failure, HE, and mortality in patients with liver cirrhosis after large-volume paracentesis[10,20-23]; however, these studies focused on specific outcomes such as large-volume paracentesis or infection-related outcomes, lacking a comprehensive evaluation of albumin’s broader clinical impact. Furthermore, the results are heterogenous and controversial. The potential effectiveness of improving circulatory and kidney function by long-term administration of albumin to patients with decompensated cirrhosis has been explored in 2 recent randomized controlled trials (RCTs), both published in abstract form, with contradictory findings[24,25]. In the recent ATTIRE trial[26], the efficacy of targeted albumin infusion for preventing infection, kidney dysfunction, or death in patients hospitalized with decompensated cirrhosis was assessed. The findings demonstrated no significant benefit of albumin over standard care, including no reduction in new infections or endpoint events, even among patients who were admitted with infections or were already receiving antibiotics. In 2023, a meta-analysis by Xu et al[27] reported that albumin infusion did not significantly reduce the mortality in decompensated cirrhosis [hazard ratio = 1.01; 95% confidence interval (95%CI) = 0.97-1.05] and was associated with an increased risk of pulmonary edema [risk ratio (RR) = 3.14]. However, albumin significantly reduced the recurrence of ascites (RR = 0.56; 95%CI = 0.46–0.68), highlighting its role in symptomatic improvement rather than survival. Similarly, Sandi et al[28], focusing on long-term albumin administration in cirrhotic patients with ascites, observed no survival benefit but reported a significant reduction in the need for paracentesis and recurrence of ascites (RR = 0.56). In contrast, Bai et al[29] conducted a more extensive meta-analysis of 42 RCTs, which reported that albumin infusion significantly reduced the overall mortality [odds ratio (OR) = 0.81; 95%CI = 0.67-0.98), particularly in patients with SBP (OR = 0.36) and HE (OR = 0.43). Additional benefits included reductions in renal impairment and ascites, but no significant effects were observed on infections or GI bleeding. The short-term use of albumin showed transient mortality benefit, whereas the long-term therapy did not significantly impact the long-term survival. However, their analysis did not stratify results based on the type of intervention in the control group, which is important, as subgroup analyses would provide greater clarity on albumin’s relative effectiveness by accounting for heterogeneity in control interventions, thus yielding more clinically meaningful and specific conclusions. Thus, there is a significant unmet need to evaluate the preventive role of albumin infusions in the management of overall patients with liver cirrhosis. With the approach of quantitatively combining the results of relevant studies meeting our study objective through meta-analysis, the uncertainty on using albumin infusion for patients with liver cirrhosis will be resolved to a greater extent, thus adding value to the topic.

Hence, this meta-analysis aimed to evaluate the clinical effectiveness of human albumin infusion compared with other therapies including standard medical therapy, VEs, antibiotics, or vasoconstrictors in preventing hyponatremia and other complications in patients with decompensated liver cirrhosis, to determine whether the use of albumin is justified in this population.

MATERIALS AND METHODS

Study design

This meta-analysis was reported as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines[30]. This statistical analysis plan was registered with the Prospective Register of Systematic Reviews (PROSPERO) database (CRD42022372709).

Search strategy

We conducted a systematic literature search using the PubMed and Embase databases from inception to June 2024 to retrieve articles comparing the efficacy and safety of human albumin with other VEs and vasoactive agents in patients with cirrhosis. The search terms included “liver cirrhosis” OR “Cirrhosis” OR “hepatic cirrhosis” OR “hepatic fibrosis” OR “liver fibrosis” OR “liver carcinoma” OR “hepatic carcinoma” OR “ascites” AND “albumin”. Specific filters such as “Clinical Study, Clinical Trial, Randomized Controlled Trial, Observational study” were incorporated into the search string for retrieving the articles. In addition, we searched reference lists of published meta-analyses and reviews related to albumin treatment in liver cirrhosis for eligible studies.

