Intensity of continuous renal replacement therapies in patients with severe sepsis and septic shock: a systematic review and meta-analysis.
Abstract: The purpose of this study was to assess the efficacy of continuous renal replacement therapies in patients with severe sepsis or septic shock, with or without acute kidney injury. We performed a systematic search in Medline, Embase, Web of Knowledge, Cochrane Library and Clinicaltrials.gov and a hand search of the retrieved studies. We included both randomised controlled clinical trials and subgroups of randomised trials that assessed the effect of continuous renal replacement therapies (at traditional or high doses) and reported clinical outcomes in adult patients with severe sepsis or septic shock. The study selection and data extraction were performed by duplicate. Analysis of heterogeneity and meta-analysis was performed according to the Cochrane Collaboration guidelines for conducting systematic reviews of interventions.

Twelve studies (1895 patients) met the inclusion criteria. Pooling of all studies resulted in a mortality risk ratio of 0.96 (95% confidence interval 0.83 to 1.12). The studies showed moderate statistical heterogeneity I (2) statistic 52%, P=0.02). The effect on mortality was not modified (interaction P values non significant) by the dose of continuous renal replacement therapies, the severity of illness or the risk of bias. The available evidence suggests that these therapies in patients with severe sepsis or septic shock are not associated with an improvement in other outcomes such as haemodynamics, pulmonary gas exchange, multiple organ dysfunction syndrome or length of stay. The best available evidence does not support the routine use of continuous renal replacement therapies (at traditional or high doses) in patients with severe sepsis or septic shock.

Key Words: sepsis, continuous renal replacement therapy, mortality, review, meta-analysis
Article Type: Report
Subject: Sepsis (Care and treatment)
Authors: Latour-Perez, J.
Palencia-Herrejon, E.
Gomez-Tello, V.
Baeza-Roman, A.
Garcia-Garcia, M.A.
Sanchez-Artola, B.
Pub Date: 05/01/2011
Publication: Name: Anaesthesia and Intensive Care Publisher: Australian Society of Anaesthetists Audience: Academic Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2011 Australian Society of Anaesthetists ISSN: 0310-057X
Issue: Date: May, 2011 Source Volume: 39 Source Issue: 3
Geographic: Geographic Scope: Australia Geographic Code: 8AUST Australia
Accession Number: 260691509
Full Text: Severe sepsis and septic shock carry a high mortality and account for a large proportion of patients admitted to intensive care units (1-4).

It is widely accepted that the release of large amounts of pro- and anti-inflammatory mediators that occurs in severe sepsis contributes to the development of multiple organ dysfunction syndrome (MODS) (5-8), including acute kidney injury (AKI). Theoretically, high-dose continuous renal replacement therapies (CRRT) could remove mediators by convection and/or adsorption (9,10) and reduce mortality, even in the absence of AKI (11). However, most current clinical practice guidelines suggest that the traditional doses of CRRT used in AKI, with or without sepsis, are insufficient to remove these mediators and recommend using at least 35 ml/kg/hour of ultrafiltration (12,13).

Recently, two large randomised clinical trials in patients with AKI (ATN study (14,15) and RENAL study (16,17)) have seriously challenged these recommendations. Additionally, four recent meta-analyses about effectiveness of CRRT in critical patients with AKI (18-21) have described no impact on the mortality or secondary outcomes of these techniques. The uncertainty regarding the effectiveness of CRRT in patients with sepsis without renal failure is even greater.

The object of this review was to assess the effectiveness of CRRT at high or conventional doses in patients with severe sepsis or septic shock, with or without (AKI).

MATERIALS AND METHODS

Inclusion and exclusion criteria

Studies were included in the review if they met the following criteria: 1) design: controlled clinical trials (including randomised controlled trials and subgroups of randomised trials); 2) patients: studies conducted in adults (16 years old or greater) with a diagnosis of severe sepsis or septic shock, with or without acute kidney failure, according to the authors' definition; 3) intervention: studies evaluating continuous veno-venous haemofiltration or haemodiafiltration (at high or standard doses) compared with continuous veno-venous haemofiltration or haemodiafiltration at standard doses or no CRRT; studies in which combined continuous and intermittent therapies were initially included, and their impact assessed with sensitivity analysis; 4) outcomes: studies that measured short-term mortality (hospital mortality or mortality at 15 to 90 days).

Along with mortality (primary endpoint), other clinical outcomes were analysed (secondary endpoints) whenever possible: haemodynamic response (mean arterial pressure or use of vasopressor drugs using an explicit protocol), pulmonary gas exchange ([P.sub.a][O.sub.2]/Fi[O.sub.2] ratio), incidence and/or evolution of MODS or length of stay in the intensive care unit.

The exclusion criteria were as follows: 1) interventional studies without external control groups (e.g. crossover trials, studies of observed versus expected mortality); 2) studies without sufficient data to assess mortality, 3) studies evaluating other renal replacement therapies (e.g. peritoneal dialysis or coupled plasma filtration immunoadsorption); 4) studies prior to 1995. No a priori language restrictions were established.

Source of data and search strategy

An electronic search was performed during October 2009 in the following databases: Medline (using PubMed), Embase (using Embase.com), Cochrane Library (CDSR and Clinical Trials Database), Clinicaltrials.gov and Web of Knowledge. The search strategies are described in Table 1.

This search was supplemented by searching the references of the retrieved full-text articles and the summaries from July to October 2009 in the following journals: New England Journal of Medicine, Lancet, Journal of the American Medical Association, Critical Care, Critical Care Medicine, Intensive Care Medicine, Journal of Critical Care, Medicina Intensiva, REMI, Kidney International, Journal of the American Society of Nephrology, Nefrologia, International Journal of Artificial Organs, Artificial Organs, Blood Purification and Nephrology Dialysis Transplantation.

The authors of potentially relevant studies were occasionally contacted to clarify the inclusion criteria (22,23), but were not used as a primary source of data. The distinction between septic and non-septic patients in three studies (24-26) was provided by a recent meta-analysis (20).

Study selection and data extraction

Studies were selected according to the the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (27). The retrieved studies were assembled into a bibliographic database. After eliminating duplicates, the articles were subjected to a screening process from the title and abstract to exclude irrelevant studies. We obtained full-text versions of the pertinent articles to determine whether they met inclusion and exclusion criteria. The results of the selection process are summarised in Figure 1.

The extracted data included the setting of the study, type of patients, renal replacement techniques, clinical outcomes and methodological quality of the studies (Tables 2 to 4). The doses used were classified as traditional (<35 ml/kg/hour), high (35 to 65 ml/kg/hour) or very high (>65 ml/kg/hour).

The methodological quality of studies was assessed according to the recommendations of the Cochrane Collaboration (28), which considers six domains, coded as 1 (no), 2 (unclear) or 3 (yes): generation of a random list to allocate the study subjects, concealed sequence of randomisation, blinding, attrition and exclusions after randomisation, selective reporting of outcomes and other. The latter domain included the following pre-defined features: study design (primary randomised controlled trial versus sepsis subgroup), early stopping by benefit (29), misbalanced baseline prognostic variables, under-dosage of haemofiltration (in the control or experimental group), observed versus expected mortality, statistical power (to detect a mortality reduction of 20%) and Jadad's scale (30).

