Discussion
This is the first national survey exploring US pharmacists’ preferences in BRT usage and treatment based on bleed risk. Pharmacists preferred the HAS-BLED tool and considered patient-specific factors and evidence-based data when weighing the risk-benefit of OACT with or without gastroprotective therapy.
Similar to our findings, where three-quarters of pharmacists used a BRT, a recent Medscape/American College of Cardiology (ACC) survey reported that 74% of cardiologists used a BRT (eg, HAS-BLED) always/most of the time or sometimes to assess a patient’s overall risk of bleeding prior to initiating DOAC therapy; 27% never or rarely used a bleed risk score before prescribing DOACs.19 Although reasons for BRT preference were not provided, they may be similar to those reported by our respondents (ie, familiarity/ease-of-use). In both surveys, rationales for not using a BRT were not obtained, but possible reasons include lack of confidence with bleed risk calculators,20 inconsistent implementation of comprehensive assessments (stroke risk, bleed risk, and medication-related issues prior to decision-making),21 and nonspecific guideline recommendations.22
More recently, a network meta-analysis found that HAS-BLED and HEMORR2HAGES had modest but balanced sensitivity (defined as the ratio between the number of major bleeding events in high-risk stratification and the total number of bleeding events) and specificity (defined as the ratio between the number of nonmajor bleeding events in the low-risk population and total nonbleeding events) for predicting major bleeding events.2,3 Several respondents did comment that, although HAS-BLED was imprecise and only studied with warfarin, it was necessary to identify bleed risk in a patient starting a high-risk medication, and that the ACC anticoagulation application uses HAS-BLED with CHA2DS2VASc along with clinical trial data to estimate stroke risk and bleed risk, with projected risk reduction (strokes) and risk increases (bleeds) expected with each treatment (www.acc.org/tools-and-practice-support/mobile-resources/features/anticoag-evaluator). The 2019 AHA/ACC/HRS atrial fibrillation guideline recommends that HAS-BLED scores be used to assess bleed risk in patients for whom anticoagulation is being considered, and that the need for and choice of OACT should be periodically reevaluated to reassess stroke and bleed risks.23
Although more than 80% of extracranial bleeds are GI bleeds,24 most BRTs are nonspecific for predicting GI bleeds. Indeed, one respondent used a spreadsheet with several BRTs to maximize treatment guidance for patients with multiple risk factors for strokes and bleeds. A comprehensive approach to determining factors that increase bleed risk should be adopted. These factors include age (HAS-BLED, HEMORR2HAGES, mOBRI, ATRIA); anemia (mOBRI, HEMORR2HAGES, ATRIA); hepatic/renal disease (HAS-BLED, HEMORR2HAGES, ATRIA, mOBRI); concomitant medications/alcohol use, including NSAIDs, corticosteroids, and antiplatelet therapy (HAS-BLED, HEMORR2HAGES); bleed history/rebleeding risk (HEMORR2HAGES, HAS-BLED, ATRIA); and GI bleeds (mOBRI).1,2 Additional risk factors for GI bleeds include being a tobacco smoker and/or being infected with Helicobacter pylori. A prospective cohort study that analyzed data from questionnaires completed by 99,359 individuals from the Copenhagen General Population Study reported that the multivariable adjusted hazard ratio for current smokers versus never smokers was 2.20 (95% CI, 1.84-2.62) for GI bleeds.25 Presence of H pylori should be investigated, with a subsequent eradication regimen implemented, as patients with warfarin-associated upper GI bleeds who were H pylori-positive had lower HAS-BLED scores versus those who were negative.26
When bleed risk was lower than stroke risk (eg, HAS-BLED < 3, CHA2DS2VASc ≥ 1), respondents appropriately initiated therapy with an OAC (predominantly apixaban); a small proportion also added gastroprotection. If the patient did not have any other GI bleed risk factors (eg, a previous GI bleed or on chronic antiplatelet or NSAID therapy), the choice of OACT depended on the attributes of each OAC and patient preference.27 Selection of warfarin was appropriate if cost, formulary restrictions, and availability of an inexpensive reversal agent were important concerns to patients and/or their health care providers. Rivaroxaban was selected because of its once-daily dosing and low risk for GI bleeding.