Autoverification of results generated by the automated analyzers is used by the clinical laboratories to increase efficiency, streamline the work flow and decrease turnaround time. However, this requires that the integrity of specimens is assured ?hemolyzed, lipemic and icteric specimens must be identified and rejection rules set to avoid verification of erroneous results. In our laboratory we measure the common chemistry analytes with Synchron LX20 Clinical System that offers index panel for identification of hemolyzed (H), lipemic (L) and icteric (I) specimens. We established the rejection rules by measuring H, I, and L for serum samples with known concentration of total hemoglobin (THgb =0-500 mg/dL, n=20), total bilirubin (TBil=1.9-28.4 mg/dL, n=36) and triglycerides (Trig=46-7370 mg/dL, n=29), and determined the effect of H, I and L values on 29 common serum chemistry analytes. None of the studied serum analytes were affected when I<= 28 and the rejection limit was set to be >28. Following analytes were affected at increased L values: HDLC, LDLC (L>2), DBil, TBil (L>4), Alb, CO2 (L>5), and at increased H values: LD and Fe (H>1), K (H>2), AST, CK, DBil, TBil (>3), AMYL (H>4), ALT (>6), Na (>8) and, therefore, these L and H values were set as the rejection limits. For the remaining analytes the rejection limits were set as L>8 and H>9. Whenever the I or H limit was exceeded the test was reported as icteric or hemolyzed, while specimen with rejected L limit, except for LDLC, were ultracentrifuged and reanalyzed. Implementation of these rejection limits on the number of affected specimen / tests in the calendar year 2003 is summarized as follows: 986 specimens (0.89% of total) were received hemolyzed (H=2, n=318; H=3, n=310, H=4, n=448, H=5, n=292, H=6, n=132, H=7, n=112; H=8, n=122; H=9, n=145; H=10, n=894) resulting in 2773 tests reported as ?hemolyzed?; 244 specimens (0.22% of total) needed to be ultracentrifuged. There were no specimens with I>28. Cost analysis: In our laboratory the serum index panel is performed automatically with each of the 29 LX20 methods. For these tests, on average, the cost per test increased by $0.013 that translates to an increase in the cost per year by 5.49%.