Comparison of the novel Maglumi ferritin immunoluminometric assay with Beckman Coulter DxI 800 ferritin

Anemia is one of the most common disorders around the globe, affecting approximately 1.62 billion people, thus corresponding to ~25% of the worldwide population [1]. Among the various causes of anemia, iron deficiency is present in approximately half of all cases. The diagnosis of iron deficiency anemia requires laboratory-confirmed anemia combined with the evidence of low iron stores. Since iron stores in the body primarily exist in form of ferritin, recent guidelines and indications suggest that the measurement of serum ferritin should be regarded as the most accurate test for diagnosing iron deficiency anemia [2, 3]. According to the World Health Organization (WHO), iron deficiency anemia is diagnosed in the presence of a serum ferritin concentration <12 ng/mL in both genders [4]. Beside iron deficiency, the measurement of serum ferritin is also essential for identifying iron overload, which is usually attributable to a number of inherited (e.g. hemochromatosis) or acquired conditions (e.g. obesity, inflammation, excessive alcohol consumption) [5, 6]. The WHO currently defines a severe risk of iron overload when the serum ferritin value exceeds 150 ng/mL in women and 200 ng/mL in men, respectively [4]. Since the clinical value of serum ferritin assessment is now indisputable, the aim of this study was to compare the newly commercialized Maglumi ferritin immunoluminometric assay with a validated commercial and fully automated technique.

The Maglumi ferritin sandwich immunoluminometric assay (Shenzhen New Industries Biomedical Engineering Co. [SNIBE], Shenzhen, China) has been developed to be used on Maglumi fully automated analyzer series, using calibrators standardized against the WHO 3rd International Standard 94/572. Briefly, the method is based on an anti-ferritin monoclonal antibody labeled with Amino-Butyl-Ethyl-Isoluminol (ABEI), a second monoclonal antibody labeled with Fluorescein Isothiocyanate (FITC), and nano magnetic microbeads coated with anti-FITC antibodies. A sample volume of 40 μL is incubated with the reagents and mixed thoroughly. After development of the immunocomplex, the starter reagents are added, a chemiluminescent reaction is started and the light signal is detected as relative light units (RLUs). The recorded optical density (OD) is proportional to the ferritin concentration present in test samples. According to manufacturers’ specifications, the functional sensitivity of this assay is 1.3 ng/mL, whereas the reference range is comprised between 13 and 232 ng/mL in women and between 25 and 350 ng/mL in men, respectively. The total imprecision is comprised between 5.4% and 9.4% in the range of values comprised between 44–680 ng/mL, and an optimal linearity has been tested between 2–3000 ng/mL.

The comparison study originally included 100 consecutive inpatient serum samples collected in evacuated blood tubes containing no additives (Becton-Dickinson, Oxford, UK), referred to the local laboratory for routine ferritin measurement, and displaying values in the clinically significant range of serum ferritin concentration (i.e. 2–1500 ng/mL). After centrifugation at 1500 g for 15 min at room temperature, the serum was separated, divided in two aliquots and immediately analyzed on both Beckman Coulter DxI 800 (Beckman Coulter Inc., Brea, CA, USA) and Maglumi 2000. The DxI ferritin, a quantitative chemiluminescent immunoassay standardized to the WHO 3rd International Standard 94/572, is used for routine ferritin assessment in the local laboratory, and displays a functional sensitivity of 0.2 ng/mL, an optimal linearity between 0.2–1500 ng/mL and a total imprecision comprised between 3.0–8.0% in the range of values comprised between 25 and 374 ng/mL [7]. According to manufacturers’ specifications, the reference range is comprised between 11–307 ng/mL in women and between 24–336 ng/mL in men, respectively. Both methods were calibrated using proprietary calibrating materials.

