POSTER ABSTRACTS

Results With in- run coefficients of var ia t ion (CVs) ranged from 3.9% to 6.0% for concentrat ions varying from 5.1 to 33.0 ~Lg/mL while total CVs (over mult iple calibrations) ranged from 6.8% to 10.6%. When using four different h igh pa t ien t samples wi th four separate serial dilutions, the assay showed excellent l ineari ty over the ent ire range. The es t imated MDC was 0.31 ~Lg/mL. The correlation evaluat ion comparing the TECHNICON IMMUNO 1 system (y) to the Abbott TDx (x) for 114 serum samples with values <47.4 ~Lg/mL was very good and the regression equat ion was: y = 0.98x-1.16; r = 0.94. Conclusions The TECHNICON IMMUNO 1 Vancomycin assay is a reliable and rapid method, provides impor tan t advantages for the clinical laboratory and demonst ra tes good precision, l ineari ty and accuracy.

19. EVALUATION OF DCL DIRECT HDL ASSAY Chu SY., Biochemistry Laboratory, Dr. E. Chalmers Hos- pital, Fredericton, N.B. E3B 5N5

The Diagnostic Chemical Ltd direct HDL kit provides a homog- enous assay for HDL-Cholesterol (HDL-C) determinat ion wi thout p re - t rea tment steps. The reagent contains sulfated ~-cyclodextrin and dext ran sulfate which selectively complex wi th VLDL, LDL and chylomicrons making them insensit ive to PEG-modified cholesterol esterase and oxidase enzymes. Objectives 1) To evaluate the analytical performance of this assay system on the Hitachi 717; 2) To make prel iminary cost assess- men t of using this procedure in comparison to t radi t ional HDL assays. Methods Analytical performance was evaluated by imprecision, l ineari ty studies. Comparison s tudy was carried out by assaying pa t ien t samples using both the direct HDL and the magnesium- phosphotungst ic acid method. Results A CV of 1.4% was obtained for samples wi th HDL-C at 1.2 mmol/L and 2.1% at 1.6 mmol/L levels. Lineari ty of the assay was obtained up to the tes ted 3.6 mmol/L level. Correlation s tudy gave the following results: Direct HDL = 0.895 Lab HDL + 0.13 wi th r = 0.992; Sy.x = 0.072. A large reduction of labour cost of 0.3 FTE was obtained in processing 80 samples per day. Conclusion The DCL direct HDL kit offers a simple, fast and reliable assay for HDL-C determinat ion wi th a much lower labour cost.

20. DIRECT HDL CHOLESTEROL METHODOLOGY, APPLIED TO THE BAYER TECHNICON DAX Blakney, G., and Higgins, T., Dynacare Kasper Medical Laboratories, Edmonton, Alberta, T5V 1B4, Canada and Rostron, P., Bayer Inc., Toronto, Ontario, M9W 1G6 Can- ada

The potent ial of labor savings, increased analytical performance, resolution of Laboratory Information System issue has made the invest igation of the recently introduced non pre - t rea tment direct HDL methodology very at tractive. Objective We evaluated the performance of two direct HDL methodologies on a Bayer Technicon DAX®. Results The resul ts for these direct methods show wi th in- run imprecision of 1.1-6.5% and between-run imprecision of 3.4-6.1% at HDL concentrat ions between 0.7 and 2.3 mmol/L. This com- pares favorably to between-run imprecision of 4.4 to 7.8% for a p re - t rea tment HDL method. Lineari ty was es tabl ished at 2.58 mmol/L. HDL cholesterol values obtained by the direct methods compared well wi th the p re - t rea tment method (r = 0.91-0.99). Conclusion Subs tan t ia l labor savings ( -0 .3 FTE for 800 speci- mens/day) and operational efficiences were achieved without

compromising analyt ical performance with the use of the direct HDL cholesterol methods.

21. A U T O M A T I O N A N D E V A L U A T I O N O F DIRECT HDL-CHOLESTEROL METHOD ON MONARCH 2000 Thibeault , D., Laboratoire de biochimie, Centre de Sant~ Maria-Chapdelaine, Dolbeau, Quebec, G8L 2R5, Canada

Curren t methods to measure HDL-cholesterol require a pre- t r e a tmen t of sample before testing. Now, with new available methods, it is possible to obtain the same results, directly, without the pre- t rea tment of samples, and s imultaneously with all other routine tests. Objectives Evaluate and automate, on the Monarch 2000 system, a commercial direct HDL-cholesterol kit featur ing the Sugiuchi formulation (Clin Chem 1995 ; 41 : 717-723). Results The automat ion of this kit on the Monarch 2000 system was not an easy task. Our first results, obtained wi th the last system parameters , gave a slope of 1.04, an intercept of -0 .0817 and a R - 0.990 for 58 samples, when compared with the s tandard method: phospho-tungstate/MgC12. A second compari- son of the same kit, us ing a different analyzer (Hitachi 717), gave a slope of 1.02, wi th an intercept of -0 .17 and R = 0.992 for 96 samples. In t ra-assay precision data showed values of 0.9, 2.36 and 2.51% for concentrat ions of 1.81, 1.07 and 0.56 mmol/L respectively. In ter-assay precision resul ts showed a C.V. of 4% using a control level of 1.03 mmol/L for 20 samples. Calibration was stable for 2 weeks on the first prel iminary results. Conclusion This direct method for measur ing HDL-cholesterol is suitable to replace the conventional method on the basis of precision and accuracy. Also, this method can effectively improve laboratory workflow and decrease turn-around t ime of results.

22. PERFORMANCE E V A L U A T I O N O F P S A BY AC- C E S S ANALYSER Sharma, S., and Lossing, I., The Mississauga Hospital, 100 Queensway West, Mississauga, Ontario L5B 1B8, Canada

PSA is widely used for screening prostate cancer and it is one of the la tes t assays available on the Access analyser developed by Pas teur Sanofi Diagnostics. Objective To evaluate the performance characterist ics of PSA by Access. Methods Sixty fresh serum specimens were analysed for PSA by Abbott IMx and Sanofi Access Systems. Regression analysis was used to compare the data. Precision, l ineari ty and min imum detection limit were determined.

Results The following data were obtained by regression analysis.

~lope: 1.2398; Intercept: -0.437; Bias: 1.302;

Corr. Coef.: 0.9994

?recision: Level 1 Level 2 Level 3

Ylean 0.31 3.20 75.37

__SD 0.01 0.10 1.76

7oCV 3.23 3.14 2.34

Linearity: up to 150 ~tg/L

~inimum Detection Limit: 0.05 ~tglL

Conclusions The data demonst ra te good performance character- istics providing a sensit ive random-access method for serum PSA.

CLINICAL BIOCHEMISTRY, VOLUME 30, J U N E 1997 373