Measurement of glycosylated ferritin with Concanavalin A: Assay design, optimization and validation

Introduction: Ferritin is the major iron-storage glycoprotein found in all tissues. Ferritin glycosylation can be assessed by the differential affinities of ferritin glycoforms for Concanavalin A (ConA), a lectin. The fraction of serum ferritin bound to ConA is called "glycosylated ferritin" (GF). Low GF reflects macrophagic activation and is an essential biomarker used in adult-onset Still's disease (AOSD), macrophage activation syndrome (MAS) and Gaucher disease diagnosis and therapeutic management. To date, no complete assay description and method validation according to the ISO 15189 standard has been published. This study aimed to describe and validate our method used for GF measurement and describe GF values observed in patients.Materials and methods: Ferritin glycoforms were separated based on their affinities for ConA using commercially available TRIS-barbital buffer, Sepharose and ConA/Sepharose 4B gels. Ferritin concentrations were measured on the Siemens Dimension Vista 1500 (R). We analysed 16,843 GF values obtained between 2000 and 2021 from our database of patients.Results: Optimal separation of ferritin glycoforms was obtained by 15-min incubation of serum with ConA/ Sepharose at pH 8. The optimized volume were 0.4 mL for total serum ferritin (TSF) 30-1000 mu g/L and 0.5 mL for TSF 1000-2500 mu g/L. Serum with higher TSF should be pre-diluted in the TRIS-barbital buffer. Reproducibility of ferritin measurement in the TRIS-barbital buffer matrix was excellent (intra-assay CV < 1%; inter-assay CV < 4%). Reproducibility of GF assay was good (intra-assay CV < 10% for low and high ferritin samples, respectively; and inter-assay CV < 10%). Inter-operator variability was 21.6% for GF < 20%. Ferritin was stable for up to 3 days in the TRIS-barbital buffer. An inter-laboratory exchange program conducted with another French hospital showed good agreement between results. In our database, <20% GF levels were scarce, compatible with the low prevalence of Still's disease, MAS, and Gaucher disease. The 95% confidence interval for GF was [26-58]%, lower than values described in the literature for healthy individuals.Conclusion: Thanks to good performances, this technique can become readily available for laboratories servicing patients with AOSD, MAS (including severe COVID-19 patients) and Gaucher disease patients.