From Dr. Aliaa Mohamed Amer, Senior Consultant Hematopathologist, Hamad Medical Corporation, Qatar: What is the value of calculating the Rosner index on the incubated plasma 1:1 mix (patient and control)? Does the cut-off used for the direct mix apply to the incubated mix? My understanding is that the Rosner index is primarily useful for the detection of LA, which is a fast-acting inhibitor, so why incubate? In our lab, we usually perform a 1:1 mix at zero hr and after a 2 hr incubation for the detection of a time-dependent inhibitor, and we apply the Chang formula after doing a 4:1 mix for the detection of a weak inhibitor. But since we started applying the Rosner index, we get a lot of discrepancy with the incubated step. Would you please advise?
Hello, Dr. Aliaa, and thank you for your question. I refer you to our February 12, 2024 discussion of mixing studies and the Rosner index. In answer to your incubation question, ~15% of LA inhibitor responses are enhanced by incubation, likewise ~15% of specific inhibitors are immediate-acting. Consequently, our facility defers directly to two-hour incubation. When the results indicate an inhibitor, we continue with the LA sequence or the Nijmegen-Bethesda assay for an inhibitor, often informed by the clinical condition when available. I discussed your question with friend and colleague Dan Kaczor, who advocates for simply using the PTT clotting interval exceeding that of the NPP by ≥10% . The local facility establishes the index and cutoff based on clinical efficacy studies.
Click here for the link that Dean Willett provides in his comment below.
The ICSH issued a guidance document last year that might be useful:
https://onlinelibrary.wiley.com/doi/10.1111/ijlh.14344
Thank you Daniel for your email. I am sure that different labs are adopting different ways for interpreting mixing study. I believe your recommendations are all applied in our lab. We’ve being doing mixing study for years without issues. We calculate the percent correction at zero hr and after a 2 hr incubation and we apply the Chang formula and we get very reasonable results. Since we started including the Rosner index in the calculation we are facing discrepant results in the sense of correction in the 1:1 mix, no inhibitor with Chang then surprisingly no correction with Rosner index or correction at zero hr then no correction after incubation which does not make sense. So we repeat mixing on another sample which sometimes matches the 1:1 and Chang and some other times does not. I was looking for some consistency or at least an error that we can work on.
From George Fritsma, Dr. Amer, you’ve probably adjusted your decision points for the Chang and Rosner indices, has that helped resolve the discrepancies you’ve been experiencing?
Hello Dr.Aliaa and George (3-3-25),
I would like to add some information in the area of LA diagnosis that comes from working with Dr Douglas Triplett. With more than 25 years experience, we always focused on keeping the process as simple and accurate as possible. Our simple process:
1. Make sure that your APTT reagent is sensitive to the presence of LA (low titer included) and that your manufacturer has great lot to lot consistency. Seems obvious but you would be surprised to know how often this simple step is overlooked.
2. Make sure that the process you use will allow you to make a very accurate normal range. Once you do this then you can be confident that any result above your upper limit of normal indicates an issue.
3. For mixing studies, we never used any index. The most important issues for your mixing reagent (NP) are twofold: it must be platelet free and the testing on this reagent should yield results for both the PT and APTT that are at the mean or lower of each test’s normal range. Doing this will assure that platelet contamination will not neutralize the LA and that the concentration of all the matrix factors will be sufficient as to negate any dilution affect.
4. Perfecting the recommendation in #3 allowed us to go directly to a 4:1 mix with great confidence. Therefore any mixing study result that was above the upper limit of the APTT suggested the presence of an immediate acting inhibitor. Reviewing patient history and meds helped us rule out interfering issues and proceed to confirmation.
I hope this helps and makes your diagnostic protocol a lot easier than applying any kind of calculation like the Rosner Index. As I mentioned because we followed the steps above we never needed to and we were very confident in our results.
We also used to use the 10% rule, and I think we once migrated to 15% at one stage; in the end we just opted to use correction into the normal range, as this was easier to incorporate for our expert rule set, as used in our large network of laboratories, and also for auto verification. No method is 100% accurate, sensitive or specific for inhibitors; they are all ~95% at best – i.e., Chang, Rosner Index and correction into the normal range.
Dr. Favaloro, my 2015 Fritsma Factor poll revealed these figures:
Limits based on a fixed PTT value such as reference interval
1:1 mix within RI upper limit (39%)
1:1 mix within RI upper limit + 5 seconds (8%)
Limits based on the pooled normal plasma PTT value
1:1 mix within NP PTT value + 5 seconds (14%)
1:1 mix within NP PTT + 10% (32%)
Other (7%): some combination of RI and Rosner.
I favor using 1:1 mix within NP PTT + 10% as it adjusts for NP and reagent variations.