COMPARISON OF THE AMES, RANDOX AND ROCHE METHODS WITH THE SYNERMED METHOD FOR THE DETERMINATION OF SERUM IRON CONCENTRATIONS ON NONDIALYSISAND DIALYSIS SPECIMENS

Objectives: To evaluate the Ames, Randox, Roche, and Synermed methods for the measurement of serum iron and to investigate patterns of possi bly discrepant results in dialysis patients. Methods: Assays were perf ormed on the Cobas Fara II analyzer. Precision and accuracy studies we re conducted; recovery studies were done by adding pooled serum from d ialysis patients to an assayed human serum-based control. Patient comp arisons included over 150 nondialysis patients and 30 dialysis patient s. Results: For the Ames, Randox, Roche, and Synermed methods, the bet ween-run precision was less than 2.80% with the normal aqueous Iron st andard; 2.00, 2.70, 0.80, and 2.00% for the four methods with the high serum iron control, respectively, and less than 2.30% with the serum pool. Using a pooled serum from dialysis patients, between-run precisi on was higher with all four methods. With an abnormal assayed human se rum-based control, accuracy was over 98% for the four methods. Recover ies were 121% for the Ames and Randox methods and 104-105% for the Poc he and Synermed methods. Accuracy as assessed with Murex EQAS specimen s ranged from 71 to 80%, 71 to 96%, 98 to 99.5%, and 42 to 50% for the four methods, respectively. For comparisons of the Ames, Randox, and Poche methods with the Synermed method, difference analyses revealed b iases (SD) for nondialysis patients of 1.9 (2.7), 1.5 (3.3), and 1.8 ( 2.2) mu mol/L, respectively; and for dialysis patients of 8.2 (13.3), 5.1 (5.4), and 1.4 (1.7) mu mol/L. Standard linear regression analyses and correlation coefficients are also provided. Conclusions: The Roch e method was slightly more precise than the other methods. Using an ab normal assayed serum-based control, all methods showed good accuracy. Recovery studies with pooled serum from dialysis patients showed inter ferences with the Ames and Randox methods and good recovery with the R oche and Synermed methods. With the bovine serum-based Murex samples, all but the Poche method yielded some low results; the Synermed method has been reported to suffer from matrix problems with bovine serum al bumin. Based on recovery studies and difference analyses, the Ames and Randox methods revealed discrepancies in iron results for samples fro m dialysis patients. The Roche and Synermed methods appeared to be sui table for measurement of serum iron in dialysis patients. Copyright (C ) 1998 The Canadian Society of Clinical Chemists.