Blood Purification and Mortality in Sepsis and Septic Shock: A Systematic Review and Meta-analysis of Randomized Trials

Today, sepsis remains one of the main causes of morbidity and mortality in the intensive care unit. Despite recent advancement in intensive care unit and sepsis management, mortality still remains high.^1–4 

The pathogenesis of sepsis involves many complex cellular and biochemical interactions between leukocytes, platelets, endothelial cells, and the complement system that trigger an inflammatory response.⁵  Inflammation is caused by the production of pro- and antiinflammatory mediators, such as cytokines, in the presence of infection and/or bacterial toxins, and the imbalance between these mediators or their excessive production may lead to multiorgan failure due to a prolonged or inadequate systemic inflammatory response syndrome.^5,6 

Extracorporeal blood purification techniques have been proposed as adjunctive therapy in sepsis. These techniques are based on the principle that removal and modulation of blood pro- and antiinflammatory mediators or bacterial toxins (or both) could attenuate the sepsis-related massive systemic inflammatory response, reducing morbidity and mortality.^7,8  Several different extracorporeal techniques have been studied for this purpose.

Hemoperfusion involves the placement of a sorbent cartridge in direct contact with blood via an extracorporeal circuit. The removal characteristics of hemoperfusion are dependent on the different types of sorbent used and could also target high-molecular-weight molecules, usually not captured by conventional hemofilters. The most studied therapy is polymyxin B immobilized fiber column hemoperfusion with Toraymyxin (Toray Industries Ltd., Japan), that could capture circulating bacterial endotoxin⁹  and modulate the inflammatory response.¹⁰  Another device is the CytoSorb (CytoSorbents Corporation, USA), a novel filter potentially able to remove both pro-inflammatory and antiinflammatory cytokines.¹¹ 

Renal replacement devices such as hemofiltration or hemodiafiltration could be used to remove part of the inflammatory mediators and toxins in septic patients without renal indication for kidney replacement therapy, by employing standard or special filters with adsorptive properties.¹²  Limited data are available on plasmapheresis, a technique based on plasma replacement with fresh frozen plasma or albumin,¹²  that has the potential to remove inflammatory cytokines and restore deficient plasma proteins.

Despite the large number of available techniques, actual evidence is scarce, and these therapies have not entered into daily clinical practice around the world yet. Several small trials were published on various devices, and the most comprehensive meta-analysis summarizing the evidence on blood purification is outdated.¹³  Some more recent meta-analyses focusing on polymyxin B immobilized fiber column hemoperfusion^14,15  or hemofiltration¹⁶  did not include some relevant trials nor the final results of the largest randomized study performed on the topic so far.¹⁷  Therefore, we performed a meta-analysis of randomized control trials in order to determine whether extracorporeal blood purification decreased mortality in patients with sepsis and septic shock.

The current systematic review was conducted in compliance with the PRISMA (Preferred Reporting Items Systematic Reviews and Meta-Analysis) guidelines¹⁸  (Supplemental Digital Content, table S1, https://links.lww.com/ALN/B977) and Cochrane methodology¹⁹  and according to a prepublished protocol (PROSPERO database, CRD42018104643).

Two investigators (A.P. and R.S.) independently searched PubMed, the Cochrane Central Register of clinical trials, and Embase up to January 1, 2019, for relevant articles (Supplemental Digital Content, table S2, https://links.lww.com/ALN/B977). The search strategy aimed to include any randomized study performed with any type of extracorporeal blood purification technique compared to conventional therapy in adult critically ill patients with sepsis and septic shock. Abstracts from recent international conferences were searched for additional studies. In addition, we hand-scanned references of retrieved articles and pertinent reviews to identify other eligible trials (backward snowballing).