Study selection process

The inclusion criteria were patients aged ≥ 18 years and diagnosed with established liver cirrhosis, irrespective of gender, race, or dose of drugs on albumin treatment. Studies that compared the efficacy and safety of albumin against a control group with desired outcomes were included in the present study. Therefore, studies comprising albumin as the intervention, and inactive/standard medical therapy (I/SMT), VEs, antibiotics, or vasoconstrictors as the control in patients with liver cirrhosis were included in this meta-analysis. We excluded meta-analyses, systematic literature reviews, narrative reviews, case reports, conference proceedings, non-comparative studies, and cost-related studies as well as the results not reporting desired objective and outcomes of interest. Studies published in languages other than English were also excluded. For the primary analysis, the studies were pooled, and albumin infusion was compared with control, whereas for the subgroup analysis, comparisons were made within categories, stratified according to the treatment groups included in the control.

Data extraction and quality assessment

Data were extracted by 2 independent reviewers from eligible studies, mainly including first author, title, study design, year of publication, sample size, characteristics of patients, and endpoints in a standard data extraction form. In the first step, the relevant articles were independently screened based on their titles and abstracts by 2 researchers to identify potential trials. The reasons for the exclusion of studies were documented as well. For secondary screening, full-text versions of selected papers were examined and assessed according to the inclusion criteria. Any differences of opinions were resolved through discussion until a consensus was reached. To minimize bias, a third reviewer was consulted who double-checked the articles and confirmed their eligibility. The quality of the included studies was assessed based on the Jadad scale and Newcastle-Ottawa scale (NOS) quality assessment tool[31,32]. The JADAD score is a 5-point scale, with studies scoring < 3 considered as low quality and those scoring ≥ 3 classified as high quality. The NOS is based on 3 factors with a total of 9 points. Studies scoring > 6 were considered as high quality, and < 6 as poor quality[33].

Study outcomes

The primary efficacy outcome was hyponatremia, others being GI bleeding, HE, severe infection, post-PICD, ascites reappearance, SBP, HRS, renal impairment, hospital stay, mortality, and peripheral edema. Safety outcomes included adverse events and in-hospital mortality.

Statistical analyses

The analysis of statistical data was performed after the completion of validation and quality checks using the R statistical software (version 4.1.2) and the metafor package[34]. The odds ratio (OR) and mean difference were used to estimate the outcome with a 95%CI, wherever applicable. Variation in the effect size because of heterogeneity between studies was quantified using the I² statistic[35]. A random-effects model was applied if the I² value is > 50%, whereas the fixed effects model was used when I² was < 50%. Publication bias was assessed using the Egger’s regression-based test[36]. Sensitivity analyses were performed for those outcomes where the publication bias was observed. Studies causing asymmetry were removed, and the articles were finalized for analysis.

RESULTS

Study characteristics

A total of 2957 studies were retrieved after the removal of duplicates through the search strategy. Furthermore, the titles and abstracts were screened, and 31 articles were found to be potentially relevant based on the full-text screening; of which, 31 were identified for evidence synthesis as per the outcomes of interest. The study process is presented in a detailed PRISMA flow diagram (Figure 1). The measured endpoints for each treatment are provided in Supplementary Table 1. Supplementary material provides the information regarding forest plots and funnel plots pertaining to outcomes and the subgroup analyses.

Figure 1 PRISMA flow diagram of this study.

Of the included articles, 8 RCTs compared albumin with alternative plasma expanders, 7 RCTs assessed albumin vs a vasoconstrictor, 12 RCT (2 studies included antibiotics in control group), and 4 observational studies compared albumin with I/SMT. The basic characteristics and quality assessment scores of the included studies along with the treatment regimens are presented in Table 1. The studies were published from inception to June 2024 and included 6255 patients, with a mean age ranging from 43.9 to 66.3 years. The quality of the studies included was found to be moderate (Table 1).

Table 1 Baseline characteristics of the included studies.

