- Email: [email protected]
Correlation Between Point-of-Care Platelet Function Testing and Bleeding After Coronary Angiography According to Two Different Definitions for Bleeding
Correlation Between Point-of-Care Platelet Function Testing and Bleeding After Coronary Angiography According to Two Different Definitions for Bleeding Manne Holm, MDa,b,*, Per Tornvall, MD, PhDc, Magnus Dalén, MDa,b, and Jan van der Linden, MD, PhDa,b Platelet function testing could be useful when assessing the risk for bleeding during treatment with antiplatelet drugs. This has been indicated in several studies, including the Antiplatelet Therapy for Reduction of Myocardial Damage During AngioplastyeBleeding (ARMYDABLEEDS) study, which demonstrated that testing with a point-of-care assay correlated with bleeding events after percutaneous coronary intervention. To standardize bleeding definitions, the Bleeding Academic Research Consortium (BARC) published a consensus report, which is in need of data-driven validation. Hence, the investigators conducted an observational, prospective, single-center study of 474 patients receiving clopidogrel and aspirin who underwent coronary angiography with or without percutaneous coronary intervention from October 2006 to May 2011. Platelet reactivity was measured with adenosine diphosphateeinduced singleplatelet function testing (Plateletworks) at the start of coronary angiography. The primary end point was the 30-day incidence of bleeding as defined by BARC and ARMYDA-BLEEDS. The aim of the present study was to investigate the relation between on-treatment platelet reactivity and the 30-day incidence of bleeding complications according to the BARC and ARMYDA-BLEEDS definitions. Patients in the first platelet aggregation quartile had a higher frequency of type 2 or higher BARC bleeding and ARMYDA-BLEEDS-defined bleeding <30 days after coronary angiography compared with the fourth quartile (16.9% vs 6.7%, p [ 0.014, and 8.5% vs 1.7%, p [ 0.016, respectively) and the third quartile (16.9% vs 7.7%, p [ 0.031, and 8.5% vs 2.6%, p [ 0.048, respectively). In conclusion, patients with low ontreatment platelet reactivity at the time of intervention had a significantly higher incidence of bleeding according to the BARC and ARMYDA-BLEEDS definitions <30 days after coronary angiography with or without percutaneous coronary intervention. Ó 2014 Elsevier Inc. All rights reserved. (Am J Cardiol 2014;114:1347e1353) The aim of this study was to investigate the relation between on-treatment platelet reactivity assessed with a point-of-care single-platelet function test and the 30-day incidence of bleeding complications, as defined by the Bleeding Academic Research Consortium (BARC)1 and Antiplatelet Therapy for Reduction of Myocardial Damage During AngioplastyeBleeding (ARMYDA-BLEEDS)2 definitions, after coronary angiography with and without percutaneous coronary intervention (PCI). Methods This observational prospective study initially included 491 patients who underwent coronary angiography with and a Department of Cardiothoracic Surgery and Anesthesiology, Karolinska University Hospital, Stockholm, Sweden; bDepartment of Molecular Medicine and Surgery, Karolinska Institutet; and cDepartment of Clinical Science and Education, Södersjukhuset, Karolinska Institutet, Stockholm, Sweden. Manuscript received June 14, 2014; revised manuscript received and accepted July 26, 2014. Financial support was provided by Karolinska Institute, Hjärt-Lungfonden, and through regional agreement on medical training and clinical research (ALF) between Stockholm County Council and the Karolinska Institute, Stockholm, Sweden. See page 1352 for disclosure information. *Corresponding author: Tel: þ46-8-51775189; fax: þ46-8-322701. E-mail address: [email protected] (M. Holm).
