Safety of Intravenous Cangrelor Administration for Antiplatelet Bridging in Hospitalized Patients: A Retrospective Study

Objective. We aimed to characterize outcomes associated with cangrelor administration used in an antiplatelet bridging strategy in real-world clinical scenarios within a large academic medical system. Background. Cangrelor has been used for antiplatelet bridging in perioperative settings or for patients unable to take oral medications. Prior studies in these settings have reported bleeding rates from 0%-40%. Methods. Patients were retrospectively identified via chart review and included if they were over 18 years old, had coronary or peripheral arterial stents, and had received at least 1 hour of cangrelor infusion during inpatient admission. The primary endpoint was Bleeding Academic Research Consortium (BARC) 3-5 bleeding during cangrelor infusion or within 48 hours of discontinuation; secondary endpoints were bleeding events defined by Thrombolysis in Myocardial Infarction (TIMI), Global Use of Strategies to Open Occluded Arteries (GUSTO), and International Society on Thrombosis and Hemostasis (ISTH) criteria, as well as BARC 2 bleeding. Results. Thirty-one patients met the inclusion criteria. Cangrelor indications were bridging to procedure in 22 patients (71.0%) and inability to take oral P2Y₁₂ inhibitors in 9 patients (29.0%). Twenty-three patients (74.2%) were men, 11 patients (35.5%) were in cardiogenic shock, and 4 patients (12.9%) were on extracorporeal membrane oxygenation (ECMO) at the time of administration. No patients received cangrelor for routine percutaneous coronary intervention. Of the 31 patients, 13 (41.9%) had BARC 3-5 bleeding and 7 (22.6%) expired during hospitalization. All 4 patients on ECMO suffered BARC 3-5 bleeding. Conclusions. We reviewed the use of cangrelor for antiplatelet bridging in real-world clinical scenarios and observed higher rates of clinically significant bleeding than seen in other similar studies. Our study suggests careful consideration when using cangrelor in a sick patient population.

J INVASIVE CARDIOL 2021;33(12):E998-E1003. Epub 2021 November 24.

Key words: antiplatelet, bleeding, high-risk PCI, oral P2Y₁₂ inhibitors

Cangrelor is a potent intravenous (non-thienopyridine) adenosine triphosphate analog that blocks the adenosine diphosphate P2Y₁₂ receptor and has a short time to therapeutic onset and short half-life. In patients undergoing percutaneous coronary intervention (PCI), cangrelor has been shown to reduce the risk of periprocedural ischemic events and was therefore approved for use during PCI in patients who have not been treated with an oral P2Y₁₂ inhibitor or a glycoprotein IIb/IIIa inhibitor.^1-3 It has also been used for the off-label indication of antiplatelet bridging prior to surgery in patients felt to be at high risk for thrombotic complications, particularly in those who have had recent coronary stenting and are therefore at higher risk for subacute stent thrombosis. Such bridging is generally used prior to cardiac surgery when continued use of oral P2Y₁₂inhibitors with longer half-lives is contraindicated due to risk of bleeding during or shortly after surgery.⁴ More recently, it has been extrapolated for use in other perioperative cases as well.⁵

One randomized study evaluating cangrelor’s safety and efficacy in perioperative bridging has been performed, as well as several smaller retrospective reports. The BRIDGE trial prospectively studied 210 patients receiving either cangrelor or placebo as a bridging strategy from oral P2Y₁₂inhibitors prior to coronary artery bypass grafting (CABG).⁴ This study demonstrated greater maintenance of platelet inhibition among patients receiving cangrelor during the study period, along with a low risk of thrombotic events in both arms and without a significant excess in bleeding complications with cangrelor use. However, like nearly all randomized controlled trials, BRIDGE enrolled a selected population (eg, glomerular filtration rate >30 mL/min, platelet count >100,000/ µL), and subsequent small observational studies investigating cangrelor’s safety in real-world perioperative bridging scenarios report significant variations in bleeding rates.^6-8 We therefore performed a retrospective analysis to characterize the real-world incidence of bleeding and ischemic events with cangrelor use as bridging therapy within a large academic medical system.