Ref.	Design	Intervention	Trial arms	Total subjects	Age (years, mean ± SD)	Baseline hyponatremia, n (%)	Quality assessment1
Ginès et al[51], 1988	RCT	Paracentesis plus IV albumin infusion	Paracentesis plus IV albumin infusion	52	56 ± 22	NA	3
			Paracentesis without IV albumin infusion	53	58 ± 22	NA
Sola-Vera et al[54], 2003	RCT	Paracentesis + albumin	Paracentesis + albumin	37	59.9 ± 10.9	8	3
			Paracentesis + saline	35	62.9 ± 10.7	13
García-Compean et al[50], 2002	RCT	Large-volume paracentesis + albumin	Large-volume paracentesis + albumin	48	58.9 ± 15.9	6 (12)	3
			Large-volume paracentesis + Dextran 40	48	59.4 ± 11.2	5 (10)
Planas et al[52], 1990	RCT	Total paracentesis + IV albumin infusion	Total paracentesis + IV albumin infusion	43	59 ± 1.52	NA	3
			Total paracentesis + IV Dextran 70 infusion	45	59 ± 1.42	NA
Bajaj et al[39], 2018	Retrospective	With albumin	With albumin	777	55.86 ± 10.24	476	7
			Without albumin	349	56.73 ± 11.31	175
Singh et al, 2006[55]	RCT	Albumin	Albumin	20	46.3 ± 12	NA	2
			Terlipressin	20	47.7 ± 8.3	NA
Moreau et al, 2002[56]	Pilot RCT	Paracentesis + albumin	Paracentesis + albumin	10	58 ± 22	NA	5
			Paracentesis + terlipressin	10	50 ± 22	NA
Salerno et al[57], 1991	RCT	Albumin infusion	Albumin infusion	27	55.9 ± 1.5	NA	2
			Hemaccel	27	53.2 ± 2	NA
García-Compeán et al[58], 1993	RCT	Total therapeutic paracentesis + albumin	Total therapeutic paracentesis + albumin	17	54.6 ± 10.9	NA	3
			Total therapeutic paracentesis without albumin	18	56.7 ± 7.8	NA
Fassio et al[59], 1992	RCT	Paracentesis with albumin	Paracentesis with albumin	21	54 ± 8	NA	2
			Paracentesis with Dextran 70	20	54 ± 7	NA
Altman et al[60], 1998	RCT	Albumin	Albumin	33	55.9 ± 11.2	NA	3
			Hydroxyethyl starch	27	56.3 ± 11.2	NA
Singh et al[61], 2006	Pilot RCT	Albumin	Albumin	20	52.5 ± 9.9	NA	2
			Noradrenaline	20	43.9 ± 9.5	NA
Appenrodt et al[37], 2008	Pilot RCT	Albumin	Albumin	13	60 (50-63)4	NA	3
			Midodrine	11	52 (48-61)	NA
Abdel-Khalek and Arif[13], 2010	Prospective RCT	Paracentesis + albumin	Paracentesis + albumin	68	47.97 ± 2.54	NA	2
			Paracentesis + hydroxyethyl starch 6%	67	46 ± 2.96	NA
Salerno et al[62], 1987	RCT	Repeated paracenteses + IV albumin infusion	Repeated paracenteses + IV albumin infusion	20	53.7 ± 2.8	NA	1
			Diuretics	21	56.4 ± 2.1	NA
Moreau et al[63], 2006	Pilot RCT	Albumin	Albumin	30	54 ± 9.0	NA	5
			Polygeline	38	56 ± 8.0	NA
Jalan et al[64], 2007	RCT	Human salt poor albumin + fluid + Na restriction	Human salt poor albumin + fluid + Na restriction	12	-	NA	1
			Fluid restriction (1.5 L) + Na restriction (80 mmol/day)	12	-	NA
Sort et al[40], 1999	RCT	Cefotaxime + IV albumin	Cefotaxime + IV albumin	63	60 ± 1.02	NA	3
			IV cefotaxime	63	62 ± 1.02	NA
Bari et al[65], 2012	Prospective, DB RCT	Albumin	Albumin	13	55 (51-61)3	NA	5
			Octreotide + midodrine	12	60 (51-65)3	NA
Solà et al[25], 2018	DB RCT	Midodrine and albumin	Midodrine and albumin	87	55 ± 10	NA	5
			Placebo	86	54 ± 11	NA
Caraceni et al[24], 2018	Open-label RCT	SMT + human albumin	SMT + human albumin	218	61.0 ± 11.4	75	3
			SMT	213	61.4 ± 10.9	74
Hu et al[38], 2023	Retrospective analysis	Albumin	Albumin	1093	58 (51-66)3	NA	6
			Non-albumin	1172	60 (52-67)3	NA
Khanna et al[66], 2024	Single-center prospective case-control	SMT + human albumin	SMT + human albumin	88	55.25 ± 11.5	NA	6
			SMT	86	55.0 ± 11.3	NA
Di Pascoli et al[67], 2019	Nonrandomized prospective	Albumin	Albumin	45	64.2 ± 10.4	NA	5
			SOC	25	65.5 ± 12.7	NA
Guevara et al[68], 2012	RCT	Albumin	Albumin	46	56 ± 10	NA	3
			Control	51	56 ± 12	NA
Thévenot et al[69], 2015	Randomized, open-label, controlled, multicenter, clinical trial	Albumin	Albumin	95	55.4 ± 8.3	NA	3
			Control	96	55.3 ± 9.0	NA
Devisetty et al[70], 2024	Prospective, non-inferiority, open-label RCT	Antibiotics and SMT+ Albumin	Antibiotics and SMT+ Albumin	58	52.52 ± 13.54	NA	3
			Antibiotics and SMT+ Albumin	116	54.34 ± 12.84	NA
Simón-Talero et al[71], 2013	Multicenter, prospective, DB, controlled trial	Albumin	Albumin	26	63.7 ± 11.3	NA	5
			Isotonic saline	30	66.3 ± 9.7	NA
Romanelli et al[72], 2006	Unblinded randomized trial	Diuretics + albumin	Diuretics + albumin	54	63 (44 - 70)3	NA	3
			Diuretics	46	63 (47-70)3	NA
Abootalebi et al[73], 2017	RCT	Albumin	Albumin	36	Total population mean age: 45.48 ± 3.95	NA	4
			HES	36		NA
			HES + albumin	36		NA
Boyer et al[74], 2016	Phase 3, multicenter, randomized, DB, placebo-controlled trial		Terlipressin + albumin	97	55.8 ± 8.4	NA	4
			Terlipressin	99	54.8 ± 8.5