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2014.07.068
without PCI at the Karolinska University Hospital from October 2006 to May 2011. The indications for coronary angiography were acute coronary syndromes, stable angina, or chest pain with high suspicion of coronary origin. Sixteen patients were excluded because of treatment with glycoprotein IIb/IIIa inhibitors before platelet function testing, and 1 patient was excluded because of treatment with prasugrel, leaving a study cohort of 474 patients. All patients not previously receiving clopidogrel and/or aspirin treatment received a loading dose of clopidogrel (150 to 800 mg) in addition to aspirin (300- to 500-mg loading dose, followed by 75 mg/day) before coronary angiography. If PCI was performed, a daily maintenance dose of clopidogrel 75 mg was postprocedurally recommended in addition to aspirin for 1 year in patients receiving drug-eluting stents, whereas 3 months of dualantiplatelet treatment was recommended to patients receiving bare-metal stents. Patients already receiving clopidogrel treatment for >5 days before coronary angiography did not receive additional loading dose but continued with their daily maintenance dose (75 mg once daily). Ten patients were receiving warfarin treatment, which was discontinued 7 days before coronary angiography. All interventions were performed according to international guidelines.3,4 The femoral approach was used in all but 32 interventions, in which the radial approach was used, and unfractionated heparin was given in weight-adjusted doses www.ajconline.org
1348
The American Journal of Cardiology (www.ajconline.org)
Table 1 Demographic characteristics Characteristic Median [range] platelet aggregation Age (years) Women BMI (kg/m2) Hypertension† Diabetes mellitus Current smokers Prior myocardial infarction Prior percutaneous coronary intervention Prior coronary bypass NCDRÒ CathPCI Bleeding Risk5 Total score (points) Low risk (<25) Medium risk (26e65) High risk (>65) Medication Clopidogrel LD (mg) 150 300 450 600 800 Time (hours) from LD to coronary angiography Time (hours) from last dose to coronary angiography Clopidogrel maintenance treatment (months) Lipid-lowering drugs Proton pump inhibitors Fondaparinux Bivalirudin
1st Quartile (n ¼ 118)
2nd Quartile (n ¼ 119)
3rd Quartile (n ¼ 117)
4th Quartile (n ¼ 120)
29% [0e46%] 63.9 8.9 28 (24%) 26.3 5.0 65 (55%) 23 (19%) 19 (16%) 33 (28%) 25 (21%) 20 (17%)
63% [47e75%] 65.0 11.2 31 (26%) 26.9 3.9 62 (52%) 35 (29%) 24 (20%) 36 (30%) 20 (17%) 18 (15%)
84% [75e91%] 65.9 11.8 23 (20%) 27.3 5.4 58 (50%) 29 (25%) 27 (23%) 36 (31%) 29 (25%) 16 (14%)
96% [91e100%] 65.5 10.9 23 (19%) 28.7 4.4 73 (61%) 32 (27%) 21 (18%) 36 (30%) 26 (22%) 15 (13%)
0.20 0.39 <0.001 0.37 0.19 0.77 0.73 0.93 0.44
63.3 20.4 — 66 (56%) 52 (44%)
65.8 20.1 1 (1%) 64 (54%) 54 (45%)
64.3 21.3 5 (4%) 63 (54%) 49 (42%)
65.0 21.7 3 (3%) 64 (53%) 53 (44%)
0.63 0.25 0.69 0.99
— 51 (43%) — 22 (19%) — 47 [24/97] 4.5 [2.5/6.0] 3 [1/3] 102 (86%) 16 (14%) 43 (36%) —
1 (1%) 59 (50%) 1 (1%) 22 (18%) 1 (1%) 69 [35/122] 5.0 [3.0/6.0] 3 [0/3] 102 (86%) 23 (19%) 47 (39%) —
1 60 2 22
(1%) (51%) (2%) (19%) — 45 [22/100] 5.5 [3.5/6.5] 3 [0/3] 96 (82%) 21 (18%) 49 (42%) 2 (2%)
1 64 2 28
(1%) (53%) (17%) (23%) — 27 [16/52] 5.0 [3.0/7.0] 3 [1/3] 102 (85%) 16 (13%) 52 (43%) 1 (1%)
p-Value*
0.32 0.12 0.50 0.38 N/A <0.001 0.107 0.75 0.96 0.28 1.00
Data are expressed by number (percentage) for categorical variables, and as mean SD or median [25th/75th percentile] for continuous variables. Quartiles were established for the percentage of platelet aggregation measured by adenosine diphosphate-induced single-platelet aggregation. LD ¼ loading dose. * Comparison between quartile 1 and quartile 4. † Defined as documented and treated hypertension.
