FALSE ALARM?

September 1, 2009 (Barcelona, Spain) — The proton-pump inhibitor (PPI) data from the large-scale TRITON-TIMI 38 trial, suggesting no effect of these agents on cardiovascular event rates in either the clopidogrel or prasugrel groups, has now been presented at the European Society of Cardiology (ESC) 2009 Congress and published online in the Lancet [1].

These data have been cited on several occasions in the controversy surrounding the issue of whether PPIs affect the action of clopidogrel, but this is the first time that they have been officially made available.

Also published in the same Lancet paper but not presented at the ESC meeting is an analysis of platelet aggregation in the much smaller PRINCIPLE-TIMI 44 study, showing that inhibition of platelet aggregation was modestly reduced with both clopidogrel and prasugrel in patients taking a PPI.

Lead author of the Lancet paper, Dr Michelle O'Donoghue (Brigham and Women’s Hospital, Boston, MA), told heartwire that the data from the TRITON trial were much more clinically relevant than the PRINCIPLE data, as the TRITON trial involved 13 600 patients and studied clinical events, while the PRINCIPLE study only involved 201 patients and focused on platelet aggregation as its primary end point.

She commented: "In TRITON-TIMI 38, we did not find use of a PPI to be associated with a higher risk of cardiovascular events for patients on either clopidogrel or prasugrel. This is in contrast to prior studies that have shown a higher risk of adverse events for patients on clopidogrel in combination with a PPI. This is also the first study to show that it appears to be safe to combine a PPI with prasugrel. However, as with prior studies, use of a PPI was not randomized in the current study. As such, these data can provide some early reassurance to clinicians that these two classes of drugs may in fact be safe to use in combination for patients who have a strong indication to be on both drugs. However, only a randomized trial of a PPI can provide definitive evidence regarding the safety of combining these two classes of drugs."

On the modest attenuation of the antiplatelet effects for both clopidogrel and prasugrel shown with PPIs in the PRINCIPLE study, O'Donoghue said: "This is consistent with findings from prior studies. The fact that this modest attenuation of the antiplatelet effects did not translate into a higher risk of clinical outcomes in TRITON highlights the fact that we should not rely solely on surrogate end points." She added: "Rather, these findings illustrate that there remains much to be understood about the exact nature of the relationship between platelet-function studies and clinical events. This is highly analogous to the situation that we observed a few years ago with atorvastatin, whereby atorvastatin appeared to modestly attenuate the antiplatelet effects of clopidogrel, but this was not associated with a higher risk of adverse outcomes."

O'Donoghue explained that the PRINCIPLE data were not presented at the ESC meeting because "it was required that we submit data from a single clinical trial. As well, our presentation was restricted to only 10 minutes." She added: "For these two reasons, we restricted our presentation to the results of the TRITON-TIMI 38 study. We believe, however, that the platelet-function studies provide additional intriguing mechanistic insight, so we included these data in the Lancet paper."

"Impressive" Data

Coauthor of an accompanying Comment article in the Lancet [2], Dr Dirk Sibbing (Deutsches Herzzentrum, Munich, Germany), who has conducted some of the platelet-function studies of PPIs and clopidogrel, told heartwire that he thought the TRITON data were "impressive."

"There have been lots of anxious discussions about an interaction, but I think these TRITON data have taken some of the hysterical reaction out of the matter," he commented.

He said that while the platelet aggregation data should not be ignored, it appears that PPIs have only a modest effect on platelet inhibition with the thienopyridines, and there would likely have to be a greater effect to have any influence on clinical events. He suggested a "two-hit" hypothesis. "I think you probably have to have another problem that can affect platelets in order for the PPI effect to matter. The average patient in TRITON did not have such problems, so I would say the TRITON data suggest that in a moderate-risk population it is safe to use PPIs with clopidogrel and prasugrel."

