Several of the studies described above, investigating variable clopidogrel response and the influence of this response on clinical outcomes, have used multiple assays to assess the degree of platelet inhibition to ADP. Even several years later, there remains little definitive opinion regarding the optimal method to apply in the clinical environment.
At the time of this research study design, comparative data was limited. The prominent studies in the literature were as outlined in Table 1. As the data summarised in this table illustrates, comparing the assays for assessing clopidogrel response was challenging. LTA in response to ADP was by far the most widely published assay, but concerns remained about utility in the clinical environment because of the demanding nature of the assay protocols. Activation of the GPIIbIIIa complex appeared promising but several of the studies seemed to use no agonist when assessing this response. Only Hochholzer et al used ADP in this circumstance and did not formally report a correlation (149). The single study evaluating VASP-PRI seemed to suggest a more promising correlation but this was a small study of only 33 patients taking clopidogrel (193).
Page | 76
Date Study n Assays Comparative
Data
Corr. p value 2002 Gurbel et al 100 LTA 5àmol/L ADP
LTA 1àg/mL collagen LTA 750àmol/L arachidonic acid WBPA 1 àg/mL collagen
Flow cyt. (GpIIb/IIIa expression, PECAM-1, no agonist)
PFA-100 ADP/Collagen
Not reported - -
2003 Gurbel et al 92 LTA 5, 20àmol/L ADP Flow cyt. (PAC-1 and P-selectin, no agonist)
LTA of 5 and 20 àmol/L ADP
0.6 Not
reported 2005 Hochholzer
et al
1001 LTA 5 and 20àmol/L ADP
Flow cyt. 20àmol/L ADP (PAC-1 and P- selectin expression)
‘Flow cytometry findings confirmed those of optical aggregation’
Not reported
2005 Serebruany et al
544 LTA 5àmol/L ADP Flow cyt. PECAM-1
Change in activated GpIIbIIIa expression vs change in 5 àmol ADP aggregation
0.72 Not reported
2005 Angiolillo 48 LTA 6àmol/L ADP Flow cyt. (P-selectin, PAC-1)
LTA vs P- selectin LTA vs Activated GpIIb/IIIa expression
0.09 0.39
.59 .015
2005 Aleil 33 LTA 5àmol/L ADP Flow cyt. (VASP-PRI)
LTA vs VASP 0.72 <.0001 2005 Gurbel et al 160 LTA 20àmol/L ADP
TEG no agonist
None reported
- -
Corr. = correlation coefficient
Table 1: Studies comparing assays to assess response to clopidogrel published before the initiation of this research.
Page | 77 Since the inception and recruitment of patients for this research, three large comparative
datasets have now been published to address the variation in measurement of clopidogrel response between a number of these assays. These studies are summarised in Table 2 (167,187,200).
Date Study n Patient
group
Assays Comparative data
2008 Lordkipanidze et al
116 Stable angina
LTA 5, 20 àmol ADP
WBPA 5, 20 àmol ADP
PFA-100 (Collagen- ADP)
VerifyNow P2Y12
Yes All between assay correlations (Figure 16 )
2009 Cuisset et al 598 ACS LTA 10 àmol ADP VASP-PRI
Yes R=0.64, p<0.01 2010 Breet et al 1069 Elective
PCI
LTA 5 and 20 àmol ADP
VerifyNow P2Y12 Plateletworks ADP Impact-R
PFA-100 collagen/ADP
No Not
reported
Table 2: Studies published since 2006 evaluating clopidogrel response using multiple assays
The study by Lordkipanidze et al provides the most detailed comparison of the majority of assays excluding flow cytometry (187). A summary of the correlation between the assays used is given in Figure 17.
Page | 78 Figure 17: Correlation coefficient for 6 assays used to evaluate response to clopidogrel
(With permission from Oxford University Press. Lordkipanidze et al (187))
This data demonstrates a strong correlation between 5 and 20àmol ADP for LTA. It also shows a strong correlation between the results of 5 and 20àmol ADP in WBPA. However, correlation between LTA and WBPA is poor. VerifyNow P2Y12 has a significant correlation with both LTA (5 and 20àmol ADP) and WBPA (20àmol ADP) although the correlation is not strong. The correlation between different concentrations of agonist using the same assay does suggest that there is a degree of reproducibility within an assay method. This suggests that different assays may be measuring different aspects of platelet aggregation and clot formation and highlights the need for outcome data to support the use of these assays in clinical practice. It also confirms the findings of previous studies suggesting that the
ADP/collagen PFA-100 cartridge system does not appear to reflect the antiplatelet effects of clopidogrel.
Which assay would best predict clinical outcome was addressed by Breet et al in their study of more than 1000 patients undergoing elective coronary stenting (167). These investigators used LTA to 5 and 20àmol ADP to assess clopidogrel response but also measured the antiplatelet effects of clopidogrel using 6 additional point-of-care assays that are commercially marketed to measure platelet function. These included VerifyNow P2Y12, Plateletworks, Impact R, IMPACT-R ADP, PFA 100 collagen/ADP and PFA 100 Innovance PFA P2Y. The primary outcome was a composite at 1 year of death, myocardial infarction, stent thrombosis and ischaemic stroke. Patients were also monitored for TIMI major and minor bleeding. A cut off threshold for high residual platelet reactivity was determined using ROC curves for each assay
Page | 79 against the primary outcome. Patients were then divided using these cut-off values and
differences in outcome assessed using the Kaplan-Meier method and the log-rank test. LTA, VerifyNow P2Y12 and Plateletworks were all predictive of future cardiovascular events. The optimal cut off for the VerifyNow P2Y12 assay in order to discriminate between those patients at risk of events was 236 PRU which is in keeping with data from Gurbel et al (67).
Interestingly, correlation coefficients between the predictive assays were not included among the results in the paper or in the additional appendix.