Evaluation of Roche Hemolysis Index for Reporting of Plasma Hemoglobin, and Considerations Regarding Hemolysis Detection in Use of Ventricular-Assist Devices

Introduction/Objective Our cardiac intensive care unit (CICU) asked the laboratory to provide plasma hemoglobin testing to monitor effects of ventricular-assist devices (VAD). Roche hemolysis index (HI) has been reported to be suitable for this purpose. Use of HI as a reportable test, however, must be validated in-house as a laboratory- developed test (LDT), according to CLIA regulations. We describe results of our evaluation, and caveats for intended use. Methods/Case Report Concentrated hemolysate was produced by osmotic lysis of packed red blood cells in water. Hemolysis in plasma was by dilution of concentrated hemolysate. Verification of HI as a measure of hemoglobin (mg/dL) used the Sigma-Aldrich spectrophotometric hemoglobin assay as a gold standard. Linearity, linear range, sensitivity, were investigated by dilution measurements; interferences by admixture experiments; reproducibility (precision, intra- and inter-assay) in each case by replicate measurements (n = 20). Results (if a Case Study enter NA) HI was confirmed to correspond to hemoglobin in mg/dL. Linearity was between 2-1000 mg/dL (r2 >0.99). Intra-assay precisions were <=2.5% (hemoglobin = 74,148 mg/dL); inter-assay precisions were <=4.9% (hemoglobin = 69,139 mg/dL). HI variability was ±2 at very low values (0-10 mg/dL). There were no interferences observed among common therapeutic drugs. Hyperlipemia (evaluated up to lipemic index (LI) = 225) and hyperbilirubinemia (evaluated up to icteric index (II) = 7) demonstrated no significant interference. Conclusion Analytically, HI exhibited acceptable performance for reporting of plasma hemoglobin. Deployment would require establishment of quality control (QC) and proficiency testing procedures for HI. Clinically, a caveat for use is that HI cannot distinguish between hemolysis in circulation vs. that due to sample collection. For the CICU, HI >10 mg/dL is characteristic of more than 20% of samples drawn in general, irrespective of use of VAD. Temporal patterns of HI must therefore be carefully evaluated in parallel with haptoglobin measurements to assist in VAD management.