Comparison of six serum ferritin immunoassays and isoferritin spectrotypes in malignancies.

Serum ferritin was measured by six enzyme immunoassays in specimens from patients with digestive cancers (n = 30) and hematologic malignancies (n = 33). Most mean comparisons show significant differences in both groups of patients. In digestive cancers correlations between any two methods are very satisfactory (r > 0.99) but a proportional bias is often observed. In hematologic malignancies, correlations are bad (r < 0.80 in 8 out of 15 correlations) because of many discrepant values. Isoelectric focusing separation of isoferritins was performed in most specimens and the pattern of each serum was compared to the between kit CV. We conclude that an 'acid' spectrotype increases between-kit analytical variability. We try to explain the results taking into account the nature of the immunological systems and the cross-reactions with tissular isoferritins. In conclusion, our results indicate that large differences may be observed in sera from hematologic malignancies (leukemias, lymphomas ... ) We recommend that monitoring be achieved by the same method of measurement.

[Multicenter study of sialic acid deficient transferrin determined by two chromatographic techniques]. / Etude multicentrique du dosage de la transferrine déficiente en acide sialique par deux techniques chromatographiques.

Serum carbohydrate-deficient-transferrin (CDT) was measured by a micro anion-exchange chromatography/enzyme immunoassay. Results obtained on 245 sera analyzed in four laboratories were compared. Moreover, one laboratory used a commercial kit with ready-to-use microcolumns and a radioimmunoassay for measuring eluted CDT. Imprecision was judged to be satisfactory. Within-assay coefficients of variation ranged from 5 to 10%, between-assay coefficients of variation ranged from 9 to 18%. Between-laboratory results were compared for 110 sera from control subjects (daily alcohol intake < 40 g), for 57 sera from chronic ethylic subjects and for 78 sera from patients suffering from non-alcoholic liver diseases. There was a large between-laboratory variation, suggesting that the method is difficult to standardize and that results are not transferable. Results of enzyme and radioimmunoassays were compared on 325 sera. The best correlation was obtained in the groups of ethylic subjects and those with non-alcoholic hepatic diseases. Finally the performance of the CDT-test was evaluated by calculating sensitivity and specificity. With both methods specificity was very high (> 85%) but sensitivity was poor (< 50%).