Background: IGF-I is a clinically relevant protein in the diagnosis and monitoring of treatment of growth disor- ders. The Growth Hormone Research Society and the International IGF Research Society have encouraged the adoption of a universal calibration for immunoassays to improve standardization of IGF-I measurements, but currently commercial assays are calibrated either against the old WHO IRR 87/518 or the new WHO 02/254. We compared two IGF-I immunochemiluminescent assays: IMMULITE® 2000 (Siemens) and LIAISON® (DiaSorin), which differ in their standardization, and verified their precision according to quality specifications based on biological variation and their linear range.
Methods: 62 patient serum samples were analyzed for both assays and compared according to standards of the Clinical and Laboratory Standards Institute (CLSI), EP9-A2-IR. Precision was verified according to CLSI EP15- A2. Optimal coefficient of variation (CVo) and desirable coefficient of variation (CVd) for IGF-I assays were calculated as quality specifications based on the biological variability, in order to assess if the interassay analytical CV (CVa1) in the two methods were appropriate. Two dilution series using the 1st WHO International Standard (WHO IS) for IGF-I 02/254 were used to verify and compare the linearity range.
Results: The regression analysis showed constant and proportional differences for serum samples (slope b = 0.8115 (CI 95% CI; 0.7575-0.8556); intercept a = 33.6873 (95% CI: 23.3613-44.0133) between assays and similar pro- portional differences for WHO IS 02/254 standard dilutions series (slope b = 0.8024 (CI 95% CI; 0.7560-0.8616); intercept a = 6.9623 (95% CI: -2.0819-18.4383) between assays. Within-laboratory coefficients of variation for low and high levels were 2.82% and 3.80% for IMMULITE® 2000 and 3.58% and 2.14% for LIAISON®, respecttively.
Conclusions: IGF-I concentrations measured by both assays are not transferable. The results emphasize the need to express IGF-I concentrations in standard deviation score (SDS) according to a matched normal population of the same age and gender. Within-laboratory precision in both methods met quality specifications derived from biological variation.