The correlation between methods was assessed with Deming fit and Spearman’s correlation, whereas the mean bias and its 95% confidence interval (CI) were estimated with Bland-Altman difference plot. The concordance between Maglumi and DxI 800 at the lower and upper limit of the respective reference ranges was calculated as percentage agreement and κ coefficient. The statistical analysis was performed with Analyse-it (Analyse-it Software Ltd., Leeds, UK). The comparison study was based on pre-existing inpatient serum samples referred for routine ferritin testing on DxI 800, and the material was obtained after analysis was completed. All samples were anonymized prior to Maglumi ferritin assessment, so that no patient informed consent was necessary. The study was carried out in accordance with the Declaration of Helsinki and was approved by the local institutional review Board.

Five out of the 100 samples originally collected ought to be excluded since their ferritin values exceeded the analytical range of DxI (i.e. >7500 ng/mL). Therefore, 95 serum samples were finally used for the statistical analysis, five of which had values above the linearity range of DxI (i.e. >1500 ng/mL, but <7500 ng/mL) and were hence measured after internal instrument dilution. An excellent Spearman’s correlation was observed between Maglumi and DxI (r=0.997; 95% CI, 0.995–0.998; p<0.001). The equation of Deming fit was: [Maglumi]=1.12×[DxI]+4 (Figure 1). The mean bias, as calculated with Bland-Altman plot analysis, was 34 ng/mL (95% CI, 10–57 ng/mL) (Figure 2). However, a much lower bias was observed after limiting the analysis to samples with DxI ferritin values <150 ng/mL in women and <200 ng/mL in men (n=61; mean bias, 10 ng/mL and 95% CI, 8–13 ng/mL), which implies that the overall bias was highly dependent upon discrepancy observed for values greater than the diagnostic thresholds for diagnosing iron overload in both genders. The strength of agreement between values obtained with Maglumi and DxI was 98% (κ, 0.94; 95% CI, 0.86–1.00; p<0.001) at the lower limit of the respective gender-specific reference range, 100% (κ, 1.00; 95% CI, 1.00–1.00; p<0.001) at the upper limit of the respective gender-specific reference range, 98% (κ, 0.93; 95% CI, 0.84–1.00; p<0.001) for diagnosing iron deficiency (i.e. serum ferritin <12 ng/mL), and 94% (κ, 0.86; 95% CI, 0.75–0.97; p<0.001) for diagnosing iron overload in both genders (i.e. serum ferritin >150 ng/mL in women and >200 ng/mL in men, respectively) (Table 1).

Figure 1:

Deming fit of SNIBE Maglumi ferritin compared with Beckman Coulter DxI 800 ferritin.

Figure 2:

Bland-Altman plot analysis of SNIBE Maglumi ferritin and Beckman Coulter DxI 800 ferritin.

The demand for routine measurement of ferritin is constantly increasing in clinical laboratories, due to the crucial role that this test plays for diagnosing iron deficiency anemia and iron overload [2–6]. Therefore, the availability of new fully-automated immunoassays that would permit the optimal integration of ferritin measurement within chemistry and immunochemistry systems should be regarded as a valuable perspective for clinical laboratories and healthcare systems plagued by increasingly shortage of funding. The results of our study attest that the results obtained with the novel Maglumi ferritin immunoluminometric assay are comparable to those of a validated immunoassay such as Beckman Coulter DxI 800. This conclusion is mainly supported by excellent correlation (i.e. r=0.997) and regression data (i.e. slope, 1.12; intercept 4 ng/mL), and especially by the high agreement at instrument-specific thresholds and WHO cut-offs for diagnosing iron deficiency and overload (i.e. between 94% and 100%). This last aspect is particularly interesting, underpinning that standardization of ferritin methods against the WHO 3rd International Standard 94/572 may be effective to achieve a much better degree of inter-assay harmonization, despite the persistence of a still meaningful bias (i.e. 34 ng/mL), which is however much limited at lower values (i.e. 10 ng/mL). In conclusion, Maglumi ferritin immunoassay may be regarded as a suitable alternative for routine and fully-automated assessment of this biomarker in clinical laboratories.