Sibbing added: "But clinicians should always err on the safe side, and I would still advise that doctors choose a PPI that has been associated with the least interactions with thienopyridines if coprescribing these drugs, ie, not omeprazole. And if a patient is at high risk of having intrinsic platelet problems, for example if they are diabetic, I would be more cautious."

Sibbing also suggested that the increase in cardiovascular events seen in some of the prior observational studies in patients taking PPIs with clopidogrel could have been due to compliance issues. "Patients taking PPIs are more likely to have symptoms of gastric pain and therefore may stop taking the thienopyridines for a few days. That might cause the event rather than the PPI itself. TRITON was a randomized trial that was closely monitored, so patients have to be more compliant, and therefore this effect would not have shown up as much in this setting."

"I would say these new TRITON data nicely demonstrate that moderate risk patients who are compliant with their thienopyridine therapy can safely be put on a PPI," he commented.

Sibbing noted that the PRINCIPLE data showing a small reduction in platelet inhibition with prasugrel as well as clopidogrel when given with PPIs was "interesting and surprising. . . . I believe this is the first time such an effect has been reported with prasugrel and shows that it is not just a clopidogrel issue, but prasugrel is a much more potent antiplatelet drug, and therefore this modest effect should not be a problem," he commented to heartwire .

TRITON Data

The current Lancet paper reports that of the 13 608 acute coronary syndrome (ACS) patients randomized to prasugrel or clopidogrel in TRITON TIMI 38, 4529 (33%) were on a PPI at randomization. O'Donoghue et al used a multivariable Cox model with propensity score to assess the association of PPI use with clinical outcomes. Results showed no association between PPI use and risk of the primary end point (a composite of cardiovascular death, myocardial infarction [MI], or stroke) for patients treated with either drug.

TRITON-TIMI 38: Rate of the Primary End Point as Related to PPI Use

Drug	PPI (%)	No PPI (%)	Unadjusted HR (95% CI)	p
Clopidogrel	11.8	12.2	0.98 (0.84–1.14)	0.80
Prasugrel	10.2	9.7	1.05 (0.89–1.23)	0.58

Patients who were treated with a PPI were significantly more likely to be older, female, and white; have an index diagnosis of unstable angina/non-ST-elevation MI; and live in North America or Western Europe than those not treated with a PPI. Patients treated with a PPI were also significantly more likely to have a history of peptic ulcer disease and have a lower baseline hemoglobin concentration at the time of randomization.

After adjustment for potential confounders and the propensity to treat with a PPI, the authors found no significant association remaining between use of a PPI and risk of the primary end point, both for patients treated with clopidogrel (HR 0.94; 95% CI 0.80–1.11; p=0.46) and for those treated with prasugrel (HR 1.00; 95% CI 0.84–1.20; p=0.97). Use of a PPI was also not associated with increased risk of MI, stent thrombosis, or a decreased risk of bleeding for patients treated with either clopidogrel or prasugrel.

Because PPIs could be started or discontinued at any time during follow-up, additional analyses were conducted to look at different time points and after varying durations of follow-up, and similar results were found. Another analysis showed no difference with different PPIs or other gastric antacid drugs, including H₂ receptor antagonists. Yet another analysis focused on patients with a single reduced-function CYP2C19 allele, and this also suggested no interaction between either clopidogrel or prasugrel and PPIs.

PRINCIPLE TIMI 44 Results

In the PRINCIPLE-TIMI 44 trial, 201 patients undergoing elective percutaneous coronary intervention (PCI) were randomized to prasugrel or clopidogrel 600-mg loading doses. In both arms, there was a modest reduction in inhibition of platelet aggregation in patients on a PPI compared with those not on a PPI.

PRINCIPLE-TIMI 44: Inhibition of Platelet Aggregation Related to PPI Use

Drug	PPI (%)	No PPI (%)	p
Clopidogrel	23.2	35.2	0.02
Prasugrel	69.6	76.7	0.054