Citrated samples from 100 patients on i.v. heparin and 20 normal patients were tested with three batches each of three activated partial thromboplastin time (APTT) reagents: Thrombosil I (Ortho); Automated APTT (Organon Teknika) and Actin FSL (Baxter). The ratio of APTT over the geometric mean normal APTT for each heparinized sample was calculated. One batch of reagent arbitrarily chosen as a reference gave the ratios APTRREF (y). The remaining reagents to be standardized against the reference system gave the ratios APTRTEST (x). The best correlation between systems was given by log vs log x. Standard curves were prepared from the APTT ratios of the 20 normal patients and 65 of the heparinized samples. On plotting log APTRTEST vs log APTRREF the y intercept was close to zero so x was expressed in terms of y using; log x = HSI. log y, where HSI (Heparin Sensitivity Index) = slope. The APTRTEST results of the remaining 35 heparinized samples were transformed using; APTRTRANS = (APTRTEST)HSI.APTRTRANS was then compared to APTRREF to determine whether the transformation brought the results closer to the reference. We conclude that although some improvement was found by using the transform, it was not possible to mathematically relate APTT results due to a high degree of variation between results using different reagents. A standard APTT reagent for the monitoring of heparin therapy is recommended. A separate APTT reagent may be required for the screening of factor deficiencies and lupus anticoagulants.
Ray et al. (1992) studied this question.