Safety, Efficacy Endpoints Must Be Balanced in Clinical Trials


When designing studies and analyzing outcomes, it is important to take into account which endpoints are most likely to have the greatest impact on patients. Composite endpoints are not always the solution, as they can potentially overemphasize safety at the expense of efficacy or vice versa, said experts at a TCT 2015 session devoted to the nuances of trial design and interpretation. 

Survival, quality of life key

Gregg StoneSurvival is most important to patients, followed by a high quality of life, said TCT Course Director Gregg W. Stone, MD, of NewYork-Presbyterian/Columbia University Medical Center, New York, N.Y. Therefore, all-cause mortality should be used as a tiebreaker if other comparisons conflict, and composite endpoints should attempt to reflect events that will have a major impact on a patient’s quality of life, he advised.

For example, “there really need to be different thresholds of diagnosis for different MIs,” Stone noted, adding that major and moderate bleeding are as “prognostically important” as MI. Bleeding definitions should be “broadened to capture both types of bleeding events. To compare small periprocedural MI with classical major or life-threatening bleeding is not necessarily fair when you are putting them in a composite endpoint,” he commented. “Adverse events of both types, if sufficiently severe, can affect mortality.”

Applying this line of thinking to studies of dual antiplatelet therapy (DAPT) duration after PCI, Stone referenced a recent meta-analysis showed that prolonged DAPT was associated with a 25% reduction in MI (P = .01), with a trend toward a 41% reduction in stent thrombosis (P = .06) but a 72% increase in bleeding (P < .0001). “So how do you decide if the patient is overall better off when you’ve got better efficacy, but also worse safety? Perhaps we can use mortality as the ultimate arbiter,” he suggested.

While DAPT duration did not have an impact on cardiac mortality in that analysis, longer therapy was associated with a 49% increase in noncardiac mortality (P = .006) and a 22% increase in all-cause mortality (P = .02).

 The lesson is not that no one should receive prolonged DAPT, Stone said. “[Rather] we need to consider individual patient characteristics to try to determine who’s at relatively high risk … for atherothrombotic ischemic events who maybe benefit from DAPT.”

Stratification may be based on whether the patient had a prior or recent MI, according to Stone. “We need an individualized patient modeling approach to see who might benefit.”

The endpoint of net adverse clinical events (NACE) balances safety and efficacy endpoints in theory, but only if the components are properly balanced, he said.

The whole story

A pitfall of NACE was seen in the MATRIX trial of heparin compared with bivalirudin (Angiomax, The Medicines Company) in patients with ACS undergoing PCI, Stuart J. Pocock, PhD, of the London School of Hygiene and Tropical Medicine, London, said during a presentation.

“The way it was reported in the New England Journal of Medicine, in the abstract — many people don’t read more — only the primary endpoints [were mentioned]. You might come away with thinking there was no significant difference in MACE or NACE, neither was significantly lower with bivalirudin, so … this was a neutral trial,” he said.

The problem, according to Pocock, is that NACE “doesn’t necessarily capture the subtleties of the tradeoff between efficacy and safety.” He expressed skepticism about NACE as a primary endpoint.

What was lost in the message of this trial, Pocock said, is that bivalirudin was associated with lower all-cause mortality (1.7% vs. 2.3%; P = .04). So, “there was quite more to be said” than what was in the abstract, he emphasized.

Disclosures:

  • Pocock and Stone report no relevant conflicts of interest. 

 

 

Comments