Measurement of lipoprotein (a): comparison of Macra and Imubind methods.

Our specific aim was to compare lipoprotein(a) [Lp(a)] measured by two enzyme-linked immunosorbent assays (the Strategic Diagnostics, Inc. Macra Lp(a) and the American Diagnostica, Inc. Imubind Lp(a) Stripwell) in sequentially referred hyperlipidemic patients. The Macra method binds Lp(a) in serum or plasma to monoclonal anti-Lp(a) in microtiter test wells followed by polyclonal anti-Lp(a) and a chromogenic reaction; the Imubind method is similar but uses two polyclonal anti-Lp(a) antibodies. Within run coefficients of variation were 1.5 percent for the Macra (Lp[a] 27.2 +/- 0.4 mg/dl) and 3.4 percent for the Imubind method (Lp[a] = 18.9 +/- 0.6 mg/dl). Between-run coefficients of variation for the Macra method were 4.7 percent (Lp[a] 14.8 +/- 0.7 mg/dl) and 8.3 percent (Lp[a] of 33.5 +/- 2.8 mg/dl). For the Imubind method in nine separate analytical runs, the between run coefficients of variation were 8.4 percent (Lp[a] of 15.4 +/- 1.3 mg/dl), 3.0 percent (18.8 +/- 0.6 mg/dl), and 5.7 percent (25.6 +/- 1.5 mg/dl). The intraclass correlation was 0.92 (p < or = 0.0001) for duplicate aliquots (n = 210) quantitated by both methods, and the lower limit of the 95 percent confidence interval of the intraclass correlation was 0.90. Comparison of the two methods revealed no systematic bias (p = 0.09), since the lower limit of the 95 percent intraclass correlation confidence interval was > or = 0.75, the two methods for measuring Lp(a) are considered interchangeable. Given the importance of Lp(a) as an independent risk factor for coronary heart disease, it is clinically important to have precise and accurate methods for its measurement.

Measurement of plasminogen activator inhibitor type 1 antigen: comparison of Tintelize and Imubind methods.

Our specific aim was to compare three plasminogen activator-inhibitor type 1 (PAI-1) antigen ELISA kit assays (the Biopool AB, Ltd, TintElize PAI-1 Strip-Well Format; the American Diagnostica, Inc., Imubind 822/1; and the second generation Imubind 822/1S). Within-run coefficients of variation (n = 6) for the TintElize, Imubind 822/1 and Imubind 822/1S methods were 5.5%, 5.9% and 6.8%, respectively. Between-run coefficients of variation for six aliquots per run were 2.9% for TintElize, 3.8% for Imubind 822/1, and 3.5% for Imubind 822/1S. Comparison of the average of duplicate aliquots from hyperlipidaemic patients demonstrated intraclass correlations of 0.75, 0.79 and 0.95 for TintElize vs Imubind 822/1 (n = 39), TintElize vs Imubind 822/1S (n = 39), and Imubind 822/1 vs 822/1S (n = 84), respectively. Lower 95% confidence interval limits of the intraclass correlation were 0.55, 0.48 and 0.93, respectively. Mean PAI-1 antigen values (n = 39) were 12.1, 15.8, 15.8 and 16.0 ng/ml, respectively, for TintElize, TintElize without using the quenching well, Imubind 822/1, and Imubind 822/1S. All three methods were easily performed and exhibited high correlation and reproducibility. A significant systematic bias (P < 0.006) existed between TintElize and TintElize without using the quenching well, Imubind 822/1, and Imubind 822/1S. However, there was no significant bias when TintElize without using the quenching well is compared with Imubind 822/1 (P > 0.8) and to 822/1S (P > 0.8) nor is there significant systematic bias between Imubind 822/1 and 822/1S (P > 0.3). By convention, interchangeability between assay methods suggests that the lower limit of the 95% intraclass correlation confidence interval be greater than 0.75.(ABSTRACT TRUNCATED AT 250 WORDS)

Comparison of the Abbott Cell Dyn 3000 SL and the Coulter STKS hematology analyzers.

A side by side comparison of the Cell Dyn 3000 SL and Coulter STKS using identical samples was performed over a four month period in the Clinical Hematology Laboratory at the University of Cincinnati Medical Center. A total of 444 samples comprised of 114 sophomore medical students and 330 randomly selected clinical patients were compared in 20 hemogram parameters, and the ability of each instrument to generate differential suspect flags was analyzed. Correlation coefficients for leukocytes, erythrocytes, hemoglobin, hematocrit, and platelets were 0.99. Correlation coefficients for the Cell Dyn and STKS compared with a 600 cell manual differential were 0.98 and 0.89 for neutrophils, 0.96 and 0.87 for lymphocytes, and 0.72 and 0.48 for monocytes, respectively. Both instruments demonstrated high precision and accuracy by internal and external quality control standards. Each analyzer exhibited strength as a screening instrument for abnormal cell populations. The Coulter STKS had overall sensitivity of 75%, specificity of 94%, 25% false negatives, and 6% false positives. Sensitivity, specificity, false negative and positive rates for the Cell Dyn 3000 SL were 68%, 92%, 32%, and 8%, respectively. Based upon this extensive side by side comparison using an identical sample population, it has been concluded that although statistically significant systematic bias (p < 0.05) exists between the two instruments, both analyzers can adequately support the needs of the clinical hematology laboratory.