Non‐pharmacological interventions for preventing clotting of extracorporeal circuits during continuous renal replacement therapy

Abstract

Background

Acute kidney injury (AKI) is a common complication amongst people who are critically ill, and it is associated with an increased risk of death. For people with severe AKI, continuous kidney replacement therapy (CKRT), which is delivered over 24 hours, is needed when they become haemodynamically unstable. When CKRT is interrupted due to clotting of the extracorporeal circuit, the delivered dose is decreased and thus leading to undertreatment.

Objectives

This review assessed the efficacy of non‐pharmacological measures to maintain circuit patency in CKRT.

Search methods

We searched the Cochrane Kidney and Transplant Register of Studies up to 25 January 2021 which includes records identified through searches of CENTRAL, MEDLINE, and EMBASE, conference proceedings, the International Clinical Trials Register (ICTRP) Search Portal, and ClinicalTrials.gov.

Selection criteria

We included all randomised controlled trials (RCTs) (parallel‐group and cross‐over studies), cluster RCTs and quasi‐RCTs that examined non‐pharmacological interventions to prevent clotting of extracorporeal circuits during CKRT. 

Data collection and analysis

Three pairs of review authors independently extracted information including participants, interventions/comparators, outcomes, study methods, and risk of bias. The primary outcomes were circuit lifespan and death due to any cause at day 28. We used a random‐effects model to perform quantitative synthesis (meta‐analysis). We assessed the risk of bias in included studies using the Cochrane Collaboration’s tool for assessing the risk of bias. Summary estimates of effect were obtained using a random‐effects model, and results were expressed as risk ratios (RR) and their 95% confidence intervals (CI) for dichotomous outcomes, and mean difference (MD) and 95% CI for continuous outcomes. Confidence in the evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.

Main results

A total of 20 studies involving 1143 randomised participants were included in the review. The methodological quality of the included studies was low, mainly due to the unclear randomisation process and blinding of the intervention. We found evidence on the following 11 comparisons: (i) continuous venovenous haemodialysis (CVVHD) versus continuous venovenous haemofiltration (CVVH) or continuous venovenous haemodiafiltration (CVVHDF); (ii) CVVHDF versus CVVH; (iii) higher blood flow (≥ 250 mL/minute) versus standard blood flow (< 250 mL/minute); (iv) AN69 membrane (AN69ST) versus other membranes; (v) pre‐dilution versus post‐dilution; (vi) a longer catheter (> 20 cm) placing the tip targeting the right atrium versus a shorter catheter (≤ 20 cm) placing the tip in the superior vena cava; (vii) surface‐modified double‐lumen catheter versus standard double‐lumen catheter with identical geometry and flow design; (viii) single‐site infusion anticoagulation versus double‐site infusion anticoagulation; (ix) flat plate filter versus hollow fibre filter of the same membrane type; (x) a filter with a larger membrane surface area versus a smaller one; and (xi) a filter with more and shorter hollow fibre versus a standard filter of the same membrane type.

Circuit lifespan was reported in 9 comparisons. Low certainty evidence indicated that CVVHDF (versus CVVH: MD 10.15 hours, 95% CI 5.15 to 15.15; 1 study, 62 circuits), pre‐dilution haemofiltration (versus post‐dilution haemofiltration: MD 9.34 hours, 95% CI ‐2.60 to 21.29; 2 studies, 47 circuits; I² = 13%), placing the tip of a longer catheter targeting the right atrium (versus placing a shorter catheter targeting the tip in the superior vena cava: MD 6.50 hours, 95% CI 1.48 to 11.52; 1 study, 420 circuits), and surface‐modified double‐lumen catheter (versus standard double‐lumen catheter: MD 16.00 hours, 95% CI 13.49 to 18.51; 1 study, 262 circuits) may prolong circuit lifespan. However, higher blood flow may not increase circuit lifespan (versus standard blood flow: MD 0.64, 95% CI ‐3.37 to 4.64; 2 studies, 499 circuits; I² = 70%). More and shorter hollow fibre filters (versus standard filters: MD ‐5.87 hours, 95% CI ‐10.18 to ‐1.56; 1 study, 6 circuits) may reduce circuit lifespan.

Death from any cause was reported in four comparisons We are uncertain whether CVVHDF versus CVVH, CVVHD versus CVVH or CVVHDF, longer versus a shorter catheter, or surface‐modified double‐lumen catheters versus standard double‐lumen catheters reduced death due to any cause, in very low certainty evidence.

Recovery of kidney function was reported in three comparisons. We are uncertain whether CVVHD versus CVVH or CVVHDF, CVVHDF versus CVVH, or surface‐modified double‐lumen catheters versus standard double‐lumen catheters increased recovery of kidney function.

Vascular access complications were reported in two comparisons. Low certainty evidence indicated using a longer catheter (versus a shorter catheter: RR 0.40, 95% CI 0.22 to 0.74) may reduce vascular access complications, however, the use of surface‐modified double lumen catheters versus standard double‐lumen catheters may make little or no difference to vascular access complications.

Authors' conclusions

The use of CVVHDF as compared with CVVH, pre‐dilution haemofiltration, a longer catheter, and surface‐modified double‐lumen catheter may prolong the circuit lifespan, while higher blood flow and more and shorter hollow fibre filter may reduce circuit life. The overall certainty of the evidence was assessed to be low to very low due to the small sample size of the included studies.

Data from future rigorous and transparent research are much needed in order to fully understand the effects of non‐pharmacological interventions in preventing circuit coagulation amongst people with AKI receiving CKRT.

Plain language summary

Non‐pharmacological interventions for preventing clotting of extracorporeal circuits during continuous kidney replacement therapy

What is the issue?

Acute kidney injury (AKI) is a major problem in people with severe illness. In cases of severe AKI, kidney replacement therapy/dialysis (KRT) using circuits is necessary. Continuous kidney replacement therapy is performed continuously over 24 hours. Clotting of the CKRT circuit can interfere with this treatment. To prevent this, a variety of non‐pharmacological (not using medication) interventions have been studied. We aimed to summarise current evidence regarding the efficacy of non‐pharmacological interventions for preventing clotting of extracorporeal circuits during CKRT.

What did we do?

We searched for available evidence from the Cochrane Kidney and Transplant Specialised Register up to 25 January 2021. Our review summarised the results of 20 randomised studies involving a total of 1143 people.

What did we find?

We found that the quality of the 20 included studies was low, and the number randomised was small. The majority of the included studies did not report death as an outcome. We found that continuous venovenous haemodiafiltration (CVVHDF), as compared with continuous venovenous haemofiltration (CVVH), may prolong circuit lifespan. In addition, pre‐dilution haemofiltration, as compared with post‐dilution haemofiltration, a longer catheter placing the tip at the right atrium, as compared with a shorter catheter placing the tip in the superior vena cava, and surface‐modified double lumen catheter, as compared with standard double lumen catheter, may extend the circuit lifespan. However, higher blood flow compared to standard blood flow rate might not affect circuit lifespan. Overall, the data was limited and of very low certainty.

Conclusions

We found that the effects of non‐pharmacological interventions in people with AKI receiving CKRT remain unclear. There is a need for studies assessing CKRT circuit lifespan as well as other clinically important outcomes.

Summary of findings

Summary of findings 1. CVVHD versus CVVH or CVVHDF for preventing clotting of extracorporeal circuits during CKRT.

CVVHD compared to CVVH or CVVHDF for preventing clotting of extracorporeal circuits during CKRT
Patient or population: AKI patients receiving CKRT Setting: ICU Intervention: CVVHD Comparison: CVVH or CVVHDF
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)
	Risk with CVVH or CVVHDF	Risk with CVVHD
Circuit lifespan	Not reported	Not reported	‐‐	‐‐	‐‐
Death from any cause Follow up: 24 hours to 90 days	536 per 1,000	413 per 1,000 (225 to 751)	RR 0.77 (0.42 to 1.40)	229 (4)	⊕⊝⊝⊝ Very low 1,2,3,4
Recovery of kidney function Follow up: ICU discharge	182 per 1,000	165 per 1,000 (67 to 402)	RR 0.91 (0.37 to 2.21)	93 (2)	⊕⊝⊝⊝ Very low 1,4
Vascular access complications	Not reported	Not reported	‐‐	‐‐	‐‐
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CVVHD: continuous venovenous haemodialysis; CVVH: continuous venovenous haemofiltration; CVVHDF: continuous venovenous haemodiafiltration; CKRT: continuous kidney replacement therapy;AKI: acute kidney injury; ICU: intensive care unit; CI: confidence interval;RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Downgraded one level due to serious risk of bias

² Downgraded one level due to serious inconsistency

³ Downgraded one level due to serious indirectness (death at the longest follow‐up)

⁴ Downgraded two levels due to very serious imprecision

Summary of findings 2. CVVHDF versus CVVH for preventing clotting of extracorporeal circuits during CKRT.

CVVHDF versus CVVH for preventing clotting of extracorporeal circuits during CKRT
Patient or population: AKI patients receiving CKRT Setting: ICU Intervention: CVVHDF Comparison: CVVH
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)
	Risk with CVVH	Risk with CVVHDF
Circuit lifespana	The mean circuit lifespan with CVVH was 8.55 hours	The mean circuit lifespan with CVVHDF was 10.15 hours higher (5.15 to 15.15 higher) than CVVH	‐	62 (1)	⊕⊕⊝⊝ Low 1,2
Death from any cause at day 28 of follow‐up	608 per 1,000	413 per 1,000 (316 to 547)	RR 0.68 (0.52 to 0.90)	206 (1 RCT)	⊕⊝⊝⊝ Very low 1,2,3
Recovery of kidney function	714 per 1,000	786 per 1,000 (614 to 1,000)	RR 1.10 (0.86 to 1.41)	96 (1 RCT)	⊕⊝⊝⊝ Very low 2,4
Vascular access complications	Not reported	Not reported	‐‐	‐‐	‐‐
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). a we calculated circuit lifespan per circuit CVVHD: continuous venovenous haemodialysis; CVVH: continuous venovenous haemofiltration; CKRT: continuous kidney replacement therapy; AKI: acute kidney injury; ICU: intensive care unit; CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Downgraded one level due to serious risk of bias

² Downgraded one level due to serious imprecision

³ Downgraded one level due to serious concern of publication bias

⁴ Downgraded two levels due to very serious risk of bias (participants were excluded from analyses if died)

Summary of findings 3. Pre‐dilution versus post‐dilution (as defined by study investigators) for preventing clotting of extracorporeal circuits during CKRT.

Pre‐dilution versus post‐dilution (as defined by study investigators) for preventing clotting of extracorporeal circuits during CKRT
Patient or population: AKI patients receiving CKRT Setting: ICU Intervention: pre‐dilution Comparison: post‐dilution
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with post‐dilution	Risk with pre‐dilution
Circuit lifespana	The mean circuit lifespan with post‐dilution was 22.5 hours	The mean circuit lifespan with pre‐dilution was 9.34 hours higher (2.60 lower to 21.29 higher) than post‐dilution	‐‐	47 (2)	⊕⊕⊝⊝ Low 1,2
Death from any cause at day 28 of follow‐up	Not reported	Not reported	‐‐	‐‐	‐‐
Recovery of kidney function	Not reported	Not reported	‐‐	‐‐	‐‐
Vascular access complications	Not reported	Not reported	‐‐	‐‐	‐‐
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). a we calculated circuit lifespan per circuit CKRT: continuous kidney replacement therapy; AKI: acute kidney injury; ICU: intensive care unit; CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Downgraded one level due to serious risk of bias

² Downgraded one level due to serious imprecision

Summary of findings 4. Higher blood flow versus standard blood flow for preventing clotting of extracorporeal circuits during CKRT.

Higher versus standard blood flow for preventing clotting of extracorporeal circuits during CKRT
Patient or population: AKI patients receiving CKRT Setting: ICU Intervention: higher blood flow Comparison: standard blood flow
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)
	Risk with standard blood flow	Risk with higher blood flow
Circuit lifespana	The mean circuit lifespan with standard blood flow was 17.2 hours	The mean circuit lifespan with higher blood flow was 0.64 hours higher (3.37 lower to 4.64 higher) than standard blood flow	‐‐	499 (2)	⊕⊕⊝⊝ Low 1,2
Death from any cause at day 28 of follow‐up	Not reported	Not reported	‐‐	‐‐	‐‐
Recovery of kidney function	Not reported	Not reported	‐‐	‐‐	‐‐
Vascular access complications	Not reported	Not reported	‐‐	‐‐	‐‐
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). a we calculated circuit lifespan per circuit CKRT: continuous kidney replacement therapy; AKI: acute kidney injury; ICU: intensive care unit; CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Downgraded one level due to serious risk of bias

² Downgraded one level due to serious inconsistency

Summary of findings 5. Long (> 20 cm) versus short (≤ 20 cm) catheter for preventing clotting of extracorporeal circuits during CKRT.

Long (> 20 cm) versus short (≤ 20 cm) catheter for preventing clotting of extracorporeal circuits during CKRT
Patient or population: AKI patients receiving CKRT Setting: ICU Intervention: long catheter (> 20 cm) Comparison: short catheter (≤ 20 cm)
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)
	Risk with short catheter (≤ 20cm)	Risk with long catheter (> 20 cm)
Circuit lifespana	The mean circuit lifespan with short catheters was 17.5 hours	The mean circuit lifespan with longer catheters was 6.5 hours higher (1.48 to 11.52 higher) than short catheters	‐‐	402 (1)	⊕⊕⊝⊝ Low 1,4
Death from any cause at day 28 of follow‐up	170 per 1,000	235 per 1,000 (104 to 529)	RR 1.38 (0.61 to 3.11)	94 (1)	⊕⊝⊝⊝ Very low 2,3
Recovery of kidney function	Not reported	Not reported	‐‐	‐‐	‐‐
Vascular access complications	511 per 1,000	194 per 1,000 (102 to 368)	RR 0.38 (0.20 to 0.72)	94 (1)	⊕⊕⊝⊝ Low 1,4
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). a we calculated circuit lifespan per circuit CKRT: continuous kidney replacement therapy; AKI: acute kidney injury; ICU: intensive care unit; CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Downgraded one level due to serious risk of bias

² Downgraded one level due to serious indirectness (used ICU death as a surrogate endpoint)

³ Downgraded two levels due to very serious imprecision

⁴ Downgraded one level due to serious imprecision

Summary of findings 6. Surface‐modified versus standard double‐lumen catheter with identical geometry and flow design for preventing clotting of extracorporeal circuits during CKRT.

Surface‐modified versus standard double‐lumen catheter with identical geometry and flow design for preventing clotting of extracorporeal circuits during CKRT
Patient or population: AKI patients receiving CKRT Setting: ICU Intervention: surface‐modified double‐lumen catheter Comparison: standard double‐lumen catheter with identical geometry and flow design
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)
	Risk with standard double‐lumen catheter with identical geometry and flow design	Risk with surface‐modified, double‐lumen catheter
Circuit lifespan	The mean circuit lifespan with standard double‐lumen catheters was 118 hours	The mean circuit lifespan with surface‐modified double‐lumen catheters was 16 hours higher (13.49 to 18.51 higher) than standard double‐lumen catheters	‐‐	262 (1)	⊕⊕⊝⊝ Low 1,2
Death from any cause at day 28 of follow‐up	331 per 1,000	357 per 1,000 (251 to 506)	RR 1.08 (0.76 to 1.53)	236 (1)	⊕⊕⊝⊝ Very low 1,2,3
Recovery of kidney function	492 per 1,000	447 per 1,000 (344 to 590)	RR 0.91 (0.70 to 1.20)	236 (1)	⊕⊕⊝⊝ Low 1,2
Vascular access complications	76 per 1,000	25 per 1,000 (7 to 92)	RR 0.33 (0.09 to 1.20)	236 (1)	⊕⊕⊝⊝ Low 1,2
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). a we calculated circuit lifespan per circuit CKRT: continuous kidney replacement therapy; AKI: acute kidney injury; ICU: intensive care unit; CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Downgraded one level due to serious risk of bias

² Downgraded one level due to serious imprecision

³ Downgraded one level due to serious indirectness (used ICU mortality as a surrogate)

Summary of findings 7. Polyethylenimine‐treated membrane (AN69ST) versus other membranes for preventing clotting of extracorporeal circuits during CKRT.