Study cohort. Patients were identified via retrospective review of inpatient pharmacy records of cangrelor use within the University of Pennsylvania Health System from January 2017 through August 2020. Patients were included in the study if they were over 18 years old, had presence of coronary or peripheral arterial stents, and had received at least 1 hour of cangrelor infusion during an inpatient admission to one of two hospitals within the health system. Baseline demographic and historical characteristics were captured via manual abstraction from electronic health record data, as were details of the hospitalization during which patients were treated with cangrelor. Data were collected for all variables included in the National Cardiovascular Data Registry’s post-PCI bleeding model, as well as the Mehran bleeding risk score.^9,10 Of note, oral antiplatelet medications are reported as being given simultaneously with cangrelor if they were administered within 48 hours of cangrelor initiation. Indication for cangrelor was classified as “bridge to procedure” if the patient was given cangrelor with intention of pursuing surgery or procedure and/or if they underwent surgical procedure; indication was classified as “other” if the patient was given cangrelor due to a non-surgical indication (eg, restricted to nothing by mouth [NPO] or without gastrointestinal [GI] access). Patients who were administered cangrelor as a bridge to surgical procedure followed our hospital’s bridging protocol, as adopted from BRIDGE, which included discontinuation of cangrelor infusion 1-6 hours prior to surgical incision (Table 1).

Authors BVJ and ERH performed chart abstraction and compared findings on randomly selected patients with no discrepancies in their findings. The study was approved by the institutional review board at the University of Pennsylvania Health System, which granted a waiver of informed consent.

Primary and secondary endpoints. The primary endpoint in this study was Bleeding Academic Research Consortium (BARC) 3-5 bleeding during cangrelor administration or within 48 hours after discontinuation.¹¹ Bleeding was assessed via review of physician and nursing notes, hemoglobin laboratory data, chest tube output, and transfusion records. Secondary outcomes included major and minor bleeding by Thrombolysis in Myocardial Infarction (TIMI) criteria,¹² moderate or life-threatening bleeding by Global Use of Strategies to Open Occluded Arteries (GUSTO) criteria,¹³ International Society on Thrombosis and Hemostasis (ISTH) severe or clinically relevant bleeding,^14,15 or BARC 2 bleeding. Unlike the BRIDGE trial, which considered major bleeding events through hospital discharge,⁴ bleeding was only considered to be associated with cangrelor administration if it occurred during cangrelor administration or up to 48 hours after discontinuation. We also report in-hospital mortality and thrombotic events, including stroke or myocardial infarction, during or shortly after cangrelor administration.

Statistical analysis. Patient characteristics are summarized with means and standard deviations for continuous variables and with total number and percentages for categorical data. Continuous variables were compared with the Student’s t-test, and categorical variables were compared using the Fisher’s exact test. Statistical analysis was conducted using Excel (Microsoft) and SPSS (IBM).

Between January 2017 and August 2020, a total of 31 inpatients received cangrelor for at least 1 hour outside of the cardiac catheterization laboratory and met the inclusion criteria for the presence of coronary or other arterial stents. Twenty-nine of 31 patients had coronary arterial stenting either during the current admission or within the previous 6 months, and the other 2 had presence of peripheral arterial stents in addition to coronary stents placed (>2 years prior). All patients were given cangrelor at a dose of 0.75 µg/kg/min consistent with the BRIDGE trial,⁴ with the exception of 1 patient who was initiated on 4 µg/kg/min during a complex PCI and was reduced to 0.75 µg/kg/min while strictly NPO after the procedure; there were no documented dosing errors. Oral P2Y₁₂ inhibitors were generally discontinued 5 days prior to planned surgeries, although were often discontinued 1-2 days prior to surgery in urgent or emergent situations. Median duration of infusion was 70 hours (minimum, 8 hours; 25th percentile, 43.9 hours; 75th percentile, 96 hours; maximum, 630 hours). Of the 31 patients, 16 (51.6%) had New York Heart Association (NYHA) class IV heart failure within 2 weeks of presentation, 11 (35.5%) were in cardiogenic shock on admission, 17 (54.8%) presented with acute coronary syndrome (ACS), and 4 (12.9%) were on extracorporeal membrane oxygenation (ECMO) at the time of cangrelor administration. In terms of indication, 22 of 31 patients were given cangrelor as bridge to a surgical procedure (10 CABG surgeries, 9 non-CABG cardiac surgeries, and 3 non-cardiac surgeries) and 9 were given cangrelor due to being strict NPO or having no GI access, categorized as “other.” Of note, 2 of the 22 patients in the bridge to surgery group did not undergo surgery; 1 patient met both criteria of being bridged to procedure and being strict NPO so was classified as bridge to procedure for summary purposes, and 1 was given cangrelor for physician preference to avoid long-acting P2Y₁₂ inhibitors due to a presumed high risk of bleeding, but required surgery and so was also classified as bridge to procedure. Table 2 displays baseline demographics, clinical characteristics, and indications for cangrelor of the patients included in the study.

