Download this article's Factoid in PDF (& PPT for Gold Subscribers)
Routine use of a bleeding risk score in patients undergoing percutaneous coronary intervention (PCI) appears to reverse the ‘risk-treatment paradox’ for bivalirudin in a real-world context, such that patients who stand to gain the most from the drug are more likely to receive it. The findings were published online March 5, 2013, ahead of print in the Journal of the American College of Cardiology.
Adnan K. Chhatriwalla, MD, of Saint Luke's Mid America Heart Institute (Kansas City, MO), and colleagues examined bivalirudin use and outcomes for 6,491 PCI patients treated at 4 centers from 2007 to 2011. They compared the periods before and after July 2009, when their health system began a program to estimate individual bleeding risk using the National Cardiovascular Data Registry (NCDR) score. The researchers retrospectively stratified patients by low (< 1%), intermediate (1-3%), or high (>3%) risk of bleeding.
According to the paper, widespread use of the score, which incorporates 9 preprocedural clinical variables, was begun “in an effort to support the more rational use of bivalirudin in patients with the greatest potential to benefit.”
Bivalirudin Use Dovetails with Risk
Bivalirudin use as a whole increased after the NCDR score was adopted. Both intermediate- and high-risk patients were more likely to receive the drug. The opposite pattern was seen in those at low risk for bleeding (table 1). Data from the NCDR Cath PCI Registry showed an overall increase in bivalirudin use at a nationwide level over the same time frame, regardless of patient bleeding risk.
Table 1. Bivalirudin Use by Bleeding Risk
After implementation of the NCDR score, bleeding complications dropped for patients considered intermediate or high risk and held steady for those at low risk (table 2).
Table 2. Bleeding Rates by Bleeding Risk
Independent predictors of bivalirudin use differed before and after implementation of the NCDR score. Before July 2009, bivalirudin was less likely to be given to patients who were smokers (OR 0.77; 95% CI 0.65-0.92), lacked insurance (OR 0.55; 95% CI 0.37-0.82), had renal failure (OR 0.45; 95% CI 0.30-0.96), or were at high bleeding risk (OR 0.94; 95% CI 0.90-0.97 for every 1% increase in risk) and more likely to be given in conjunction with hypertension (OR 1.38; 95% CI 1.12-1.68) or cerebrovascular disease (OR 1.32; 95% CI 1.11-1.67). After July 2009, only bleeding risk (OR 1.07; 95% CI 1.04-1.10) and prior CABG (OR 1.30; 95% CI 1.07-1.56) were independently associated with bivalirudin use.
Among patients at high risk of bleeding, STEMI presentation most strongly predicted lack of bivalirudin use (OR 0.43; 95% CI 0.30-0.64).
Use of the NCDR score varied substantially among physicians. Even after July 2009, “if 2 patients with identical clinical characteristics presented to 2 random interventional cardiologists at our institution, there was, on average, an approximately 3-fold greater probability of receiving bivalirudin with 1 physician as compared to another,” the researchers note.
More Calculation, Less Intuition
In an e-mail communication with TCTMD, Dr. Chhatriwalla said that bleeding scores are not used routinely in US practice, “with the exception of a few centers in the country. . . . I cannot stress how novel this approach is—the majority of interventionalists make treatment decisions based on subjective intuition rather than objective risk stratification.” This characterization is not meant to sound insulting, he explained, “but facts are facts.”
A patient’s bleeding risk cannot be judged intuitively, Dr. Chhatriwalla stressed. “The NCDR model incorporates 9 variables, and physicians cannot just do this in their heads.” A lack of objective risk stratification may explain the risk-treatment paradox, he added. “If you do not correctly identify the high-risk patients to start with, you have no hope of tailoring therapy appropriately for them.”
Deepak L. Bhatt, MD, MPH, of Brigham and Women’s Hospital (Boston, MA), agreed on the rarity of bleeding scores in clinical practice and described the NCDR risk estimate as “a very good one,” though others are available.
He told TCTMD in an e-mail communication that cardiologists “should probably get in the habit of calculating bleeding (and for that matter, ischemic) risk scores. This knowledge could lead to medicine being practiced a bit more objectively and even more cost effectively.”
Both physicians cited cost as the main obstacle to simply giving bivalirudin to everyone undergoing PCI. “In my opinion, the data for bivalirudin are very strong across the full spectrum of PCI,” Dr. Bhatt said. “I use it in the vast majority of cases that I do.”
Dr. Chhatriwalla pointed out that “there is tremendous physician-level variability in many aspects of US health care, not just with respect to bleeding avoidance strategies during PCI.
“Physicians fall across the spectrum of using bivalirudin for every patient, heparin for every patient, and everywhere in between,” he continued. “Many things are necessary to overcome this [variability.] First, physicians need to accept the data [showing that] bleeding is an important complication that should be avoided, that bleeding risk stratification can identify high-risk patients, and that bleeding avoidance strategies are most effective and cost-effective in high-risk patients. Even then, it is notoriously difficult to change physician behavior without standardized protocols, guideline recommendations, financial incentives, etc.”
Some variability can be “attributed to the art of medicine,” Dr. Bhatt said, noting that even when quality improvement initiatives work, “it doesn’t disappear, not that we would want it to.”
Variables comprising the NCDR bleeding risk score include: ACS type, cardiogenic shock, gender, prior heart failure, prior PCI, New York Heart Association class IV heart failure, peripheral vascular disease, age, and estimated glomerular filtration rate