Comparing different lots of activated partial thromboplastin time reagent: analysis of two methods

This study was conducted to compare the performance of 2 different lots of activated partial thromboplastin time (APTT) reagent from the same manufacturer. Two hundred plasma samples were analyzed for APTT on the same instrument using 2 separate APTT reagent lots. Results were compared with those obtained from duplicate APTT determinations with 1 reagent lot on an additional 267 plasma samples. Regression analysis of paired APTT data was suboptimal because the disagreement of the results increased as the APTT values became higher. A better approach was the use of the Bland and Altman assessment of agreement method on logarithmically transformed data, which showed that the new reagent lot produced values that were 88.9% of the values produced by the old lot. This importantly reduced the APTT therapeutic range for unfractionated heparin (UH). In comparison, duplicate APTT values were 97.6% of a first APTT value. However, 95% limits of agreement for duplicate APTT determinations were +/-10% over the range of values considered therapeutic for UH. The variability of APTT results within and between reagent lots is one reason that APTT may not be optimum for monitoring UH therapy.