Hemoadsorption during VA-ECMO: results from the international prospective COSMOS registry

Background/Introduction

The International COSMOS Registry prospectively collects real-world data of hemoadsorption (HA) use in critically ill cardiac patients. In VA-ECMO patients, HA may reduce systemic inflammation, but its clinical benefit remains uncertain.

Purpose

To analyze the use of HA in VA-ECMO patients, focusing on timing, clinical effects, and safety.

Methods

The Registry protocol was approved by institutional Ethics Committees and all participants provided informed consent. Data were prospectively collected at screening, 24 hours pre and post HA, ICU discharge, and 90-day follow-up. Analyses followed a predefined statistical analysis plan. Results are reported as mean ± SD or median [IQR]. HA initiation relative to ECMO start was stratified into early (≤24 hrs) and late (>24 hrs).

Results

50 patients (mean age 57±13 years; 26% female) with median APACHE II score 25 [19, 31] from 10 centers (Germany, Portugal, Spain) were included. VA-ECMO indications were cardiogenic shock (62%), post-cardiotomy shock (5%), refractory ventricular tachycardia (8%), and periprocedural complications (11%); 57% had cardiac arrest before ECMO. HA was integrated into renal replacement therapy (80%), used as standalone hemoperfusion (12%), or connected to an ECMO bypass loop (8%). Median devices per patient were 4 [2, 7]; 30% received ≥6 devices, with 15 [9, 24] hours per device.

Twenty-four hours after HA, interleukin-6 decreased (524 [187, 4,877] to 83 [49, 224] pg/mL; p<0.0001), norepinephrine dose declined (0.13 [0.06, 0.29] to 0.04 [0, 0.14] µg/kg/min; p=0.0005), and fluid balance improved (+875 [-246, 2,549] to +190 [-849, 1,108] mL/day; p=0.011). Platelet counts were unchanged (p=0.058). SOFA scores improved from 14.5 [12, 16] to 12.5 [11, 15] (p=0.032). ICU mortality was 34% and 90-day mortality 44.7%.

Early (n=27) versus late (n=23) HA initiation was associated with shorter ECMO duration (7 [4, 11] vs 14 [10, 29] days; p<0.001). Among survivors, early HA was linked to shorter ICU stay (23 [18, 26] vs 35 [25, 57] days; p=0.024) and hospital stay (23 [19, 40] vs 49 [28, 68] days; p=0.017). No HA-related adverse events were reported.

Conclusion

In VA-ECMO patients, concomitant hemoadsorption was associated with reduced vasopressor needs, lower inflammatory markers, and improved organ dysfunction within 24 hours. Early initiation (≤24 hrs) was linked to shorter ECMO support and faster recovery. Controlled prospective studies are needed to confirm these findings.Key parameter relative percentage changeFor image description, please refer to the figure legend and surrounding text.