Cannula Position Makes or Breaks Carbon Dioxide Removal in High-Flow ECCO2R

For patients drowning in their own carbon dioxide, an artificial lung outside the body can be a lifeline. Extracorporeal carbon dioxide removal, or ECCO2R, draws blood out of a large vein, strips CO2 across a membrane oxygenator, and returns it to the circulation, easing the burden on failing lungs and allowing gentler mechanical ventilation. Yet the technology has struggled to prove itself in large clinical trials, and one of the most persistent suspicions has been that many devices simply do not remove enough CO2 to matter. A new study from a French intensive care unit now points to a surprisingly simple culprit: where exactly the draining cannula sits inside the venous system.

The research, conducted at Lapeyronie University Hospital in Montpellier and published in the Journal of Artificial Organs, analyzed 17 patients treated with high-flow ECCO2R between January 2018 and May 2024. All patients received extracorporeal blood flows of at least 2 liters per minute, delivered through an ECMO console with a large membrane oxygenator and two single-lumen cannulas inserted on the same side of the body. Because the drainage and reinfusion cannulas are separate, the team could deliberately vary the position of the draining cannula, creating four distinct configurations: central and distal versions of both a jugular-jugular and a femoro-femoral setup. In the central configurations, the multistage draining cannula was advanced through the right atrium, while in the distal configurations it remained close to the reinfusion cannula in the distal portions of the vena cava.

The physiological variable at the heart of the study is recirculation, a phenomenon familiar to anyone who works with extracorporeal circuits. When freshly reinfused blood is immediately sucked back into the draining cannula, the machine processes the same blood twice while the patient’s CO2-rich venous return flows past untouched. The team quantified this with a recirculation coefficient calculated from oxygen saturations in the inlet and outlet lines of the oxygenator, using an assumed mixed venous saturation of 65 percent, and validated the approach with sensitivity analyses across a range of plausible values. They also computed the effective extracorporeal blood flow, the portion of the pump flow that actually represents new blood from the patient, and measured CO2 elimination across the membrane in standard-temperature-and-pressure-dried milliliters per minute.

The results were stark. In the central drainage configurations, the median recirculation coefficient was just 9 percent, whereas the distal configurations produced a median of 59 percent, more than six times higher. Effective blood flow collapsed accordingly: about 2.1 liters per minute in the central setups versus 0.9 liters per minute in the distal ones, with the distal femoro-femoral configuration falling to a median of only 0.5 liters per minute. Because CO2 transfer depends on both how much blood crosses the membrane and how much CO2 that blood carries, this hidden shunting translated directly into lost decarboxylation. Median CO2 elimination was 122 milliliters per minute with central drainage but only 58 milliliters per minute with distal drainage, despite comparable nominal pump flows between the groups.

The statistical analysis went beyond simple comparison. Using multivariable linear mixed models that accounted for repeated measurements in each patient, the researchers examined how much CO2 elimination increased for each unit of the decarboxylation index, a product of blood flow and sweep gas flow that they had previously validated as a predictor of performance. For every additional unit of that index, CO2 elimination rose 5.2 times less under a distal drainage configuration than under a central one. Compared with the central jugular-jugular reference, the distal jugular-jugular setup lost 3.2 milliliters per minute of CO2 elimination per unit of index, and the distal femoro-femoral setup lost 5.2 milliliters per minute, both highly significant. The central femoro-femoral configuration performed nearly as well as the jugular reference, a difference that was not statistically significant.

The mechanics behind this failure of distal drainage are rooted in fluid dynamics and venous anatomy. In the distal portions of the superior or inferior vena cava, the available blood volume is small, the side holes of the cannula can be occluded by the venous wall, and the negative pressures generated by spontaneous breathing further reduce the suction gradient. As a result, the proportion of drainage drawn from freshly reinfused blood rapidly becomes dominant, and recirculation soars. High recirculation also lowers the CO2 content of the blood entering the oxygenator, and because diffusion across the membrane follows Fick’s law, a lower inlet CO2 concentration means less driving force for transfer. The machine appears to be working hard while accomplishing progressively less.

These findings carry an uncomfortable implication for standard practice. The authors note that their distal drainage configurations behave much like the single dual-lumen cannulas used in most ECCO2R programs worldwide, in which drainage and reinfusion ports coexist within one catheter. If distal drainage inherently produces high recirculation, the modest decarboxylation performance that has dogged dual-lumen ECCO2R may be a structural feature rather than a fixable flaw. Recent trials have suggested that only veno-venous systems running high blood flows, generally above 2 liters per minute, can meaningfully manage severe respiratory acidosis or reduce ventilator settings, and this study shows that even a nominally high-flow circuit can be quietly sabotaged by cannula geometry. No CO2 elimination value in the distal group reached 110 milliliters per minute.

The study also offers practical guidance for clinicians. The central jugular-jugular configuration, which routes drainage from the inferior vena cava through the right atrium, delivered the best rheological conditions but requires sedation, ultrasound guidance to advance the cannula across the atrium, and head positioning that is uncomfortable for awake patients. The central femoro-femoral configuration emerged as a compelling compromise, providing high and stable flows with acceptable recirculation even in seated, spontaneously breathing patients, thanks to wire-reinforced cannulas and proximal placement near the right atrium. The authors suggest it is particularly suited to awake patients undergoing respiratory weaning with non-invasive ventilation. A further advantage of the two-cannula approach is upgradeability: in case of refractory hypoxemia, the same access sites can be reconfigured into full veno-venous ECMO without new punctures.

Safety data temper the enthusiasm. Nine of the 17 patients, 53 percent, developed one or more thromboembolic events, all of them cannula-associated deep vein thromboses that occurred exclusively in the central drainage group and were detected by systematic CT scanning after decannulation. The authors suspect their anticoagulation targets were too lenient, particularly during the decannulation phase, when endothelial injury from catheter removal activates the coagulation cascade, and they point to percutaneous closure devices as a possible mitigation. The study has other limitations that the authors acknowledge candidly: it is a single-center retrospective analysis with a small and imbalanced cohort, only the first 72 hours of therapy were analyzed to avoid membrane fouling effects, and the recirculation calculation necessarily relies on an assumed mixed venous saturation. Clinical outcomes such as ventilation duration were not assessed. Still, the internal consistency of the findings, replicated across sensitivity analyses and mixed models, makes the central message hard to dismiss. In extracorporeal CO2 removal, the dose delivered to the patient is not set by the pump alone; it is set by the plumbing, and a cannula placed a few centimeters too distal can halve the therapy before it begins.

Subject of Research: Effect of draining cannula position on recirculation and CO2 elimination in high-flow extracorporeal carbon dioxide removal

Article Title: Impact of draining cannula position and configuration under high-flow extracorporeal CO2 removal (ECCO2R) on recirculation and extracorporeal decarboxylation

Article References: Charbit, J., Muller, C., Girard, M., Courvalin, E., Dagod, G., Bouchdoug, K., Deras, P., Weber, H., & Capdevila, X. (2026). Impact of draining cannula position and configuration under high-flow extracorporeal CO2 removal (ECCO2R) on recirculation and extracorporeal decarboxylation. Journal of Artificial Organs, 29(4), Article 69. https://doi.org/10.1007/s10047-026-01582-7

Image Credits: AI Generated

DOI: 10.1007/s10047-026-01582-7

Keywords: ECCO2R, extracorporeal CO2 removal, recirculation, cannula position, ECMO, decarboxylation, respiratory failure, ARDS, COPD, membrane oxygenator, veno-venous support, intensive care