[FIGURE 1 OMITTED]

The screening and selection of articles and data extraction were performed in duplicate, and disagreements were resolved by consensus.

Statistical methods

Due to the clinical heterogeneity of the studies, the effect of haemofiltration on mortality was analysed using a random effects meta-analysis (DerSimonian-Laird (31)). Statistical heterogeneity was assessed by the Q test, the I (2) statistic (32) and the Galbraith plot (33). The risk of publication bias and/or small study effect was explored by constructing a funnel plot with enhanced contours (34) and Harbord's test (35).

Random effects meta-regression was performed by the residual maximum likelihood method (36). The following pre-selected variables were analysed in the meta-regression model: study design (randomised controlled trial versus randomised trial subgroup), methodological quality of the study (the six domains recommended by the Cochrane Collaboration), renal replacement therapy dose (prescribed or applied), epidemiological design of the study (randomised controlled trial in patients with sepsis versus sepsis subgroup included in a randomised controlled trial), use of renal replacement therapy in the control group, severity of illness (mean Acute Physiology and Chronic Health Evaluation II score in the control group), year of publication and financial support.

Additionally, several exploratory sensitivity analyses were conducted to examine the impact of meta-analysis model (fixed versus random effects), selected association measure (odds ratio vs relative risk) and type of intervention (continuous vs mixed continuous-intermittent techniques).

Given the diversity of measures used, the effects of CRRT on other outcomes (haemodynamics, [P.sub.a][O.sub.2]/ [F.sub.i][O.sub.2] , MODS and length of stay) were analysed using a narrative synthesis of the evidence.

The analyses were performed using the software Stata/IC 11.0, Review Manager 5.0.23, StatsDirect 2.7.7, Reference Manager 12 and GPower 3.0.

RESULTS

Included and excluded studies

Initially, 3776 potentially relevant articles were screened for inclusion in the review. We determined that 3741 of these were not relevant after examination of the title or abstract. After full-text review of the remaining 35 articles, eight studies were excluded because they lacked an external control group (37-44), 10 studies had a control group different from CVVH or no dialysis (45-54) and five additional studies were excluded due to insufficient raw data to assess mortality using a two-by-two table (22,23,55-57). Therefore, 12 studies were finally included in the review (8,9,14,16,24-26,58-62) (Figure 1, Tables 2 and 3).

Three of the included studies (8,58,62) assessed the effect of CVVH versus no CRRT, while the remaining nine studies (9,14,16,24-26,59-61) assessed the effect of higher versus lower doses. Most of these studies included exclusively patients with AKI (9,14,16,24-26,59,61). Only three studies included patients without AKI (8,58,62), two of them (8,58) used low doses of therapy versus standard medical treatment.

Quality of the studies

Whereas all the included studies were described as 'randomised', only six studies described how the random list was generated (8,14,16,25,26,62) (Table 4). The sequence of allocation was concealed in five studies (8,14,16,25,26), and unclear in the remaining seven studies. Due to the nature of the intervention, all of the included studies were non-blinded.

Nine studies were free of significant attrition or exclusions after randomisation (8,14,16,24-26,59,60,62). The risk of bias due to exclusions was considered high in two studies (58,61). One additional study that reported complete follow-up (9) was found later to have excluded patients with septic shock (63), so its risk of bias due to exclusion after randomisation was considered unclear. Four studies had study protocol available and were considered free of selective outcome reporting bias (14,16,26,62). The risk of selective reporting bias was considered high in one study which reported an unusual primary outcome (survival at 15 days after discontinuation of treatment)9 and unclear in the remainder of the studies. Most of the included studies showed other pre-defined limitations, such as the existence of misbalanced groups (58-61), excessive mortality in the control group (9,60), under-dosage in the experimental group (8,26,58) or early stopping by benefit (Saudan et al (25), as reported by Van Wert et al (20)).

[FIGURE 3 OMITTED]

Effect on mortality

There was no evidence of benefit in studies which used renal replacement therapy at traditional doses (less than 35 ml/kg/hour) or high doses (35 to 65 ml/kg/hour) (Figure 2). The relative risk was slightly lower in studies that used more than 65 ml/kg/hour (60,61) (relative risk=0.84, 95% confidence interval 0.59 to 1.19), however there was no statistical interaction between trial group and dose (P=0.237).

Taken as a whole, the included studies showed a moderate degree of statistical heterogeneity (I (2) =52%, P=0.02) (Figure 2) with a pooled risk ratio (random effects) of 0.96 (0.83 to 1.12). The Galbraith plot identified the study of Saudan et al (25) as an outlier. After the exclusion of this study, the heterogeneity was low (I (2) 20%, P=0.25) with a pooled risk ratio of 1.0 (0.90 to 1.11). There was no suggestion of publication bias and/or small studies effect in the funnel plot (Figure 3). These results were robust to the exclusion of the ATN study that used a combination of continuous and intermittent renal replacement therapies (14).

The univariate meta-regression analysis did not detect significant effect of any of the predefined variables including prescribed or applied dose, type of epidemiological study (randomised controlled trials versus subgroups of randomised trials) or risk of bias (Table 5).

Other outcomes

In addition to mortality, several of the randomised studies included in the review assessed other clinical outcomes. However, the heterogeneity of measures used did not allow the conducting of a statistical synthesis.

The haemodynamic effect of haemofiltration was assessed in seven randomised studies (8,14,58-62). One (61) reported a higher proportion of 'responders' (defined as a decreased noradrenaline dose of more than 75% in 24 hours) in the high volume haemofiltration group (P=0.004). The remaining six studies did not find any systematic benefit of haemofiltration on the haemodynamic parameters or vasopressor support.

Pulmonary function was examined in five randomised studies (16,24,59,61,62). No association was detected between haemofiltration and [P.sub.a][O.sub.2]/[F.sub.i][O.sub.2] ratio (59,61,62) or duration of mechanical ventilation (16,24). Five studies (8,59-62) reported the effect of haemofiltration on the evolution of MODS. None showed benefit and one reported a more rapid deterioration of the SOFA scores in the experimental group (P=0.027). Length of stay in the intensive care unit or the hospital was reported in six studies (14,16,24,25,61,62) with negative results. The study of Saudan et al (25) reported a tendency (P=0.06) toward a longer intensive care unit stay in the experimental group. Various studies reported adverse effects of the treatment. Payen et al (62) reported a higher incidence and severity of organ failure in the experimental group. The RENAL study (16) reported an increased incidence of hypophosphataemia in the experimental group (P <0.001). The ATN study (14) reported a higher incidence of hypotension requiring vasopressor therapy (P=0.02), hypophosphataemia (P=0.001) and hypokalaemia (P=0.03) in the experimental group.