Polyethylenimine‐treated AN69ST versus other membranes for preventing clotting of extracorporeal circuits during CKRT
Patient or population: AKI patients receiving CKRT Setting: ICU Intervention: polyethylenimine‐treated AN69ST Comparison: other membranes
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)
	Risk with other membrane	Risk with polyethylenimine‐treated AN69ST
Circuit lifespan	The mean circuit lifespan with other membranes was 14.1 hours	The mean circuit lifespan with polyethylenimine‐treated AN69ST was 1.54 hours higher (4.53 lower to 7.6 higher) than other membranes	‐‐	56 (2)	⊕⊕⊝⊝ Low 1
Death from any cause at day 28 of follow‐up	Not reported	Not reported	‐‐	‐‐	‐‐
Recovery of kidney function	Not reported	Not reported	‐‐	‐‐	‐‐
Vascular access complications	Not reported	Not reported	‐‐	‐‐	‐‐
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). a we calculated circuit lifespan per circuit CKRT: continuous kidney replacement therapy; AKI: acute kidney injury; ICU: intensive care unit; CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Downgraded two levels due to very serious imprecision

Summary of findings 8. Flat plate versus hollow fibre filter of the same membrane type for preventing clotting of extracorporeal circuits during CKRT.

Flat plate versus hollow fibre filter of the same membrane type for preventing clotting of extracorporeal circuits during CKRT
Patient or population: people with AKI receiving CKRT Setting: ICU Intervention: flat plate filter Comparison: hollow fibre filter (of the same membrane type)
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)
	Risk with hollow fibre filter of the same membrane type	Risk with flat plate filter
Circuit lifespan	The mean circuit lifespan with hollow fibre filters was 17.1 hours	The mean circuit lifespan with flat plate filters was 1.4 hours lower (12.12 lower to 9.32 higher) than hollow fibre filters	‐‐	38 (1)	⊕⊝⊝⊝ Very low 1,2
Death from any cause at day 28 of follow‐up	Not reported	Not reported	‐‐	‐‐	‐‐
Recovery of kidney function	Not reported	Not reported	‐‐	‐‐	‐‐
Vascular access complications	Not reported	Not reported	‐‐	‐‐	‐‐
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CKRT: continuous kidney replacement therapy; AKI: acute kidney injury; ICU: intensive care unit; CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Downgraded one level due to serious risk of bias

² Downgraded two levels due to very serious imprecision

Summary of findings 9. More and shorter versus standard hollow fibre filter of the same membrane type for preventing clotting of extracorporeal circuits during CKRT.

More and shorter versus standard hollow fibre filter of the same membrane type for preventing clotting of extracorporeal circuits during CKRT
Patient or population: AKI patients receiving CKRT Setting: ICU Intervention: more and shorter hollow fibre filter Comparison: standard hollow fibre filter
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)
	Risk with standard hollow fibre filter of the same membrane type	Risk with more and shorter fibre filters
Circuit lifespan	The mean circuit lifespan with standard hollow fibre filters was 21.27 hours	The mean circuit lifespan with more and shorter fibre filters was 5.87 hours lower (10.18 to 1.56 lower) than standard hollow fibre filters	‐‐	6 (1)	⊕⊕⊝⊝ LOW 1 2
Death from any cause at day 28 of follow‐up	Not reported	Not reported	‐‐	‐‐	‐‐
Recovery of kidney function	Not reported	Not reported	‐‐	‐‐	‐‐
Vascular access complications	Not reported	Not reported	‐‐	‐‐	‐‐
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CKRT: continuous kidney replacement therapy; AKI: acute kidney injury; ICU: intensive care unit; CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Downgraded one level due to serious risk of bias

² Downgraded one level due to serious imprecision

Background

Description of the condition

Acute kidney injury (AKI) is defined as an abrupt decline in kidney function with increased serum creatinine or decreased urine output (KDIGO Acute Kidney Injury Work Group 2012). AKI occurs in nearly half of the admitted cases in intensive care units (ICUs) and is related to increased death (Bouchard 2015; Fujii 2018; Hoste 2015; Nisula 2013). For people with severe AKI, kidney replacement therapy (KRT) is required to manage electrolytes, fluid balance, and waste products. Previous epidemiological studies revealed that the need for KRT accounted for 16% to 24% of people with AKI in the ICUs (Fujii 2018; Nisula 2013).

Administration of KRT can be achieved either intermittently or continuously. Although intermittent KRT removes fluids more rapidly, it has been suggested to cause hypotension and further damage to kidney function (Manns 1997; Silversides 2014). Moreover, deranged electrolyte levels, which are frequently observed in people with AKI, are normalised more successfully when KRT is administered continuously (Uchino 2003). Accordingly, continuous kidney replacement therapy (CKRT) is recommended for people with AKI who are haemodynamically unstable, and it is a technique used most frequently in the ICU setting (Bouchard 2015; Fujii 2018; KDIGO Acute Kidney Injury Work Group 2012). A multicentre survey conducted in 33 countries showed that 75.2% of KRT sessions were conducted using the CKRT approach (Hoste 2015).

The choice of CKRT modalities varies across the world. An international observational study in 2007 reported that continuous venovenous haemofiltration (CVVH) had been widely used (Uchino 2007). However, another small observational study showed comparable small and mid‐sized molecular solute removal between CVVH and continuous venovenous haemodialysis (CVVHD) (Ricci 2006). Evidence supporting a clear choice of one CKRT modality over others remains inconclusive although the use of various diffusion‐based techniques, such as CVVHD and continuous venovenous haemodiafiltration (CVVHDF), has been reported in a recent observational study (Fujii 2018).

In clinical practice, CKRT is typically provided through a double‐lumen venous catheter as a continuous 24‐hour therapy; however, CKRT may also operate for 21‐23 hours a day in different clinical settings (Mehta 2001; Uchino 2003; Vesconi 2009). Reasons for the interruption of CKRT could be clotting of the extracorporeal circuits, clogging of the membrane, or the need for transportation to outside of the ICU settings, such as to the operation theatre or the radiology department for imaging tests. Such treatment interruptions decrease the delivered CKRT dose, leading to insufficient uraemic control (Fealy 2002; Mitchell 2003). A small single‐centre study indicated that circuit clotting was the primary reason for the shortened circuit life (Venkataraman 2002). Undoubtedly, exchanging the circuits due to circuit failure leads to increased medical costs and the workload of healthcare professionals (Fealy 2002; Mehta 2001). Thus, treatment interruption is widely considered a key quality indicator for CKRT (Rewa 2017).

Description of the intervention

The primary intervention to maintain the patency of the CKRT circuit is anticoagulation. A previous multicentre trial showed that as many as half of the critically‐ill study subjects were treated with CKRT without the use of any anticoagulation (RENAL 2006). Indeed, anticoagulation therapy is associated with a substantial risk of bleeding, which in itself is a common feature amongst people who are critically ill. Research efforts have recently shifted to identify effective and safe anticoagulants, e.g. citrate, as well as strategies to minimise the risk of clotting without using anticoagulants during CKRT. Since there is an existing Cochrane review exploring the effects of pharmacological interventions in CKRT (Tsujimoto 2020), our present review aimed to focus on non‐pharmacological strategies for circuit survival.

Clotting of CKRT circuits is attributed to stasis or turbulence of blood flow, haemoconcentration, or activation of the intrinsic coagulation system by blood–tube, blood–air, or blood–filter contact (RENAL 2006). Therefore, non‐pharmacological interventions to prevent clotting of the CKRT circuit generally include the strategic selection of a catheter or access site, optimising the blood flow rate, CKRT modalities, and methods of haemodilution.

How the intervention might work

The selection of a catheter and access site may play a significant role in determining circuit life. According to Poiseuille's Law, for fluid/blood flow through an intravenous (IV) catheter a thick and short catheter may be theoretically preferable to avoid stasis. In addition, to avoid kinking or curving of the catheter, which may lead to impaired blood flow, the right jugular venous route may be preferable as it is straight and easily monitored by bedside clinical staff.

In CKRT, blood flow rates are typically set at 100 to 200 mL/minute or based on a filtration ratio. In fact, the pump used in CKRT delivers blood with peristaltic revolutions, and the flow rate is, in effect, the rate of the pump revolution. An observational study revealed a forward and backward blood flow path between the pump and filter, which may cause blood flow stasis (Baldwin 2004). Considering the potential blood flow fluctuations due to the peristaltic roller pump, maintaining the blood flow at a high rate may be useful to prevent CKRT filter clotting.

Regarding the choice of CKRT modalities, filtration may shorten the circuit lifetime compared to dialysis by haemoconcentration due to its ultrafiltration process (Ricci 2006). During filtration, haematocrit levels increase and consequently elevate the risk of coagulation in the filter. In CVVH and CVVHDF, substitution fluids can be administered before (pre‐dilution) or after (post‐dilution) filtration. Pre‐dilution CKRT aims to decrease haemoconcentration and improve blood flow. This may improve filter lifespan and efficiency, leading to less treatment interruptions and better clearance. However, a critical aspect of pre‐dilution therapies is the relationship between the blood flow rate and replacement fluid rate (Clark 2017). Pre‐dilution may also reduce molecular clearance when the blood flow is low, and its precise clinical impact upon people who are critically ill is currently unclear.

Why it is important to do this review

Circuit failure during CKRT affects the actual delivery of CKRT, which leads to decreased treatment efficiency. Moreover, it increases the associated medical costs and workload of clinical staff. Thus, maintaining the patency of the CKRT circuit is particularly crucial in ICU settings. This review aimed to provide a clinically relevant and solid body of evidence to assist healthcare decision‐making when delivering CKRT in critical care medicine.

Objectives

This review assessed the efficacy of non‐pharmacological measures to maintain circuit patency during CKRT.

Methods

Criteria for considering studies for this review

Types of studies

We included all parallel‐group and cross‐over randomised controlled trials (RCTs), cluster‐randomised trials and quasi‐randomised trials (studies in which allocation to treatment was obtained by alternation, use of alternate medical records, date of birth or other predictable methods) investigating non‐pharmacological interventions for preventing clotting of extracorporeal circuits during CKRT. We excluded observational studies, case reports, reviews, editorials, and commentaries.

Types of participants

We included people with AKI who received CKRT in the ICU settings regardless of age or sex. In this review, we defined AKI according to the Kidney Disease: Improving Global Outcomes (KDIGO) definition and staging system (KDIGO Acute Kidney Injury Work Group 2012).

Types of interventions

Dialysis modalities

• CVVHD versus CVVH or CVVHDF

• CVVHDF versus CVVH

• Pre‐dilution versus post‐dilution (as defined by study investigators)

Blood flow rate

• Higher (≥ 250 mL/minute) blood flow versus standard (< 250 mL/minute) blood flow

Catheter types

• Long (> 20 cm) catheter versus short (≤ 20 cm) catheter

• Left‐sided catheters versus right‐sided catheter

• Surface‐modified double‐lumen catheter versus standard double‐lumen catheter with identical geometry and flow design

Membrane types

• Polyethylenimine‐treated polyacrylonitrile membrane (AN69ST) versus other types of membrane

• Heparin‐grafted membrane versus other membranes

• A flat plate filter versus a hollow fibre filter of the same membrane type

• A filter with a larger membrane surface area versus a smaller one

• A filter with more and shorter hollow fibres versus a standard filter of the same membrane type

Access sites

• Internal jugular access versus femoral access

• Subclavian venous versus superior vena cava as the line tip portion in the same access catheter insertion site

• Subclavian venous versus right atrium as the line tip portion in the same access catheter insertion site

• Superior vena cava versus right atrium as the line tip portion in the same access catheter insertion site

Other types of non‐pharmacological interventions

• Single‐site infusion anticoagulation versus double‐site infusion anticoagulation

• Infusion anticoagulation from access line versus pre‐filter single site

• Saline flushing versus no saline flushing.

We intended to compare the effects of one non‐pharmacological intervention to another. We included studies that used or did not use anticoagulants in both intervention and control groups. We excluded studies comparing non‐pharmacological interventions to pharmacological interventions.

Types of outcome measures

Primary outcomes

• Circuit lifespan (commences at the point of starting CKRT and concludes upon circuit cessation for any reason)

• Death (any cause) at day 28 of follow‐up or last reported time point

Secondary outcomes

• Recovery of kidney function: numbers of participants assessed as free of KRT at day 28, 90, and 180 of follow‐up

• Vascular access complications: functions (e.g. ability to use CKRT, uninterrupted use without the need for any intervention, percentage change in access blood flow) or access‐site infections (e.g. those suspected to be catheter‐related) requiring antibiotic therapy

• Costs related to health care services (unit: USD, calculated rate 1 USD = 0.93 Euro)

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid

Search methods for identification of studies

Electronic searches

We searched the Cochrane Kidney and Transplant Register of Studies up to 25 January 2021 through contact with the Cochrane Kidney and Transplant Information Specialist using search terms relevant to this review. The Register contains studies identified from the following sources.

1. Monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL)

2. Weekly searches of MEDLINE OVID SP

3. Searches of kidney and transplant journals, and the proceedings and abstracts from major kidney and transplant conferences

4. Searching of the current year of EMBASE OVID SP

5. Weekly current awareness alerts for selected kidney and transplant journals

6. Searches of the WHO International Clinical Trials Register (ICTRP) Search Portal and ClinicalTrials.gov.

Studies contained in the Register are identified through searches of CENTRAL, MEDLINE, and EMBASE based on the scope of Cochrane Kidney and Transplant. Details of search strategies, as well as a list of handsearched journals, conference proceedings and current awareness alerts, are available on the Cochrane Kidney and Transplant website under CKT Register of Studies.

See Appendix 1 for search terms used in strategies for this review.

Searching other resources

We screened reference lists of narrative review articles, relevant studies, and clinical practice guidelines for additional relevant information. Where appropriate, we contacted experts and professional organisations in the field for information on unpublished or incomplete/ongoing studies.

Due to resource constraints, we did not consider grey literature sources (e.g. conference abstracts, dissertations, and theses) additional to those already included in the Cochrane Kidney and Transplant Register of Studies.

Data collection and analysis

Selection of studies

Two authors (YT and SM) independently screened the titles and abstracts obtained from the literature search in duplicate, and clearly irrelevant records were excluded. Studies that might include relevant data or information were retained initially for subsequent assessment where the two authors independently assessed the full texts of these potentially eligible studies against our prespecified inclusion/exclusion criteria. We arranged for non‐English and non‐Japanese study reports to be translated before full assessment. Disagreements were resolved by discussion with a third author acting as an arbiter (TF).

Data extraction and management

Data extraction was carried out independently by three pairs of review authors (YT and SM, YK and HS, HY and HT) in duplicate, using pre‐standardised data extraction forms. We resolved disagreements by discussion with another author acting as an arbiter (TF). Where more than one record from the same study existed, records were grouped together and the record with the most complete data was identified as the primary reference of the study. We specifically highlighted discrepancies between available study protocols and final published reports where appropriate.

Assessment of risk of bias in included studies

Three pairs of review authors (YT and SM, YK and HS, HY and HT) assessed the risk of bias in included studies using the Cochrane Collaboration’s tool for assessing the risk of bias (Higgins 2011) (see Appendix 2). Any disagreement was resolved by discussion with another author acting as an arbiter (TF).

The Cochrane Collaboration’s tool for assessing the risk of bias considered six aspects of bias: (i) selection bias; (ii) performance bias; (iii) detection bias; (iv) attrition bias; (v) reporting bias; and (vi) other bias.

• Was there adequate sequence generation (selection bias)?