References obtained from searches were first independently examined at the abstract level by two authors (A.P. and R.S.) and then collected as full-text articles if potentially relevant. Eligible studies met the following PICOS criteria: (1) Population: adult critically ill patients with sepsis with or without septic shock; (2) Intervention: any extracorporeal blood purification technique (hemoperfusion, renal replacement therapy techniques, plasmapheresis); (3) Comparison intervention: conventional therapy; (4) Outcome: mortality at longest follow-up available; and (5) Study design: randomized controlled trial. The exclusion criteria were blood purification for renal failure indication at randomization, trials with overlapping populations with a previously included article (e.g., manuscripts with different follow-up or subanalyses of a previously published trial), and pediatric studies. Two authors (A.P. and R.S.) independently assessed selected studies for the final analysis, with disagreements resolved by consensus with a third author (G.L.). If the article did not include data on mortality or was not full-text, the corresponding author was contacted for further data. No language restrictions were imposed.

One author (A.P.) extracted relevant information from each selected study. These data were checked by another author (R.S.). Disagreement was resolved by consensus with a third author (G.L.). We specifically extracted potential sources of significant clinical heterogeneity (e.g., study design, clinical setting, inclusion and exclusion criteria, blood purification regimen).

The primary endpoint of this review was mortality at the longest follow-up available, and the secondary endpoint was mortality at 28 to 30 days.

Two authors (A.P. and R.S.) independently assessed the internal validity of each included trial according to the Cochrane Collaboration methods.^19,20  We assessed the risk of bias associated with the random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, completeness of outcome data, selective reporting, and other bias. The other bias domain included the classic items reported by the “Cochrane Handbook for Systematic Reviews of Interventions”¹⁹  but also the presence of an intention-to-treat analysis, sample size calculation, and ethical approval of the trial. If one or more of the domains were judged as having a high or unclear risk of bias, we classified the trial as having a high risk of bias. Due to the nature of the intervention, blinding of participants and personnel seemed difficult and was therefore not judged as crucial for bias assessment. We evaluated the potential risk of bias by applying a rating of “Low,” “High,” or “Unclear” to each study.

Two authors (A.P. and R.S.) independently reviewed the presence of authors’ possible conflict of interest and the funding source for each study, then rated each trial as of “Low,” “High,” or “Unclear” risk regarding those specific points.

The certainty of the body of evidence was assessed using the grading of recommendations assessment, development, and evaluation framework.^21,22  The grading of recommendations assessment, development, and evaluation framework characterizes the certainty of a body of evidence on the basis of study limitations, imprecision, inconsistency, indirectness, and other considerations.

Individual trial and summary results were reported as relative risk with 95% CI. We used a random-effects model except in cases where few trials dominated the available evidence or where significant publication bias was present, as random-effects meta-analysis applied in these contexts may give inappropriately high weight to smaller studies. Statistical heterogeneity was explored by the Cochran Q statistic and characterized using the I² metric. Publication bias was assessed by visually inspecting the funnel plot for the primary outcome. Statistical significance was set at P = 0.05. The meta-analysis was performed using Review Manager (RevMan, version 5.3; The Nordic Cochrane Center, The Cochrane Collaboration, Copenhagen, Denmark, 2014).

The primary analysis was stratified by blood purification technique: hemoperfusion, hemofiltration, hemoperfusion combined with hemofiltration, or plasmapheresis. Hemoperfusion subgroup analyses including trials on polymyxin B immobilized fiber column hemoperfusion or hemoperfusion with other devices were carried out. To explore the sources of heterogeneity, we performed some subgroup analyses: (1) low risk of bias versus unclear/high risk of bias trials; (2) trials conducted in Asia versus Europe and America; (3) trials from the Nakamura group versus other trials; and (4) trials published after 2010 versus older trials.

To explore the relationship between treatment effect and disease severity, we performed various analyses: (1) a random-effects meta-regression on the APACHE II (Acute Physiology, Age, Chronic Health Evaluation II) score,²³  sepsis-related organ failure assessment score,²⁴  and control group mortality;¹⁴  (2) subgroup analyses according to conventional therapy group mortality: low-risk group (mortality rate less than 30%), intermediate-risk group (30 to 60%), and high-risk group (greater than 60%).¹⁴  We also performed a meta-regression for age to investigate a possible influence on outcome estimates. Finally, sensitivity analyses were performed by analyzing the data with a fixed or random effects model and using other summary statistics.