(50 to 100 IE/kg). The sheath size was 6Fr. A vascular closure device (Angio-Seal; St. Jude Medical, St. Paul, Minnesota) was used in 247 patients. A compression assist device (Femostop; St. Jude Medical) was used in the rest of the cohort and in patients with vascular closure devices when required for hemostasis. Use of periprocedural antiplatelet agents other than clopidogrel and aspirin, for example, glycoprotein IIb/IIIa inhibitors, was at the discretion of the interventionist. At the start of each coronary angiographic procedure, a 4-ml blood sample was drawn from the arterial line. Assessment of adenosine diphosphateeinduced platelet aggregation was performed by using single-platelet counting with the Plateletworks assay (Helena Laboratories, Beaumont, Texas). The test was always performed <10 minutes after blood sampling. The baseline platelet count was obtained by the addition of 1 ml whole blood to the first Plateletworks tube, primed with the synthetic anticoagulant ethylenediaminetetraacetic acid. One milliliter of whole blood was then added to the second Plateletworks tube, containing citrate and adenosine diphosphate (20 mmol), inducing platelet aggregation. For each tube, the platelet count was then measured with a cell counter (ABX Micros 60; Horiba ABX Diagnostics, Holliston, Massachusetts).
Because platelet aggregates exceed normal platelet size, it is possible for the cell counter to discriminate between aggregated and nonaggregated platelets on the basis of size. The difference in platelet count between the 2 samples was used as a measurement of platelet aggregation. A research nurse, who was well familiar with the testing equipment and had received training from the manufacturer, conducted all blood sampling and subsequent platelet function testing. The physicians responsible for the patients during their hospital stays were not aware of the platelet function test results. Written informed consent was obtained from all patients. The regional human research ethics committee in Stockholm, Sweden, approved the study. Data on in-hospital bleeding events were prospectively acquired, including location and extent, laboratory data, imaging data, medications, and treatment. The data on outof-hospital bleeding events that did not require direct visits to health care professionals were registered at later routine follow-up visits. The study database and patients’ medical records were reexamined for every bleeding event by 2 researchers blinded to platelet aggregation to classify them according to the bleeding definitions. The primary outcome was the 30-day incidence of bleeding complications after coronary angiography in relation to quartile distribution of
Coronary Artery Disease/Bleeding After Coronary Angiography
1349
Table 2 Procedural characteristics Characteristic Median [range] platelet aggregation Procedural characteristics Femoral approach Radial approach Vascular closing device PCI performed GP IIb/IIIa inhibitors Laboratory data Platelet count, 109/L Hemoglobin (g/L) Hematocrit Creatinine clearance (ml/min)† Diagnosis at discharge STEMI NSTEMI/unstable angina Stable angina Unspecific chest pain Otherz
1st Quartile (n ¼ 118)
2nd Quartile (n ¼ 119)
3rd Quartile (n ¼ 117)
4th Quartile (n ¼ 120)
29% [0e46%]
63% [47e75%]
84% [75e91%]
96% [91e100%]
114 4 58 70 14
(97%) (3%) (49%) (59%) (12%)
220.9 59.5 138.0 16.4 0.39 0.04 86.6 25.9 2 68 36 5 6
(2%) (58%) (31%) (4%) (5%)
112 7 57 56 9
(94%) (6%) (48%) (47%) (8%)
220.4 51.4 136.8 16.4 0.38 0.05 83.1 29.2 2 66 31 12 6
(2%) (55%) (26%) (10%) (5%)
108 9 64 64 16
(92%) (8%) (55%) (55%) (14%)
219.1 69.3 138.6 16.1 0.38 0.05 87.5 35.0 5 67 31 6 8
(4%) (57%) (26%) (5%) (7%)
110 10 68 75 14
(92%) (8%) (57%) (63%) (12%)
221.1 53.8 140.8 16.5 0.40 0.04 93.3 31.7 11 68 34 3 4
(9%) (57%) (38%) (3%) (3%)
p-Value*
0.11 0.11 0.25 0.62 0.96 0.98 0.19 0.16 0.11 0.01 0.88 0.71 0.50 0.54
Data are expressed by number (percentage) for categorical variables, or as mean SD for continuous variables. Quartiles were established for the percentage of platelet aggregation measured by adenosine diphosphate-induced single-platelet aggregation. GP ¼ glycoprotein; NSTEMI ¼ non ST-elevation myocardial infarction; PCI ¼ percutaneous coronary intervention; STEMI ¼ ST-elevation myocardial infarction. * Comparison between quartile 1 and quartile 4. † Estimated with Cockcroft-Gault equation. z Includes heart failure, aortic stenosis, arterio-ventricular block, and tachycardia.