• Was allocation adequately concealed (selection bias)?

• Was knowledge of the allocated interventions adequately prevented during the study?

  • Blinding of participants and personnel (performance bias)

  • Blinding of outcome assessors (detection bias)

• Were incomplete outcome data adequately addressed (attrition bias)?

• Was the study report free of potential selective outcome reporting (reporting bias)?

For assessing other potential threats to validity ('Other bias'), we assumed that the use of anticoagulation therapy or sedative agents were important co‐interventions in this review. Additionally, we judged how baseline imbalance of anticoagulation affected treatment effect estimates.

• Was the study apparently free of other problems that could put it at risk of bias (e.g. co‐intervention of anticoagulation)?

• Did baseline imbalance regarding anticoagulation therapy exist?

Measures of treatment effect

For dichotomous outcomes (e.g. death or recovery of kidney function), treatment effect measures were expressed in risk ratios (RRs) with 95% confidence intervals (CIs). Where continuous scales of measurement were used to assess the effects of treatment interventions (e.g. circuit lifespan, or costs), we calculated mean differences (MDs) with their respective 95% CIs.

Unit of analysis issues

We assessed unit of analysis issues in three possible ways in which they might arise.

1. Multiple measurements of the same participants from either individually randomised or cross‐over studies

If there were multiple measurements of the outcomes such as circuit lifespan (e.g. several circuits were used by the same individual due to repeated CKRT using new circuits), unit of analysis issues might arise. We considered addressing these issues by first assessing each included study for any evidence of multiple enrolments (e.g. number of circuits exceeded the number of participants). We then planned to exclude those with multiple measurements by entering the data of only the first circuit, but such information was not available. We, therefore, performed our analysis based on available data as reported, using the total number of circuits as the denominator according to methods employed in a previous study (Tsujimoto 2020).

2. Clustering at the level of the enrolled units in cluster‐randomised studies

In dealing with cluster‐RCTs, for dichotomous data, we would have applied the design effect formula and calculated effective sample size and number of events using the intra‐cluster correlation coefficient (ICC) among each unit and the average cluster size, as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). If the ICC had not been reported, we would have used the ICC of a similar study as a substitute. For continuous data, only the sample size would be reduced; means and standard deviations would remain unchanged (Higgins 2011).

3. Randomised cross‐over studies

We considered only data from the first period (before crossing over) where possible. However, for cross‐over studies that did not report first‐phase data, we decided to include the study and analysed the data as if the trials were parallel trials. We rated its validity according to the Cochrane Collaboration’s tool for assessing the risk of bias as being at 'unclear risk of bias' based on baseline imbalance due to the cross‐over design (Assessment of risk of bias in included studies).

Multiple comparisons

All intervention groups that were assessed to be relevant to this review as per our a priori eligibility criteria were included (Types of interventions).

Dealing with missing data

Where possible, we requested further information from the study corresponding authors via electronic or postal mail and any such requested, relevant information obtained were included in the review. We diligently evaluated important numerical data, such as the numbers of screened/randomised participants as well as intention‐to‐treat, as‐treated and per‐protocol populations were reported. Attrition rates, including drop‐outs, losses to follow‐up and withdrawals, were also investigated. Issues of missing data and imputation methods (for example, use of ‘last observation carried forward’ (LOCF)) employed by study investigators were identified and critically appraised (Higgins 2011).

If studies did not report mean or standard deviation for a continuous outcome, we imputed them from the median and range (Hozo 2005) or from the median and interquartile range (Wan 2014). When there was no information on variability, we imputed the standard deviation from other studies in the same meta‐analysis (Furukawa 2006).

Assessment of heterogeneity

We assessed the presence of statistical heterogeneity by visual inspection of the forest plots. We quantified statistical heterogeneity using the I² statistic, which describes the percentage of total variation across studies that is due to heterogeneity rather than sampling error (Higgins 2003). We followed the guide to the interpretation of I² values as indicated in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011)/

• 0% to 40%: might not be important

• 30% to 60%: may represent moderate heterogeneity

• 50% to 90%: may represent substantial heterogeneity

• 75% to 100%: considerable heterogeneity.

The importance of the observed value of I² depends on the magnitude and direction of treatment effects and the strength of evidence for heterogeneity (e.g. P‐value from the Chi² test, or confidence intervals for I²) (Higgins 2011).

Assessment of reporting biases

In this review, since our analyses did not include more than 10 studies, and we could not perform Egger's test or funnel plot symmetry inspection to explore possible small‐study and publication biases (Egger 1997). Instead, we investigated the scale of ongoing studies in the field based on the search results from the Cochrane Kidney and Transplant Register of Studies, which includes records from trial registries (WHO ICTRP and ClinicalTrials.gov).

Data synthesis

We conducted data synthesis using a random‐effects analytical model.

Subgroup analysis and investigation of heterogeneity

In the protocol, we planned to perform the following subgroup analyses where possible for the primary outcomes to explore possible sources of heterogeneity. However, due to insufficient data, we were unable to pursue the following analyses.

• Age of the participants (< 18 years versus ≥ 18 years)

• Presence or absence of sepsis (Levi 2013)

• Pharmacological co‐intervention (citrates, unfractionated heparins, other types of anticoagulation, no anticoagulation).

Sensitivity analysis

At the protocol stage, we decided to pursue the following sensitivity analyses for the primary outcomes. However, these were not performed for the full review due to limited data.

• Repeating the analysis by excluding unpublished studies

• Repeating the analysis by taking into account the risk of bias in included studies

• Repeating the analysis by excluding large‐scale studies or studies of long duration

• Repeating the analysis by excluding studies using the following filters: diagnostic criteria, language of publication, source of funding (industry versus other), and country

• Repeating the analysis by including only studies that used the same dosage of anticoagulants in both arms.

Summary of findings and assessment of the certainty of the evidence

We presented the main results of the review as 'Summary of findings' tables, created by the GRADEpro GDT software. These tables present key information concerning the quality/certainty of the evidence, the magnitude of the intervention effects, and the sum of the available data for the main outcomes (Schunemann 2011a). The 'Summary of findings' tables also include an overall grading of the evidence related to each of the main outcomes using the GRADE approach (Guyatt 2008; Guyatt 2011b), which defines the quality/certainty of a body of evidence as to the extent to which one can be confident that an estimate of effect or association is close to the true quantity of specific interest. To assess a body of evidence‐based on the GRADE approach, within‐trial risk of bias (methodological quality), directness of evidence, heterogeneity, the precision of effect estimates, and risk of publication bias are important domains for consideration (Schunemann 2011b).

Our 'Summary of findings' tables illustrated evidence relating to the following outcome measures.

• Circuit lifespan

• Death from any cause at day 28, or last reported time point

• Recovery of kidney function

• Vascular access complications.

Results

Description of studies

Results of the search

Our study selection process of this review is illustrated as a flow diagram in Figure 1. Our search of the Cochrane Kidney and Transplant Register of Studies identified 69 studies (179 records). After full‐text assessment, 26 studies (37 records) met the eligibility criteria and we excluded 43 studies (142 records). Of these 20 studies were included, five studies are ongoing (Weidhase 2020; NCT01062984; NCT01779635; NCT01790620; NCT03426943), and one study is awaiting classification (ISRCTN02674550). These six studies will be assessed in a future update of this review.

1.

Study flow diagram.

Included studies

Details of our included studies are summarised in the 'Characteristics of included studies' tables. These 20 studies randomised 1143 participants. Of these, four studies (57 participants) did not report any of our prespecified outcome measures (Alamartine 1994a; Broman 2019; Maxvold 2000; Wynckel 2004) and were thus excluded from subsequent quantitative synthesis (meta‐analysis). We, therefore, eventually included 16 studies (1072 participants) in our quantitative synthesis. Of these, one study was published only as a conference abstract (Plata‐Menchaca 2017).

Study design

Of the 20 included studies, nine were parallel‐group RCTs (Alamartine 1994a; Daud 2006; Fealy 2017; Meier 2011a; OMAKI 2012; Plata‐Menchaca 2017; RADICAL 2012; Ramesh Prasad 2000; Saudan 2006); and 11 were cross‐over RCTs (Baldwin 2002; Broman 2019; Davies 2008; de Pont 2006; Dungen 2001; Kellum 1998; Maxvold 2000; Schetz 2012; van der Voort 2005; Wynckel 2004; Yin 2015).

Sample size

Included studies were small in scale, with sample size ranging from 6 to 236 participants, with five studies involving 100 or more participants (Fealy 2017; Meier 2011a; Plata‐Menchaca 2017; RADICAL 2012; Saudan 2006).

Settings

The majority of the included studies were conducted in the USA, Europe, and Australia in the early 2000s. All the reported study settings were ICUs.

Participants

All but one study included adults (mean age 50.5 to 72.7 years) with kidney failure requiring CKRT; Maxvold 2000 included only paediatric participants. In nine studies, APACHE II was used for defining severity score, for which the mean score ranged from 18.5 to 27.  The causes of AKI differed between studies, with sepsis being the major cause in 11 studies (Broman 2019; Daud 2006; de Pont 2006; Dungen 2001; Meier 2011a; OMAKI 2012; Plata‐Menchaca 2017; RADICAL 2012; Schetz 2012; van der Voort 2005; Yin 2015); in two studies, the main cause of AKI was surgical procedures (Davies 2008; Fealy 2017).

Interventions and comparators

The comparisons and the corresponding included studies are as follows.

• CVVHD versus CVVH or CVVHD: 7 studies (254 participants) (Alamartine 1994a; Daud 2006; Kellum 1998; Maxvold 2000; OMAKI 2012; Plata‐Menchaca 2017; Wynckel 2004).

• CVVH versus CVVHDF: 2 studies (251 participants) (Davies 2008; Saudan 2006).

• Higher versus standard blood flow: 2 studies (134 participants) (Fealy 2017; Ramesh Prasad 2000).

• Pre‐dilution versus post‐dilution: 2 studies (48 participants) (de Pont 2006; van der Voort 2005).

• Long (> 20 cm) versus short catheter (≤ 20 cm): 1 study (100 participants) (RADICAL 2012).

• Surface‐modified double‐lumen catheter versus standard double‐lumen catheter of identical geometry and flow design: 1 study (236 participants) (Meier 2011a).

• Polyethylenimine‐treated AN69 membrane (AN69ST) versus other membranes: 3 studies (76 participants) (Broman 2019; Schetz 2012; Yin 2015).

• Filter with more and shorter (21 cm) hollow fibre versus standard (30 cm) hollow fibre: 1 study (6 participants) (Dungen 2001)

Baldwin 2002 (38 participants) included three separate randomised crossover trials and compared the following non‐pharmacological interventions.

• Single‐site versus double‐site infusion anticoagulation

• Flat plate versus hollow fibre filter of the same membrane type

• Filter with a larger membrane surface area versus a smaller one.

We could not identify any evidence regarding the other prespecified comparisons as listed in the section 'Types of interventions'.

Outcomes

The number of studies and participants that assessed our prespecified outcome measures are summarised as follows (Types of outcome measures).

• Circuit lifespan: 11 studies (663 participants) (Baldwin 2002; Davies 2008; de Pont 2006; Dungen 2001; Fealy 2017; Meier 2011a; RADICAL 2012; Ramesh Prasad 2000; Schetz 2012; van der Voort 2005; Yin 2015).

• Death from any cause

  • One study (206 participants) reported death (any cause) at 28 days (Saudan 2006)

  • Three studies (357 participants) reported ICU death (Daud 2006; Meier 2011a; RADICAL 2012)

  • Other death‐related outcome measures reported were death at 24 hours (Kellum 1998), death at 60 days (OMAKI 2012), and death at 90 days (Plata‐Menchaca 2017);

  • In three cross‐over trials, the numbers of deaths were reported; however, the numbers of deaths in each study arm were unavailable due to the nature of a cross‐over design (Maxvold 2000; Wynckel 2004; Yin 2015).

• Recovery of kidney function: 3 studies (304 participants) (Daud 2006; OMAKI 2012; Saudan 2006).

• Vascular access complications: 2 studies (336 participants) (Meier 2011a; RADICAL 2012).

• Costs to health care services: 2 studies (336 participants) (Meier 2011a; RADICAL 2012).

None of the eligible studies reported the following outcomes: types and number of dialysis filters, circuits and catheters, and consumption of dialysate.

Excluded studies

See Characteristics of excluded studies. Reasons for exclusion were wrong population (14 studies) and wrong intervention (29 studies).

Risk of bias in included studies

See Figure 2 and Figure 3 for findings of our risk of bias assessment in the included studies. Overall, the risk of bias across our 20 included studies was judged to be unclear to high.

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation

Four studies were judged to be at low risk of bias: two used computer‐generated randomisation numbers (RADICAL 2012; Saudan 2006), one study used a web‐based randomisation approach (Fealy 2017), and one study used a random number table (Yin 2015). The remaining studies did not clearly report the method of randomisation and were judged to have an unclear risk of bias.

Allocation concealment

Four studies were judged to be at low risk of bias: three used sealed opaque envelopes for the concealment (Broman 2019; RADICAL 2012; Saudan 2006), and one used a web‐based randomisation approach (Fealy 2017). The remaining included studies did not clearly report their allocation methods to allow for a precise judgment of their risk of selection bias.

Blinding

Two studies blinded the study participants, personnel, and outcome assessors to treatment allocation (Schetz 2012; Yin 2015) and were judged to be at low risk of bias for both performance and detection bias. Five studies were judged to be at high risk of both performance and detection bias (Fealy 2017; Meier 2011a; OMAKI 2012; RADICAL 2012; Ramesh Prasad 2000), and nine studies did not provide sufficient information to determine these risks of bias (Alamartine 1994a; Baldwin 2002; Davies 2008; de Pont 2006; Dungen 2001; Kellum 1998; Maxvold 2000; Plata‐Menchaca 2017).

Broman 2019 blinded study participants and personnel and was judged to be at low risk of performance bias but had unclear detection bias. Three studies were judged to be at high risk of performance bias but at low risk of detection bias (Saudan 2006; van der Voort 2005; Wynckel 2004).

Incomplete outcome data

Sixteen studies followed all the participants from randomisation to end of the study and were judged to be at low risk of attrition bias (Baldwin 2002; Daud 2006; Dungen 2001; Fealy 2017; Kellum 1998; Maxvold 2000; Meier 2011a; OMAKI 2012; Plata‐Menchaca 2017; RADICAL 2012; Ramesh Prasad 2000; Saudan 2006; Schetz 2012; van der Voort 2005; Wynckel 2004; Yin 2015). We judged two studies to be at high risk of attrition bias due to missing data of more than 10% of the study participants (Davies 2008; de Pont 2006), and two studies did not provide sufficient information to determine the risk of bias (Alamartine 1994a; Broman 2019).

Selective reporting

We were able to find the study protocols for only 5 of the 18 included studies (Broman 2019; Fealy 2017; OMAKI 2012; Plata‐Menchaca 2017; RADICAL 2012). We judged Plata‐Menchaca 2017 to be at high risk of reporting bias because the study did not report our prespecified outcome of death.

Other potential sources of bias

For this review, we also defined two potential sources of bias: baseline imbalance and co‐interventions. Ten studies described baseline characteristics and we judged them to be at low risk of other bias regarding baseline imbalance (Alamartine 1994a; Daud 2006; Fealy 2017; Maxvold 2000; Meier 2011a; OMAKI 2012; RADICAL 2012; Ramesh Prasad 2000; Saudan 2006; Schetz 2012).