We performed a predefined random-effects trial sequential analysis,^25–27  with the intent of maintaining an overall 5% risk of type I error and a 10% risk of type II error. We assumed a relative risk reduction of 15% and derived the control event proportion from the actual dataset. The resulting required information size was further diversity (D²)-adjusted. In case of D² = 0 we performed a sensitivity analysis assuming a D² = 25%. We used the trial sequential analysis software (TSA Viewer [Computer program], version 0.9.5.5 Beta, Copenhagen Trial Unit, Center for Clinical Intervention Research, Rigshospitalet, Copenhagen, Denmark, 2016). Deviations from the initial protocol are reported in the supplement (Supplemental Digital Content, eMethods 1, https://links.lww.com/ALN/B977).

The search strategy identified 6,933 citations and, after exclusion of inadequate reports (Supplemental Digital Content, table S3, https://links.lww.com/ALN/B977), 37 trials with 2,499 patients were included in the meta-analysis (fig. 1).^17,28–63 

The characteristics of the included studies are shown in table 1 and in the supplement (Supplemental Digital Content, tables S4-S6, https://links.lww.com/ALN/B977). Two trials had four treatment arms.^43,58  Twenty trials used a hemoperfusion technique, 13 used hemofiltration or hemodiafiltration, 4 trials combined hemofiltration with hemoperfusion, and 2 trials used plasma exchange. In three cases we received further information from corresponding authors.^17,59,62 

Three trials were judged to be at low risk of bias,^17,30,46  20 at unclear risk, and 14 at high risk (Supplemental Digital Content, figs. S1 and S2, https://links.lww.com/ALN/B977). The grading of recommendations assessment, development, and evaluation assessment is reported in table S7 in the Supplemental Digital Content (https://links.lww.com/ALN/B977).

Hemoperfusion (20 trials and 1,548 patients), which comprises various techniques differing among other things on the presence or absence of polymyxin B in the treatment regimen, was associated with a lower mortality compared to the control group (relative risk = 0.87 [95% CI, 0.78 to 0.98], P = 0.02, trial sequential analysis inconclusive, very low-certainty evidence) the analysis was limited by publication bias, small trial effects, and high heterogeneity (Supplemental Digital Content, figs S3-S5 and table S8, https://links.lww.com/ALN/B977). Subanalyses are reported in the supplement (Supplemental Digital Content, figs. S6—S9 and eResults 1, https://links.lww.com/ALN/B977).

Polymyxin B immobilized fiber column hemoperfusion (13 trials and 1,163 patients) was associated with a lower mortality at longest follow-up available compared to control (relative risk = 0.87 [95% CI, 0.77 to 0.98], P = 0.03, very low-certainty evidence), although the analysis was limited by very high heterogeneity (I² = 74%, P_{heterogeneity} < 0.001) (fig. 2). No significant difference in 30-day mortality was found (Supplemental Digital Content, fig. S9, https://links.lww.com/ALN/B977).

Low risk of bias trials (three trials and 745 patients) found no difference in mortality with polymyxin B immobilized fiber column hemoperfusion versus control (relative risk = 1.14 [95% CI, 0.96 to 1.36], P = 0.12, moderate-certainty evidence; fig. 2). Recent trials published after 2010 (three trials and 740 patients) showed that polymyxin B immobilized fiber column hemoperfusion was associated with higher mortality than conventional therapy (relative risk = 1.22 [95% CI, 1.03 to 1.45], P = 0.02, I² = 0%, low-certainty evidence), while trials published before 2011 were associated with a mortality benefit (relative risk = 0.58 [95% CI, 0.49 to 0.69], P < 0.001, I² = 8%; P_groups < 0.001). Studies conducted in Asia (seven trials in Japan and one in Thailand, with a total of 367 patients) showed that polymyxin B immobilized fiber column hemoperfusion decreased mortality (relative risk = 0.62 [95% CI, 0.52 to 0.75], P < 0.001, I² = 57%, P_{heterogeneity} = 0.02), while aggregate data from trials conducted in the United States and Europe (five trials and 796 patients) found no difference (relative risk = 1.11 [95% CI, 0.94 to 1.32], P = 0.21, I² = 50%, P_{heterogeneity} = 0.09), (P_groups < 0.001). Similarly, when excluding trials performed in Japan by the Nakamura group (five trials and 162 overall patients), polymyxin B immobilized fiber column hemoperfusion was associated with no difference in mortality compared to conventional therapy (relative risk = 0.98 [95% CI, 0.86 to 1.12], P = 0.80; Supplemental Digital Content, figs. S10-S13, https://links.lww.com/ALN/B977).