DISCUSSION

In contrast to the study of Van Wert (20) which included septic patients with AKI, our study tried to respond to the question of effectiveness of CRRT in relevant clinical outcomes in these patients with or without AKI. Our results suggest that the addition of CRRT or its use at high doses does not improve the clinical outcomes of patients with severe sepsis or septic shock with or without AKI and irrespective of the technique used or the definition of AKI. Albeit conventional haemofiltration, haemofiltration using high cut-off filters, high volume haemofiltration and haemodiafiltration are clearly different, the results are consistent and homogeneous, evidencing a lack of effect. With regard to mortality, only one trial (25) reported a significant reduction in mortality. However this was a small study (based on 28 events) (64), which was stopped early by benefit (28,29,65), which reported an unusual reduction in mortality (risk ratio of 0.31), and that was identified as an outlier in the tree plot (Figure 2) and funnel plot (Figure 3). Therefore, there is a high probability that it was a false positive. After exclusion of this trial, the heterogeneity was greatly reduced and the pooled relative risk was 1.

A specific consideration should be done with respect to three studies comparing conservative treatment versus CVVH or high volume haemofiltration (58,62), or in patients without AKI (8) respectively. Although it is doubtful whether these studies should be analysed together due to differences in design, a subgroup analysis did not reveal any subgroup effect.

One concern with our study could be the mixing of different types of renal replacement therapies, specially continuous and intermittent. Only one study (14) included both types of techniques and this issue was specifically addressed in the meta-regression analysis. The effect on mortality did not change if this study was included or not, showing that the schedule of application of renal replacement therapies was not a factor capable to modifying our conclusions.

With respect to other outcomes such as improvement in haemodynamic status or pulmonary oxygenation, much of the available evidence comes from animal and non-randomised studies (mainly pre-post studies without external control groups (37-39,41)) not included in this review. However, the evidence based on randomised controlled trials is consistent with that of mortality. Only one study with significant methodological limitations reported a reduction in the use of vasopressors in the experimental group (61), and none of the trials reviewed reported an improvement in gas exchange, duration of mechanical ventilation, development of MODS or length of stay. Respect to other outcomes, two recent meta-analyses (18,20) found no effect of high-dose renal replacement therapy on dialysis dependence or length of stay in patients with AKI.

We did not detect any difference of effect of haemofiltration according to the three groups of doses used. However, only two small studies used doses higher than 65 ml/kg/hour. Therefore our study does not preclude the efficacy of these doses in patients with severe sepsis or septic shock. The dose for attaining a sepsis could very likely be different from the dose used for renal support in AKI. Currently there is an ongoing randomised clinical trial (66) addressing this issue. In any case, the results of our review do not support the routine use of doses higher than 35 ml/kg in patients with severe sepsis with or without AKI.

Similarly, this review is limited to studies comparing high-dose haemofiltration-haemodiafiltration or standard haemofiltration-haemodiafiltration versus traditional dosage or no haemofiltration. Thus, the study results cannot be generalised to other haemofiltration techniques with dialysis (e.g. high adsorption filters, filters of high porosity or plasmapheresis).

A further limitation of our study is that six of the 12 studies which met the inclusion criteria were actually not designed to study patients with severe sepsis and septic shock. These studies evaluated patients with AKI and some had very low numbers of septic patients. Furthermore, these groups of septic patients may not have been defined in the same way across studies. Therefore, the external validity of our study is limited by the scarcity of randomised controlled trials addressing specifically clinical outcomes of renal replacement therapies in septic patients. Indeed, almost all the studies that compared high versus low doses were performed in patients with AKI. The effect of high doses in septic patients without acute kidney injury therefore cannot be fully evaluated until well-designed and powered trials are performed.

Finally, the efficacy of haemofiltration in patients with non-infectious systemic inflammatory response syndrome is beyond the scope of this review. It is possible that patients with systemic inflammatoryresponse syndrome (post-cardiac arrest syndrome (67), severe trauma (68,69), pancreatitis (23), severe burns (44)) experience a massive release of mediators and therefore may benefit from early haemofiltration. In contrast, patients with sepsis undergo haemofiltration at a later stage in the course of the disease. It can be hypothesised that the haemofiltration in patients with sepsis is performed outside the therapeuticwindow when organ damage has already occurred. Further research is needed to address this issue.

CONCLUSION

The best evidence available does not support the routine use of CRRT in patients with sepsis. Further research is necessary regarding the efficacy of early high-dose CRRT in patients with severe systemic inflammatory response syndrome of non-infectious origin.

Web of knowledge

(Sepsis OR SIRS [topic]) AND (haemofiltration OR renal dialysis OR renal replacement therapy [topic]) AND (mortality OR survival OR organ dysfunction OR arterial pressure OR vasoactive drug OR shock reversal OR hypoperfusion OR lactate [topic])

Clinicaltrials.gov

(Sepsis OR SIRS) AND (haemofiltration OR renal dialysis OR renal replacement therapy) AND (mortality OR survival OR organ dysfunction OR arterial pressure OR vasoactive drug OR shock reversal OR hypoperfusion OR lactate)

SIRS=systemic inflammatory response syndrome.

FINANCIAL DECLARATION

Dr E. Palencia-Herrejon has received financial support from Baxter Laboratories for attending conferences.

REFERENCES

(1.) Vincent J-L, Sakr Y, Sprung CL, Ranieri VM, Reinhart K, Gerlach H et al. Sepsis in European intensive care units: results of the SOAP study. Crit Care Med 2006; 34: 344-353.

(2.) Brun-Buisson C, Meshaka P, Pinton P, Vallet B. EPISEPSIS: a reappraisal of the epidemiology and outcome of severe sepsis in French intensive care units. Intensive Care Med 2004; 30: 580-588.

(3.) Esteban A, Frutos-Vivar F, Ferguson ND, Penuelas O, Lorente JA, Gordo F et al. Sepsis incidence and outcome: contrasting the intensive care unit with the hospital ward. Crit Care Med 2007; 35: 1284-1289.

(4.) Adrie C, Alberti C, Chaix-Couturier C, Azoulay E, De Lassence A, Cohen Y et al. Epidemiology and economic evaluation of severe sepsis in France: age, severity, infection site, and place of acquisition (community, hospital, or intensive care unit) as determinants of workload and cost. J Crit Care 2005; 20: 46-58.

(5.) Pinsky MR, Vincent JL, Deviere J, Alegre M, Kahn RJ, Dupont E. Serum cytokine levels in human septic shock. Relation to multiple-system organ failure and mortality. Chest 1993; 103: 565-575.

(6.) Bone RC, Grodzin CJ, Balk RA. Sepsis: a new hypothesis for pathogenesis of the disease process. Chest 1997; 112: 235-243.

(7.) Koch T. Origin and mediators involved in sepsis and the systemic inflammatory response syndrome. Kidney Int Suppl 1998; 64: S66-69.

(8.) Cole L, Bellomo R, Hart G, Journois D, Davenport P, Tipping P et al. A phase II randomized, controlled trial of continuous hemofiltration in sepsis. Crit Care Med 2002; 30: 100-106.