For co‐interventions, three studies were judged to be at high risk of bias: the amount of heparin used was different between the two groups in Alamartine 1994a and Meier 2011a, and anticoagulant usage and frequency of saline flushing varied between the two groups in Ramesh Prasad 2000. Ten studies were judged to be at low risk of bias (Baldwin 2002; Davies 2008; de Pont 2006; Dungen 2001; Fealy 2017; RADICAL 2012; Schetz 2012; van der Voort 2005; Wynckel 2004; Yin 2015)

For the remaining studies, we judged them to be at unclear risk of bias due to their cross‐over design. We could not determine whether the baseline characteristics were balanced between the intervention and control groups amongst these cross‐over trials since group‐based baseline information was unavailable. Additionally, we planned to include studies that allocated anticoagulants or sedative agents to both intervention and control arms according to a predefined protocol; however, the majority of the included studies did not report the administration procedures of these treatment options. We decided to include these studies in our review but judged their risk of 'other bias' regarding co‐interventions as per the Cochrane Collaboration's tool for assessing the risk of bias.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7; Table 8; Table 9

Please refer to Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7; Table 8; Table 9 for further information on our review findings and corresponding GRADE assessment to assess the certainty of included evidence.

Dialysis modalities

CVVHD versus CVVH or CVVHDF

Four studies were included in this comparison (Daud 2006; Kellum 1998; OMAKI 2012; Plata‐Menchaca 2017).

Death from any cause

None of the included studies reported death at 28 days, therefore we synthesised data from the four studies that reported death at the longest follow‐up period (range 24 hours to 90 days). It is uncertain whether CVVHD reduced the risk of death from any cause at the longest follow‐up because the certainty of the evidence was very low (Analysis 1.1 (4 studies, 216 participants): RR 0.78, 95% CI 0.41 to 1.39; I² = 78%).

1.1. Analysis.

Comparison 1: CVVHD versus CVVH or CVVHDF, Outcome 1: Death from any cause

Recovery of kidney function

Two studies reported recovery of kidney function at ICU discharge (Daud 2006; OMAKI 2012). It is uncertain whether CVVHD increased recovery of kidney function because the certainty of the evidence was very low (Analysis 1.2 (2 studies, 93 participants): RR 0.91, 95% CI 0.37 to 2.21; I² = 0%).

1.2. Analysis.

Comparison 1: CVVHD versus CVVH or CVVHDF, Outcome 2: Recovery of kidney function

Other outcomes

The following outcomes were not reported in the four included studies.

• Circuit lifespan

• Vascular access complications

• Costs

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid.

CVVHDF versus CVVH

Two studies were included in this comparison (Davies 2008; Saudan 2006).

Circuit lifespan

CVVHDF may increase circuit lifespan when compared with CVVH (Analysis 2.1 (1 study, 62 circuits): MD 10.15 hours, 95% CI 5.15 to 15.15; low certainty evidence).

2.1. Analysis.

Comparison 2: CVVHDF versus CVVH, Outcome 1: Circuit lifespan

Death from any cause at day 28

It is uncertain whether CVVHDF decreased the risk of death from any cause at 28 days compared to CVVH because the certainty of the evidence was very low (Analysis 2.2 (1 study 206 participants): RR 0.68, 95% CI 0.52 to 0.90).

2.2. Analysis.

Comparison 2: CVVHDF versus CVVH, Outcome 2: Death from any cause at 28 days

Recovery of kidney function

It is uncertain whether CVVHDF improved the recovery of kidney function compared to CVVH because the certainty of the evidence was very low (Analysis 2.3 (1 study, 96 participants): RR 1.10, 95% CI 0.86 to 1.41).

2.3. Analysis.

Comparison 2: CVVHDF versus CVVH, Outcome 3: Recovery of kidney function

Other outcomes

The following outcomes were not reported in the two included studies.

• Vascular access complications

• Costs

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid.

Pre‐dilution versus post‐dilution (as defined by the study investigators)

Two studies were included in this comparison (de Pont 2006; van der Voort 2005).

Circuit lifespan

Pre‐dilution may lead to an extended circuit lifespan compared to post‐dilution (Analysis 3.1 (2 studies, 47 circuits): MD 9.34 hours, 95% CI ‐2.60 to 21.29; I² = 13%; low certainty evidence).

3.1. Analysis.

Comparison 3: Pre‐dilution versus post dilution (as defined by study investigators), Outcome 1: Circuit lifespan

Costs

van der Voort 2005 reported the costs of one‐week treatment of CVVH using pre‐ or post‐dilution were 1,817 Euros and 2,146 Euros, respectively.

Other outcomes

The following outcomes were not reported in the two included studies.

• Death

• Vascular access complications

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid.

Blood flow rate

Higher (≥ 250 mL/min) versus standard blood flow (< 250 mL/min)

We included two studies for quantitative synthesis (Fealy 2017; Ramesh Prasad 2000). Ramesh Prasad 2000 compared standard blood flow (125 mL/min) with 100 mL bolus administration once/hour and augmented blood flow (200 to 250 mL/min) with 100 ml fluid bolus administration once every 30 minutes. We considered this comparison to be eligible for inclusion but rated the study as being at high risk of other bias in terms of co‐interventions.

Circuit lifespan

Higher blood flow may make little or no difference to prolonging circuit lifespan compared to standard blood flow (Analysis 4.1 (2 studies, 499 circuits): MD 0.64 hours, 95% CI ‐3.37 to 4.64; I² = 70%; low certainty evidence).

4.1. Analysis.

Comparison 4: Higher blood flow versus standard blood flow, Outcome 1: Circuit lifespan

Other outcomes

The following outcomes were not reported in the two included studies.

• Death

• Recovery of kidney function

• Vascular access complications

• Costs

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid.

Catheter types

Longer (> 20 cm) versus shorter (≤ 20 cm) catheter

One study compared the effects for preventing circuit clotting by either placing the tip of a longer catheter targeting the right atrium versus placing a shorter catheter targeting the tip in the superior vena cava (RADICAL 2012).

Circuit lifespan

Compared with a shorter catheter, using a longer catheter may prolong circuit lifespan (Analysis 5.1 (1 study, 402 circuits): MD 6.50 hours, 95% CI 1.48 to 11.52; low certainty evidence).

5.1. Analysis.

Comparison 5: Longer (> 20 cm) versus shorter (≤ 20 cm) catheter, Outcome 1: Circuit lifespan

Death from any cause at day 28

It is uncertain whether a longer catheter reduced the risk of death due to any cause at day 28 compared to a shorter catheter because the certainty of the evidence was very low (Analysis 5.2 (1 study, 94 participants): RR 1.38, 95% CI 0.61 to 3.11).

5.2. Analysis.

Comparison 5: Longer (> 20 cm) versus shorter (≤ 20 cm) catheter, Outcome 2: Death from any cause at 28 days

Vascular access complications

Compared with shorter catheters, the use of a longer catheter may reduce the risk of vascular access complications (Analysis 5.3 (1 study, 94 participants): RR 0.40, 95% CI 0.22 to 0.74; low certainty evidence).

5.3. Analysis.

Comparison 5: Longer (> 20 cm) versus shorter (≤ 20 cm) catheter, Outcome 3: Vascular access complications

Costs

RADICAL 2012 reported that potential cost reduction using a longer catheter targeting the right atrium was 427 USD per person (95% CI 61 to 794) when compared with a shorter catheter targeting the superior vena cava.

Other outcomes

The following outcomes were not reported.

• Recovery of kidney function

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid.

Surface‐modified versus standard double‐lumen catheter (based on identical geometry and flow design)

One study compared surface‐modified double‐lumen catheters with standard double‐lumen catheters (Meier 2011a).

Circuit lifespan

Compared with a standard double‐lumen catheter, the use of a surface‐modified double‐lumen catheter may extend circuit lifespan (Analysis 6.1 (1 study, 262 circuits): MD 16.00 hours, 95% CI 13.49 to 18.51; low certainty evidence).

6.1. Analysis.

Comparison 6: Surface‐modified versus standard double‐lumen catheter with identical geometry and flow design, Outcome 1: Circuit lifespan

Death from any cause at day 28

Meier 2011a reported ICU death. It is uncertain whether surface‐modified double‐lumen catheter reduced the risk of ICU deaths compared to standard double‐lumen catheter because the certainty of the evidence was very low (Analysis 6.2 (1 study, 236 participants): RR 1.08, 95% CI 0.76 to 1.53; very low certainty evidence).

6.2. Analysis.

Comparison 6: Surface‐modified versus standard double‐lumen catheter with identical geometry and flow design, Outcome 2: Death from any cause at 28 days

Recovery of kidney function

Compared with a standard double‐lumen catheter, a surface‐modified double‐lumen catheter may not lead to the recovery of kidney function (Analysis 6.3 (1 study, 236 participants): RR 0.91, 95% CI 0.70 to 1.20; low certainty evidence).

6.3. Analysis.

Comparison 6: Surface‐modified versus standard double‐lumen catheter with identical geometry and flow design, Outcome 3: Recovery of kidney function

Vascular access complications

Compared with standard double‐lumen catheters, surface‐modified double‐lumen catheters may make little or no difference to vascular access complications (Analysis 6.4 (1 study, 236 participants): RR 0.32, 95% CI 0.08 to 1.20; low certainty evidence).

6.4. Analysis.

Comparison 6: Surface‐modified versus standard double‐lumen catheter with identical geometry and flow design, Outcome 4: Vascular access complications

Costs

Compared with standard double‐lumen catheters, surface‐modified double‐lumen catheters may lead to reduced medical costs (exchange costs of temporary catheters) (Analysis 6.5 (1 study, 236 participants): MD ‐5508.00 USD, 95% CI ‐5612.98 to ‐5403.02; low certainty evidence)

6.5. Analysis.

Comparison 6: Surface‐modified versus standard double‐lumen catheter with identical geometry and flow design, Outcome 5: Costs

Other outcomes

The following outcomes were not reported.

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid.

Membrane types

Polyethylenimine‐treated AN69 membrane (AN69ST) versus other membranes

Three studies compared AN69ST with other types of membranes as a non‐pharmacological intervention for preventing circuit clotting during CKRT (Broman 2019; Schetz 2012; Yin 2015). Broman 2019 but did not examine the prespecified outcome measures of this review.

Circuit lifespan

AN69ST may make little or no difference to circuit lifespan compared to other membranes (Analysis 7.1 (2 studies, 56 circuits): MD 1.54 hours, 95% CI ‐4.53 to 7.60; I² = 0%; low certainty evidence).

7.1. Analysis.

Comparison 7: Polyethylenimine‐treated AN69 membrane (AN69ST) versus other membrane, Outcome 1: Circuit lifespan

Other outcomes

The following outcomes were not reported in the three included studies.

• Death

• Recovery of kidney function

• Vascular access complications

• Costs

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid.

Flat plate versus hollow fibre filters of the same membrane type

One study compared flat plate and hollow fibre filters for preventing circuit clotting during CKRT (Baldwin 2002).

Circuit lifespan

It is uncertain whether the use of a flat plate haemofilter extended circuit lifespan compared to hollow fibre filters as the certainty of the evidence was very low (Analysis 8.1 (1 study, 38 participants): MD ‐1.40 hours, 95% CI ‐12.12 to 9.32).

8.1. Analysis.

Comparison 8: Flat plate versus hollow fibre filter of the same membrane type, Outcome 1: Circuit lifespan

The following outcomes were not reported.

• Death

• Recovery of kidney function

• Vascular access complications

• Costs

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid

More and shorter hollow fibre filters versus standard filters

One study compared more and shorter hollow fibre filters with standard filters (Dungen 2001).

Circuit lifespan

Compared with standard filters, the use of a filter with more and shorter hollow fibres may reduce circuit lifespan (Analysis 9.1 (1 study, 6 participants): MD ‐5.87 hours, 95% CI ‐10.18 to ‐1.56; low certainty evidence).

9.1. Analysis.

Comparison 9: More and shorter versus standard hollow fibre filter of the same membrane type, Outcome 1: Circuit lifespan

The following outcomes were not reported.

• Death

• Recovery of kidney function

• Vascular access complications

• Costs

• Types and number of dialysis filters, circuits, and catheters

• Consumption of dialysate/dialysis fluid

Filters with a larger membrane surface area versus a smaller membrane surface area

Baldwin 2002 reported circuit lifespan did not differ between larger versus smaller membrane surface areas (15.8 hours (SE 14.3) versus 16.8 hours (SE 13.1), respectively). However, the precise numbers of participants involved in these two analyses were not reported. We contacted the corresponding author in an attempt to obtain additional information for further clarification, but the respective data were not available.

Discussion

Summary of main results

This review included 20 studies involving 1143 randomised participants receiving CKRT that reported the effects of a variety of non‐pharmacological interventions on circuit clotting‐related outcomes. Most of the included studies were judged to be at unclear risk of bias. All but one study involved adult participants. The causes of AKI varied across studies but the major cause (as indicated in 11 included studies) was sepsis. According to the GRADE approach, we found that the certainty of the available evidence regarding non‐pharmacological interventions for preventing clotting of extracorporeal circuits during CKRT was low to very low. CVVHDF (as compared with CVVH), pre‐dilution haemofiltration (as compared with post‐dilution haemofiltration), placing the tip of a longer catheter targeting the right atrium (as compared with placing a shorter catheter targeting the tip in the superior vena cava, and surface‐modified double‐lumen catheter (as compared with a standard catheter) may extend circuit lifespan. It is worth noting that the majority of the included studies reported findings on circuit lifespan but not death.

Overall completeness and applicability of evidence

We conducted the present review using standard Cochrane review methodology in order to present a comprehensive assessment of available evidence indicating the effects of non‐pharmacological interventions for preventing circuit clotting during CKRT; we included eligible studies irrespective of participants' age, sex, or underlying/primary conditions. Our approach yielded evidence on two primary outcomes and five secondary outcomes across 10 treatment comparisons. However, as our prespecified outcomes of interest were often not reported, we were only able to present limited evidence for a few outcomes in each comparison. Furthermore, only one RCT contributed to each estimate of circuit lifespan in the following comparisons: CVVHDF versus CVVH (Davies 2008), long versus short catheter (RADICAL 2012), surface‐modified versus standard double‐lumen catheter with identical geometry and flow design (Meier 2011a), flat plate versus hollow fibre filter of the same membrane type (Baldwin 2002), and more and shorter versus standard hollow fibre filter (Dungen 2001). For a number of comparisons, we found only one eligible study reporting the outcome measures of interest. Cross‐over studies, accounting for more than half of the included studies in this review, did not provide data from the first phase i.e. before crossing over. We thus decided instead to use the results of within‐group comparisons. Due to limited data, we could not perform predetermined subgroup or sensitivity analyses, nor did we identify sufficient numbers of studies to assess publication bias. We identified five ongoing studies (NCT01062984; NCT01779635; NCT01790620; NCT03426943; Weidhase 2020); of these, three planned to compare the modes of haemodialysis: CVVH, CVVHD, and CVVHDF (Weidhase 2020; NCT01062984; NCT01790620), but we were unable to identify any of our prespecified outcome measures in the study registers. NCT01779635 and NCT03426943 plan to examine the effects of AN69ST membrane, which falls into the scope of our review, on the circuit life and death. This information will be incorporated in future updates. 

We are aware that the applicability of our review evidence is limited because the majority of the included studies were performed more than 10 years ago and, understandably, clinical practice in this field has since changed. As an example, a number of the included studies did not use specific treatment criteria but instead relied on physicians' judgment for the initiation of CKRT (Davies 2008; Fealy 2017; Schetz 2012). Moreover, the appropriate timing of CKRT initiation is yet to be fully investigated (Gaudry 2020; Wald 2014). Regional citrate anticoagulation has been widely used and it was recommended in international guidelines (KDIGO Acute Kidney Injury Work Group 2012). The choice of anticoagulants should have an impact on the filter life in the current era. It should be highlighted that the evaluation of pharmacological interventions to prevent clotting is covered by a separate Cochrane review, as such, the findings from this review should be read in combination with the review on pharmacological interventions to prevent clotting of extracorporeal circuits during CKRT (Tsujimoto 2020).

