Factors Differentially Associated with Low PFA-100 Closure Times in Patients on Chronic Aspirin Therapy

Background: The PFA-100 device tests platelet function in an in vitro simulated high-shear environment. Low closure times (CTs) on the collagen/ADP (C/ADP) cartridge indicate platelet hyper-reactivity; shortened CTs on the collagen/epinephrine (C/EPI) cartridge suggest residual platelet function after aspirin administration. The impact of patient-specific and pre-analytic variables such as citrate concentration on C/ADP and C/EPI CTs has not been well characterized for a single patient population.

Methods: PFA-100 CTs were measured in 289 patients on chronic aspirin therapy who had undergone cardiac surgery 6 months previous and in whom there was complete inhibition of arachidonic acid (AA)-induce platelet aggregation. CTs were measured using C/ADP and C/EPI agonist cartridges in whole blood collected in both 3.2% and 3.8% citrate. Simultaneous measurements of hematologic parameters (complete blood count, reticulated platelet count by flow cytometry, ABO blood group,) as well as markers of inflammation (CRP, fibrinogen and vWF levels), oxidative stress (urine 8-iso-PGF_2α) and aspirin responsiveness (urine 11-dehydro-TXB₂, arachidonic acid platelet aggregometry) were also performed. Demographic data were recorded, as were medications that could potentially affect assay results. Univariate logistic regression was performed to predict odds of low closure time despite aspirin treatment: <= 71 sec for C/ADP and <=193 sec for C/EPI. Predictors were log transformed or treated categorically if necessitated by distribution of the data. Variables with P <= 0.20 on unadjusted analysis were included in the four initial multivariate models to predict CT. Thereafter, model fit was optimized using the Akaike Information Criteria, by testing stepwise removal of each variable.

Results: Multivariate analysis identified several factors that differentially correlated with odds of low CT depending on whether the C/ADP or C/EPI agonist cartridges were used and whether the blood was collected in 3.2% vs. 3.8% citrate. Predictors of low CT are shown below in descending order of strength of evidence of association as assessed by t-statistic and importance to model fit:

Conclusion: The significance of patient-specific, hematologic, inflammatory and oxidative indices on the performance of the PFA-100 varies differentially between cartridges and citrate concentrations. vWF antigen levels > 150% were the strongest predictor of low CT for three of four cartridge-citrate combinations. Odds of low CT are associated with a number of factors potentially extrinsic to platelet function. The collection of blood in 3.8% rather than 3.2% citrate may limit the influence of extrinsic variables on PFA-100 results. These results may have implications for how PFA-100 tests are analyzed and interpreted.