¹Quality assessment of the randomized controlled trials was performed done using the Jadad scale and observational non-randomized controlled trial using the Newcastle-Ottawa scale.

²Values are expressed as mean (SE).

³Values are expressed as median (range).

⁴values are expressed in mean (range).

DB: Double-blind; HES: Hydroxyethyl starch; IV: Intravenous; RCT: Randomized controlled trial; SOC: Standard of care; SMT: Standard medical therapy.

Efficacy outcomes

With respect to the primary efficacy outcome for hyponatremia (Supplementary Table 1), 14 studies were included in analysis. Pooled OR with 95%CI for hyponatremia in the albumin-treated group vs the control group was 0.67 (0.53-0.85; I² = 0%; Figure 2A). In the subgroup analysis, a favorable improvement was observed with albumin infusion vs I/SMT (OR = 0.54; 95%CI = 0.27-1.09; Figure 2B); however, this reduction was not statistically significant. For PICD, pooled data from 5 studies showed that the use of albumin significantly reduced incidences compared with the control group (Figure 3A; Supplementary Table 1). In the subgroup analysis, the improvement with albumin was significant compared with other VEs (OR = 0.31; 95%CI = 0.11-0.85; Supplementary Figure 1) but not with vasoconstrictors (OR = 0.63; 95%CI = 0.21-1.91). From 9 studies reporting severe infection (Supplementary Table 1), pooled data of 8 studies showed a lower incidence of severe infection with the albumin-treated group compared with the control group; however, this reduction was not statistically significant (OR = 0.55; 95%CI = 0.28-1.07; I² = 0%; Figure 3B).

Figure 2 Forest plot for hyponatremia outcome comparing albumin with control and hyponatremia comparing albumin and control across other treatment groups. A: Hyponatremia outcome comparing albumin with control; B: Hyponatremia comparing albumin and control across other treatment groups. 95%CI: 95% confidence interval; df: Degrees of freedom; FE: Fixed effect.

Figure 3 Forest plot for post-paracentesis circulatory dysfunction outcome comparing albumin with control and severe infection outcome comparing albumin with control. A: Post-paracentesis circulatory dysfunction outcome comparing albumin with control; B: Severe infection outcome comparing albumin with control. 95%CI: 95% confidence interval; df: Degrees of freedom; FE: Fixed effect.