Quality of the evidence

We assessed the certainty of included randomised evidence using the GRADE approach (Guyatt 2008). Overall, the certainty of our review evidence was judged to be low to very low due to the high risk of bias and small sample size amongst the included studies. For the majority of these studies, random sequence generation, allocation concealment, and blinding domains were assessed to be at unclear or high risk of bias (Figure 3). The total numbers of participants in all meta‐analyses were less than the optimal information size of 400, as recommended by the GRADE framework (Guyatt 2011a). With regards to the thresholds for appreciable treatment benefits and harms, we used a relative risk reduction of 0.25 for binary outcomes, and a standard deviation of 0.2 for continuous outcomes (Cohen 1988; Guyatt 2011a). Regarding the timing of mortality measurements, only one study in this review reported the prespecified 28‐day mortality and the others were downgraded for indirectness. As aforementioned, identified ongoing studies did not plan to include our review outcomes of interest and we judged that publication bias was highly unlikely.

Potential biases in the review process

We used the effect estimates from within‐group comparisons in the included cross‐over trials since outcome data from the first phase (before crossing over) were unavailable. However, it is unlikely that participant status before and after crossing over was similar since health conditions, in general, often change drastically within a short period of time in the ICU setting (Leloup 2020). Additionally, one could foresee the circuit lifespan during the second phase (after crossing over) given that the study interventions were not blinded to study personnel or participants. This, in effect, might lead to the use of co‐interventions, such as saline flushes, in the second phase. Therefore, we explicitly considered baseline imbalance and co‐interventions as potential sources of 'Other risk of bias' as per the Cochrane Collaboration's tool for assessing the risk of bias (Higgins 2011). As a result, half of the included studies were judged to be at high or unclear risk of co‐interventions. The lack of adjustment for anticoagulation methods should also be acknowledged as a limitation of this review. Additionally, we did not consider censoring effect on circuit life as this often needs individual participant data. The robustness of our findings should be further confirmed by comparing survival censoring for not clotting or for catheter reasons. We were unable to locate study plans/protocols for numerous included studies and consequently rated the risk of reporting bias as unclear. 
 

Agreements and disagreements with other studies or reviews

A published non‐Cochrane systematic review included both RCTs and observational studies assessing the effects of non‐pharmacological interventions (Brain 2017). The review authors reported results similar to ours, where longer catheter or CVVHDF were superior in terms of prolonging circuit lifespan. On the other hand, as opposed to findings reported in Brain 2017, we found that higher blood flow exerted no effects on circuit lifespan. This might be due to the inclusion of non‐randomised evidence in Brain 2017. It should be noted that the included trials examined higher blood flow versus standard blood flow. As such, lowering the blood flow rate below the target, which is based on filtration fraction when filtration is used is not recommended for improving circuit life in this review. However, as mentioned above, our review outcomes, including death due to any cause and kidney function recovery, were often not reported.

Brain 2017 assessed the effects of several non‐pharmacological interventions that intended to prolong circuit life on several outcomes, such as different access sites and types of catheters however, the evidence came from observational studies and it was not our intention to consider observational studies in this review. It is worth highlighting some novel findings presented in this Cochrane review, where the surface‐modified double‐lumen catheter (as compared with standard double‐lumen catheter and the use of a long catheter (as compared with a short catheter) are generally cheaper and may prolong circuit lifespan.

As for the other included non‐pharmacological interventions, the currently available evidence is insufficient to allow for comprehensive assessment on the effects of a non‐pharmacological approach on circuit clotting as well as prognosis for people with AKI undergoing CKRT. We would like to highlight that a major strength of our review is adopting the robust Cochrane methodology during the entire review development process, in particular the use of a prespecified protocol as well as the GRADE approach to assess of certainty of evidence.

Authors' conclusions

Implications for practice.

Current evidence shows that the following non‐pharmacological strategies: CVVHDF, pre‐dilution haemofiltration, longer catheter, and surface‐modified double‐lumen catheter, may extend circuit lifespan during CKRT. On the other hand, a filter with more and shorter hollow fibres may reduce circuit lifespan. Analyses revealed that applying a higher blood flow may not alter circuit life. The available evidence is limited and thus we are unable to draw any solid conclusions in regards to the role and impact of the other non‐pharmacological interventions for preventing circuit clotting amongst people receiving CKRT.

Implications for research.

Recent clinical practice guidelines did not address the use of non‐pharmacological interventions for maintaining circuit patency (KDIGO Acute Kidney Injury Work Group 2012). Current evidence is limited and thus further high‐quality RCTs are needed to provide a more comprehensive assessment before anyone particular non‐pharmacological intervention is to be recommended for clinical practice. Our review illustrates that any research into non‐pharmacological methods should be prospective, randomised, adequately powered, and transparently reported. Furthermore, future studies should not only consider circuit lifespan as an outcome measure but also other patient‐important outcomes including death. We remain cautious as to the appropriateness of a cross‐over study design for future studies in this field due to systematic differences between the two periods of the cross‐over studies are highly likely, which may lead to a higher risk of bias; a potentially higher rate of drop‐outs/losses to follow‐up after the first period is also likely due to high death rates in the critically ill patients in the ICU settings.

What's new

Date	Event	Description
8 November 2021	Amended	Addition of third comparison for recovery of kidney function to Abstract results

History

Protocol first published: Issue 5, 2019
Review first published: Issue 9, 2021

Date	Event	Description
17 September 2021	Amended	Minor edit to PLS
15 September 2021	Amended	Author affiliation updated

Appendices

Appendix 1. Electronic search strategies

Database	Search terms
CENTRAL	((continuous next venovenous next h*emodialysis) or CVVHD):ti,ab,kw (Word variations have been searched) ((continuous next venovenous next h*emodiafiltration) or CVVHDF):ti,ab,kw (Word variations have been searched) pre‐dilution or post dilution:ti,ab,kw (Word variations have been searched) (femoral or jugular) and catheter*:ti,ab,kw (Word variations have been searched) (blood next flow*):ti,ab,kw (Word variations have been searched) saline flush*:ti,ab,kw (Word variations have been searched) blood air contact*:ti,ab,kw (Word variations have been searched) membrane*:ti,ab,kw (Word variations have been searched) length and catheter*:ti,ab,kw (Word variations have been searched) (circuit failure):ti,ab,kw (Word variations have been searched) (circuit life):ti,ab,kw (Word variations have been searched) {or #1‐#11} MeSH descriptor: [Acute Kidney Injury] 1 tree(s) exploded ("acute kidney injury" or "acute renal injury"):ti,ab,kw ("acute renal failure" or "acute kidney failure"):ti,ab,kw "critically ill":ti,ab,kw "renal replacement therapy":ti,ab,kw {or #13‐#17} {and #12, #18}
MEDLINE	exp Acute Kidney Injury/ (acute kidney failure or acute renal failure).tw. (acute kidney injur$ or acute renal injur$).tw. (acute kidney insufficie$ or acute renal insufficie$).tw. acute tubular necrosis.tw. (ARI or AKI or ARF or AKF or ATN).tw. or/1‐6 Renal Replacement Therapy/ (continuous venovenous h$emofiltration or CVVH).tw. (continuous venovenous h$emodialysis or CVVHD).tw. (continuous venovenous h$emodiafiltration or CVVHDF).tw. (continuous renal replacement or crrt).tw. (circuit adj 2 life).tw. circuit failure.tw. (circuit adj3 clot$).tw. (pre‐dilution or predilution).tw. (post‐dilution or postdilution).tw. flush$.tw. blood air contact.tw. venous bubble$.tw. air‐trap.tw. Catheters/ Membranes, Artificial/ membrane$.tw. blood flow$.tw. Equipment Failure/ Equipment Design/ Blood Flow Velocity/ (femoral or jugular).tw. saline flush$.tw. or/8‐30
EMBASE	acute kidney failure/ acute kidney tubule necrosis/ (acute kidney failure or acute renal failure).tw. (acute kidney injur$ or acute renal injur$).tw. (acute kidney insufficie$ or acute renal insufficie$).tw. acute tubular necrosis.tw. (ARI or AKI or ARF or AKF or ATN).tw. or/1‐7 exp continuous renal replacement therapy/ (continuous venovenous h$emofiltration or CVVH).tw. (continuous venovenous h$emodialysis or CVVHD).tw. (continuous venovenous h$emodiafiltration or CVVHDF).tw. exp continuous renal replacement therapy/ circuit failure.tw. (circuit adj2 life).tw. (circuit adj3 clot$).tw. (pre dilution or predilution).tw. (post dilution or postdilution).tw. exp catheter/ (jugular$ and access).tw. (femoral$ and access).tw. length.tw. or/20‐22 and/19,23 blood flow velocity/ saline flush$.tw. blood air contact.tw. air‐trap.tw. dialysis membrane/ or "filters and membranes"/ or/9‐2

Appendix 2. Cochrane Collaboration's tool for assessing risk of bias

Potential sources of bias	Assessment criteria
Random sequence generation Selection bias (biased allocation to interventions) due to inadequate generation of a randomised sequence	Low risk of bias: Random number table; computer random number generator; coin tossing; shuffling cards or envelopes; throwing dice; drawing of lots; minimisation (minimisation may be implemented without a random element, and this is considered to be equivalent to being random).
	High risk of bias: Sequence generated by odd or even date of birth; date (or day) of admission; sequence generated by hospital or clinic record number; allocation by judgement of the clinician; by preference of the participant; based on the results of a laboratory test or a series of tests; by availability of the intervention.
	Unclear: Insufficient information about the sequence generation process to permit judgement.
Allocation concealment Selection bias (biased allocation to interventions) due to inadequate concealment of allocations prior to assignment	Low risk of bias: Randomisation method described that would not allow investigator/participant to know or influence intervention group before eligible participant entered in the study (e.g. central allocation, including telephone, web‐based, and pharmacy‐controlled, randomisation; sequentially numbered drug containers of identical appearance; sequentially numbered, opaque, sealed envelopes).
	High risk of bias: Using an open random allocation schedule (e.g. a list of random numbers); assignment envelopes were used without appropriate safeguards (e.g. if envelopes were unsealed or non‐opaque or not sequentially numbered); alternation or rotation; date of birth; case record number; any other explicitly unconcealed procedure.
	Unclear: Randomisation stated but no information on method used is available.
Blinding of participants and personnel Performance bias due to knowledge of the allocated interventions by participants and personnel during the study	Low risk of bias: No blinding or incomplete blinding, but the review authors judge that the outcome is not likely to be influenced by lack of blinding; blinding of participants and key study personnel ensured, and unlikely that the blinding could have been broken.
	High risk of bias: No blinding or incomplete blinding, and the outcome is likely to be influenced by lack of blinding; blinding of key study participants and personnel attempted, but likely that the blinding could have been broken, and the outcome is likely to be influenced by lack of blinding.
	Unclear: Insufficient information to permit judgement
Blinding of outcome assessment Detection bias due to knowledge of the allocated interventions by outcome assessors	Low risk of bias: No blinding of outcome assessment, but the review authors judge that the outcome measurement is not likely to be influenced by lack of blinding; blinding of outcome assessment ensured, and unlikely that the blinding could have been broken.
	High risk of bias: No blinding of outcome assessment, and the outcome measurement is likely to be influenced by lack of blinding; blinding of outcome assessment, but likely that the blinding could have been broken, and the outcome measurement is likely to be influenced by lack of blinding.
	Unclear: Insufficient information to permit judgement
Incomplete outcome data Attrition bias due to amount, nature or handling of incomplete outcome data.	Low risk of bias: No missing outcome data; reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk not enough to have a clinically relevant impact on the intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardised difference in means) among missing outcomes not enough to have a clinically relevant impact on observed effect size; missing data have been imputed using appropriate methods.
	High risk of bias: Reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk enough to induce clinically relevant bias in intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes enough to induce clinically relevant bias in observed effect size; ‘as‐treated’ analysis done with substantial departure of the intervention received from that assigned at randomisation; potentially inappropriate application of simple imputation.
	Unclear: Insufficient information to permit judgement
Selective reporting Reporting bias due to selective outcome reporting	Low risk of bias: The study protocol is available and all of the study’s prespecified (primary and secondary) outcomes that are of interest in the review have been reported in the prespecified way; the study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were prespecified (convincing text of this nature may be uncommon).
	High risk of bias: Not all of the study’s prespecified primary outcomes have been reported; one or more primary outcomes is reported using measurements, analysis methods or subsets of the data (e.g. sub‐scales) that were not prespecified; one or more reported primary outcomes were not prespecified (unless clear justification for their reporting is provided, such as an unexpected adverse effect); one or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta‐analysis; the study report fails to include results for a key outcome that would be expected to have been reported for such a study.
	Unclear: Insufficient information to permit judgement
Other bias Bias due to problems not covered elsewhere in the table	Low risk of bias: The study appears to be free of other sources of bias. (e.g. co‐intervention of anticoagulation)
	High risk of bias: Had a potential source of bias related to the specific study design used; stopped early due to some data‐dependent process (including a formal‐stopping rule); had extreme baseline imbalance; has been claimed to have been fraudulent; had some other problem.
	Unclear: Insufficient information to assess whether an important risk of bias exists; insufficient rationale or evidence that an identified problem will introduce bias.

Data and analyses

Comparison 1. CVVHD versus CVVH or CVVHDF.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Death from any cause	4	216	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.41, 1.39]
1.1.1 Death at 24 hours	1	13	Risk Ratio (M‐H, Random, 95% CI)	0.43 [0.05, 3.64]
1.1.2 Death at 60 days	1	77	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.66, 1.46]
1.1.3 Death at 90 days	1	106	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.23, 0.68]
1.1.4 Death in ICU	1	20	Risk Ratio (M‐H, Random, 95% CI)	1.17 [0.79, 1.74]
1.2 Recovery of kidney function	2	93	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.37, 2.21]

Comparison 2. CVVHDF versus CVVH.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Circuit lifespan	1	62	Mean Difference (IV, Random, 95% CI)	10.15 [5.15, 15.15]
2.2 Death from any cause at 28 days	1	206	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.52, 0.90]
2.3 Recovery of kidney function	1	96	Risk Ratio (M‐H, Random, 95% CI)	1.10 [0.86, 1.41]

Comparison 3. Pre‐dilution versus post dilution (as defined by study investigators).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Circuit lifespan	2	47	Mean Difference (IV, Random, 95% CI)	9.34 [‐2.60, 21.29]

Comparison 4. Higher blood flow versus standard blood flow.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Circuit lifespan	2	499	Mean Difference (IV, Random, 95% CI)	0.64 [‐3.37, 4.64]

Comparison 5. Longer (> 20 cm) versus shorter (≤ 20 cm) catheter.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Circuit lifespan	1	402	Mean Difference (IV, Random, 95% CI)	6.50 [1.48, 11.52]
5.2 Death from any cause at 28 days	1	94	Risk Ratio (M‐H, Random, 95% CI)	1.38 [0.61, 3.11]
5.3 Vascular access complications	1	94	Risk Ratio (M‐H, Random, 95% CI)	0.38 [0.20, 0.72]

Comparison 6. Surface‐modified versus standard double‐lumen catheter with identical geometry and flow design.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Circuit lifespan	1	262	Mean Difference (IV, Random, 95% CI)	16.00 [13.49, 18.51]
6.2 Death from any cause at 28 days	1	236	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.76, 1.53]
6.3 Recovery of kidney function	1	236	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.70, 1.20]
6.4 Vascular access complications	1	236	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.09, 1.20]
6.5 Costs	1	236	Mean Difference (IV, Random, 95% CI)	‐5.51 [‐5.61, ‐5.40]

Comparison 7. Polyethylenimine‐treated AN69 membrane (AN69ST) versus other membrane.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Circuit lifespan	2	56	Mean Difference (IV, Random, 95% CI)	1.54 [‐4.53, 7.60]

Comparison 8. Flat plate versus hollow fibre filter of the same membrane type.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
8.1 Circuit lifespan	1	38	Mean Difference (IV, Random, 95% CI)	‐1.40 [‐12.12, 9.32]

Comparison 9. More and shorter versus standard hollow fibre filter of the same membrane type.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
9.1 Circuit lifespan	1	12	Mean Difference (IV, Random, 95% CI)	‐5.87 [‐10.18, ‐1.56]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Alamartine 1994a.

