Selection bias may have played role in SAFE-PCI trial
Women with lower risk may have been selected for the SAFE-PCI for Women trial, which may have contributed to the lower predicted risk for mortality and bleeding, according to a study published in Circulation: Cardiovascular Interventions.
“This allowed us to assess whether selective enrollment could have explained the lower than expected event rate that led to the early termination of the SAFE-PCI trial,” Jennifer A. Rymer, MD, MBA, fellow at the Duke Clinical Research Institute, and colleagues wrote.
As Cardiology Today previously reported, the SAFE-PCI for Women trial found that radial access may be preferable in women undergoing PCI or diagnostic catheterization, but clinicians should be aware that a proportion of patients will require conversion to femoral access.
Researchers analyzed data from 496 women (mean age, 64 years) enrolled in the SAFE-PCI for Women trial, which was conducted with data from the CathPCI Registry.
In addition to trial data, information from the registry itself was used to compare in-hospital bleeding and vascular complications in women 1 year before enrollment of the SAFE-PCI trial (Sept. 1, 2010, to Sept. 1, 2011; n = 13,315; mean age, 68 years), during the enrollment period (n = 15,904; mean age, 68 years) and 1 year after enrollment for the trial (July 2013 to July 2014; n = 11,643; mean age, 68 years). There were also 12,212 patients who were linked to CMS.
Women enrolled in the SAFE-PCI trial were younger, had lower rates of vascular complications, lower rates of bleeding after PCI and lower predicted risks for mortality and bleeding compared with women from the registry who were not enrolled in the trial. These women also had fewer bleeding events within 72 hours after PCI or before discharge (P for both < .01).
Of the women linked to CMS, there were no differences among the four groups for unplanned revascularization and death at 30 days.
“These data demonstrate the value of a registry-based trial structure to examine generalizability,” Rymer and colleagues wrote. “Further iterations in registry-based trials may facilitate identification of selection bias in real time.” – by Darlene Dobkowski
Disclosures: Rymer reports no relevant financial disclosures. Please see the study for all other authors’ relevant financial disclosures.