Of the 13, 14, and 18 studies reporting GI bleeding, HE, and renal impairment, respectively (Supplementary Table 1), the pooled data showed no significant differences between albumin and control in the GI bleeding and renal impairment, whereas the subgroup analysis showed comparable improvement in the aforementioned outcomes between albumin and other treatment modalities. Specifically, the pooled analysis showed a significant reduction in HE with albumin compared with control. For renal impairment, although the overall analysis did not show a statistically significant difference, the subgroup analysis comparing albumin with I/SMT demonstrated a favorable effect. Similarly, the results for GI bleeding were favorable towards albumin but not statistically significant (Supplementary Figures 2, 3, 4 and 5). The incidence of hospital stay days was reported in 9 studies (Supplementary Table 1). Pooled data showed that the standardized mean difference between albumin and control was comparable (Supplementary Figure 6A). A similar observation was seen in the subgroup analysis with albumin vs other VEs and the I/SMT group (Supplementary Figure 6B). There were 11, 8, 7, and 12 studies mentioning ascites reappearance, HRS, SBP, and mortality, respectively, during the follow-up visits (Supplementary Table 1). The results for the pooled analysis as well as the subgroup study are mentioned in Supplementary Figures 7, 8, 9 and 10. Four studies included peripheral edema as an outcome in their studies (Supplementary Table 1). The analysis of pooled data showed albumin and control to be comparable (OR = 0.97; 95%CI = 0.55-1.70; I² = 0%). Similar results were observed between albumin and I/SMT (OR = 1.07, 95%CI = 0.51-2.24; Supplementary Figure 11).

Safety outcomes

From the 4 studies reporting overall adverse events (Supplementary Table 1), the outcome between albumin and control, for which the OR was 1.05 (95%CI = 0.71-1.56; I² = 0%; Supplementary Figure 12A). Similar observation was seen in the subgroup analysis with albumin vs vasoconstrictor (OR = 1.01; 95%CI = 0.44-2.34), whereas with the albumin group compared with the I/SMT group the odds ratio was 1.05 (0.67-1.66); Supplementary Figure 12B). Pooled data from 14 studies included in the meta-analysis showed a reduction in the hospital mortality using albumin therapy compared with the control but not significant (OR = 0.88; 95%CI = 0.59-1.32; I² = 0%; Supplementary Figure 13A). Comparable in-hospital mortality was observed with albumin vs other VEs (OR = 1.02; 95%CI = 0.39-2.72). There was an improvement with the albumin group compared with the I/SMT group, but not significant (OR = 0.72; 95%CI = 0.36-1.45; Supplementary Figure 13B).

Publication bias

Publication bias of studies was assessed using funnel plots and the Egger's test as well as visual inspection. A sensitivity analysis was performed during the follow-up for hyponatremia, in-hospital mortality, and mortality. Of the 15 hyponatremia studies, 1 study causing asymmetry was removed[37], and the remaining 14 articles were finalized for analysis. Similarly, of the 16 studies pertaining to in-hospital mortality, 2 studies causing asymmetry were removed[38,39], and the remaining 14 articles were finalized for analysis. Likewise, of the 12 studies relating to mortality during the follow-up, 1 study[38] was removed, and the remaining 11 articles were finalized for the final analysis (Supplementary Figure 14). One study was removed from SBP[39] and one from severe infection[40].

DISCUSSION

This study has provided a comprehensive and up-to-date overview of the comparative effectiveness and safety of albumin in reducing complications of liver cirrhosis in both RCTs and observational studies by comparing it with different treatment modalities such as VEs (haemaccel, hydroxyethyl starch, dextran 70, and synthetic colloid), vasoconstrictors (terlipressin, octreotide, midodrine, and noradrenaline), I/SMT, and antibiotics (cefotaxime) in patients with liver cirrhosis. This is the first study to report peripheral edema as an efficacy outcome.

The results of this meta-analysis have many noteworthy contributions. Among efficacy measures, albumin was significantly more effective than the alternative treatments evaluated in resolving hyponatremia and PICD, whereas its effect on severe infection showed a favorable trend but was not significant. This is a major clinical implication of this study because it highlights the potential of human albumin infusion as an essential therapeutic agent in improving clinical outcomes, thereby informing treatment protocols and optimizing patient management strategies.

