NEJM Feature: GI bleeds reduced with clopidogrel-PPI mix, no rise in CV events
 | |
| Deepak L. Bhatt, MD, chief of cardiology, VA Boston Healthcare System & director of integrated interventional cardiovascular program at Brigham and Women’s Hospital Image source: Thrombosis Clinic |
The finding that PPIs do not interfere with the antiplatelet effects of clopidogrel (Plavix, Bristol-Myers Squibb/Sanofi-Aventis) still holds now that the study has been verified by two academic statisticians using the full adjucated data set, Deepak L. Bhatt, MD, from the Veterans Affairs Boston Healthcare System and Brigham & Women's Hospital in Boston, told Cardiovascular Business News.
"What was not emphasized when we presented those preliminary findings at TCT.09 and not fully apparent until now was the substantial reduction in the primary composite of GI event rates, but, more importantly, the stand alone GI bleeding rate as an endpoint," he said.
"This is the first randomized trial to demonstrate that prophylactic PPI administration reduces clinical intestinal tract bleeding in patients with low or average risk of bleeding," said Bhatt, adding that other studies have shown a reduction at endoscopically-defined endpoints.
Previous data are conflicting regarding the effects of PPIs to decrease the risk of GI complications associated with antithrombotic therapy, particularly in patients receiving dual-antiplatelet therapy. Some data also suggest that PPIs reduce the efficacy of clopidogrel.
To help clarify the matter, Bhatt and others initiated the COGENT (Clopidogrel and the Optimization of Gastrointestinal Events Trial) study. Investigators had intended to enroll 5,000 patients, but settled for 3,761 in the final analysis because the trial's sponsor, Cogentus Pharmaceuticals, filed for bankruptcy in January 2009. Patients were enrolled at 393 sites in 15 countries from January 2008 to December 2008.
The randomized double-blind, double-dummy, placebo-controlled parallel group phase III study evaluated the efficacy and safety of CGT-2168, a fixed-dose once-daily combination of clopidogrel (75 mg) and omeprazole (20 mg) (Prilosec, AstraZeneca) compared with clopidogrel alone. All patients received aspirin. There were 1,876 in the omeprazole arm and 1,885 in the placebo arm.
The primary GI endpoint was a composite of overt or occult bleeding, symptomatic gastroduodenal ulcers or erosions, obstruction or perforation. The primary cardiovascular endpoint was a composite of death from CV causes, non-fatal MI, revascularization or stroke.
Researchers found no significant differences between the two arms in primary or secondary endpoints. Overall, 51 patients had a GI event: 1.1 percent with omeprazole and 2.9 percent with placebo at 180 days. The hazard ratio (HR) with omeprazole was 0.34, with a good confidence interval. The rate of overt upper GI bleeding was also reduced with omeprazole compared with placebo (HR 0.13).
"Despite COGENT being terminated prematurely, we still found reductions in GI bleedings that were highly statistically significant and if the trial had gone its full duration, that result would have been even more statistically significant," Bhatt said.
"The GI findings are remarkable and have the potential to be practice changing," he said. "If one interprets these data honestly, it shows that more patients should be on PPIs, and not less. That message is a correct one."
While COGENT investigators did not conduct a formal cost-effectiveness analysis, Bhatt said there is room for that debate. "Is it worthwhile to prescribe PPIs to everyone getting aspirin and clopidogrel, or for that matter, more potent antiplatelet therapy such as aspirin and prasugrel or aspirin and ticagrelor, where the benefits could be even larger because the bleeding risks are higher?" he asked.
"We've opened the door pretty wide to this concept of GI protection and its importance in patients with cardiovascular disease," Bhatt said, adding that companies might now realize there is an unmet need for this vulnerable patient population.
CV events
In COGENT, Bhatt and colleagues found that 109 patients had a cardiovascular event, with event rates of 4.9 percent with omeprazole and 5.7 percent with placebo. The hazard ratio with omeprazole was 0.99, with the confidence interval crossing 1.00.
The trial was not powered for cardiovascular events, Bhatt said. However, as investigators became aware of the emerging data with respect to a clopidogrel-PPI ex vivo platelet interaction, they prespecified examination of CV safety as well.
"Even our data, which showed an upper limit of 95 percent confidence interval at 1.44, are comparable to what the hazard ratios of 1.3 that the FDA now mandates for diabetes drugs outcomes trials. We're close in COGENT to excluding the degree of risk that is acceptable in those diabetes outcomes trials," he said.
The Kaplan-Meier curves for cardiovascular events were fairly close. "On a population level, we can't see any cardiovascular interaction, and my guess is that a longer trial would have shown the same," Bhatt said.
He explained that the highest risk of ischemic events is shortly after a stent is placed in a patient with acute coronary syndrome. That is when COGENT researchers acquired their data—during this most vulnerable time. "If we didn't see an interaction between clopidogrel and PPIs during that high risk time period, I'm skeptical we would have seen an interaction if we had more time to follow out the patients."
Bhatt said that the FDA will likely not revise its boxed warning on clopidogrel regarding cardiovascular events because the confidence interval was too wide. "The data don't completely rule out the potential interaction in particular patients that have a genotype associated with reduced response to clopidogrel," he said. "My response is those criticisms are valid."
"Nevertheless, this is the first and only randomized clinical trial that was able to evaluate whether there is or isn't a clopidogrel-PPI interaction and it doesn't find an interaction. And randomized controlled data, even if not perfect, are better than observation data, which in turn carry more weight than ex vivo platelet function testing."