1. Introduction Methods Results Conclusion
Network meta-analysis of clinical trials of fluid treatments in sepsis demonstrates improved survival with albumin compared to crystalloid and hydroxyethyl starch
M Bansal1 ; A Farrugia2 ; S Balboni1; G Martin3; M Bult1 1Plasma Protein Therapeutics Association, Annapolis, United States; 2Department of Surgery, Crawley, Australia;  3Emory University School of Medicine, Atlanta, United States
Introduction
Table#1: Basic characteristics of the included trials
Study
Population
Treatment arms
Follow-up duration
ALBIOS
Sepsis or septic shock
Crystalloid, Albumin
hospital discharge
Myburgh et al. CHEST study
New England Journal of Medicine. 2012
Sepsis as a sub-group
Crystalloid, HES
90 day
Brunkhorst (VISEP)
New England Journal of Medicine. 2008
Severe sepsis
Charpentier et al. (EARSS) Intensive Care Med
Septic shock
28 day
Dolecek et al. Hepatogastroenterology
Albumin, HES
Finfer S et al. (SAFE study) BMJ
Guidet et al. (CRYSTMAS) Critical Care.16 (R94) 2012
90 days
Maitland et al. British Journal of Haematology
Severe malaria (sepsis as a sub-group)
Maitland et al. Clinical infectious diseases
Maitland K, (FEAST Trial Group)
New England Journal of Medicine
Severe febrile illness & impaired perfusion (sepsis as a sub-group)
30 day
Mcintyre (FINESS) Can J Anaesth
28 days
Perner (6S) N Engl J Med
Veneman Wiener klinische Wochenschrift
Severely ill with sepsis
Crystalloid, Albumin, HES
In-hospital
Sepsis is a systemic, deleterious host response to infection leading to severe sepsis (acute organ dysfunction secondary to documented or suspected infection) and septic shock (severe sepsis plus hypotension not reversed with fluid resuscitation) [1]. Mortality is about 33.3% in the patients with severe sepsis who are hospitalized [2] treated with the current standard of care in the US. Hypovolemia is a feature of sepsis which is treated with fluids. Among several fluid treatments available for treating sepsis, the first line of treatment is generally crystalloids in a range of formulations. Colloids, including albumin and hydroexythl starch are also widely used. Recent large randomized clinical trials (RCTs) provide insight in determining the efficacy of the fluids used in treatment of sepsis. A number of adverse events have been associated with HES including haemostatic - and renal - dysfunction.
In this study, we have conducted a network meta-analysis [3] and have combined the mortality outcomes from selected randomized clinical trials. A Bayesian analysis ranked the treatments and provided probability of being the most effective treatment
Methods
A literature search of human clinical trials was conducted in PubMed, ClinicalTrials.gov and within the bibliographies of other recent relevant systematic reviews. In addition, data for mortality for treatment and control arms for the few trials for albumin in septic shock were extracted from the results presented in conferences.
The search terms used were sepsis, septicemia, systemic inflammatory response syndrome, septic shock, fluid therapy, resuscitation, plasma substitute, albumin, serum albumin, starch, hydroxyethyl starch, hetastarch, pentastarch, tetrastarch and mortality. The studies were first screened by their titles and abstracts done by a reviewer (MB). In the second screening, two reviewers (AF, MB) were involved. In this screening, full texts were reviewed and articles were excluded on the basis of no relevant comparators, not a randomized clinical trial, no outcomes for only sepsis patients or no mortality results (Figure 1). The chosen studies were validated. The trials with saline, Ringer’s lactate, Ringer’s acetate or which mentioned crystalloids were included to represent crystalloids in the analysis.
In the base case, all formulations of crystalloid, HES and albumin were assumed to be the pharmaceutically equivalent. The effectiveness of the treatment was assumed to be same across different age groups. The mortality data for the longest follow-up in a trial were included in the analysis. The trials which had data for patients with severe sepsis, septic shock and sepsis as a sub-group were included.
The indirect comparison between crystalloid, albumin and HES trials was conducted using the Bayesian method for binomial likelihood, fixed effects network meta-analysis [3]. The analysis was conducted in OpenBUGS version Studies with crystalloid as a reference treatment compared to any composition of colloid fluid (albumin or HES) treatments in septic patients and head-to-head trials of colloid fluids were included; no trials which originated from the group of Joachim Boldt were included.
The choice for fixed or random effect model was made by assessing model fit using deviance information criteria (DIC) and standard deviation between the trials [3]. The consistency for risk of bias was also checked [3] along with an assessment of the quality of all the studies. The odds ratio and corresponding 97.5% credible intervals were obtained with 50,000 iterations and convergence was seen around 15,000.
Figure 2: Forest plot of results of Bayesian fixed effect network meta-analysis
391 studies identified through database search and all other sources
336 excluded as not relevant in first screening
55 full text articles were reviewed in second screening
Excluded: 42
Not relevant comparator n=6
Not sepsis n=11
Not a randomized clinical trial n=3
No mortality results n=19
Duplicates/Repeats n=3
13 studies included in network meta-analysis
Figure 1: Study Flow Diagram
Figure 3: Bar plots for the ranking probabilities of competing fluid treatments. Horizontal axis is the possible rank of each treatment (from best to worse). The size of each bar corresponds to the probability of each treatment to be at specific rank.
Results
In the literature search, 391 studies were identified. 336 of these were excluded in the first screening. 13 studies were finally included in the analysis. DIC was for random effects model and for fixed effect. The posterior mean of the residual deviance for random effects model was more than fixed effect at 17.2 vs The standard deviation with random effect was also quite small at 0.09 thus exhibiting very low heterogeneity between trials. All these indicate that fixed effect model would fit the data better. The assessment showed similar effect sizes and DICs for the consistency and inconsistency models in case of the fixed suggesting systematic consistency.
Conclusion
This is the first network meta-analysis on risk of mortality in fluid resuscitation among patients with severe sepsis comparing the three widely used fluids. Albumin as a fluid therapy in sepsis is associated with the lowest mortality of the three modalities studied. These results are relevant to strategies for fluid treatment in intensive care.
Bibliography
1] Dellinger RP et al. Surviving Sepsis Campaign: International guidelines for management of severe sepsis and septic shock: Critical Care Medicine [2] Healthcare Cost and Utilization Project (HCUP) Available from: [3] Dias S et al. NICE DSU Technical Support Document 2, 3, 4: A Generalized Linear Modelling Framework for Pairwise and Network Meta-analysis of Randomised Controlled Trials; Heterogeneity: Subgroups, Meta-Regression, Bias and Bias-Adjustment; Inconsistency in networks of evidence based on randomized controlled trials. 2012