Of the 31 patients in the study, 13 (41.9%) had BARC 3-5 bleeding during cangrelor infusion or within 48 hours of discontinuation and 7 (22.6%) died during hospitalization. Nine of the 21  patients (42.9%) who were bridged to a procedure exhibited BARC 3-5 bleeding peri- or postoperatively, including 3 of the 10 patients (30.0%) who underwent CABG. Four of the 9 patients (44.4%) who were in the “other” group had BARC 3-5 bleeding events. Of 4 patients on ECMO, 2 were bridge to non-CABG cardiac surgery and 2 were other, and all (100%) had BARC 3-5 bleeding. Table 3 summarizes bleeding by definition and death during hospitalization broken down by indication subgroups. Supplemental Table S1 displays bleeding scores by definition and death of all patients on ECMO by indication subgroups.

On univariable analyses, only ECMO was significantly associated with BARC 3-5 bleeding. Obesity was significantly associated with a lower incidence of bleeding. Time of cangrelor infusion (hours) trended toward significance. No other baseline demographic, historical, or clinical criteria were associated with the development of BARC 3-5 bleeding (Table 4).

No patient in the study developed in-stent thrombosis or other thrombotic complication while receiving cangrelor. However, 2 patients had thrombotic events within 2 hours of stopping therapy; 1 suffered a right posterior inferior cerebral artery ischemic stroke while undergoing planned intracardiac mass excision 3.5 months after receiving a coronary stent for ST-elevation myocardial infarction and 1 had in-stent thrombosis en route to the operating room for planned CABG.

In this retrospective study of patients who received cangrelor as part of an antiplatelet bridging strategy at a large academic medical center, over 40% of patients overall had BARC 3-5 bleeding, including 30% of CABG patients, a rate that is substantially higher than the 11.8% of patients with BARC 3-5 bleeding seen in the BRIDGE trial.⁴ In patients with no planned surgical intervention, 14.3% of non-ECMO patients and 100% of ECMO patients had BARC 3-5 bleeding. Overall, 80.6% of patients had at least 1 type of bleeding event, including minor, during or within 48 hours of discontinuing cangrelor.

Prior studies investigating cangrelor use have reported significant variability of major bleeding rates. The CHAMPION PHOENIX studies examined cangrelor use in PCI; they found patients to be at a higher risk of non-significant bleeding events (eg, hematoma) than the placebo group with no difference in major bleeding, and noted significantly reduced ischemic events during PCI in patients receiving cangrelor compared with placebo.^1-3 The BRIDGE trial prospectively examined 210 patients bridged with cangrelor infusion vs placebo prior to undergoing CABG surgery;⁴ notably, it found no significant difference in bleeding events between the cangrelor group and the placebo group, with approximately 10% of patients in both arms having bleeding events. Like most randomized controlled trials, BRIDGE enrolled a selected population, specifically excluding patients with GFR <30 mL/min; therefore, the applicability of these results to an all-comers population of patients treated with cangrelor as a bridging strategy is uncertain. Other studies have investigated cangrelor use in severely ill patients, primarily for those in cardiogenic shock, but not necessarily for the purpose of bridging. Vaduganathan et al found GUSTO mild or moderate bleeding in 11 of 38 patients (29%) with cardiogenic shock treated with cangrelor during the first 48 hours of infusion, but reported no major bleeding events.¹⁶ Droppa et al described 136 cardiogenic shock patients treated with cangrelor (± glycoprotein IIb/IIIa inhibitors) and found in-hospital GUSTO moderate and severe bleeding in 33 patients (24.6%); when 88 of these patients were matched to “control” patients from IABP-SHOCK II who did not receive cangrelor, patients treated with cangrelor did not exhibit statistically higher rates of GUSTO moderate or severe bleeding (21.6% in the cangrelor group vs 19.3% in the control group).¹⁷