Hemoperfusion with devices other than polymyxin B–immobilized filter column (seven trials and 385 patients) was not associated with a difference in mortality compared to conventional therapy (relative risk = 0.81 [95% CI, 0.53 to 1.21], P = 0.30, very low-certainty evidence). The hemoperfusion devices included were Adsorba-300 filter (one trial, relative risk = 0.50 [95% CI, 0.22 to 1.14], P = 0.10); Alteco endotoxin hemoadsorber (one trial, relative risk = 0.57 [95% CI, 0.06 to 5.03], P = 0.61); CytoSorb (two trials, relative risk = 0.94 [95% CI, 0.14 to 6.49], P = 0.95); HA330 resin cartridge (two trials, relative risk = 0.61 [95% CI, 0.31 to 1.19], P = 0.15); and Matisse EN 500 endotoxin adsorber (one trial, relative risk = 1.13 [95% CI, 0.66 to 1.96], P = 0.65; Supplemental Digital Content, fig. S3, https://links.lww.com/ALN/B977).

The use of hemofiltration with a blood purification aim was associated with lower mortality compared to control (relative risk = 0.79 [95% CI, 0.63, 1.00], P = 0.05, trial sequential analysis inconclusive, very low-certainty evidence) in 13 trials and 596 patients without acute kidney injury requiring renal replacement therapy (fig. 3 and Supplemental Digital Content, fig. S14, https://links.lww.com/ALN/B977). On subgroup analysis, hemofiltration was not associated with a difference in mortality in trials conducted in Europe and the United States (relative risk = 0.94 [95% CI, 0.74 to 1.19], P = 0.61, I² = 0%, five trials and 223 patients) but was associated with a decrease in mortality in trials conducted in Asia (relative risk = 0.58 [95% CI, 0.40 to 0.82], P = 0.002, I² = 0%, eight trials and 373 patients; P_groups = 0.02); other analyses are reported in the supplement (Supplemental Digital Content, figs. S15-S18, table S9 and eResults 2, https://links.lww.com/ALN/B977).

The association of hemoperfusion and hemofiltration was not associated with a significant difference in mortality compared to control (relative risk = 0.63 [95% CI, 0.36 to 1.13], P = 0.12, trial sequential analysis inconclusive, very low-certainty evidence) in four trials including a total of 247 patients without acute kidney injury requiring renal replacement therapy (fig. 3).

Plasmapheresis was associated with a lower mortality compared to standard treatment (relative risk = 0.63 [95% CI, 0.42 to 0.96], P = 0.03, trial sequential analysis inconclusive, very low-certainty evidence) with two trials and 128 patients included (fig. 3).

We performed a comprehensive systematic review and meta-analysis on the mortality effects of blood purification with extracorporeal techniques in sepsis. The certainty of evidence underlying the use of blood purification therapies in sepsis is very low, and does not support their systematic use in patients with sepsis with or without septic shock.

