Extracorporeal life support (ECLS), formerly known as extracorporeal membrane oxygenation (ECMO), is increasingly used for patients with severely impaired cardiac and/or pulmonary function. Risks associated with ECLS include clot formation in the system, since contact of the blood with the ECLS system results in activation of coagulation. Despite intensive anticoagulation therapy, prevention of clot formation in the system remains a challenge. Clot formation as well as shear stress and pressure variations in the ECLS system are associated with flow limitation, risk for embolism, and complications related to hemolysis. The risk for clot formation is further increased as hemolysis can result in oxidative stress, activation of the immune system, and activation of platelets. Monitoring of D-dimers and free hemoglobin (fHb) levels is therefore indicated during ECLS. Conventional methods for fHb analysis are laborious and not readily available in many laboratories. We evaluated the suitability of the hemolysis-index (H-index), an internal quality control parameter of clinical chemistry platforms, as a clinical parameter for ECLS patients.

Our research (1) shows that the H-index (of the Roche Cobas platform) significantly correlates with conventionally analyzed fHb (r = 0.99, P < 0.01) and shows acceptable analytical performance with a CV of 1.3% for control material. In our retrospective study of 56 ECLS patients, 19.6% of the ECLS patients had an H-index above 20 (a level deemed to reflect clinically relevant hemolysis) in at least two consecutive blood draws. Since preanalytical (ex vivo) factors can also result in hemolysis, it is essential to analyze trends in fHb rather than single elevations. In two patients with clot formation in the pumphead of the ECLS system, the H-index peaked above 100. Visible clots at other locations in the ECLS system did not always correlate with H-index elevations. The duration of ECLS and the type of ECLS (venovenous vs. venoarterial) did not significantly affect the H-index level. Interestingly, H-index peaks (H-index ≥20, repeatedly) were more prevalent in patients that died during ECLS support (91.0% vs. 57.1% in the total study population, P=0.016).

Our studies demonstrated that the routinely measured H-index is a suitable, readily available (24/7) and cost-efficient alternative for the conventional fHb analysis with acceptable analytic performance. Since the H-index can easily be measured multiple times a day, the H-index helps in the early detection of hemolysis in patients with ECLS. The H-index has now been implemented in the routine care of these patients. To ensure early detection and follow-up of hemolysis, we would advise at least daily H-index monitoring with a consecutive 2nd confirmatory measurement on the same day when an H-index ≥20 is observed. In our study, a repeated H-index ≥20 was associated with increased mortality, although this level (4xURL) was arbitrarily chosen. The best threshold for the H-index serving as a warning and/or action signal will be determined in follow-up studies.

This study illustrates the power of interdisciplinary collaboration, where clinical chemistry can offer solutions for unmet clinical needs. Further studies have been initiated to establish cut-off values for free hemoglobin in ECLS patients and, in a broader context, to improve clinical chemistry testing in this complex patient group.


  1. Bosma M, Waanders F, Van Schaik HP, Van Loon D, Rigter S, Scholten E, Hackeng C. Automated and cost-efficient early detection of hemolysis in patients with extracorporeal life support: Use of the hemolysis-index of routine clinical chemistry platforms. J Crit Care 2019, 51:29-33.