Following the BRIDGE trial, 3 articles have summarized cangrelor use in bridging, with all studies reporting 100% efficacy of cangrelor (no thrombotic events during infusion) and variable bleeding complication rates.^6-8 One study focused exclusively on cangrelor as a bridge to LVAD implantation in 5 patients with cardiogenic shock and recent stent implantation, and found major postoperative bleeding events in 2 patients (40%), although no preoperative bleeding during the bridging period.⁶ The other 2 studies, like ours, focused on cangrelor use for bridging in real-world scenarios. Bowman et al examined 11 patients receiving cangrelor as bridge to various surgical procedures (3 CABG surgeries), and 3 patients (27.3%) were found to have suffered bleeding complications; while the study does not comment on major vs minor bleeding, a maximum of 1 of the 11 patients (9.1%) appeared to exhibit BARC 3 bleeding based on descriptions within the study.⁷ Stern et al described a relatively sick population of 22 patients (54% in cardiogenic shock) who received cangrelor bridging therapy to various surgeries (9 CABG); they reported no GUSTO major bleeding and no BARC 4 bleeding in CABG patients, but did report GUSTO moderate bleeding in 5 patients (22%).⁸ In summary, rates of significant bleeding events (BARC 3-5) observed in our study were higher than in prior studies, which reported rates of significant bleeding between 0%-11.8% in patients undergoing CABG or other cardiac surgery procedures,^4,7,8 40% significant bleeding in patients undergoing left ventricular assist device placement,⁶ and no significant bleeding in extrathoracic surgical procedures.^7,8 Compared with these prior observational real-world bridging studies, our study includes a greater number of patients, adjudicated bleeding events by multiple standardized bleeding definitions, including CABG-specific bleeding definitions, and included a cohort of patients with very serious illness, including 35.5% with cardiogenic shock, 12.9% with renal failure, and 12.9% on ECMO. For this reason, the rates of bleeding we report may better reflect bleeding risk among high-risk patients treated with cangrelor than those observed in prior bridging studies. This is further supported by relatively higher rates of significant in-hospital bleeding observed in severely ill patients by Droppa et al.¹⁷

Regarding surgical bleeding, although the BRIDGE trial demonstrated that platelet function should nearly normalize by the time of procedure when cangrelor is held 6 hours, it is possible that the duration may be longer in some patients or that other factors may have contributed to increased bleeding, including comorbidities or persistent effects from prior oral P2Y₁₂ inhibitors.

Patients treated with cangrelor who were not proceeding to surgery also had a relatively high bleeding rate during infusion. All 4 patients on ECMO who received cangrelor had BARC 3-5 bleeding. These high rates of observed significant bleeding suggest a need for careful evaluation of patient selection criteria for cangrelor use and/or possible evaluation of alternative dosing strategies in critically ill patients at high risk of bleeding. Use of ECMO was associated with a higher risk of BARC 3-5 bleeding, and obesity was associated with a lower effect. Although it did not reach statistical significance, a longer duration of cangrelor infusion was also associated with higher risk of bleeding. These findings suggest that careful consideration should be used when considering cangrelor use in patients on ECMO or with low body weight, and that duration of infusion should be as short as possible. Furthermore, as recommendations for duration of dual-antiplatelet therapy after coronary intervention in high bleeding risk patients continue to evolve, indications for cangrelor bridging after coronary intervention may become limited to those with very recent coronary stents.

Finally, the 2 thrombotic events that occurred in the study reported here did so within 2 hours of cessation of cangrelor. Early cessation prior to operative intervention is a risk for thrombotic events. Hospitals should have a clear strategy outlined for discontinuation of cangrelor infusion prior to surgical procedures to mitigate this risk, and should be prepared for the occurrence of thrombotic events.

Study limitations. Several limitations of this study may affect its interpretation. First, it was conducted within a single health system with a small study population, including just 10 patients undergoing CABG. This is considerably smaller than the BRIDGE trial population and even 1-2 bleeding events could therefore have a substantial impact on our data. This study is retrospective, so that documentation of bleeding was derived only from later chart review, which may impact data collection. This study did not include monitoring of platelet inhibition, so duration of cangrelor effects after discontinuation is unknown.

This retrospective study from a large, academic medical center demonstrated high rates of clinically significant bleeding in patients who were treated with parenteral cangrelor therapy during an antiplatelet bridging strategy, particularly in those patients undergoing cardiac surgery, both CABG and non-CABG, or treated with ECMO. While cangrelor for bridging therapy remains an important treatment approach, continued studies of patient selection and bleeding outcomes are warranted.

• Association of Iso-Osmolar vs. Low-Osmolar Contrast with Major Adverse Renal or Cardiovascular Events in High Risk Patients Undergoing Endovascular Abdominal Aortic Aneurysm Repair
• Percutaneous Treatment of Symptomatic Left Ventricular Assist Device Outflow Graft Obstruction
• Clinical Outcomes in Patients Treated With Biodegradable-Polymer Biolimus-Eluting Stents and 6 Months of Dual-Antiplatelet Therapy: The French eBiomatrix 6-Month DAPT Registry
• Comparison of Different Stenting Techniques in Left Main Bifurcation Disease: Evidence From a Network Meta-Analysis
• Use of the CardioHELP Device for Temporary Hemodynamic Support During High-Risk Percutaneous Coronary Intervention