Here is another excellent question from Sheila Minor: We use the Rosner index for both interpretation of a PT or PTT mixing study and for LA testing, in which we follow up the DRVVT confirm and ratio with a 50:50 DRVVT mix to resolve discrepancies. Most references state that a Rosner index value of 15 or more indicates LA. What is the appropriate cut-off for factor deficiencies, or the presence of other inhibitors like heparin? Also, there are a variety of cut-offs for LA non-positive results. I have seen <11 for factor deficiencies, and 11–14 being described as indeterminate. What do you recommend?
Hi, Sheila, I have attached a Powerpoint PDF of a mixing study talk I presented August 13, 2015 that provides details for the Rosner index and the Chang ratio, both used to compute limits that distinguish between correction and no correction. The most recent information I’ve found establishes a Rosner index ratio of 11 as the limit, with no indeterminate range, but recommends using normals and LA specimens to establish a local limit (slide 28). The Chang formula (slide 27) is similar to the Rosner and generates result expressed in percent, with 75% as its recommended cut-off. In either case, the limit delineates correction, implying a coagulopathy, from non-correction, implying an inhibitor, usually LA.
There is no established limit for unfractionated heparin or the DOACs, however a thrombin time tells you if heparin or dabigatran, the oral direct thrombin inhibitor, is present. Dabigatran, used extensively for outpatients with atrial fibrillation, profoundly prolongs the PTT and slightly prolongs the PT. Likewise, the oral direct anti-Xa antithrombotics rivaroxaban, apixaban, and edoxaban prolong the PT, confounding PT mixing study results. The anti-Xa DOACs may be detected using the chromogenic anti-Xa as a screen, though you can’t currently use the assay to distinguish among nor measure the direct anti-Xas. Here is the Powerpoint presentation:
I’m sure I’ve mentioned it
I’m sure I’ve mentioned it here before, but my favorite paper on this topic is from a colleague–Kershaw G, Orellana D. Mixing tests: diagnostic aides in the investigation of prolonged prothrombin times and activated partial thromboplastin times. Semin Thromb Hemost. 2013; 39:283–90, who investigated this for many scenarios. The upshot–no cut-off is 100% sensitive and specific for any event; so, you need to chose a value that gives the best ‘compromise’ sensitivity and specificity for the event(s) you want to identify.