The importance of using an acceptable linear calibration

The importance of using an acceptable linear calibration curve for factor assays Elaine Bromidge, Clare Dunsmore, Gary W. Moore Viapath Analytics, Diagnostic Haemostasis & Thrombosis, Guy’s & St. Thomas’ NHS Foundation Trust, London, United Kingdom Introduction Generation and scrutiny of parallel line bioassay data are crucial to determine accurate factor assays 1. At least three dilutions of test plasma are needed to determine linearity and parallelism. The coagulation data must be checked by looking at the slope value R 2 and the shape of the curve. By using the parallel line method to determine relative factor potency this may show evidence of inhibition (most commonly the presence of antibodies to coagulation factors or lupus anticoagulants). A point to point curve or a single point curve can give a misleading result and inhibitors can be missed. Observations • Analyses were carried out on the Sysmex CA-1500 and CS-2100 i using Siemens Factor deficient plasma and SHP calibrator and Actin FS. *NEQAS sample 15/24 was reanalysed using the updated Factor IX protocol described below and the results are shown in Table 1. NEQAS Survey 212 Factor IX 15/24 NEQAS Median 106 u/d. L Original Result (mean) 121. 6 u/d. L Grade D Updated Result * (mean) 104. 8 u/d. L Grade A Table 1: NEQAS results Standard Curve for Ig. M Assay R 2 = 0, 9998 100 10 Figure 1: Test samples containing inhibitory activity show an apparent increase in relative potency as the samples is diluted OD (Sysmex CA-1500). 1 Original Protocol Sysmex UK Raw Data Calibration - Linear Secs 1/2 53. 0 1/4 59. 6 1/8 66. 6 1/16 73. 4 1/32 79. 4 Initial sample dilution 1/5 Table 2: Protocol amendments to the Factor IX assay. Updated Protocol Sysmex UK Raw Data Calibration - Linear Secs 2/1 53. 9 3/2 56. 2 1/1 60 1/2 66. 5 1/4 73. 4 1/8 79. 7 1/16 85. 1 1/16 72. 7 1/32 79. 6 1/50 82. 6 1/95 87. 3 Low Curve The importance of using linear curves has been highlighted in NEQAS sample 15. 24 (Table 1). Results achieved at St Thomas’ were 117. 3 (1/1), 122. 2 (1 -2) and 125. 3 (1/4) (mean = 121. 6 u/d. L). The discrepancy across the dilutions revealed a loss of linearity at the top of the curve. Close scrutiny of the associated calibration curve showed that the assay was no longer linear below approximately 15% Factor IX activity 2 and therefore the resulting curve skewed when forced into linear interpolation, emphasizing the importance of scrutiny of dose response data. 1000 Concentration (MPL) • For the sample displayed in Figure 1, if single point analysis had been used, the above would have been underestimated at the 1/1 dilution equivalent to 24. 8 iu/d. L, with no inhibitor effect observed. Initial sample dilution 1/20 and 1/5 for low curve Figure 2: Factor IX calibration curves on old and new protocols. Green = new protocol, blue = original protocol Calibration curve protocol amendments & Outcomes An evaluation was carried out by Sysmex and Viapath Analytics and it was determined that the CS-2100 i protocol for Factor IX assays should be adjusted, as detailed in Table 2. The initial sample dilution was increased from 1/5 to 1/20 and additional dilutions were added to the standard curve and in consideration of the reduction in linearity at lower concentrations; a low curve was introduced to cover factor IX concentrations of less than 10%. The multi dilution approach was kept in order to maintain the ability to observe potential presence of inhibitors across a range of dilutions whilst maintaining accuracy. Each batch of Factor IX deficient plasma has dose response analysis as part of the batch acceptance process and the data are closely scrutinized to assess the limits of the central linear component and alter standard curve dilutions as appropriate to maximise linearity. As a consequence of this investigation UK NEQAS now request the lot number detail of the reference and deficient plasma used as well as the sample dilutions used to generate the result. Acknowledgements The authors would like to acknowledge the support and advice of Ian Holding at Sysmex UK and UK NEQAS. References 1. I Mackie, P Cooper, A Lariw, S Kitchen, E Gray, M Laffan ‘Guidelines on the laboratory aspects of assays used in haemostasis and thrombosis(2012) British Committee for Standards in Haematology 2. R Mellick Factor IX investigation (12. 06. 2015) Document reference: SUKBMS 13 -411 Sysmex UK.