(9.) Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P et al. Effects of different doses in continuous veno-venous haemofiltration on outcomes of acute renal failure: a prospective randomised trial. Lancet 2000; 356: 26-30.

(10.) Ronco C, Tetta C, Mariano F, Wratten ML, Bonello M, Bordoni V et al. Interpreting the mechanisms of continuous renal replacement therapy in sepsis: the peak concentration hypothesis. Artif Organs 2003; 27: 792-801.

(11.) Bagshaw SM, George C, Bellomo R. Early acute kidney injury and sepsis: a multicentre evaluation. Crit Care 2008; 12: R47.

(12.) Bellomo R, Honore PM, Matson J, Ronco C, Winchester J. Extracorporeal blood treatment (EBT) methods in SIRS/Sepsis. Int J Artif Organs 2005; 28: 450-458.

(13.) Bouman CSC, Oudemans-van SHM, Schultz MJ, Vroom MB. Hemofiltration in sepsis and systemic inflammatory response syndrome: the role of dosing and timing. J Crit Care 2007; 22: 1-12.

(14.) Palevsky PM, Zhang JH, O'Connor TZ, Chertow GM, Crowley ST, Choudhury D et al. Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med 2008; 359: 7-20.

(15.) Bonventre JV Dialysis in acute kidney injury--more is not better. N Engl J Med 2008; 359: 82-84.

(16.) Bellomo R, Cass A, Cole L, Finfer S, Gallagher M, Lo S et al. Intensity of continuous renal-replacement therapy in critically ill patients. N Engl J Med 2009; 361: 1627-1638.

(17.) Palevsky PM. Renal support in acute kidney injury--how much is enough? N Engl J Med 2009; 361: 1699-1701.

(18.) Casey ET, Gupta BP, Erwin PJ, Montori VM, Murad MH. The dose of continuous renal replacement therapy for acute renal failure: a systematic review and meta-analysis. Ren Fail 2010; 32: 555-561.

(19.) Jun M, Heerspink HJL, Ninomiya T, Gallagher M, Bellomo R, Myburgh J et al. Intensities of renal replacement therapy in acute kidney injury: a systematic review and meta-analysis. Clin J Am Soc Nephrol 2010; 5: 956-963.

(20.) Van Wert R, Friedrich JO, Scales DC, Wald R, Adhikari NKJ. High-dose renal replacement therapy for acute kidney injury: Systematic review and meta-analysis. Crit Care Med 2010; 38: 1360-1369.

(21.) Zhongheng Z, Xiao u, Hongyang Z. Intensive- vs less-intensive-dose continuous renal replacement therapy for the intensive care unit-related acute kidney injury: a meta-analysis and systematic review. J Crit Care 2010; 25: 595-600.

(22.) Gasparovic V, Filipovic-Grcic I, Merkler M, Pisl Z. Continuous renal replacement therapy (CRRT) or intermittent hemodialysis (IHD)--what is the procedure of choice in critically ill patients? Ren Fail 2003; 25: 855-862.

(23.) Jiang H-L, Xue W-J, Li D-Q, Yin A-P, Xin X, Li C-M, Gao J-L. Influence of continuous veno-venous hemofiltration on the course of acute pancreatitis. World J Gastroenterol 2005; 11: 4815-4821.

(24.) Bouman CSC, Oudemans-Van SHM, Tijssen JGP, Zandstra DF, Kesecioglu J. Effects of early high-volume continuous venovenous hemofiltration on survival and recovery of renal function in intensive care patients with acute renal failure: a prospective, randomized trial. Crit Care Med 2002; 30: 2205-2211.

(25.) Saudan P, Niederberger M, De Seigneux S, Romand J, Pugin J, Perneger T et al. Adding a dialysis dose to continuous hemofiltration increases survival in patients with acute renal failure. Kidney Int 2006; 70: 1312-1317.

(26.) Tolwani AJ, Campbell RC, Stofan BS, Lai KR, Oster RA, Wille KM. Standard versus high-dose CVVHDF for ICU-related acute renal failure. J Am Soc Nephrol 2008; 19: 1233-1238.

(27.) Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009; 6: e1000097.

(28.) Higgins J, Altman De. Assessing risk of bias in included studies. In: Higgins J, Green Se, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 501 (updated September 2008). The Cochrane Collaboration 2008.

(29.) Montori VM, Devereaux PJ, Adhikari NKJ, Burns KEA, Eggert CH, Briel M et al. Randomized trials stopped early for benefit: a systematic review. JAMA 2005; 294: 2203-2209.

(30.) Jadad A. Randomised Controlled Trials. BMJ Books, London 1999.

(31.) DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177-188.

(32.) Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557-560.

(33.) Galbraith RF. A note on graphical presentation of estimated odds ratios from several clinical trials. Stat Med 1988; 7: 889-894.

(34.) Peters JL, Sutton AJ, Jones DR, Abrams KR, Rushton L. Contour-enhanced meta-analysis funnel plots help distinguish publication bias from other causes of asymmetry. J Clin Epidemiol 2008; 61: 991-996.

(35.) Harbord RM, Egger M, Sterne JAC. A modified test for small-study effects in meta-analyses of controlled trials with binary endpoints. Stat Med 2006; 25: 3443-3457.

(36.) Thompson SG, Sharp SJ. Explaining heterogeneity in meta-analysis: a comparison of methods. Stat Med 1999; 18: 2693-2708.

(37.) Lange R, Erhard J, Sander A, Scherer R, Eigler F. [Use of continuous veno-venous high volume hemofiltration (CVVHVH) in patients with multiple organ failure. Principle and applications]. Zentralbl Chir 1996; 121: 535-540.

(38.) Honore PM, Jamez J, Wauthier M, Lee PA, Dugernier T, Pirenne B et al. Prospective evaluation of short-term, high-volume isovolemic hemofiltration on the hemodynamic course and outcome in patients with intractable circulatory failure resulting from septic shock. Crit Care Med 2000; 28: 3581-3587.

(39.) Klouche K, Cavadore P, Portales P, Clot J, Canaud B, Beraud JJ. Continuous veno-venous hemofiltration improves hemodynamics in septic shock with acute renal failure without modifying TNFalpha and IL6 plasma concentrations. J Nephrol 2002; 15: 150-157.

(40.) Kellum JA, Bellomo R, Mehta R, Ronco C. Blood purification in non-renal critical illness. Blood Purif 2003; 21: 6-13.

(41.) Cornejo R, Downey P, Castro R, Romero C, Regueira T, Vega J et al. High-volume hemofiltration as salvage therapy in severe hyperdynamic septic shock. Intensive Care Med 2006; 32: 713-722.

(42.) Piccinni P, Dan M, Barbacini S, Carraro R, Lieta E, Marafon S et al. Early isovolaemic haemofiltration in oliguric patients with septic shock. Intensive Care Med 2006; 32: 80-86.

(43.) Bagshaw SM, Berthiaume LR, Delaney A, Bellomo R. Continuous versus intermittent renal replacement therapy for critically ill patients with acute kidney injury: a meta-analysis. Crit Care Med 2008; 36: 610-617.

