engleski

Comparison of automated versus visual detection of lipemic, icteric and hemolysed specimens : who is doing better?

Results from hemolyzed, icteric, and lipemic samples may be inaccurate and can lead to some kind of medical error. Preanalytical interferences may be detected either visually or by automated assessment. Modern state-of the art clinical chemistry analyzers are equipped with automated systems for detection of lipemic, icteric, and hemolytic samples. Visual inspection is time-consuming, highly subjective and non-standardized. Our aim was to assess the comparability of automated spectrophotometric detection and visual inspection of lipemic, icteric and hemolytic samples. The comparability assessment was performed on a consecutive series of 1727 routine biochemistry serum samples. Automated detection was performed on Olympus AU2700 clinical chemistry analyzer, using commercially available Olympus LIH reagent. We assessed: (1) comparability of visual and automated detection of lipemic, icteric and hemolytic samples, (2) precision of Olympus LIH reagent for all three interfering substances at two concentration levels, and (3) inter-observer variability for visual inspection, between five observers. We observed moderate comparability of visual and automated detection for lipemic and icteric samples (weighted kappa=0.555 and 0.529, respectively) whereas agreement for hemolytic samples was good (weighted kappa=0.638), but still not perfect. Precision of Olympus LIH reagent was high (≥ 90%) for all interferences, with the exception of precision for slightly lipemic samples. Best overall agreement between observers was present in assessing lipemia (mean weighted kappa=0.698), whereas worse agreement was observed in assessing icteria (mean weighted kappa=0.476). Visual inspection of lipemic, icteric and hemolytic samples is highly unreliable and should be replaced by automated systems reporting serum indices for every sample.

hemolysis; icteria; lipemia; automated detection; preanalytical; interference; unsuitable specimens; quality

nije evidentirano