A variety of hemoperfusion techniques exists. Only a few randomized clinical trials were published on hemoperfusion techniques other than polymyxin B–immobilized filter column (e.g., CytoSorb, Alteco endotoxin hemoadsorber), suggesting the need for further clinical trials. However, polymyxin B immobilized fiber column hemoperfusion emerged as a promising therapy in septic shock with elevated endotoxin levels, and several studies were published on the topic in the past 20 yr. This technique consists of using a sorbent cartridge containing fibers coated with polymyxin B, an antibiotic with high affinity for lipopolysaccharide.⁹  Lipopolysaccharide is a cell wall component in Gram-negative bacteria that acts as an endotoxin by stimulating the production of inflammatory mediators by macrophages in a dose-dependent way and enhancing the inflammatory response.^9,64  Endotoxemia seems to be more pronounced when tissue hypoperfusion is present and lipopolysaccharide blood levels seem to correlate with sepsis severity.^9,65  Promising results in pilot studies showed improvement in inflammatory mediators,¹⁰  cardiac and renal dysfunction,⁵⁶  hemodynamics, organ dysfunction, and 28-day mortality³⁰  in patients with abdominal septic shock. All these promising findings, together with the significant increase in arterial pressure after therapy initiation,^15,17,30  made polymyxin B immobilized fiber column hemoperfusion an attractive therapy for clinicians. Conversely, recent large high-quality trials such as the EUPHRATES (Evaluating the Use of Polymyxin B Hemoperfusion in a Randomized Controlled trial of Adults Treated for Endotoxemia and Septic Shock)¹⁷  and the ABDO-MIX (Effects of Hemoperfusion With a Polymyxin B Membrane in Peritonitis With Septic Shock) group⁴⁶  trials yielded inconclusive results and reported a nonsignificant increase in mortality with polymyxin B immobilized fiber column hemoperfusion at the longest follow-up assessed.

The EUPHRATES trial, which is the largest and highest-quality randomized clinical trial performed so far, randomized 450 patients with septic shock and a high endotoxin activity assay level to two sessions of polymyxin B immobilized fiber column hemoperfusion of 90 to 120 min at a 24 h distance or to a sham treatment aiming at reducing 28-day mortality. The trial found no significant difference in the primary endpoint in the overall population or in the higher disease severity subgroup.¹⁷ 

Our meta-analysis including 13 trials on polymyxin B immobilized fiber column hemoperfusion is the largest and most comprehensive to date. Recently, some meta-analyses^14,15,66  on polymyxin B immobilized fiber column hemoperfusion have appeared but failed to include some old and new randomized clinical trials. A meta-analysis concluded that this therapy may reduce mortality in patients with severe sepsis and septic shock in high disease severity subgroups based upon the aggregate analysis of 12 nonrandomized trials and 5 small randomized clinical trials representing a very low-quality evidence.¹⁴  Two other meta-analyses respectively including only six and seven randomized clinical trials concluded that only low-quality evidence supported polymyxin B immobilized fiber column hemoperfusion for mortality reduction in sepsis.^15,66  Since the release of EUPHRATES and other trials, a more comprehensive analysis was made possible. The positive results previously reported regarding polymyxin B immobilized fiber column hemoperfusion were driven by small randomized clinical trials conducted in Asia of low methodologic quality. Interestingly, when limiting the analysis to trials published after 2010 and including the two largest randomized clinical trials performed on the topic,^17,46  polymyxin B immobilized fiber column hemoperfusion is associated with a higher mortality rate at the longest follow-up available. This together with inconclusive results on trial sequential analysis suggests that the current aggregate randomized evidence cannot consistently refute potential positive or detrimental effects on mortality. These findings do not support the use of polymyxin B immobilized fiber column hemoperfusion in sepsis and septic shock.

The use of hemofiltration techniques as a blood purification treatment in patients without renal failure has also been suggested, with controversial results and insufficient evidence to recommend its use outside of experimental clinical settings.^16,67  High-volume hemofiltration, further increasing plasma exchanges, was also investigated with limited results in patients with or without renal failure.^29,68  We found a positive survival trend associated to hemofiltration, although those results are driven by small, low-quality randomized trials, and further investigation is therefore warranted.

The first randomized clinical trial to ever address plasmapheresis as a blood purification technique reported a decrease in the intensity of acute-phase response.⁵⁰  A second randomized clinical trial with a larger sample population found an absolute mortality risk reduction of 20.5%.²⁸  Despite those promising results, the evidence is still too weak to recommend the use of plasmapheresis for blood purification in sepsis.⁶⁹ 

Previous meta-analyses found that hemoperfusion was associated with a large positive effect in trials with a control group mortality rate greater than 60%, suggesting that hemoperfusion could be useful in the setting of higher disease severity.^14,66  Our study yielded similar findings and also found a trend toward increased mortality in the lower disease severity subgroup (mortality less than 30%). Meta-regressions on APACHE II and sepsis-related organ failure assessment scores, both predictors of sepsis mortality, did not find any significant trend supporting those findings. Furthermore, most trials with a greater than 60% control group mortality are at unclear/high risk of bias, are small in size, and were conducted in Asia. In the EUPHRATES trial, the per-protocol subgroup analysis with high disease severity, including patients with a Multiple Organ Dysfunction Score greater than 9 at randomization and a control group 1-year mortality rate of 50%, was inconclusive and did not suggest any trend favoring polymyxin B immobilized fiber column hemoperfusion.¹⁷ 

