The importance of locally derived reference ranges and standardized calculation of dilute Russell's viper venom time results in screening for lupus anticoagulant

Three commercial dilute Russell's viper venom time (DRVVT) kits were evalua ted at four UK centres experienced at performing lupus anticoagulant (LA) t ests. Each centre established a normal reference range for the DRVVT ratio calculated against local pooled normal plasma from 20 healthy normal subjec ts, Plasma from LA-positive patients and LA-negative thrombophilia patients was also tested. DRVVT ratios and the degree of correction were assessed i n a variety of ways to reflect not only the UK national Guidelines, but als o the manufacturers' recommendations. The reference range data showed a nor mal distribution in each case, but considerable variation in the mean and S D between the centres and reagents, with the mean +2SD value ranging from 1 .06 to 1.19. The use of an arbitrary DRVVT ratio of <1.1 as the cut-off val ue for normality, which is applied in many laboratories, is therefore inapp ropriate. Although no single kit had a clear overall advantage in terms of sensitivity and specificity, the way in which the screen and confirmation d ata were analysed had a major impact on the interpretation of the results. A data analysis method employing a mean plus two standard deviations (SDs) cut-off for normality, and judgement regarding confirmation of LA based on a percentage correction of DRVVT ratio, was the simplest and most consisten t, with overall sensitivity and specificity values of 81% and 94%, respecti vely, for uncomplicated LA-positive and LA-negative thrombophilia samples. We conclude that the 1991 BSCH Guidelines are in need of revision, each lab oratory should establish its own normal reference range for the DRVVT ratio and a common method should be used for calculating the degree of correctio n with confirmation reagents, so that LA results can be correctly interpret ed between laboratories. Standardizing DRVVT interpretation in this way sho uld improve the consistency of LA detection.