(44.) Chung KK, Juncos LA, Wolf SE, Mann EE, Renz EM, White C et al. Continuous renal replacement therapy improves survival in severely burned military casualties with acute kidney injury. J Trauma 2008; 64: S179-185.

(45.) Bellomo R, Farmer M, Parkin G, Wright C, Boyce N. Severe acute renal failure: a comparison of acute continuous hemodiafiltration and conventional dialytic therapy. Nephron 1995; 71: 59-64.

(46.) Dillon JJ. Continuous renal replacement therapy or hemodialysis for acute renal failure? Int J Artif Organs 1999; 22: 125-127.

(47.) John S, Griesbach D, Baumgartel M, Weihprecht H, Schmieder RE, Geiger H. Effects of continuous haemofiltration vs intermittent haemodialysis on systemic haemodynamics and splanchnic regional perfusion in septic shock patients: a prospective, randomized clinical trial. Nephrol Dial Transplant 2001; 16: 320-327.

(48.) Phu NH, Hien TT, Mai NTH, Chau TTH, Chuong LV, Loc PP et al. Hemofiltration and peritoneal dialysis in infection-associated acute renal failure in Vietnam. N Engl J Med 2002; 347: 895-902.

(49.) Augustine JJ, Sandy D, Seifert TH, Paganini EP. A randomized controlled trial comparing intermittent with continuous dialysis in patients with ARF. Am J Kidney Dis 2004; 44: 1000-1007.

(50.) Uehlinger DE, Jakob SM, Ferrari P, Eichelberger M, Huynh-Do U, Marti H-P et al. Comparison of continuous and intermittent renal replacement therapy for acute renal failure. Nephrol Dial Transplant 2005; 20: 1630-1637.

(51.) Vinsonneau C, Camus C, Combes A, Costa de Beauregard MA, Klouche K, Boulain T et al. Continuous venovenous haemodiafiltration versus intermittent haemodialysis for acute renal failure in patients with multiple-organ dysfunction syndrome: a multicentre randomised trial. Lancet 2006; 368: 379-385.

(52.) Faulhaber-Walter R, Hafer C, Jahr N, Vahlbruch J, Hoy L, Haller H et al. The Hannover Dialysis Outcome study: comparison of standard versus intensified extended dialysis for treatment of patients with acute kidney injury in the intensive care unit. Nephrol Dial Transplant 2009; 24: 2179-2186.

(53.) Lins RL, Elseviers MM, Van der Niepen P, Hoste E, Malbrain ML, Damas P et al. Intermittent versus continuous renal replacement therapy for acute kidney injury patients admitted to the intensive care unit: results of a randomized clinical trial. Nephrol Dial Transplant 2009; 24: 512-518.

(54.) Zhao S, Wu J, Ai Y, Sun B, Xu D, Guo Q. [Comparison of clinical efficacy between continuous veno-venous hemofiltration (CVVH) and CVVH combined with hemoperfusion for the treatment of multiple organ dysfunction syndrome]. Zhongguo Wei Zhong Bing Ji Jiu Yi Xue 2009; 21: 373-374.

(55.) Bauer M, Marzi I, Ziegenfuss T, Riegel W. Prophylactic hemofiltration in severely traumatized patients: effects on posttraumatic organ dysfunction syndrome. Intensive Care Med 2001; 27: 376-383.

(56.) Yang Z-h, Yang J, Wang Y. [Protective effect of continuous veno-venous hemofiltration on tissue and organ damage in patients with severe acute pancreatitis]. Zhongguo Wei Zhong Bing Ji Jiu Yi Xue 2004; 16: 232-234.

(57.) Kawanishi H, Shintaku S, Shishida M, Moriishi M, Tsuchiya S. The Effect of Higher Dialysis Dosing in Critical Ill With AKI and Dialysis Patients. Blood Purification 2009; 28: 305.

(58.) Sander A, Armbruster W, Sander B, Daul AE, Lange R, Peters J. Hemofiltration increases IL-6 clearance in early systemic inflammatory response syndrome but does not alter IL-6 and TNF alpha plasma concentrations. Intensive Care Med 1997; 23: 878-884.

(59.) Morgera S, Slowinski T, Melzer C, Sobottke V, Vargas-Hein O, Volk T et al. Renal replacement therapy with high-cutoff hemofilters: impact of convection and diffusion on cytokine clearances and protein status. Am J Kidney Dis 2004; 43: 444-453.

(60.) Ghani RA, Zainudin S, Ctkong N, Rahman AFA, Wafa SRWSH, Mohamad M et al. Serum IL-6 and IL-1-ra with sequential organ failure assessment scores in septic patients receiving high-volume haemofiltration and continuous venovenous haemofiltration. Nephrology (Carlton) 2006; 11: 386-393.

(61.) Boussekey N, Chiche A, Faure K, Devos P, Guery B, d'Escrivan T et al. A pilot randomized study comparing high and low volume hemofiltration on vasopressor use in septic shock. Intensive Care Med 2008; 34: 1646-1653.

(62.) Payen D, Mateo J, Cavaillon JM, Fraisse F, Floriot C, Vicaut E. Impact of continuous venovenous hemofiltration on organ failure during the early phase of severe sepsis: a randomized controlled trial. Crit Care Med 2009; 37: 803-810.

(63.) Piccinni P, Ronco C. Early isovolemic hemofiltration in oliguric patients with septic shock [author reply]. Intensive Care Med 2006; 32: 1097.

(64.) Pogue J, Yusuf S. Overcoming the limitations of current meta-analysis of randomised controlled trials. Lancet 1998; 351: 47-52.

(65.) Bassler D, Briel M, Montori VM, Lane M, Glasziou P, Zhou Q et al. Stopping randomized trials early for benefit and estimation of treatment effects: systematic review and meta regression analysis. JAMA 2010; 303: 1180-1187.

(66.) Haemofiltration Study: IVOIRE (High Volume in Intensive Care). From http://www.clinicaltrials.gov/ct2/show/ NCT00241228?term = IVOIRE&rank=1 Accessed August 2010.

(67.) Laurent I, Adrie C, Vinsonneau C, Cariou A, Chiche J-D, Ohanessian A et al. High-volume hemofiltration after out-of-hospital cardiac arrest: a randomized study. J Am Coll Cardiol 2005; 46: 432-437.

(68.) Sanchez-Izquierdo JA, Perez Vela JL, Lozano Quintana MJ, Alted Lopez E, Ortuno de Solo B, Ambros Checa A. Cytokines clearance during venovenous hemofiltration in the trauma patient. Am J Kidney Dis 1997; 30: 483-488.

(69.) Sanchez-Izquierdo RJA, Alted E, Lozano MJ, Perez JL, Ambros A, Caballero R. Influence of continuous hemofiltration on the hemodynamics of trauma patients. Surgery 1997; 122: 902-908.

