engleski

Analytical verification of Combur10 Test® M strips and urine analyzer Cobas u411

Introduction: Our aim was to: a) verify urinary test strips for urine chemistry analysis Combur10Test® M (Roche Diagnostics GmbH, Mannheim) on analyzer Cobas u411 ; b) asses the comparability of Cobas u411 and Miditron Junior II (both Roche Diagnostics, GmbH, Mannheim). Materials and Methods: Analytical verification was performed according to CLSI guideline EP12- A. Repetability was tested on 10 consecutively repeated measurements on negative control (BIORAD, Liquicheck Urinalysis Control Level 1) and in two patient samples with pathological values. Precision from day-to-day was assesed on control samples (Level 1 and 2) in duplicate for 10 days. Comparability was evaluated comparing 87 patient›s samples on analyzers Cobas U411 and Miditron Junior II. Accuracy for glucose, bilirubin and proteins was tested by comparing the results of test strips with values measured by reference method on analyzer Architect c8000 (Abbott, IL, USA). Precision was estimated by the degree of agreement of repeated measurements, accuracy by the degree of agreement with result of reference method. Comparability was expressed with kappa coefficient with 95% confidence interval. Acceptance criteria were: precision: 90% agreement of repeated measurements, pH: acceptable bias ±1 pH unit, specific gravity: acceptable bias ± 0.005 and comparability: kappa coefficient ≥0.6. MedCalc (v12.7.2.0, Ostend, Belgium) was used for statistical analysis. Results: Precision was acceptable for all parameters, except for leukocytes in one patient›s sample (agreement for only 8/10 measurements). Comparability between analyzers was satisfactory, except for bilirubin (kappa=0.370 (0.112 to 0.628)). Accuracy for bilirubin was satisfactory (agreement 10/10). The observed differences in proteins (8/10) and glucose (6/10) were not clinically significant. Conclusion: Analytical performance of Combur10Test® M on Cobas u411 analyzer are within acceptance limits. Observed bias for glucose and proteins are not clinically significant and are caused by differences in methodology. Analyzers are comparable, except for bilirubin measurement.

method verification; method comparability; precision

nije evidentirano