# Algorithm for Reporting Free Hemoglobin in ECMO Patients: Need for a Multidisciplinary Approach

## TL;DR

This study proposes an algorithm to improve the accuracy of free hemoglobin reporting in ECMO patients by integrating hemolysis indices and laboratory methods.

## Contribution

The study introduces a novel algorithm combining hemolysis indices and the Harboe method to enhance free hemoglobin reporting in ECMO patients.

## Key findings

- The Harboe method yielded significantly higher median free hemoglobin levels compared to the Abbott Alinity estimated method.
- A moderate correlation was found between the two methods, with a notable negative bias in the estimated method.
- An algorithm was developed to integrate hemolysis indices and bilirubin levels for improved reporting accuracy.

## Abstract

Background: Intravascular hemolysis is a common complication in patients undergoing extracorporeal membrane oxygenation (ECMO), with plasma free hemoglobin (pfHb) serving as a biomarker for detection. Without standardized protocols, laboratories face challenges in interpreting and reporting results. Hemolysis indices may enhance reporting accuracy. Methods: This retrospective observational study at University Hospital Center Zagreb included 61 lithium heparin plasma samples from ECMO patients. pfHb was measured using the Harboe method (fHb) and estimated from hemolysis indices on Abbott Alinity c analyzer (efHb). Total and conjugated bilirubin, hemolysis, icterus, and lipemia indices (HIL) were recorded. Method comparison used Passing-Bablok regression and Bland–Altman analysis. An algorithm for pfHb reporting accounting for HIL interferences was developed. Results: Significant differences were observed between methods, with Harboe yielding higher median fHb (261 mg/L) versus efHb (58 mg/L). Regression analysis showed constant negative bias of −91 mg/L (95% CI: −143 to −16) for efHb relative to fHb. Bland–Altman analysis demonstrated wide limits of agreement. Correlation between fHb and efHb was moderate (Spearman’s rho = 0.618, p < 0.001). The delta between methods increased with higher bilirubin concentrations. An algorithm integrating HIL indices with the Harboe method was developed to guide result validation and reporting. Conclusions: Accurate hemolysis assessment in ECMO patients requires careful interpretation, appropriate method selection, and laboratory–clinician collaboration. The proposed algorithm improves the clinical utility of pfHb testing by accounting for analytical interferences and supporting informed decision-making.