Hyponatremia is a clinically significant marker of the severity of cirrhosis and is correlated with an increased risk of mortality[17]. In our study, the improvement in the prevention of hyponatremia was statistically significant in patients with liver cirrhosis in the overall analysis (OR = 0.67; 95%CI = 0.53-0.85) and favorable but not significant when compared with I/SMT (OR = 0.54; 95%CI = 0.27-1.09), a favorable trend towards improvement was observed with VEs (OR = 0.67; 95%CI = 0.37-1.21) or vasoconstrictors (OR = 0.45; 95%CI = 0.05-3.75) with neither being statistically significant. Thus, patients with cirrhosis will benefit from albumin infusions as decreased incidences of hyponatremia will improve the clinical outcomes arising out of frequent hospitalizations. The results are of clinical significance as patients with hyponatremia are highly susceptible to developing HE, refractory ascites, less responsive to diuretics, and having an urgent need for paracentesis[20]. In a meta-analysis of RCTs by Bernardi et al[20], albumin reduced the incidence of hyponatremia compared with VE (OR = 0.61; 95%CI = 0.40-0.93) and vasoconstrictor (OR = 0.37; 95%CI = 0.09-1.49) in patients undergoing large-volume paracentesis. Another meta-analysis showed that human albumin was superior to albumin substitutes in lowering the incidence of hyponatremia (OR = 0.62; 95%CI = 0.42-0.94) in patients with ascites caused by cirrhosis undergoing drainage[23]. The results from the present study were aligned with Shrestha et al[10] and Kütting et al[22]. Patients with cirrhosis without hyponatremia had a significantly lower incidence of hyponatremia (OR = 0.55; 95%CI = 0.38-0.80) in comparison with patients with hyponatremia (OR = 1.50; 95%CI = 1.17-1.92)[17]. Liver cirrhosis is associated with complications such as ascites, HE, and SBP, each requiring tailored treatment approaches to achieve specific therapeutic goals. For example, in the case of PICD, albumin infusion has demonstrated efficacy in improving circulatory function and preventing complications during large-volume paracentesis for the management of ascites. Similarly, in HE, albumin administration leads to a reduction in ammonia levels and improves neurologic function. In addition, albumin plays a critical role in managing severe infections such as SBP by enhancing antibiotic effectiveness and reducing systemic inflammation. This highlights the capability of albumin in managing various complications associated with liver cirrhosis. Previous meta-analyses[10,20-23] specifically evaluated the prevention of hyponatremia in patients with liver cirrhosis after paracentesis; however, this study has not segregated patients based on paracentesis, severe infection, ascites, or HE but have included all under the umbrella of “liver cirrhosis” to gauge whether the use of albumin can be generalized in addressing the pressing issue of hyponatremia, which affects approximately 5% of adults and 35% of hospitalized patients, and other liver cirrhosis endpoints[41].

PICD is a silent syndrome occurring as a complication of large-volume paracentesis and is the most vital predictor of mortality[42,43]. Although the use of albumin significantly reduced PICD as compared with the control (OR = 0.38; 95%CI = 0.20-0.71) and VEs [OR = 0.31 (0.11-0.85)], it was not significant when compared with vasoconstrictors (OR = 0.63; 95%CI = 0.21-1.91). The results support the findings of Shrestha et al[10] who used albumin to reduce the odds of PICD by 60% (OR = 0.40; 95%CI = 0.27-0.58) and was significant against VEs (OR = 0.34; 95%CI = 0.22-0.52) but not against vasoconstrictors (OR = 0.93; 95%CI = 0.35-2.45), Bernardi et al[20] (albumin vs VE: OR = 0.34; 95%CI = 0.23-0.51; albumin vs vasoconstrictor: OR = 0.79; 95%CI = 0.32-1.92), and Kwok et al[21] (OR = 0.26; 95%CI = 0.08-0.93). Therefore, administering albumin infusions can be a cost-effective option in preventing not only PICD but also associated consequences such as acute kidney injury, hyponatremia, and mortality[44].

Infections in cirrhosis present a challenge to healthcare practitioners, since the mortality from sepsis is on the widespread increase worldwide[45]. However, in the case of severe infection reduction, no significant difference was observed between the albumin and I/SMT or albumin and VEs groups. Thus, the benefits of albumin treatment on the reduction in the incidences of hyponatremia and PICD is a major finding of our study, its benefit in reducing infections remains suggestive but inconclusive, that can pave the path for reducing long-term clinical outcomes. We found significant benefits with albumin in preventing ascites reappearance, SBP, HRS, and not in renal impairment, mortality during the follow-up, and in-hospital mortality in the overall analysis; however, in the subgroup analysis, the use of albumin was slightly favored in comparison with other subgroups for the outcomes excepting ascites reappearance. Owing to the insufficient number of articles for SBP and overall adverse events, comparison of albumin with certain subgroups was not possible. With respect to GI bleeding, HE, and hospital stay days, albumin treatment was comparable between the treatments in the main as well as overall subgroup group analyses.

