Measurement of Insulin in Different Automated Platforms: Still Need to Improve?

Presentation Number: MON 576
Date of Presentation: April 3rd, 2017

Andrea Kozak*, Gabriela Fernanda Ruibal, Cecilia Andrea Fenili, Mariana Dicugno, Lorena Gomez, María Laura Romano, Patricia Otero, Hilda Farelo and Ana Mari­a Sequera
Argentine Society for Endocrinology and Metabolism, CABA, Argentina


Measurement of circulating insulin (INS) may improve the classification and management of diabetes mellitus and assist in treating people with Insulin resistance. Several institutions in the world (CDC, ADA, EASD, etc) are working on the harmonization process INS to improve the performance of the methodologies currently in use. Agreement among methods can be improved by establishing traceability to the IDMS procedure. From these developments it is necessary to verify the improvements made by the diagnostic industry in this sense. Objective: Analyze the behavior of INS in currently available methods and to quantify the degree of dispersion among them. Assess in basal samples the impact of differences in the calculation of metabolic indexes as HOMA-IR.
Materials and Methods: We analyzed 168 samples of 84 OGTT (WHO protocol) with the aim of covering a wide range of Ins concentrations. Samples were classified in three groups following INS levels (uUI/ml), G1 (n=53, 48 INS-0 (B) +5 INS post)= INS<12, G2 (n=62, 34 B+28 INS post): INS: 12.1-50.0 and G3 (n=53, 53 INS post):> 50. INS was measured by four immunoassays (all with IRP 66/304): ECLIA:Cobas-e411(Reference Value(RV): 5.0-24.9,Roche)=C and three CLIAs: Architect I-2000,(RV: Not reported , Abbott)=A, IMMULITE 2000,(RV: <2.0 -29.1,Siemens)=I and Liaison,(RV:Not reported, DiaSorin)=L. The dispersion in each sample between methods of INS was calculated by: (xi – median /median) x 100. HOMA-IR (Glucose (mg / dl) x INS (uIU/ml) / 405) on all baseline samples was calculated discarding diabetic patients. Statistic: In each group, in each sample between methods: Friedman test, Student-Newman-Keuls. Results: INS (uUI/ml): (median and range):G1: A: 6,6 (1,5-11,6)*; C: 6,9 (3,2-11,6)*; I: 3,8 (2,0-9,4)* y ** and L: 9.7 (4,4-26,8). G2: A: 17,4 (9,6-81,5)* y**; C: 20,0 (12,3-48,8); I: 14,1 (5,0-91,6)* y** and L: 23,3 (4,4-82,7).G3: A: 75,0 (37,1-239,9)*y **; C: 96,2 (50,1-325,5)*; I: 66,6 (30,9-193,0)* y ** and L: 121,0 (49,8-344,7).*p<0.05 vs LIAISON; **p<0.05 vs COBAS. INS dispersions (%, in range): G1: A: -35 a 18%; C: -18 a 22%; I: -75 a 8% and L: 0 a 305%. G2: A: -30 a 22%; C: -25 a 23%; I: -75 a 108% and L: -73 a 208%. G3: A: -24 a 8%; C: -9 a 11%; I: -51 a 6% and L: -3 a 110%. We observed (in the same basal samples) HOMA-IR > 2.0 in: 50% samples measured by A, 53% by C; 35% by I and 70% by L. Conclusions: We observed a considerable dispersion among INS results measured in all levels, prone to higher values in Liaison and lower in Immulite. RV reported by manufacturers should be reviewed, because it isn’t representative for the healthy population. HOMA are affected by these dispersions so that the same patient could be characterized differently depending on the method used.


Nothing to Disclose: AK, GFR, CAF, MD, LG, MLR, PO, HF, AMS