Study characteristics
Methods	Publication type: full article Study design: parallel RCT Follow‐up period: not reported Duration of study: not reported
Participants	Country: France Setting: not reported ICU: yes Number: treatment group (7); control group (6) Mean age ± SD: not reported Sex (M/F): not reported Mean severity ± SD (APACHE 2 score): treatment group (24 ± 6); control group (20 ± 4) Cause of AKI: not reported Inclusion criteria: not reported Exclusion criteria: not reported
Interventions	Treatment group CVVHDF with convection that was obtained by a 2 L/hour pre‐dilutional infusion Control group CVVHD without convection
Outcomes	None of our prespecified outcomes of interest were reported
Notes	Funding source: Fresenius‐Smad
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Selective reporting (reporting bias)	Unclear risk	Study protocol was not available; none of our prespecified outcomes were reported
Other bias: baseline imbalance	Low risk	We did not identify any substantial differences between groups
Other bias: co‐interventions	High risk	The amount of heparin use in the two groups was substantially different

Baldwin 2002.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: 18 months
Participants	Country: Australia Setting: not reported ICU: yes Number: 38 Mean age ± SD (years): not reported Sex (M/F): not reported Severity: not reported Cause of AKI: not reported Inclusion criteria: critically ill patients with AKI requiring CKRT Exclusion criteria: not reported
Interventions	Treatment group 1 Flat plate filter Control group 1 Hollow fibre filter of the same membrane type as the intervention group Treatment group 2 Administration of heparin equally divided into the venous air‐chamber (50%) and into the standard prefilter port (50%) Control group 2 Complete (100%) infusion into the standard prefilter port alone Treatment group 3 Filter with a larger membrane surface area (Filtral 12, Hospal, Lyon, France; surface area: 1.3 m²) Control group 3 Filter with a smaller membrane surface area (Filtral 8, Hospal, Lyon, France; surface area: 0.75 m²)
Outcomes	Circuit lifespan
Notes	Funding source: unclear We could not extract data from the first period (before crossing over)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comparison 1 Low risk: No missing outcome data Comparison 2 Unclear: Insufficient information to permit judgement Comparison 3 Unclear: Insufficient information to permit judgement
Selective reporting (reporting bias)	Unclear risk	Trial protocol was not available
Other bias: baseline imbalance	Unclear risk	Cross‐over RCT; unable to assess baseline imbalance
Other bias: co‐interventions	Low risk	Study investigators predefined the treatment protocol for heparin during CVVH

Broman 2019.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: February 2016 to February 2018
Participants	Country: Sweden Setting: multicentre (number of sites not reported) ICU: yes Number (randomised/analysed): 20/16 Mean age ± SD: 69.4 ±11.2 years Sex (M/F): 10/6 Mean severity ± SD (SAPS 3): 73.8 ± 11.5 Cause of AKI (number): septic shock (16) Inclusion criteria: adult ICU patients who had septic shock, with a blood culture positive for a gram‐negative bacteria or suspected to be caused by a gram‐negative agent; endotoxin level > 0.03 EU/mL; associated KDIGO stage 3 AKI Exclusion criteria: < 18 years; known HIV; hepatitis B or C infection
Interventions	Treatment group CVVHDF with an oXiris™ (AN69ST) filter Control group CVVHDF with a standard filter
Outcomes	None of our prespecified outcomes of interest were reported
Notes	Funding source: Baxter Healthcare We could not extract data from the first period (before crossing over)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Low risk	simple sealed opaque envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The treating physician and staff as well as the participant were blinded to the type of filter by covering the brand marks on the front and the bar code on the back using non‐transparent tape
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement; no outcome of interest was reported
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Insufficient information to permit judgement; no outcome of interest was reported
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; no outcome of interest was reported
Other bias: baseline imbalance	Unclear risk	Cross‐over design; could not assess baseline imbalance
Other bias: co‐interventions	Unclear risk	Insufficient information to permit judgement

Daud 2006.

Study characteristics
Methods	Publication type: full article Study design: parallel RCT Follow‐up period: during ICU stay Duration of study: September 2001 to December 2003
Participants	Country: Malaysia Setting: multicentre ICU: yes Number: treatment group (11); control group (9) Mean age, IQR (years): treatment group (49.9, 36.00); control group (52.0, 26.50) Sex (males): treatment group (63.6%); control group (55.6%) Mean severity ± SD (SAPS 2): treatment group (65.1 ± 15.5); control group (66.5 ± 11) Cause of AKI (number): sepsis (15); surgical (5) Inclusion criteria: critically ill on ventilatory support in ICU or coronary care unit with AKI; AKI defined by a sudden rise in SCr to > 2.3 mg/dL (200 μmoI/L) in participants with prior normal kidney function, and/or a rise by > 2.3 mg/dL (200 μmoI/L) in participants with prior kidney dysfunction plus one of the following:‐ (i) oliguria (urine output < 200 mL in the preceding 12 hours) despite fluid resuscitation and furosemide administration, (ii) severe acidaemia (pH < 7.1) due to metabolic acidosis, (iii) hyperkalaemia (potassium > 6.5 mmol/L), and (iv) suspected uraemic organ involvement (pericarditis, encephalopathy) Exclusion criteria: aged < 13 years; pregnant; kidney failure with KRT on chronic dialysis or calculated CrCl < 10 mL/min; presence of more than 3 major organ failures at study recruitment; SBP < 90 mm Hg at study recruitment despite ionotropic support
Interventions	Treatment group CVVH with ultrafiltration of 2 L/hour Control group CVVHD with the dialysate outflow rate of 1.7 L/hour
Outcomes	ICU death Complete kidney recovery at ICU discharge
Notes	Funding sources: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data identified
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Low risk	No substantial differences between the two groups
Other bias: co‐interventions	Unclear risk	Insufficient information to permit judgement

Davies 2008.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: December 2004 to July 2006
Participants	Country: Australia Setting: single centre ICU: yes Number: 45 Mean age ± SD: 56.7 ± 17.2 years Sex (M/F): 24/21 Mean severity ± SD (APACHE 2): 25.53 ± 6.2 Cause of AKI (number): pneumonia (12); septic shock (9), surgical (15); other medical conditions (9) Inclusion criteria: ≥ 18 years of age admitted to the ICU requiring CKRT Exclusion criteria: unable to receive the unit standard heparin regimen for anticoagulation of the extracorporeal circuit, such as recent trauma or surgical intervention; an existing coagulopathy and/or thrombocytopenia
Interventions	Treatment group CVVH with ultrafiltrate dose of 35 mL/kg/hour Control group CVVHDF with a fixed pre‐dilution volume of 600 mL/hour replacement fluid with a dialysate dose of 1 L/hour
Outcomes	Circuit lifespan
Notes	Funding source: Tyco Healthcare We could not extract data from the first period (before crossing over)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Incomplete outcome data (attrition bias) All outcomes	High risk	A total of 14 participants who did not cross‐over were eventually excluded from study analysis without further information of the intervention they received in the first phase
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; trial protocol was not available
Other bias: baseline imbalance	Unclear risk	Cross‐over design; could not assess baseline imbalance
Other bias: co‐interventions	Low risk	Study investigators predefined the protocol for administering heparin during CVVH

de Pont 2006.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: not reported
Participants	Country: Netherlands Setting: not reported ICU: yes Number: 8 Mean age ± SD: 63 ± 13 years Sex (M/F): 3/5 Mean severity ± SD (APACHE 2): 23 ± 8 Cause of AKI (number): infection (6), others (2) Inclusion criteria: critically‐ill adults with an indication for CKRT Exclusion criteria: recent bleeding; treatment with aspirin within 1 week before enrolment; treatment with therapeutic doses of unfractionated or LMWH within 12 hours before enrolment; results of routine coagulation tests (PT, APTT) exceeding twice the upper limit of normal
Interventions	Treatment group pre‐dilution with the total flow through the haemofilter constant at 200 mL/min Control group Post‐dilution with the total flow through the haemofilter constant at 200 mL/min
Outcomes	Circuit lifespan
Notes	Funding source: unclear We could not extract data from the first period (before crossing over)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Incomplete outcome data (attrition bias) All outcomes	High risk	In one participant, kidney function recovered after the first haemofiltration run, obviating the need for a second run
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Unclear risk	Cross‐over trial; could not assess baseline imbalance
Other bias: co‐interventions	Low risk	Heparin was equally administered to both arms. Before starting CVVH, a loading dose of 2850 IU nadroparin was administered intravenously, followed by a continuous prefilter infusion of 456 IU hourly

Dungen 2001.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: December 1996 to May 1997
Participants	Country: Germany Setting: single centre ICU: yes Number: 6 Median age (range): 66 years (33 to 76) Sex (M/F): 5/1 Median severity (range (APACHE 3): 57 (43 to 72) Cause of AKI (number): septic shock (6) Inclusion criteria: adults in ICU with AKI due to septic shock Exclusion criteria: pregnancy; < 18 years or > 80 years
Interventions	Treatment group CVVH with handmade filter with more and shorter hollow fibres (number of hollow fibres, 7168; and length 21 cm) Control group CVVH with standard filter (number of hollow fibres, 4608; and length 30 cm)
Outcomes	Circuit lifespan ICU death
Notes	Funding source: the Fresenius Company We could not extract data from the first period (before crossing over)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Unclear risk	Cross‐over trial; we could not assess baseline imbalance
Other bias: co‐interventions	Low risk	No difference in anticoagulants

Fealy 2017.

Study characteristics
Methods	Publication type: full article Study design: parallel RCT Follow‐up period: not reported Duration of study: June 2013 to August 2014
Participants	Country: Australia Setting: single centre ICU: yes Number: treatment group (49); control group (51) Mean age ± SD (years): treatment group (60.77 ± 18.31); control group (61.08 ± 15.96) Sex (M/F): treatment group (24/23); control group (24/15) Mean severity ± SD (APACHE 3): treatment group (87.21 ± 26.28); control group (85.65 ± 23.17) Cause of AKI (number): infection (4); surgical (42); other medical conditions (50) Inclusion criteria: critically ill in ICU and fulfilled 3 criteria: 1) ≥ 18 years; 2) AKI requiring CKRT; and 3) vascular access was via the femoral vein for standardization Exclusion criteria: required citrate anticoagulation (citrate protocol requires set BFR of 150 mL/min); and 2) expected to stay in the ICU < 24 hours
Interventions	Treatment group BFR: 250 mL/min Control group BFR: 150 mL/min
Outcomes	Circuit lifespan
Notes	Funding sources: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Web‐based central randomisation service
Allocation concealment (selection bias)	Low risk	Web‐based central randomisation service
Blinding of participants and personnel (performance bias) All outcomes	High risk	Protocol stated: open (masking not used)
Blinding of outcome assessment (detection bias) All outcomes	High risk	Protocol stated: open (masking not used)
Incomplete outcome data (attrition bias) All outcomes	Low risk	Missing data as reported was minimal (4 of 100) and the reasons for such missing data were well balanced (randomised but CKRT not performed)
Selective reporting (reporting bias)	Low risk	All predefined outcomes were reported
Other bias: baseline imbalance	Low risk	No substantial difference between the 2 groups
Other bias: co‐interventions	Low risk	Anticoagulation was provided according to a predefined ICU protocol, which stated that no anticoagulation in participants at risk of bleeding from coagulopathy or thrombocytopenia

Kellum 1998.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: not reported
Participants	Country: USA Setting: single centre ICU: yes Number: 13 Mean age ± SD (years): not reported Sex (M/F): not reported Severity: not reported Cause of AKI: not reported Inclusion criteria: both severe SIRS and AKI who needed KRT; fulfilled all 3 of the following criteria for severe SIRS with haemodynamic instability: SBP < 90 mm Hg, and/or vasopressor requirement (use of norepinephrine, epinephrine or neosynephrine at any dose, or dopamine > 5 µg/kg/min Cardiac index > 3 L/min/min² At least one of the following: (1) WCC > 11 x 109/L (or < 2.0 x 109/L with > 20% bands); (2) bacteraemia or known site of infection; or (3) whole blood lactate concentration > 2.0 mmol/L Participants were also required to have at least one criterion for acute dialysis: Volume overload with oliguria (urine output < 30 mL/hour), unresponsive to diuretics BUN > 100 mg/dL Hyperkalaemia (potassium > 6.0 mmol/L). Exclusion criteria: not reported
Interventions	Treatment group CVVH with haemofiltration rate of 2 L/hour using a 0.6‐m AN69 haemofilter Control group CVVHD with the dialysate outflow rate of 2 L/hour using a 0.6‐m AN69 haemofilter
Outcomes	24‐hour death
Notes	Funding sources: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Unclear risk	Cross‐over trial; we could not assess baseline imbalance
Other bias: co‐interventions	Unclear risk	Insufficient information to permit judgement

Maxvold 2000.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: not reported
Participants	Country: USA Setting: single centre/paediatric ICU ICU: yes Number: 6 Mean age ± SD: 131.5 ± 81.2 months Sex (M/F): 2/4 Severity: not report Cause of AKI (number): sepsis (2), other (4) Inclusion criteria: unclear, paediatrics undergoing CKRT Exclusion criteria: not reported
Interventions	Treatment group CVVH Control group CVVHD
Outcomes	Death within observation period was reported; however, the number of deaths in each group was not reported due to the cross‐over design
Notes	Funding sources: unclear We could not extract data from the first period (before crossing over)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to permit judgement
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Low risk	Cross‐over study; we could not assess baseline imbalance
Other bias: co‐interventions	Unclear risk	insufficient information to permit judgement

Meier 2011a.

Study characteristics
Methods	Publication type: full article Study design: parallel RCT Follow‐up period: not reported Duration of study: May 2006 to October 2009
Participants	Country: Switzerland Setting: single centre ICU: yes Number: treatment group (118); control group (118) Mean age ± SD (years): treatment group (55.4 ± 19.2); control group (58.4±16.6) Sex (M/F): treatment group (67/51); control group (70/48) Mean severity ± SD (APACHE 2): treatment group (29 ± 8); control group (28 ± 5) Cause of AKI (number): septic shock (104), surgical (23), other (109) Inclusion criteria: critically ill; ≥ 18 years; AKI defined according to the RIFLE classification; had to be deemed by the treating clinician to require CKRT and to meet at least one of the following criteria: oliguria (urine output < 100 mL in a 6‐hour period) that was unresponsive to fluid resuscitation measures; a serum potassium concentration exceeding 6.5 mmol/L; severe acidaemia (pH < 7.2); BUN > 70 mg/dL (25 mmol/L); SCr > 3.4 mg/dL (300 μmol/L); the presence of clinically significant organ oedema (e.g. pulmonary oedema) Exclusion criteria: received any previous KRT during the same hospital admission or who were on maintenance dialysis for ESKD; previous catheterization of the vein during the same ICU stay and/or catheter introduced by guide‐wire exchange before study entry; history of deep venous thrombosis and local trauma at the insertion site
Interventions	Treatment group Film‐coated domain structured (surface‐modified) double lumen catheter (GamCath Dolphin® Protect 1320, Gambro, Hechingen, Germany) Control group Standard double lumen catheter (GamCath® GDK‐1320, Gambro, Hechingen, Germany)
Outcomes	Circuit lifespan ICU death Vascular access complications: includes function (e.g. ability to use CKRT, uninterrupted use without the need for any intervention, percentage change in access blood flow), or infections requiring antibiotic therapy suspected to be catheter related Costs to health care services
Notes	Funding sources: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	High risk	Only single‐blinded (participants)
Blinding of outcome assessment (detection bias) All outcomes	High risk	Unblinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Low risk	No substantial baseline imbalances were detected.
Other bias: co‐interventions	High risk	Continuous heparin dose was different between two groups

OMAKI 2012.