on-treatment platelet reactivity measured by Plateletworks. Bleeding was defined according the criteria of BARC1 and of the ARMYDA-BLEEDS study (>10-cm hematoma, pseudoaneurysm, arteriovenous fistula, or major Thrombolysis In Myocardial Infarction [TIMI] bleeding criteria).2 The National Cardiovascular Data Registry (NCDR) CathPCI Registry model was used for retrospective evaluation of preprocedural risk for bleeding. This predictive risk score is based on the variables ST-segment elevation myocardial infarction, age, body mass index, previous PCI, chronic kidney disease, shock, cardiac arrest <24 hours, gender, hemoglobin level, and PCI status.5 After returning to a referring hospital, 1 patient died from momentary iatrogenic lung bleeding during a chest drainage procedure, which aimed to evacuate a pleural effusion. This bleeding event was not regarded as an outcome for this study, because according to the autopsy protocol, it was a direct result of a surgical trauma. A power calculation was done on the basis of the ARMYDA-BLEEDS study, which showed 10.1% and 1.3% incidence rates of bleeding in the first and fourth platelet aggregation quartiles, respectively. For the present study, assuming that the bleeding event rate would be similar, it would require 85 patients in each platelet aggregation quartile (a total of 340 patients) to show a statistical difference between the first and the fourth quartiles (p <0.05, 80% power). Patients were divided into quartiles according to adenosine diphosphateeinduced platelet aggregation. Normal distribution for continuous variables was tested with the Shapiro-Wilk test. Continuous variables are described as mean SD or medians with interquartile ranges (IQRs) and
were compared with Student’s t test when normally distributed and with the Mann-Whitney U test when not normally distributed. Categorical variables are described as percentages and were compared with either the chi-square test or Fisher’s exact test. The optimal cutoff for each bleeding definition was assessed with receiver-operating characteristic (ROC) curve. To identify possible variables associated with bleeding, an initial univariate Cox regression analysis was performed. Variables with p values <0.10 in the univariate analysis were considered for the stepwise forward and backward manual multivariate Cox regression analysis. The bleeding events in the different quartiles over time were visualized with Kaplan-Meier curves. Two-sided p values <0.05 were considered significant. All statistical analyses were performed with SPSS version 21.0 (SPSS, Inc., Chicago, Illinois). Results Detailed clinical and procedural patient data are listed in Table 1 and Table 2, including comparisons between the lowest and the highest platelet aggregation quartiles. As listed in Table 1, the different clopidogrel loading doses were evenly distributed among quartiles. Median time from clopidogrel loading dose to coronary angiography was higher in quartile 1 compared with quartile 4. At 30-day follow-up, data on bleeding were available for all patients, except for 1 patient who had moved abroad (0.2%). Three patients had gastrointestinal bleedings, 2 had bladder or urethral bleedings, 1 had an intracranial bleeding, 1 had an eye bleeding, and the remaining bleeding complications were entry-site bleedings. If a patient had >1 bleeding event