J. LATOUR-PEREZ *, E. PALENCIA-HERREJON ([dagger]), V. GOMEZ-TELLO([dagger])([dagger]), A. BAEZA-ROMAN ([section]), M. A. GARCIA-GARCIA **, B. SANCHEZ-ARTOLA ([dagger])([dagger])

Intensive Care Unit, Elche University General Hospital, Elche, Spain

* M.D., Ph.D., Specialist in Intensive Care Medicine and Clinical Head.

([dagger]) M.D., Ph.D., Specialist in Intensive Care Medicine and Clinical Head, Intensive Care Unit, Hospital Infanta Leonor, Madrid.

([dagger])([dagger])M.D., Ph.D., Intensivist, Intensive Care Unit, Hospital Moncloa.

([section]) M.D., Resident in Intensive Care Medicine.

** M.D., Specialist in Intensive Care Medicine and Consultant, Intensive Care Unit, Hospital de Sagunt, Sagunto.

([dagger])([dagger]) M.D., Specialist in Intensive Care Medicine and Consultant, Department of Internal Medicine, Hospital Infanta Leonor.

Address for correspondence: Dr V. Gomez-Tello, Hospital Moncloa, Unidad de Cuidados Intensivos, Av. Valladolid, 83 28008 Madrid, Spain.

Email: vgtello@gmail.com

Accepted for publication on December 2, 2010.
TABLE 1

Search strategy

PubMed

#1        (SIRS OR systemic inflammatory response syndrome
          OR infection OR sepsis OR septic shock OR shock)

#2        (cardiac arrest OR trauma OR acute pancreatitis
          OR burns OR acute renal failure)

#3        (#1) OR (#2)

#4        (renal dialysis OR renal replacement therapy OR
          haemofiltration OR haemodiafiltration)

#5        (mortality OR survival OR organ dysfunction OR
          arterial pressure OR vasoactive drug OR shock
          reversal OR hypoperfusion OR lactate)

#6        ((#3) AND (#4) AND (#5))

#7        "1995" [pdat] : "2009" [pdat]

#8        (Clinical Trial [ptyp] OR Meta-Analysis [ptyp] OR
          Randomised Controlled Trial [ptyp] OR Comparative
          Study [ptyp])

#9        ("humans"[MeSH Terms])

#10       (# 6 AND #7 AND #8 AND #9)

Embase

#1        SIRS OR systemic AND inflammatory AND
          response AND ("syndrome"/exp OR syndrome)
          OR "infection"/exp OR infection OR "sepsis"/exp
          OR sepsis OR septic AND ("shock"/exp OR shock)
          OR "shock"/exp OR shock OR (cardiac AND arrest
          OR "trauma"/exp OR trauma OR acute AND
          ("pancreatitis"/exp OR pancreatitis) OR "burns"/
          exp OR burns OR acute AND renal AND failure)
          AND [1995-2009]/py

#2        renal AND ("dialysis"/exp OR dialysis) OR renal
          AND replacement AND ("therapy"/exp OR therapy)
          OR "haemofiltration"/exp OR haemofiltration OR
          "haemodiafiltration"/exp OR haemodiafiltration OR
          "dialysis'/exp OR dialysis OR "haemodialysis"/exp
          OR haemodialysis OR dialytic AND [1995-2009]/py

#3        "mortality"/exp OR mortality OR "survival"/exp OR
          survival OR "organ"/exp OR organ AND dysfunction
          OR arterial AND ("pressure"/exp OR pressure) OR
          vasoactive AND ("drug"/exp OR drug) OR "shock"/
          exp OR shock AND reversal OR "hypoperfusion"/
          exp OR hypoperfusion OR "lactate"/exp OR lactate
          AND [1995-2009]/py

#4        #1 AND #2 AND #3

Web of knowledge

TABLE 2

Included studies (patients and setting)

Study,        Patients                 Country, number
year                                   of centres

Sander,       Surgical                 Germany, 1
1997 (58)

Ronco,        AKI (sepsis subgroup)    Italy, 1
2000 (9)

Bouman,       ICU patients with AKI    Netherlands, 2
2002 (24)     (predominant post-
              cardiac surgery)
              (sepsis subgroup)

Cole,         Sepsis                   Australia, 1
2002 (8)

Morgera,      Sepsis                   Germany, 1
2004 (59)

Ghani,        Sepsis                   Malaysia, 1
2006 (60)

Saudan,       ICU patients with AKI    Switzerland, 1
2006 (25)     (sepsis subgroup)

Tolwani,      ICU patients with AKI    USA, 1
2008 (26)     (sepsis subgroup)

ATN,          AKI (sepsis subgroup)    USA, 27
2008 (14)

Boussekey,    Septic shock             France, 1
2008 (61)

Payen,        Sepsis                   France, 12
2009 (62)

RENAL,        AKI (sepsis subgroup)    Australia and
2009 (16)                              New Zealand,
                                       35

Study,        Financial    Inclusion criteria
year          support

Sander,       Not          Sepsis
1997 (58)     reported

Ronco,        Not          AKI
2000 (9)      reported

Bouman,       Not          Oliguric acute kidney
2002 (24)     reported     failure plus mechanical
                           ventilation

Cole,         Mixed        Severe sepsis or septic
2002 (8)      public-      shock
              private
Morgera,      Not          Sepsis plus acute kidney
2004 (59)     reported     failure and MODS

Ghani,        Private      Severe sepsis or septic
2006 (60)                  shock

Saudan,       Not          Clinical diagnosis of acute
2006 (25)     reported     renal failure

Tolwani,      Mixed        Clinical diagnosis of acute
2008 (26)     public-      renal failure
              private

ATN,          Public       Adult patients with AKI
2008 (14)                  and failure of one or more
                           non-renal organ systems
                           (SOFA score [greater than
                           or equal to] 2) or sepsis

Boussekey,    Hospital     Septic shock with AKI
2008 (61)

Payen,        Private      Severe sepsis or septic
2009 (62)                  shock <24 h and SAPS
                           score 35-63

RENAL,        Mixed        Critical adult patients
2009 (16)     public-      AKI and need for renal
              private      replacement therapy

Study,        Main exclusion criteria            Mean APACHE-II
year                                             (predicted mortality)

Sander,       1) Age <18 or >80; 2) pregnancy;   14 (18.6%)
1997 (58)     3) recent sepsis; 4) chronic
              renal failure; 5)
              contraindications against
              systemic anticoagulation; 6)
              immunosuppression or
              immunodeficiency

Ronco,        No informed consent; septic        23 (46.0%)
2000 (9)      shock (63)

Bouman,       1) Previous renal failure; 2)      23 (46.0 %)
2002 (24)     renal failure not secondary to
              acute tubular necrosis; 3)
              severe comorbidity (post-cardiac
              arrest encephalopathy, AIDS,
              grade-C cirrhosis)

Cole,         1) End-stage renal failure; 2)     22 (42.4%)
2002 (8)      malignancy; 3) AIDS; 4) life
              expectancy <6 months; 5)
              possible life support withdrawal

Morgera,      Not reported                       31 (%)
2004 (59)