Study characteristics
Methods	Publication type: full article Study design: parallel RCT Follow‐up period: 61 days Duration of study: May 2008 to October 2010 (study completion data)
Participants	Country: Canada Setting: multicentre (6 sites) ICU: yes Number (randomised/analysed): treatment group (39/38); control group (39/35) Mean age ± SD (years): treatment group (64.3 ± 14.3); control group *58.8 ± 13.7) Sex (M/F): treatment group (22/16); control group (25/14) Severity: not reported Cause of AKI (number): sepsis (63), other (14) Inclusion criteria: critically ill adults (≥ 16 years) with AKI, defined as SCr increase ≥ 50% from baseline; at least one of the following indications for KRT initiation needed to be present: (i) oliguria (urine output < 100 mL in the preceding 4 hours); (ii) metabolic acidosis (serum bicarbonate < 15 mmol/L and pH < 7.25); (iii) refractory hyperkalaemia (serum potassium > 6 mmol/L despite medical efforts at potassium removal); (iv) serum urea > 50 mmol/L, or (v) suspected uraemic organ involvement (pericarditis, encephalopathy, neuropathy or myopathy). Participants needed to be haemodynamically unstable, defined as SOFA‐ cardiovascular score ≥ 1 on the day of screening which required the participants to have MAP < 70 mm Hg or receipt of at least one vasopressor or inotrope. Exclusion criteria: receipt of any KRT within the previous 2 months; presence of an obstructive aetiology for AKI; receipt of a kidney transplant within the preceding year; diagnosis of rapidly progressive glomerulonephritis, vasculitis, or acute interstitial nephritis; a clinical indication for intermittent HD (for example, the presence of a dialyzable toxin); terminal illness with an associated life expectancy less than 2 months; moribund status (life expectancy < 48 hours as per judgment of the attending physicians); prior enrolment in this study; enrolment in a competing ICU interventional study; non‐availability of a CKRT machine, or administration of KRT for > 36 hours prior to eligibility assessment
Interventions	Treatment group CVVH Control group CVVHD
Outcomes	Death at 60 days
Notes	Funding sources: Gambro
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded
Blinding of outcome assessment (detection bias) All outcomes	High risk	Unblinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	The number of study participants with missing data was < 10%
Selective reporting (reporting bias)	Low risk	All reported outcomes in the protocol were subsequently reported in the full report
Other bias: baseline imbalance	Low risk	No substantial differences between groups
Other bias: co‐interventions	Unclear risk	Insufficient information to permit judgement

Plata‐Menchaca 2017.

Study characteristics
Methods	Publication type: abstract and protocol Study design: parallel RCT Follow‐up period: 90 days Duration of study: May 2013 to August 2017
Participants	Country: Spain Setting: multicentre (2 sites) ICU: yes Number: treatment group (49); control group (57) Mean age ± SD: 64 ± 13 years Sex (M/F): 70/36 Mean severity ± SD (APACHE 2): 25 ± 9 Cause of AKI: sepsis (106) Inclusion criteria: critically ill adults with sepsis and AKI who met CKRT initiation criteria; within the first 72 hours from ICU admission; clinical diagnosis of severe sepsis or septic shock Exclusion criteria: kidney failure with KRT; received previous CKRT or HD in the last 3 months, or inclusion in another ongoing study
Interventions	Treatment group CVVH + adsorption membrane (AN69‐ST‐150) Control group CVVHD + adsorption membrane (AN69‐ST‐150)
Outcomes	Survival at 90 days Recovery of kidney function at 90 days (stated in protocol only)
Notes	Funding sources: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	"Double blind" written in protocol, but "single blind" written in abstract of full study report
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	"Double blind" written in protocol, but "single blind" written in abstract of full study report
Incomplete outcome data (attrition bias) All outcomes	Low risk	The rate of missing data was less than 10%
Selective reporting (reporting bias)	High risk	The results of mortality was not shown in the publication
Other bias: baseline imbalance	Unclear risk	Insufficient information to permit judgement
Other bias: co‐interventions	Unclear risk	Insufficient information to permit judgement

RADICAL 2012.

Study characteristics
Methods	Publication type: full article Study design: parallel RCT Follow‐up period: not reported Duration of study: February 2010 to February 2011
Participants	Country: Australia Setting: single centre ICU: yes Number: treatment group (50); control group (50) Mean age ± SD (years): treatment group (58 ± 18); shorter group (60 ± 16) Sex (M/F): treatment group (30/17); control group (25/22) Mean severity ± SD (APACHE 2): treatment group (25 ± 7); control group (26 ± 8) Cause of AKI (number): sepsis (32); surgical (15); other (47) Inclusion criteria: critically ill adults requiring CKRT were screened for eligibility for the study; if the treating physicians intended to use one of the great thoracic veins (IJV or subclavian vein) as access for short‐term dialysis catheter placement Exclusion criteria: indwelling long‐term tunnelled dialysis catheter; required catheter access through the femoral vein; received systemic anticoagulation therapy for other medical reasons; planned to receive regional citrate anticoagulation due to excessive bleeding risk; with repeated clotting (defined as 3 consecutive dialyser lifespans 8 hours) had their anticoagulation strategy changed and thus were discontinued from the study at this point because they subsequently received either higher heparin doses or citrate dialysis, which would bias the outcome
Interventions	Treatment group Longer catheter: 20 cm if inserted in a right great thoracic vein and 24 cm if inserted in a left great thoracic vein Control group Shorter catheter: 15 cm if inserted through a right great thoracic vein and 20 cm if inserted through a left great thoracic vein
Outcomes	Circuit lifespan ICU death Hospital death Number of dialysers and dialysis circuits taken down due to vascular access problem Costs related to health care service access
Notes	Funding sources: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	QUOTE "...used a randomization table generated by a computer; sequential numbered and sealed in envelopes to achieve randomization and allocation concealment, respectively"
Allocation concealment (selection bias)	Low risk	QUOTE "...used a randomization table generated by a computer; sequential numbered and sealed in envelopes to achieve randomization and allocation concealment, respectively"
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded
Blinding of outcome assessment (detection bias) All outcomes	High risk	Unblinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	Missing data was less than 10%
Selective reporting (reporting bias)	Low risk	All outcomes were as reported in the protocol
Other bias: baseline imbalance	Low risk	No substantial difference between groups
Other bias: co‐interventions	Low risk	No difference in anticoagulation

Ramesh Prasad 2000.

Study characteristics
Methods	Publication type: full article Study design: parallel RCT Follow‐up period: not reported Duration of study: August 1995 to March 1997
Participants	Country: USA Setting: single centre ICU: yes Number: treatment group (16); control group (18) Mean age ± SD (years): treatment group (55.3 ± 9.6); control group (57.7 ± 10.6) Sex (M/F): treatment group (12/4); control group (14/4) Severity: not reported Cause of AKI: not reported Inclusion criteria: receiving CVVHD in an ICU Exclusion criteria: not reported
Interventions	Treatment group BFR was set at 200 to 250 mL/min and 100 mL fluid boluses were administered at 30‐minute intervals Control group BFR was set at 125 mL/min and a prefilter flush with 0.9% saline or other electrolyte solution infused as a 100 mL bolus once/hour
Outcomes	Circuit lifespan Terminated treatment due to death
Notes	Funding sources: unclear
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding was not feasible due to the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	High risk	Thrombosis was diagnosed by unblinded nursing staff
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were no drop‐outs
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Low risk	No substantial differences between groups
Other bias: co‐interventions	High risk	Use of anticoagulants was left to the discretion of the attending physician, and the reported frequency of saline flush was different between groups.

Saudan 2006.

Study characteristics
Methods	Publication type: full article Study design: parallel RCT Follow‐up period: 90 days Duration of study: October 2000 to December 2003
Participants	Country: Switzerland Setting: single centre ICU: yes Number: treatment group (104); control group (102) Mean age ± SD (years): treatment group (62 ± 15); control group (65 ± 12) Sex (M/F): treatment group (57/47); control group (65/37) Mean severity SD (APACHE 2): treatment group (24 ± 9); control group (26± 9) Cause of AKI: sepsis (71), other (135) Inclusion criteria: adults treated in the medical and surgical ICUs of Geneva University Hospital: oliguria (urine output < 200 mL/12 hours) despite fluid resuscitation and intravenous diuretic treatment; and/or azotaemia (BUN > 430 mmol/L) with urine output <1500 mL/12 hours Exclusion criteria: pre‐kidney failure (reversibility of oliguria/uraemia with fluid administration or with improvement of cardiac output); post‐kidney failure (on kidney ultrasound examination and/or antegrade or retrograde contrast studies if high clinical suspicion); suspicion of glomerular disease (if high clinical suspicion and compatible urinalysis and/or serologic tests); kidney failure with KRT (baseline SCr > 4300 mmol/L or CrCl < 20 mL/min); receiving ACEi (24‐hour temporary exclusion owing to the possibility of ACEi‐induced anaphylactoid reaction with the polyacrylonitrile AN69 filter
Interventions	Treatment group In addition to the ultrafiltration rate, dialysate flow rate was added between 1 and 1.5 L/hour for participants weighing ≤ 70 kg. Owing to the effluent bag capacity, 2.5 L was the maximum hourly ultrafiltration rate if participants were randomised in the CVVHDF group and also received 1.5 L/hour of dialysate Control group Ultrafiltration flow rate (1 to 2.5 L/hour) was determined according to participant’s estimated urea distribution volume (60% of their body weight at enrolment), so that this estimated volume could be cleared within 24 hours, for example, an 80 kg subject will have an ultrafiltration flow rate of 2 L/hour. For the purposes of simplification, hourly ultrafiltration flow rate was rounded off to the upper 500 mL level within the interval from 1 to 2.5 L
Outcomes	Death from any cause at day 28 Recovery of kidney function at day 90
Notes	Funding source: unclear
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	QUOTE "...computer‐generated randomization list, in random blocks of four and six participants"
Allocation concealment (selection bias)	Low risk	QUOTE "...corresponding treatment allocation cards were placed in consecutively numbered opaque envelopes. Each time a participant was enrolled in the study, the next available envelope was opened by the nephrologist on call, and the allocated treatment option was communicated to the treatment team"
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding was impossible for "logistic reasons"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcomes were objective by nature
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were no drop‐outs
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Low risk	No substantial differences between groups
Other bias: co‐interventions	Unclear risk	Insufficient information to permit judgement

Schetz 2012.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: 72 hours Duration of study: November 2005 to June 2007
Participants	Country: Belgium Setting: single centre ICU: yes Number: treatment group (19); control group (20) Mean age ± SD: 59.2 ± 15.4 years Sex (M/F): 30/9 Mean severity ± SD (APACHE 2): 26 ± 5 Cause of AKI: sepsis (20), other (19) Inclusion criteria: adults requiring CKRT and weighing between 30 and 120 kg Exclusion criteria: suspicion of heparin‐induced thrombocytopenia; pregnancy; indication for systemic anticoagulation; anticipated need for transport; poor short‐term prognosis; absence of consent
Interventions	Treatment group AN69ST surface‐treated membrane (ST100) Control group AN69 original membrane (M100)
Outcomes	Circuit lifespan Vascular access complications
Notes	Funding sources: none We extracted the data from the first period (before cross‐over)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Methods for generating random sequence were unclear. Study authors only reported that the sequence was "determined by randomization (numbered and sealed allocation)"
Allocation concealment (selection bias)	Unclear risk	Methods regarding allocation concealment were unclear. Study authors only reported that the sequence was "determined by randomization (numbered and sealed allocation)"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Clinicians and investigators were blinded to the nature of the membrane
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Clinicians and investigators were blinded to the nature of the membrane
Incomplete outcome data (attrition bias) All outcomes	Low risk	Rate of missing data was < 10%
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Low risk	No substantial differences between groups
Other bias: co‐interventions	Low risk	Co‐interventions were consistent amongst the two groups

van der Voort 2005.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: not reported This study included two separate trials pre‐dilution versus post‐dilution Nadroparin infused by prefilter versus heparin infused by prefilter and protamine infused by post filter
Participants	Country: Netherlands Setting: single centre ICU: yes Number (randomised/analysed): study 1 (20/16); study 2 (20/15) Median age, IQR (years): study 1 (75, 66 to 77); study 2 (68, 60 to 77) Sex (M/F): study 1 (11/9); study 2 (9/11) Median severity, IQR (APACHE 2): study 1 (27, 22 to 37); study 2 (31, 23 to 37) Cause of AKI (number) Study 1: sepsis (6); surgical (4); other (10) Study 2: sepsis (7); surgical (2); other (11) Inclusion criteria: mechanically ventilated, in ICU with AKI who had not been treated with CVVH or HD before Exclusion criteria: in need of a specific CVVH mode, for instance, because of active bleeding; with planned surgery within 96 hours
Interventions	Treatment group Study 1: one filter run in pre‐dilution CVVH Study 2: one filter run with systemic nadroparin (475 IU/hour by continuous infusion prefilter) Control group Study 1: one filter run in post‐dilution CVVH Study 2: one filter run with heparin (prefilter) and protamine (post‐filter)
Outcomes	Study 1 Circuit lifespan Costs of 1 week of CVVH in euros Study 2 Circuit lifespan
Notes	Funding sources: unclear We could not extract data from the first period (before crossing over)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding was not feasible due to the nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcomes were objective by nature
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were no drop‐outs
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement
Other bias: baseline imbalance	Unclear risk	Cross‐over trial; we could not assess baseline imbalance
Other bias: co‐interventions	Low risk	Standardised protocol in terms of dialysis machine, ultrafiltration rate, blood flow, filter, anticoagulation, substitution fluid, and catheter was used,

Wynckel 2004.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: not reported
Participants	Country: France Setting: single centre ICU: yes Number: 18 Mean age ± SD: 62.1 ± 16.7 years Sex (M/F): not reported Mean severity ± SD (SAPS 2): 59.5 ± 14.3 Cause of AKI (numbers): sepsis (4); surgical (5); other (9) Inclusion criteria: critically ill patients with SAPS II scores < 85 and who were suffering from AKI Exclusion criteria: not reported
Interventions	Treatment group CVVHD Control group Pre‐dilution CVVH Post‐dilution CVVH
Outcomes	Number of death was reported; however, the number in each group was not reported due to the cross‐over trial design
Notes	Funding sources: unclear We could not extract data from the first period (before crossing over).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to permit judgement
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	High risk	Impossible to conduct blinding due to the nature of the study intervention
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcomes were objective by nature
Incomplete outcome data (attrition bias) All outcomes	Low risk	No drop‐outs
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Unclear risk	Cross‐over trial; unable to assess baseline imbalance
Other bias: co‐interventions	Low risk	Study investigators predefined the treatment protocol for anti‐coagulation

Yin 2015.