1350
The American Journal of Cardiology (www.ajconline.org)
Table 3 Quartile distribution of bleeding events
Median [range] of platelet aggregation BARC bleeding Type 1 Type 2 Type 3a 3b 3c Type 4 Type 5a 5b Type 1 Type 2 ARMYDA-BLEEDS defined bleeding Major TIMI bleeding Entry site complications >10 cm hematoma Pseudoaneurysm AV-fistula Combined end pointz
Total (n ¼ 474)
1st Quartile (n ¼ 118)
2nd Quartile (n ¼ 119)
3rd Quartile (n ¼ 117)
4th Quartile (n ¼ 120)
75% [0e100%]
29% [0e46%]
63% [47e75%]
84% [75e91%]
96% [91e100%]
76 39 5 5 1 1
(16%) (8%) (1%) (1%) (1%) (1%) — — 127 (27%) 51 (11%)
28 14 3 2
(24%) (12%) (3%) (2%) — 1 (1%) — — 48 (41%) 20 (17%)
20 12 1 1
(17%) (10%) (1%) (1%) — — — — 34 (29%) 14 (12%)
p-Value*
(14%) (4%) (1%) (2%) (1%) — — — 25 (21%) 9 (8%)†
12 (10%) 8 (7%) — — — — — — 20 (17%) 8 (7%)
<0.01 0.17 0.12 0.25 N/A 0.50 N/A N/A <0.001 0.014
16 5 1 2 1
6 (1%)
4 (3%)
1 (1%)
1 (1%)
—
0.06
14 (3%) 6 (1%) — 22 (5%)
5 (4%) 1 (1%) — 10 (8%)
4 (3%) 3 (3%) — 7 (6%)
3 (3%) — — 3 (3%)x
2 (2%) 2 (2%) — 2 (2%)
0.28 1.0 N/A 0.016
Data are expressed by number (percentage) for categorical variables, or as mean SD for continuous variables. Quartiles were established for the percentage of platelet aggregation measured by adenosine diphosphate-induced single-platelet aggregation. AV ¼ arterio-venous; TIMI ¼ Thrombolysis in Myocardial Infarction. * Comparison between quartile 1 and quartile 4. † p ¼ 0.031 (quartile 1 vs quartile 3). z >10 cm hematoma, pseudo-aneurysm, AV-fistula, or major TIMI bleeding (4 patients had 2 events each). x p ¼ 0.048 (quartile 1 vs quartile 3).
Figure 1. Frequency of type 2 BARC bleeding over time according to platelet aggregation quartiles.
within 30 days, the most extensive bleeding was included in the analysis. The incidence of any BARC bleeding (type 1) was 26.8% (127 of 474 patients; Table 3). Patients who experienced any BARC bleeding within 30 days had a higher median NCDR CathPCI risk score compared with patients without bleeding (70 [IQR 55 to 80] vs 60 [IQR 50 to 75], p ¼ 0.019). Median platelet aggregation was lower in patients
with any BARC bleeding compared with patients without bleeding (51.0% [IQR 32.8% to 84.0%] vs 79.1% [IQR 52.5% to 92.0%], p ¼ 0.001). Patients in the first quartile had a higher incidence of any BARC bleeding compared with the fourth quartile and the third quartile, as listed in Table 3. The incidence of type 2 BARC bleeding was 10.8% (51 of 474 patients; Table 3). Patients who experienced type 2 BARC bleeding within 30 days had a higher median NCDR CathPCI risk score compared with patients with type 0 or type 1 bleeding (70 [IQR 60 to 85] vs 65 [IQR 50 to 75], p ¼ 0.035). Median platelet aggregation was lower in patients with type 2 BARC bleeding compared with patients with type 0 or type 1 bleeding (59.9% [IQR 31.5% to 83.8%] vs 76.9% [IQR 48.0% to 91.1%], p ¼ 0.005). Patients in the first platelet aggregation quartile had a higher frequency of type 2 BARC bleeding within 30 days compared with the fourth quartile and the third quartile, as listed in Table 3. There was no significant difference in type 2 BARC bleeding between clopidogrel-naive patients and those already on a clopidogrel maintenance dose (10.7% vs 10.9%, p ¼ 0.932). The frequency of type 2 BARC bleeding complications over time is depicted in Figure 1. As listed in Table 4, multivariate Cox regression showed that platelet aggregation in the first quartile predicted increased bleeding risk for type 2 BARC bleeding within 30 days, when adjusted for NCDR risk score, concurrent antiplatelet treatment, and anticoagulant treatment. Moreover, the NCDR risk score significantly predicted type 2 BARC bleeding complications in the multivariate Cox regression model. The ROC curve analysis showed a significant correlation with type 2 BARC bleeding and platelet aggregation, with
