The effect of heparin infusion intensity on outcomes for bridging hospitalized patients with atrial fibrillation

Abstract Background Perioperative bridging in atrial fibrillation (AF) is associated with low thromboembolic rates but high bleeding rates. Recent guidance cautions the practice of bridging except in high risk patients. However, the practice of bridging varies widely and little data exist regarding appropriate anticoagulation intensity when using intravenous unfractionated heparin (UFH). Hypothesis To determine if high intensity UFH infusion regimens are associated with increased bleeding rates compared to low intensity regimens for bridging patients with AF. Methods We conducted a single center retrospective cohort study of admitted patients with non‐valvular AF receiving UFH for ≥24 hours. UFH intensities were chosen at the providers' discretion. The primary endpoint was the rate of bleeding defined by the International Society on Thrombosis and Hemostasis during UFH infusion or within 24 hours of discontinuation. The secondary endpoint was a composite of cardiovascular events, arterial thromboembolism, venous thromboembolism, myocardial infarctions and death during UFH infusion. Results A total of 497 patients were included in this analysis. Warfarin was used in 82.1% and direct acting oral anticoagulants in 14.1% of patients. The rate of any bleed was higher among high intensity compared to low intensity UFH regimens (10.5% vs 4.9%, odds ratio = 2.29, 95% confidence interval = 1.07‐4.90). Major bleeding was significantly higher among high intensity compared to low intensity UFH regimens. There was no difference in composite thrombotic events or death. Conclusions Low intensity UFH infusions, targeting lower anticoagulation targets, were associated with decreased bleeding rates without a signal of increased thromboembolic events in hospitalized AF patients.


| INTRODUCTION
Patients with atrial fibrillation (AF) have a 4-to 5-fold increased risk of ischemic stroke compared to patients in sinus rhythm. 1 The annual risk of stroke is 1.69% to 4.2% in patients with AF, with 15% of strokes in the United States attributed to AF. 2 To mitigate this risk, hospitalized patients are often treated with parenteral anticoagulation during perioperative interruption of oral anticoagulation or while oral anticoagulation is sub-therapeutic.
Recent studies suggest perioperative bridging in patients with AF provides limited benefit in terms of preventing thromboembolic events, but increases bleeding rates. [3][4][5][6][7] Most studies, including the BRIDGE trial, 3 the only randomized placebo-controlled trial, evaluated perioperative bridging strategies for elective surgeries, whereas the data for bridging hospitalized patients with acute issues has until recently not been addressed. 5 These studies enrolled predominantly low to moderate thromboembolic risk patients (mean CHADS 2 = 2.1-2.4 and CHA 2 DS 2 -VASc = 3.6-4.1) and limited patients at higher thromboembolic risk. Studies have also shown that perioperative bridging in AF is associated with a 4-fold risk of bleeding compared to non-bridging strategies (any bleed rates: 5%-34% and major bleed rates: 3%-9%) with undifferentiated thromboembolic events (0%-4%). [3][4][5][6][7] Current guidelines and expert consensus support perioperative bridging with subcutaneous low molecular weight heparin or intravenous unfractionated heparin (UFH) for patients taking oral vitamin K antagonists who are at high risk for thromboembolism, indicated by a CHA 2 DS 2 -VASc score greater than 5 to 6 or a prior thromboembolic event. 8 The median duration of UFH was 86 and 83 hours in patients who received high and low intensity UFH regimens, respectively, P = .69.
Duration of hospitalization was a median of 12.1 days in the high, and 11.6 days in the low intensity UFH regimens, respectively, P = .19.
During this time, 76% and 76.6% of patients in the high and low intensity UFH regimens, respectively, underwent a procedure, P = .971.
There was no difference in the classification (non-surgical vs surgical) or individual type of procedures.
The median percent time in therapeutic range was 50% without significant differences between the two groups. Overall, patients in the high intensity group had a greater total exposure to UFH than did the lower intensity group (Figure 2). All UFH parameters are shown in Table 2.
The rate of any bleeding event was significantly higher in patients  Table S4.
The rate of secondary endpoints of composite cardiovascular events, stroke, myocardial infarction, arterial thromboembolism, venous thromboembolism and death was low and not significantly different in patients receiving high and low intensity UFH infusions ( Figure 2). The rates of primary and secondary endpoints are shown in It is well established that bleeding occurs at a much higher rate than thromboembolism in patients who receive perioperative bridging, at an approximate bleed to thrombosis ratio of 13:1 with a marked increase in the risk of bleeding (OR = 3.6, 95% CI = 1.52-8.50). 17 Prior studies in the low to moderate risk AF population show increased bleeding events and similar thromboembolic events in those who received perioperative bridging compared to those who did not. [3][4][5][6][7] In 2017, an expert consensus recommended the use of parenteral anticoagulation for bridging patients with AF only in patients with a CHA 2 DS 2 -VASc score ≥5 and/or a prior thromboembolic event. 8  absolute reduction in any bleed and 3.7% absolute reduction in major bleeds, would estimate the bleed events averted as noted above.
Given the large and growing population of patients with AF, the frequency at which bridge therapy is required and the associated bleeding risk, strategies to minimize adverse outcomes remains a public health priority. Therefore, large multicenter prospective studies are needed to validate these findings, as well as investigations into the optimal UFH infusion dosing intensity for bridging in AF patients with the highest thromboembolic risk.
Limitations of this study include the single center and retrospective nature and potential confounders inherent with such a design.
Measures taken to limit confounding include a well-balanced and similar baseline patient population and the exclusion of patients on intravenous UFH for non-AF conditions as their primary indication for bridging. UFH protocols and medical practice may vary among institutions. Differences in physician, pharmacy and nursing practice can affect the degree of UFH infusion prescribing, dosing, titration, and monitoring. Studies have suggested that monitoring UFH infusions with anti-Xa levels compared to aPTT is superior in maintaining values within goal range. 21,22 However, therapeutic dosing of UFH with regards to safety and efficacy has been guided by aPTT, not anti-Xa levels in clinical trials, mostly acute coronary syndrome. This study did not assess the difference in bleeding events and the type of monitoring used (aPTT or anti-Xa). The type of monitoring was chosen at the discretion of the providers and the baseline characteristics were well balanced between the two groups. In addition, as can be seen in the heparin protocols viewed in Tables S1 and S2, there were some overlap in anticoagulation intensities as measured by therapeutic goal ranges (ie, mechanic valve protocol [high intensity] and heart failure protocol [low intensity] have a goal anti-Xa levels of 0.35-0.7 and 0.35-0.5 units/mL, respectively). The small sample size limited the ability to detect differences in the secondary endpoints. However, prior studies were also unable to detect differences in thromboembolic events given the inherent low rate of occurrence in bridging and large numbers needed to show such a small difference. [3][4][5][6][7] The results of this study demonstrate that use of a low intensity UFH regimen in hospitalized AF patients can decrease major morbidity by decreasing both any and major bleeding event rates without a signal of increasing systemic thromboemboli. Low intensity UFH should now be the preferred dosing strategy for this population of patients in clinical practice.