Study characteristics
Methods	Publication type: full article Study design: cross‐over RCT Follow‐up period: not reported Duration of study: 1 March to 31 December 2013
Participants	Country: China Setting: single centre ICU: yes Number: 17 Mean age ± SD: 63.1 ± 6.93 years Sex (M/F): 15/2 Mean severity ± SD (APACHE 2): 18.53 ± 2.67  Cause of AKI (number): sepsis (8); other (9) Inclusion criteria: ICU aged ≥ 16 years; meet CKRT treatment indications and cannot respond anticoagulation with whole body; body weight 30 to 120 kg Exclusion criteria: suspected heparin‐induced thrombocytopenia exists; age < 16 years; need Indications for systemic anticoagulation; non‐compliance with CKRT indications; poor short‐term prognosis or need to be transferred; weight < 30 kg or > 120 kg
Interventions	Treatment group Each study subject was treated with 4 filters: 2 AN69 ST membrane filters (A) and 2 conventional AN69 membrane filters (B), divided into group Ⅰ according to the random number table method (according to A‐B‐A‐B sequential treatment) Control group Group Ⅱ (treatment in the order of B‐A‐B‐A)
Outcomes	Circuit lifespan Vascular access complications Number of death was reported; however, the number in each group was not reported due to the cross‐over study design
Notes	Funding sources: not reported We could not extract data from the first period (before crossing over)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number table was used
Allocation concealment (selection bias)	Unclear risk	Insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Clinicians and participants were blinded to the nature of the membrane
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcomes were objective by nature
Incomplete outcome data (attrition bias) All outcomes	Low risk	No drop‐outs
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement; study protocol was not available
Other bias: baseline imbalance	Unclear risk	Due to the cross‐over study design, we were unable to assess baseline imbalance
Other bias: co‐interventions	Low risk	Study investigators predefined the protocol for using anticoagulants

ACEi ‐ angiotensin‐converting enzyme inhibitor; AKI ‐ acute kidney injury; APACHE ‐ Acute Physiology And Chronic Health Evaluation; APPT ‐ activated partial thromboplastin time; BFR ‐ blood flow rate; BUN ‐ blood urea nitrogen; CKRT ‐ continuous kidney replacement therapy; CrCl ‐ creatinine clearance; CVVH ‐ continuous venovenous haemofiltration; CVVHD ‐ continuous venovenous haemodialysis; CVVHDF ‐ continuous venovenous haemodiafiltration; ESKD ‐ end‐stage kidney disease; HD ‐ haemodialysis; HIV ‐ human immunodeficiency virus; ICU ‐ intensive care unit; IQR ‐ interquartile range; KRT ‐ kidney replacement therapy; LMWH ‐ low molecular weight heparin; M/F ‐ male/female; MAP ‐ mean arterial pressure; PT ‐ prothrombin time; RCT ‐ randomised controlled trial; RIFLE ‐ Risk Injury Failure Loss ESKD; SAP ‐ simplified acute physiology; SBP ‐ systolic blood pressure; SCr ‐ serum creatinine; SD ‐ standard deviation; SIRS ‐ systemic immune response syndrome; SOFA ‐ sequential organ failure assessment; WCC ‐ white cell count

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abe 2010c	Wrong intervention: sustained HDF versus CVVHDF
Abe 2011d	Wrong intervention: sustained HDF versus CVVHDF
Albino 2014	Wrong intervention: prolonged intermittent KRT sessions (10 versus 6 hours)
Allen 2005	Wrong population: included participants without AKI
Anstey 2016	Wrong intervention: dilute citrate solutions (15 versus 18 mmol/L)
Atan 2014	Wrong intervention: high cut‐off versus standard HF
Atapour 2006	Wrong population: included participants without AKI
ATN 2005	Wrong population: included intermittent HD
Augustine 2004	Wrong intervention: CVVHD versus intermittent HD
Badawy 2013	Wrong intervention: extended daily dialysis versus CVVHDF
Baldwin 2007	Wrong intervention: CVVH and extended daily dialysis with filtration
CATHEDIA 2008	Wrong population: included intermittent HD
CONVINT 2014	Wrong population: included intermittent HD
Davenport 1993a	Wrong intervention: intermittent versus continuous HF
ENDoX 2013	Wrong population: included participants without AKI
Farkas 1992	Wrong population: included participants without AKI
Feliciani 2007	Wrong population: included intermittent HD
George 2011	Wrong intervention: CVVHDF versus PD
Haase 2007b	Wrong intervention: 9 hours of HF
Hakim 1994	Wrong intervention: biocompatible versus bioincompatible membrane
Hassan 2013	Wrong intervention: coupled plasma filtration and adsorption plus CVVH or CVVFH alone
Humes 2010	Wrong intervention: immunomodulatory membranes
ISRCTN01121161	Wrong intervention: compared two different heparinisation strategies
Jeffrey 1994	Wrong intervention: compared HF and HD for 30 minutes
Kielstein 2004	Wrong intervention: extended dialysis versus CVVH
Klouche 2007	Wrong population: included intermittent HD
Kumar 2004	Wrong intervention: extended daily dialysis versus continuous HD
Leslie 1996	Wrong intervention: diluted heparin with different injection points
Libetta 2013	Wrong population: participants with AKI who did not undergo CKRT
Mishra 2017	Wrong intervention: sustained low‐efficiency dialysis (SLED) versus CKRT
Nand 2010	Wrong intervention: CAVHDF versus CVVHDF
NCT03496935	Wrong population: included intermittent HD
Opatrny 2002b	Wrong intervention: pharmacological intervention
Oudemans‐van Straaten 2009	Wrong intervention: filtrate flow of 4 L/hour and converted to 2 L/hour after 60 min in group 1, and vice versa in group 2
Park 2016	Wrong intervention: conventional (40 mL/kg/hour) versus high (80 mL/kg/hour) doses of CVVHDF
Pedrini 2009	Wrong population: included intermittent HD
RENAL 2006	Wrong intervention: conventional (25 mL/kg/hour) versus high (40 mL/kg/hour) doses of CVVHDF
RESCUE 2012	Wrong population: not CKRT but SLED
Su 2017	Wrong population: included intermittent HD
Tolwani 2008	Wrong intervention: standard (20 mL/kg/hour) versus high‐dose (35 mL/kg/hour) CVVHDF
Uehlinger 2005	Wrong intervention: CVVHDF versus intermittent HD
Weidhase 2019	Wrong intervention: high‐flux dialyser versus high‐cut‐off dialyser
Zhang 2004a	Wrong intervention: CVVH versus high‐volume HF; only 8 hours of treatment

AKI ‐ acute kidney injury; CKRT ‐ continuous kidney replacement therapy; CAVHDF ‐ continuous arteriovenous haemodiafiltration; CVVHDF ‐ continuous venovenous haemodiafiltration; CVVH ‐ continuous venovenous haemofiltration; ICU ‐ intensive care unit; KRT ‐ kidney replacement therapy; HD ‐ haemodialysis; HDF ‐ haemodiafiltration; HF ‐ haemofiltration; PD ‐ peritoneal dialysis; SLED ‐ sustained low‐efficiency dialysis

Characteristics of studies awaiting classification [ordered by study ID]

ISRCTN02674550.

Methods	Registered information Study design: RCT Follow‐up period: not reported Duration of study: November in 2005 to December 2007
Participants	Country: France Setting: multicentre ICU: not reported Inclusion criteria: adults (≥ 18 years) requiring CKRT; weighing 30 to 120 kg Number: 30 (planned) Exclusion criteria: history of heparin antibodies or heparin‐induced thrombocytopenia; known hypersensitivity to any dialysis membrane; pregnancy; current enrolment in another trial which could impact the successful completion of this study; under guardianship; anticipated to require transportation outside the unit for diagnostic or therapeutic procedures in the coming first week; unconscious, for whom no relative or person of trust can give consent for treatment. In the absence of any relative or person of trust, the study participant in question cannot be included in the study
Interventions	Treatment Prismaflex filter (coated with heparin) without any addition of heparin Control group Not reported
Outcomes	Circuit lifespan
Notes	Funding sources: Gambro Industries No full text publication identified

CKRT ‐ continuous kidney replacement therapy

Characteristics of ongoing studies [ordered by study ID]

NCT01062984.

Study name	Continuous venovenous hemofiltration versus continuous venovenuous hemodialysis
Methods	Publication type: trial registration / protocol Study design: parallel RCT Follow‐up period: not reported Duration of study: start January 2009
Participants	Country: USA Setting: not reported Inclusion criteria: ≥ 18 years; able to sign consent (or surrogate); must have been referred to the inpatient nephrology consult service for evaluation of AKI; expected survival of at least 48 hours. Number: 20 Exclusion criteria: Hb < 8 g/dL; serum potassium ≥ 6.5 mEq/L; weight > 120 kg
Interventions	Treatment group CVVH using the NxStage Cartridge Express dialyzer (polyethersulfone membrane). BRF 250 mL/min, effluent flow rate of 35 mL/kg/hour Control group CVVHD using the NxStage Cartridge Express dialyzer (polyethersulfone membrane). BFR 250 mL/min; effluent flow rate 35 mL/kg/hour
Outcomes	None of the prespecified outcomes of interest were reported
Starting date	Jan 2009
Contact information	Jay Koyner, MD University of Chicago Chicago, Illinois, United States, 60637
Notes	Funding sources: NxStage Medical

NCT01779635.

Study name	Efficacy and safety of heparin‐grafted membrane for CRRT (CARROM)
Methods	Publication type: trial registration/protocol Study design: cross‐over RCT Follow‐up period: not reported
Participants	Country: Singapore Setting: ICU Inclusion criteria: ≥ 21 years admitted to ICU or CCU and requiring CKRT for AKI or ESKD patients at moderate bleeding risk Number: 20 Exclusion criteria: patients at very high bleeding risk (for which they should also fall outside of the below inclusion criteria) Heparin‐induced thrombocytopenia or allergic to heparin Other medical conditions for which heparin is contraindicated Require systemic anticoagulation for medical indications (We will accept patients who are on prophylactic doses of anticoagulation for DVT prophylaxis) Pregnancy Decline to consent
Interventions	Treatment group Heparin‐grafted AN69 membrane (oXiris, Gambro) Control group Conventional AN69 membrane (M150, Gambro)
Outcomes	None of the prespecified outcomes of interest were reported
Starting date	August 2013 to December 2016
Contact information	Horng‐Ruey Chua, MBBS National University Health System
Notes	Funding sources: unclear

NCT01790620.

Study name	Impact of CVVHD with adsorption capacity membranes in septic acute kidney injury
Methods	Publication type: trial registration/protocol Study design: parallel RCT Follow‐up period: not reported
Participants	Country: Spain Setting: mixture of ICU and non‐ICU Inclusion criteria: Less than 72 hours from ICU admission to inclusios Clinical diagnosis of severe sepsis or septic shock Correct therapeutic initial management of septic process Number: 110 Exclusion criteria Kidney failure with KRT Received previous CKRT or HD in the last 3 months Inclusion in other ongoing study
Interventions	Treatment group CVVHD Control group CVVH
Outcomes	Circuit lifespan 91 days death 90 days recovery of kidney function
Starting date	May 2013 to Oct 2017
Contact information	Joan Sabater Riera, MD  Hospital Universitari de Bellvitge
Notes	Funding sources: unclear

NCT03426943.

Study name	Endotoxins and cytokines removal during continuous hemofiltration with oXiris™ (ECRO)
Methods	Publication type: trial registration/protocol Study design: parallel RCT Follow‐up period: not reported
Participants	Country: France Setting: ICU Inclusion criteria:  ≥ 18 years admitted to ICU with the diagnosis of secondary to a community‐acquired or a nosocomial peritonitis; AKI KDIGO ≥ stage 2 or another indication for KRT Number: 40 Exclusion criteria: inability to obtain informed consent from the study subject or next of kin; actual participation in another interventional study, contraindications to citrate; allergy to heparin; pregnant or breastfeeding woman; neutropenia < 0.5 G/L resulting from chemotherapy or other iatrogenic causes; receiving immunosuppressive therapy; long‐term corticosteroids; therapeutic antibodies, chemotherapy in the last 6 months (whatever the dose); innate or acquired immune deficiency; onco‐haematological disease (lymphoma, leukaemia, myeloma) treated within the last 5 years (but inclusion of a participant with solid cancer who did not receive chemotherapy during the past 6 months is possible); expected ICU length of stay < 48 hours; limitation of active care was pronounced at the time of enrolment; no social security insurance, with restricted liberty, or under legal protection
Interventions	Treatment group CVVH using oXiris™ (AN69ST) filter Control group CVVH using PrismafleX HF1400 (standard) filter
Outcomes	Survival at day 30
Starting date	December 21, 2018
Contact information	Thomas RIMMELE, MD, PhD Hospices Civils de Lyon
Notes	Funding source: Baxter Healthcare Corporation, Hospices Civils de Lyon

Weidhase 2020.

Study name	Myoglobin clearance in CVVHD and CVVHDF with regional citrate anticoagulation
Methods	Publication type: trial registration/protocol Study design: parallel RCT Follow‐up period: not reported Duration of study: not reported
Participants	Country: Germany Setting: Single centre Inclusion criteria: critical ill with indication for KRT Number: 70 Exclusion criteria: indication for therapeutic anticoagulation for different reasons; < 18 years; high risk of citrate accumulation (e.g. liver dysfunction)
Interventions	Treatment group Dialysis using CVVHDF CiCa Ultraflux AV 1000S Control group Dialysis using CVVHD CiCa Ultraflux EMiC2
Outcomes	None of the prespecified outcomes of interest were reported
Starting date	2017/05/02
Contact information	Universitätsklinikum Leipzig AöR Interdisziplinäre Internistische Intensivmedizin Leiter Prof. Dr. med. Sirak Petros Dr. med. Lorenz Weidhase Liebigstraße 20 04103 Leipzig Germany
Notes	Funding: Fresenius Medical Care Deutschland GmbH

AKI ‐ acute kidney injury; BFR ‐ blood flow rate; CCU ‐ coronary care unit; CKRT ‐ continuous kidney replacement therapy; CVVHD ‐ continuous venovenous haemodialysis; CVVHDF ‐ continuous venovenous haemodiafiltration; CVVH ‐ continuous venovenous haemofiltration; ESKD ‐ end‐stage kidney disease; DVT ‐ deep vein thrombosis; Hb ‐ haemoglobin; ICU ‐ intensive care unit

Differences between protocol and review

We planned to include only participants with AKI but we decided to expand the scope to include all participants who received CKRT. This decision was derived from the fact that some studies did not report the reasons behind administering CKRT.

We included the following comparisons that were not predefined in the published Cochrane protocol.

1. Surface‐modified versus standard double‐lumen catheter with identical geometry and flow design

2. Flat plate versus a hollow fibre filter of the same membrane type

3. Filter with a larger membrane surface area versus a smaller one

4. Filter with more and shorter hollow fibres versus a standard filter.

We planned to include studies that allocated anticoagulant or sedative agents to both intervention and control arms according to a predefined protocol; however, the majority of the included studies did not report the administration procedures of these treatment options. We decided to include these studies in our review but judged their risk of 'other bias' regarding co‐interventions as per the Cochrane Collaboration's tool for assessing the risk of bias.

We did not consider 'unit of analysis' issues regarding multiple measurements of outcome measures in our published protocol. We added our approach on handling these issues in the current full review (Unit of analysis issues).

Contributions of authors

• Drafting the protocol: SM, HT, YT, HS, HY, YK, TF

• Study selection: YT, SM, HT, HS, HY, YK, TF

• Extracting data from studies: YT, SM, HT, HS, HY, YK

• Entering data into RevMan: YT

• Carrying out the analysis: YT

• Interpreting the analysis: YT, SM, HT, HS, HY, YK, TF

• Drafting the final review: YT, SM, HT, HS, HY, YK, TF

• Disagreement resolution: TF

• Updating the review: YT, SM, HT, HS, HY, YK, TF

Sources of support

Internal sources

• Kyoritsu Hospital, Japan

  Salary to YT

• Sumitomo Hospital, Japan

  Salary to SM

• Hyogo Prefectural Amagasaki General Medical Center, Japan

  Salary to HS and YK

• Kameda Medical Center, Japan

  Salary to HY

• Australian and New Zealand Intensive Care Research Centre, Department of Epidemiology and Preventive Medicine, Monash University, Australia

  Salary to TF in 2019

• Jikei University Hospital, Jikei University School of Medicine, Japan

  Salary to TF in 2020

External sources

• No sources of support provided

Declarations of interest

• Yasushi Tsujimoto has declared that they have no conflict of interest

• Sho Miki has declared that they have no conflict of interest

• Hiraku Tsujimoto has declared that they have no conflict of interest

• Hiroki Shimada has declared that they have no conflict of interest

• Hideto Yasuda has declared that they have no conflict of interest

• Yuki Kataoka has declared that they have no conflict of interest

• Tomoko Fujii: supported by Japan Society for the Promotion of Science (JSPS) and has received a grant from JSPS. JSPS played no role in the review design, collection, management, analysis, and interpretation of data, writing of the review or the decision to submit the review for editorial approval.

Edited (no change to conclusions)