A recent meta-analysis reported the administration of albumin to significantly reduce ascites reappearance (OR = 0.56; 95%CI = 0.46-0.68) but not the mortality of patients with decompensated liver cirrhosis[27]. The occurrence of ascites, renal failure, and HE was reduced in another study and the administration of short-term albumin had no effect on mortality[46]. However, the mortality was lower in patients receiving albumin than alternative treatments (OR = 0.64; 95%CI = 0.41-0.98), with low pooled risk for in-patient mortality and overt HE[20,47]. No significant difference was observed between the long-term albumin and the control groups with respect to mortality, refractory ascites, SBP, HE, GI bleeding, or adverse events[28]. Similarly, no significant difference in encephalopathy, readmission, renal impairment, and mortality was observed in patients on albumin use in paracentesis[21]. This holds true regarding mortality and renal dysfunction in patients with cirrhosis having extraperitoneal infections on albumin treatment[48]. The findings resonate with those from Shrestha et al[10]. However, in a study by Salerno et al[49], renal impairment was prevented and mortality was reduced in patients with SBP. This may be due to the low statistical power as only 4 RCTs were included. The mean follow-up period was 14 ± 11.8 months[50] and 6 months[13] in the 2 respective studies and ranged from 23.7 to 34.4 months in the remaining studies[51,52]. In this study, we have elucidated the benefits of albumin infusions on significant clinical outcomes, namely hyponatremia, PICD, and severe infection. This would play an important role in enhancing the quality of life and improving the incidences of HE, neurologic complications, renal dysfunction, ascites, and infection, ultimately resulting in fewer hospitalizations and lowering the risks of morbidity and mortality[53]. Sandi et al[28] observed no significant difference between the long-term albumin and the control groups, whereas Xu et al[27] observed an increased risk of pulmonary edema adverse reaction. To the best of our knowledge, this is the first report on the clinical endpoint peripheral edema. Among safety results, albumin administration was comparable with control in the overall analysis and against I/SMT. Comparable improvement in the overall adverse events was also observed with albumin.

One of the major strengths of our meta-analysis is the comprehensive to date review of studies to evaluate the clinical effectiveness and safety of human albumin in patients with cirrhosis. A predefined inclusion criteria was used to shortlist the studies. As a result, the overall heterogeneity was low for the outcomes. The inclusion of 31 trials in this meta-analysis strengthens the study further. In addition, comparisons with albumin were made within the categories, which were stratified according to the treatment groups. However, our study has certain limitations. Owing to the inclusion of only English articles, a possible language bias cannot be eliminated, and the results cannot be generalized to a wider population. Most of the included studies lacked statistical power and heterogeneity existed in the patient population, treatment, and outcomes across various studies, which may contribute to inconsistencies in findings and may limit the generalizability of the results. Another limitation of this study is the lack of data regarding the specific dosage, volume, percentage of albumin used, and frequency of administration. These missing parameters can significantly impact the interpretation of the findings, as variability in albumin administration could influence patient outcomes and limits the ability to assess dose-response relationships. This weakens the capacity to draw accurate clinical recommendations regarding optimal therapeutic regimens, and future studies should ensure standardized reporting of these variables to enhance comparability across trials.

CONCLUSION

This meta-analysis demonstrates that albumin infusions are associated with a significant reduction in the incidence of hyponatremia and PICD in patients with cirrhosis. Although a potential benefit was observed for reducing severe infections, this finding did not reach statistical significance. These results support the targeted use of albumin in selected cirrhosis-related complications where a clear benefit has been demonstrated, rather than its broad application across all clinical scenarios, in order to avoid adverse reactions and unnecessary use of medical resources. Given the heterogeneity of existing studies and limitations in data quality, further high-quality evidence is needed. However, conducting large, multicenter RCTs in cirrhotic populations poses practical and ethical challenges, including patient frailty, disease variability, and resource constraints. Future research should aim to optimize study designs that are both rigorous and feasible in this vulnerable population.

ACKNOWLEDGEMENTS

Medical writing support was provided by Sankara Doddam and Tanushree Goswami; statistical analysis support was provided by Ashwini Patil of Indegene Limited, Bangalore. This work was supported by Takeda (China) International Trading Company and complied with Good Publication Practice 2022 guidelines.