A Canadian study has found that many primary care physicians (PCPs) incorrectly estimate stroke or bleeding risk in patients with atrial fibrillation (AF). In the study, PCPs used guideline-recommended risk scores to evaluate stroke risk in only about half of their patients and used the guidelines to determine bleeding risks in just one quarter of their patients.
As a result, many patients were receiving suboptimal oral anticoagulant dosages, which could increase their risk for stroke or bleeding. Moreover, PCPs did not estimate the risk of stroke and bleeding at all in 15% and 25% of patients, respectively.
The study, by Paul Angaran, MD, from the Terrence Donnelly Heart Centre, St Michael’s Hospital, University of Toronto, Ontario, Canada, and colleagues, was published online October 14 in the Canadian Journal of Cardiology.
“The Canadian Cardiovascular Society (CCS) AF Guidelines recommend that all patients with AF should be stratified using a predictive index for the risk of stroke and for the risk of bleeding, and that most patients should receive antithrombotic therapy,” lead investigator Shaun G. Goodman, MD, from the Terrence Donnelly Heart Centre and the Canadian Heart Research Centre, Toronto, said in a journal news release.
“Despite these recommendations, the uptake of these evidence-based therapies was suboptimal,” he continued, “Among those who did receive anticoagulation with warfarin, as many as four in 10 patients spent less time in the therapeutic range we know is optimal to reduce the risk of stroke.”
Most patients with AF need antithrombotic therapy, which decreases their stroke risk but increases bleeding risk. The extent to which PCPs use evidence-based guidelines to balance these two risks has been unclear, according to the investigators.
In the study, researchers used chart audit data from the Canadian Facilitating Review and Education to Optimize Stroke Prevention in Atrial Fibrillation (FREEDOM AF), a quality assurance program funded by Boehringer Ingelheim Canada aimed at using evidence-based practices to improve the management of patients with AF by PCPs.
The study included data on 4670 adult patients (39% women) who did not have significant valvular heart disease and who were seen by 474 PCPs between February and September 2011.
PCPs filled out standardized forms about patient demographics and treatment. They also estimated risk of stroke and bleeding risk as low (<3% per year), intermediate (3% – 6% per year), high (>6% per year), or “do not know.”
Researchers then used data on the collection forms to calculate guideline-recommended predictive risk estimates for stroke and bleeding. For stroke risk, they used the CHADS2 score (congestive heart failure, hypertension, age, diabetes, stroke/transient ischemic attack) and the CHA2DS2VASc score (congestive heart failure, hypertension, age [≥75 years], diabetes, stroke/transient ischemic attack, vascular disease, age [65 – 74 years], sex [female] score). For bleeding risk, they used the HAS-BLED score (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly [>65 years], drugs/alcohol concomitantly).
Finally, they compared the guideline-recommended scores to estimates for stroke and bleeding risk provided by the PCPs.
Results showed that PCPs did not estimate stroke risk at all in 15% of patients, yet 67% of these people had an intermediate or higher risk for stroke, according to guideline-recommended risk estimates. Likewise, PCPs did not estimate bleeding risk in 25% of patients, yet 42% of these individuals had high bleeding risk.
Physicians who provided risk estimates used guideline-recommended risk assessments for stroke in 50% of patients, and for bleeding in just 26% of patients.
Overall, PCPs underestimated stroke risk in 20.2% of patients and bleeding risk in 61.4% of patients. They overestimated stroke risk in 25.4% of patients and bleeding risk in 8.0% of patients.
There were 96 patients receiving antithrombotic therapy, of whom 90% received warfarin. Many patients had international normalized ratios that were suboptimal according to the time in the therapeutic range (TTR). Among these patients, 24% had a TTR lower than 50%, 9% were between 50% and 60%, 11% were between 60% and 70%, and 56% had a TTR of 70% or higher.
Participating PCPs volunteered and received a small remuneration, so they may have been more likely to provide better care than PCPs in the general population. PCPs also selected the patients to evaluate, which could have affected results, the authors point out.
“This large real-world contemporary study of primary care specialists offers an important insight into current practice patterns for patients with AF,” write Laurent Macle, MD, from the Montreal Heart Institute, Université de Montréal, Montréal, Québec, Canada, and Jason Andrade MD, from the Vancouver General Hospital, University of British Columbia, Canada, in an accompanying editorial.
Still, the results should be viewed with “some caution,” they suggest, because oral anticoagulant use in this study was greater than reported community rates of 50% to 60%. The study also took place before the advent of newer oral anticoagulants such as apixaban, dabigatran, and rivaroxaban, which have now become first-line treatment for AF. Current rates of oral anticoagulant use may differ from those in the study, they point out. Finally, current Canadian Cardiovascular Society guidelines do not recommend oral anticoagulants in patients with an estimated stroke risk of 1.0% to 1.5%, not the risk of lower than 3% used in this study. The use of the higher rate could have placed more patients in the low-risk category, potentially skewing the results.
Nevertheless, the study “highlights the discordance that exists between empirical risk assessment tools and subjective physician assessments,” Dr Macle and Dr Andrade write.