Ghani,        1) end-stage renal disease; 2)     Not reported
2006 (60)     malign neoplasm; 3) AIDS; 4)
              life expectancy <6 months

Saudan,       1) Pre-or post-renal failure; 2)   25 (53.3%)
2006 (25)     Suspicion of glomerular disease;
              3) end-stage renal failure; 4)
              patients on angiotensin-
              converting enzyme inhibitors

Tolwani,      1) End-stage renal disease; 2)     26 (56.9%)
2008 (26)     previous intermittent
              haemodialysis; 3) >24 h of CRRT
              at time of enrollment; 4) Body
              weight >125 kg or <50 kg

ATN,          1) advanced nephropathy; 2)        26 (56.9%)
2008 (14)     acute kidney failure not due to
              acute tubular necrosis; 3) >72 h
              from the beginning of the AKI
              and BUN >100 mg/dl

Boussekey,    1) Severe chronic renal failure;   32 (76.0%)
2008 (61)     2) patients included in another
              study; 3) severe
              immunosuppression; 4) moribund;
              limitation of therapy; septic
              shock or renal failure >5 days
              after ICU admission; 5) absence
              of written consent; 6) patients
              who died within the first day
              after randomisation

Payen,        1) Pregnancy; 2) age <18; 3)       Not reported
2009 (62)     moribund; 4) chronic renal
              failure; 5) immunosuppression

RENAL,        Previous dialysis, chronic         APACHE-III 102
2009 (16)     dialysis, moribund, weight <60     (approximate
              or >100-120 kg                     mortality 70-80%)

APACHE=Acute Physiology and Chronic Health Evaluation score,
AKI=acute kidney injury, ICU=intensive care unit, AIDS=acquired
immune deficiency syndrome, MODS=multiple organ dysfunction
syndrome, CRRT=continuous renal replacement therapies,
SOFA=sequential organ failure assessment, BUN=blood urea nitrogen,
SAPS=Simplified Acute Physiology Score.

TABLE 3

Renal replacement techniques

Study,       Renal replacement           Blood flow
year         therapy                     (ml/min)

Sander,      CVVH vs no                  150
1997 (58)    haemofiltration

Ronco,       HVHF vs CVVH                120-240
2000 (9)

Bouman,      HVHF vs CVVH                100-200
2002 (24)

Cole,        CVVH vs no                  200
2002 (8)     haemofiltration

Morgera,     CVVHDF vs CVVH              Not reported
2004 (59)

Ghani,       HVHF vs CVVH                250-350
2006 (60)

Saudan,      CVVHDF vs CVVH              100-125
200625

Tolwani,     CVVHDF                      100-150
2008 (26)

ATN,         IHD+SLED (each day)         360 (IHD)/
2008 (14)    + CVVHDF-CVVH               220-210 (CRRT)
             (high dose) vs IHD + SLED
             (alternate days) +
             CVVHDF-CVVH
             (standard dose)

Boussekey,   HVHF (very high dose) vs    180-300
2008 (61)    CVVH (high dose)

Payen,       CVVH vs no                  150
2009 (62)    haemofiltration

RENAL,       CVVHDF                      >150
2009 (16)

Study,       Dose prescribed            Dose applied
year

Sander,      1 l/h                      Not recorded
1997 (58)

Ronco,       35/45 vs 20 ml/kg/h        33, 6/42, 4 vs 18,
2000 (9)                                9 ml/kg/h

Bouman,      72/96 vs 24/36 l/24 h      48.2 vs 19.5 ml/kg/h
2002 (24)

Cole,        2 l/h                      Not reported
2002 (8)

Morgera,     2.5 vs 1 l/h               Not reported
2004 (59)

Ghani,       100 ml/kg/h (or 6 l/h)     Not reported
2006 (60)    6 h vs 2 l/h
             (35 ml/kg/h approx)

Saudan,      42 ml/kg/h (CVVHDF)        34.9 vs 21.8 ml/kg/h
200625       vs 25 ml/kg/h (CVVH)

Tolwani,     35 ml/kg/h                 29 vs 17 ml/kg/h
2008 (26)

ATN,         Kt/V >1.2 (IHD) + 36,      Kt/V 1, 32 (IHD) + 35,
2008 (14)    2 ml/kg/h (CRRT) vs        8 ml/kg/h (CRRT) vs
             Kt/V >1.2 (IHD) + 21,      Kt/V 1, 31 (IHD) + 22,
             5 ml/kg/h (CRRT)           0 ml/kg/h (CRRT)

Boussekey,   65 vs 35 ml/kg/h           62 vs 32 ml/kg/h
2008 (61)

Payen,       25 ml/kg/h                 24, 7 ml/kg/h
2009 (62)

RENAL,       40 vs 25 ml/kg/h           33.4 vs 22.0 ml/kg/h
2009 (16)

Study,       Filter ([m.sup.2])    Anticoagulation
year

Sander,      PAN (0.6 [m.sup.2])   Heparin
1997 (58)

Ronco,       Polysulfone           Heparin
2000 (9)     (0.7-1.3 [m.sup.2])

Bouman,      Cellulose             Heparin/nadroparin/no
2002 (24)    triacetate
             (1.9 [m.sup.2])

Cole,        PAN (1.2 [m.sup.2])   Heparin (regional)
2002 (8)

Morgera,     Polyamide             Heparin
2004 (59)    (1.1 [m.sup.2])

Ghani,       Polysulfone           Heparin/no
2006 (60)    (1.4 [m.sup.2])

Saudan,      PAN (0.9 [m.sup.2])   Not reported
200625

Tolwani,     PAN (0.9 [m.sup.2])   No/citrate
2008 (26)

ATN,         Various               No/heparin/citrate
2008 (14)    (polysulfone,
             PAN)

Boussekey,   Polysulfone           No/heparin
2008 (61)    (1.4 [m.sup.2])

Payen,       Polysulfone           Heparin
2009 (62)    (1.2 [m.sup.2])

RENAL,       PAN                   48% heparin
2009 (16)                          prefilter, 19%
                                   heparin+protamine,
                                   49% no anticoagulation

Study,       Reposition     Substitution
year                        fluid

Sander,      Not reported   Ringer
1997 (58)

Ronco,       Postdilution   Lactate
2000 (9)

Bouman,      Postdilution   Bicarbonate
2002 (24)

Cole,        Predilution    Lactate
2002 (8)

Morgera,     Postdilution   Bicarbonate
2004 (59)

Ghani,       Pre and        Bicarbonate
2006 (60)    postdilution
             (1/2)

Saudan,      Predilution    Bicarbonate
200625

Tolwani,     Predilution    Not
2008 (26)                   reported

ATN,         Predilution    Bicarbonate
2008 (14)    dominant       (except
                            citrate)

Boussekey,   Pre (1/3) or   Bicarbonate
2008 (61)    postdilution
             (2/3)
Payen,       Not recorded   Bicarbonate
2009 (62)

RENAL,       Postdilution   Bicarbonate
2009 (16)

CVVH=continuous veno-venous haemofiltration, PAN=polyacrylonitrile,
HVHF=high-volume haemofiltration, CVVHDF=continuous veno-venous
haemodiafiltration, IHD=intermittent haemodialysis, SLED=sustained
low-efficiency dialysis, CRRT=continuous renal replacement therapy.

