Due to a request from our local physicians treating covid patients experiencing severe dic, our laboratory was asked to begin reporting d-dimer values that were above the analytical measurement range for the assay.This situation requires a dilution of the original sample for accurate assay and quantitation.Prior to implementing this procedure, our laboratory performed a manual and automated (instrument) dilution verification study using high patient specimens with a known value (previously assayed and within the upper limit of the amr), with the intent to verify the instrument (acl top 550) was diluting properly.The diluted (manual and automated) specimens replicated, but both were unacceptably biased when compared to the expected (known) value.To investigate this phenomenon, we performed patient serial dilution studies.The results suggest that the assay does not dilute in a linear fashion at high dilution ratios (approx 1:8 or above).The manufacturer's standard procedure performs a 1:15 dilution automatically on the instrument; in addition, a manual dilution is approved for a ratio up to 1:100.Based on our data, use of the automated dilution factor will lead to a 30-50% positive bias in the determined patient sample concentration.This bias increases with dilution factor, with dilutions up to 1:100 having a positive bias of > 50%.Information provided by il that showed linear dilution (data stated to have been submitted as part of the fda 510k for the assay) could not be replicated.This concern was discussed with the vendor to identify if there was an internal issue present causing these discrepancies.However, our laboratory felt initially that the issue was being marginalized and instead of considering the issue as potentially serious, our staff were effectively accused of being the cause of the problem.We have since replicated this issue at another laboratory on a different system (acl top 750).We believe we have evidence that there is an issue with the analytical performance of this assay and have been provided with no data or assistance to date that effectively rules out this concern.The vendor did eventually request information which we have submitted.(k160276).Fda safety report id# (b)(4).
|