Combining renal replacement therapy (RRT) with extracorporeal membrane oxygenation (ECMO) creates a practical question for perfusion and critical care teams: should the therapies use separate vascular access, or should renal support connect directly to the ECMO circuit? A scoping review by Eden Kharrouba and colleagues maps the comparative evidence and finds no consistently superior configuration. Its central contribution is identifying what has been studied, where technical differences matter, and why stronger comparisons are still needed.
What the review examined
The investigators searched PubMed, the Cochrane Library, Web of Science, Embase, and ClinicalTrials.gov, with a final search on February 9, 2026. The review followed JBI and PRISMA-ScR methods, used a registered protocol, and involved independent screening by two reviewers. Eligible reports compared at least two ECMO–RRT approaches and linked those configurations to reported outcomes.
From 6,806 retrieved records, 20 studies were included: 12 clinical studies, three laboratory studies, one computational study, and four practice surveys. Only one clinical study was a randomized controlled trial. Two clinical studies involved pediatric patients, and adult studies enrolled between 12 and 170 patients. Differences in populations, equipment, anticoagulation, and outcome definitions limited direct comparison.
Parallel and integrated support are not single uniform strategies
In a parallel arrangement, ECMO and RRT run through separate vascular access. This separates the devices operationally but adds another access site and its associated management burden. Integrated approaches connect an RRT device or an in-line hemofilter to the ECMO circuit. They can avoid a separate dialysis catheter, but expose the renal support circuit to pressures that depend on the exact connection points.
The review describes numerous connection schemes involving locations before or after the pump and oxygenator. Grouping all integrated circuits together can therefore conceal important engineering differences. A study of one inlet-and-return arrangement cannot automatically establish the performance of another. The authors emphasize clear reporting of the actual configuration, rather than relying only on the labels “integrated” and “parallel.”
Filter life: promising observations, inconsistent findings
Some observational studies reported longer filter life with integration, while others found comparable filter performance. One retrospective comparison reported mean filter life of 47 ± 14 hours with integrated support versus 12 ± 3 hours with parallel support. However, differences in anticoagulation and treatment delivery make it difficult to isolate the effect of the connection itself.
The included 80-patient randomized trial did not demonstrate a significant difference in its primary filter-lifespan outcome. The review reports median filter life of approximately 70 hours in both groups in the as-treated comparison, with p = 0.39. This finding tempers the assumption that integrating the therapies necessarily prolongs filter life. It also illustrates why favorable results from individual observational cohorts should not be treated as a universal advantage.
Delivered therapy matters as much as prescribed therapy
Across the included evidence, no consistent advantage emerged for solute clearance or ultrafiltration efficiency. Some reports found similar creatinine changes and fluid-removal performance. Another reported greater blood urea nitrogen reduction with integration, but did not adequately report the prescribed CRRT dose, leaving an important alternative explanation for the difference.
One pediatric comparison favored an automated CRRT device over a free-flow in-line hemofilter for ultrafiltration accuracy. That result concerns a specific equipment comparison and should not be generalized to every integrated configuration. More broadly, the review highlights uncertainty about whether the intended prescription and actual delivered treatment remain aligned when the renal device is exposed to unusual access and return pressures.
Pressure alarms, complications, and mortality
Pressure interactions are among the clearest technical differences between configurations. Depending on the connection sites, integrated CRRT can encounter strongly positive or negative pressures, influencing alarms, monitoring, and treatment interruptions. Surveys described persistent pressure alarms and changes in configuration when those alarms could not be resolved. These reports demonstrate a workflow problem, but do not establish which arrangement is safest for every device combination.
The clinical literature reported bleeding, thromboembolic events, hemolysis, bloodstream infection, and air embolism without consistently identifying a safer configuration. Small cohorts, inconsistent definitions, and differences in anticoagulation complicate interpretation. Survey reports of an event are also different from prospectively measured complication rates; perceived hazards should not be presented as quantified comparative risks.
Similarly, the review found no significant short-term mortality difference in the cited comparisons of parallel and integrated approaches. This is not proof that the approaches are equivalent. The available studies may be too small or too confounded to detect meaningful differences, and their mixed ECMO populations make pooled clinical conclusions especially uncertain.
Implications for perfusion practice and research
For perfusion teams, the review supports careful documentation of the full ECMO–RRT interface: device models, connection sites, pressures, flows, renal-support modality, and anticoagulation strategy. It also identifies useful quality measures beyond survival, including delivered versus prescribed dose, filter life, treatment downtime, alarm burden, workload, and cost. These are areas for multidisciplinary review rather than a prescribed circuit recipe.
Pediatric applications deserve separate attention because extracorporeal volume and additional vascular access have different practical consequences in smaller patients. The authors also call for laboratory and computational comparisons that examine blood trauma, embolic hazards, and treatment performance before translating findings into clinical trials. Future clinical studies should account for why a configuration was selected and document any crossover between approaches.
Limitations and interpretation
This is an evidence-mapping review, not a pooled demonstration of treatment effectiveness. Requiring comparative configurations excluded many single-configuration reports, while heterogeneity and methodological weaknesses prevented definitive conclusions. Several descriptions and reference details in the publication are internally inconsistent, reinforcing the value of consulting individual source studies for precise technical or numerical claims. The review is most useful as a map of knowledge gaps and a framework for better research.
The authors reported no funding and no competing interests. The practical conclusion is that current comparative evidence does not support choosing a universal ECMO–RRT configuration on the basis of proven outcome superiority. Local device compatibility, approved protocols, and multidisciplinary expertise remain essential while more rigorous evidence develops.
Source: Kharrouba E, Moyon Q, Cove ME, Arens J, Martins Costa A, Fraser JF, Suen JY. Comparison of the approaches to combine Renal Replacement Therapy with Extracorporeal Membrane Oxygenation – a Scoping Review. Annals of Intensive Care. 2026;16:100165. doi:10.1016/j.aicoj.2026.100165.
Attribution: © 2026 Published by Elsevier Masson SAS on behalf of Société de Réanimation de Langue Française. The source article is available under CC BY-NC-ND 4.0. This independently written IPA educational summary is separate from the unchanged source PDF. Cover artwork is conceptual, not a circuit-connection guide.