TABLE 4

Included studies (risk of bias) *

Study        Design             Random       Concealed
                                list         randomisation
                                generation

Sander,      Randomised         Unclear      Unclear
1997 (58)    study

Ronco,       Subgroup of        Unclear      Unclear
2000 (9)     randomised
             controlled trial

Bouman,      Subgroup of        Unclear      Unclear
2002 (24)    randomised
             controlled trial

Cole,        Randomised         Yes          Yes
2002 (8)     study

Morgera,     Randomised         Unclear      Unclear
2004 (59)    study

Ghani,       Randomised         Unclear      Unclear
2006 (60)    study

Saudan,      Subgroup of        Yes          Yes
2006 (25)    randomised
             controlled trial

Tolwani,     Subgroup of        Yes          Yes
2008 (26)    randomised
             controlled trial

ATN,         Subgroup of        Yes          Yes
2008 (14)    randomised
             controlled trial

Boussekey,   Randomised         Unclear      Unclear
200861       study

Payen,       Randomised         Yes          Unclear
2009 (62)    study

RENAL,       Subgroup of        Yes          Yes
2009 (16)    randomised
             controlled trial

Study        Attritions-   Other domains
             exclusions
             addressed

Sander,      No #          No: baseline misbalance favouring
1997 (58)                  experimental group. High control
                           group mortality (92%). Sub-dosage
                           in the experimental group

Ronco,       Unclear       No: undefined population.
2000 (9)     ([dagger])    Misbalanced groups. High control
                           group mortality (expected 42%,
                           observed 75%). Unusual definition
                           of mortality

Bouman,      Yes           Unclear: misbalanced groups
2002 (24)

Cole,        Yes           Unclear: small study. Sub-dosage
2002 (8)                   in the experimental group

Morgera,     Yes           Unclear: misbalanced groups
2004 (59)

Ghani,       Yes           No: misbalanced groups. High
2006 (60)                  control group mortality

Saudan,      Yes           No: early stopped by benefit
2006 (25)

Tolwani,     Yes           No: misbalanced groups (higher
2008 (26)                  proportion of mechanical
                           ventilation in the experimental
                           group); sub-dosage in both arms

ATN,         Yes           Unclear: late initiation of
2008 (14)                  haemofiltration. Low renal
                           recovery rate. Relative high doses
                           in the experimental group

Boussekey,   No            No: misbalanced groups. Relative
200861                     low mortality in the experimental
                           group

Payen,       Yes           Unclear: early stopped by harm
2009 (62)

RENAL,       Yes           Yes
2009 (16)

Study        Statistical power   Jadad's
             (to detect a        scale
             20% mortality       score
             reduction)

Sander,      26%                 2
1997 (58)

Ronco,       19%                 2
2000 (9)

Bouman,      14%                 2
2002 (24)

Cole,        6%                  3
2002 (8)

Morgera,     8%                  2
2004 (59)

Ghani,       18%                 2
2006 (60)

Saudan,      39%                 3
2006 (25)

Tolwani,     36%                 3
2008 (26)

ATN,         80%                 3
2008 (14)

Boussekey,   8%                  2
200861

Payen,       15%                 2
2009 (62)

RENAL,       79%                 3
2009 (16)

* All the studies were non-blinded. No serious selective reporting
was detected. # Results analysed in this review as intention to treat.

([dagger]) No-specified exclusion of septic shock patients:
see Piccinni 2006 (63).

TABLE 5

Subgroup effect (meta-regression analysis)

                                         Relative
                                         odds               Residual
Variable                                 ratio      P       [I.sup.2]

CRRT dose experimental group             0.576      0.237   55.3%
(prescribed) (>65/35-65/<35 ml/kg/h)

CRRT dose experimental group             0.998      0.955   60.5%
(prescribed) (continuous variable)

CRRT dose experimental group (applied)   0.982      0.644   70.4%
(continuous variable)

CRRT control group (yes vs no)           0.868      0.840   59.7%

APACHE-II score control group            1.065      0.514   65.7%
(continuous variable)

Type of study (RCT sepsis subgroup)      0.879      0.824   60.0%
(yes vs no)

Random list generation                   0.899      0.856   60.3%
(yes/unclear/no)

Concealed randomisation                  0.697      0.515   59.3%
(yes/unclear/no)

No post/randomisation exclusions         1.795      0.196   55.5%
(yes/unclear/no)

Other (yes/unclear/no)                   1.686      0.185   60.3%

Jadad's scale (continuous variable)      0.697      0.515   59.3%

Year of publication                      1.02       0.791   60.1%
(continuous variable)

Financial support                        1.38       0.551   51.9%
(disclosed/not disclosed)

CRRT=continuous renal replacement therapy, APACHE=Acute Physiology
and Chronic Health Evaluation, RCT=randomised controlled trial.

FIGURE 2: Forest plot (all studies). RR=relative risk,
CI=confidence interval.

                                       Events,     Events,
Study       Year   RR (95% CI)         treatment   control   % weight

Traditional dose   0.79 (0.56, 1.12)   11/15       12/13     10.40
(<35 ml/kg/h)
1997

Sander
Cole        2002   1.00 (0.32, 3.10)   4/12        4/12      1.67
Payen       2009   1.29 (0.82, 2.03)   22/39       17/39     7.51

Sub-total          0.99 (0.66, 1.48)   37/66       33/64     19.57
([I.sup.2]=44.2%,
P=0.167)

High dose
(35-65 ml/kg/h)

Ronco       2000   092 (0.65, 1.29)    22/32       15/20     10.36
Bauman      2002   2.00 (0.78, 5.14)   8/14        4/14      2.32
Morgera     2004   2.00 (0.82, 4.89)   8/12        4/12      2.56
Saudan      2006   0.31 (0.15, 0.63)   7/37        21/34     3.74
Tolwani     2008   1.09 (0.83, 1.43)   37/54       34/54     12.90
ATN         2008   1.08 (0.95, 1.24)   204/358     184/350   18.83
RENAL       2009   0.91 (0.79, 1.06)   168/359     186/363   18.20

Sub-total          0.98 (0.80, 1.21)   454/866     448/847   68.90
([I.sup.2]=67.3%,
P=0.005)

Very high dose
(>65 ml/kg/h)

Ghani       2006   0.88 (0.61, 1.27)   11/15       15/18     9.62
Boussekey   2008   0.56 (0.19, 1.59)   3/9         6/10      1.91

Sub-total          0.84 (0.59, 1.19)   14/24       21/28     11.52
([I.sup.2]=0.0%,
P=0.376)

Sub-total          0.96 (0.83, 1.12)   505/956     502/939   100.00
([I.sup.2]=51.7%,
P=0.019)
Gale Copyright: Copyright 2011 Gale, Cengage Learning. All rights reserved.