Those inconsistencies make a beneficial effect of hemoperfusion or polymyxin B immobilized fiber column hemoperfusion in high-disease-severity patients unlikely. This specific question merits further investigation.

Current randomized evidence cannot support the use of extracorporeal blood purification techniques; further trials are warranted before systemic implementation of these techniques. Furthermore, an increase in mortality related to extracorporeal therapies should not be excluded. Some randomized clinical trials describe a trend toward higher mortality with polymyxin B immobilized fiber column hemoperfusion⁴⁶  or CytoSorb-HP⁵⁴ , for example. Numerically higher adverse events with polymyxin B immobilized fiber column hemoperfusion^17,46  and greater disease severity with hemofiltration⁴⁵  were also reported. Our meta-analysis found an increased mortality at the longest follow-up available with polymyxin B immobilized fiber column hemoperfusion in a post hoc subgroup analysis including only the trials published after 2010. The unspecific removal of cytokines may remove mediators necessary to the function of the immune system, eventually provoking a deleterious outcome. Furthermore, the complex interaction between extracorporeal devices and inflammatory systems, micronutrients,⁷⁰  trace elements, electrolytes, and antibiotics levels and activity remain uninvestigated. Only few studies assessed the impact of those therapies on antibiotics, the only proven therapy in sepsis. A recent study on in vitro removal of antiinfective agents by CytoSorb-HP showed that all tested antibiotics were adsorbed by the cartridge in substantial amounts.⁷¹  The authors speculated that an additional dose within the first hours of treatment and therapeutic drug monitoring should be considered in this population.⁷¹  Similarly, an in vitro study assessing the effects of polymyxin B immobilized fiber column hemoperfusion on nine antibiotics reported adsorption of linezolid, suggesting a necessity for the monitoring of blood antimicrobial concentrations during polymyxin B immobilized fiber column hemoperfusion.^72,73  A larger literature is present on hemofiltration, suggesting an increased antibiotic clearance with these devices.^74–76 

We performed a comprehensive meta-analysis on blood purification techniques in sepsis and septic shock, which represents an important update to the literature in comparison to previous meta-analyses.^13–16  Limitations of this study may appear similar to those of previous meta-analyses. Most included studies are small in size and at unclear or high risk of bias. Some studies assessed technical feasibility, but side effects were rarely reported, and a systematic assessment of adverse events is warranted in future trials. Heterogeneity in sepsis management, blood purification regimens (e.g., modality, hemofiltration volume, duration of the session, filter and cartridge change, and so forth), and populations across different centers is evident, but we made an attempt at an exploration through several subanalyses in order to further assess the clinical potential of blood purification modalities in sepsis. The positive treatment effect found in trials conducted in Asia was also reported elsewhere¹³  and could be explained by publication bias, small studies effect, low methodologic quality, or a higher burden of disease as suggested by the high control group mortality. Furthermore, seven polymyxin B immobilized fiber column hemoperfusion trials from Japan were performed before 2005, and the progress in conventional therapy management and outcome in the past years could have diluted or cancelled the beneficial effects of this treatment.

Very low-quality randomized evidence demonstrates that the use of hemoperfusion, hemofiltration, or plasmapheresis may reduce mortality in sepsis or septic shock. Moderate-certainty evidence supports that polymyxin B immobilized fiber column hemoperfusion is not associated with any significant difference in mortality in comparison to conventional treatment regimen. Detrimental effects on survival could not be excluded by aggregate randomized evidence. Further high-quality randomized controlled trials adequately powered for mortality are needed to assess the real impact of blood purification techniques before such therapies can be systematically implemented in clinical practice.

Support was provided solely from institutional and/or departmental sources.

The authors declare no competing interests.