Coronary Artery Disease/Bleeding After Coronary Angiography
1351
Table 4 Cox regression analysis Variable
Univariate Analysis OR (95% CI)
BARC type 2 bleeding Quartile 1 (platelet aggregation) NCDR risk score (10-point increase)* Fondaparinux Periprocedural GP IIb/IIIa inhibitors Diabetes Bilivarudin Angioseal ARMYDA-BLEEDS defined bleeding Quartile 1 (platelet aggregation) NCDR risk score (10-point increase)* Fondaparinux Periprocedural GP IIb/IIIa inhibitors Diabetes Bilivarudin Angioseal
Multivariate Analysis p-Value
OR (95% CI)
p-Value
1.99 1.17 1.23 1.26 0.54 0.05 1.11
(1.14e3.50) (1.03e1.33) (0.71e2.14) (0.57e2.79) (0.26e1.15) (0.0e2.01 106) (0.64e1.93)
0.016 0.014 0.457 0.572 0.112 0.698 0.71
2.08 1.18 1.04 1.19
(1.19e3.66) (1.03e1.35) (0.59e1.83) (0.54e2.65) — — —
0.011 0.014 0.902 0.666 — — —
2.55 1.01 1.02 1.77 0.47 0.05 0.63
(1.10e5.89) (0.83e1.24) (0.44e2.40) (0.60e5.22) (0.14e1.58) (0.0e7.1 108) (0.27e1.46)
0.034 0.900 0.956 0.303 0.220 0.801 0.28
2.54 1.02 1.02 1.74
(1.10e5.90) (0.82e1.26) (0.42e2.50) (0.59e5.14) — — —
0.029 0.868 0.964 0.318 — — —
* The NCDR Cath PCI Bleeding Risk Score5 is based on the following variables: STEMI, age, BMI, Previous PCI, chronic kidney disease, shock, cardiac arrest <24 h, gender, hemoglobin levels, and PCI status.
Figure 2. Frequency of ARMYDA-BLEEDS-defined bleeding over time according to platelet aggregation quartiles.
an area under the curve of 0.62 (95% confidence interval [CI] 0.54 to 0.70, p ¼ 0.005). The optimal platelet aggregation cutoff was 76.7%, with 71% sensitivity and 51% specificity. The incidence of type 2 BARC bleeding within 30 days was 14.7% (36 of 245 patients) in patients with 76.7% platelet aggregation and 6.6% (15 of 229 patients) in patients with >76.7% platelet aggregation (relative risk 2.2, 95% CI 1.3 to 4.0, p ¼ 0.004). As listed in Table 3, the incidence of ARMYDABLEEDS-defined bleeding was 4.6% (22 of 474 patients). Patients who experienced bleeding according to the ARMYDA-BLEEDS-defined bleeding end point within 30 days had an equal median NCDR CathPCI risk score compared with patients without bleeding (65 [IQR 60 to 76.25] vs 65 [IQR 50 to 80], p ¼ 1.00). Median platelet
aggregation was lower in patients with ARMYDABLEEDS-defined bleeding compared with patients without bleeding (50.1% [IQR 31.8% to 75.3%] vs 76.4% [IQR 47.0% to 91.0%], p ¼ 0.016). Patients in the first platelet aggregation quartile had a higher frequency of ARMYDABLEEDS-defined bleeding within 30 days compared with the third quartile and the fourth quartile, as listed in Table 3. There was no significant difference in ARMYDA-BLEEDSdefined bleeding between clopidogrel-naive patients and those already receiving clopidogrel maintenance doses (4.2% vs 5.8%, p ¼ 0.429). The frequency of ARMYDA-BLEEDS-defined bleeding events over time is presented in Figure 2. As listed in Table 4, multivariate Cox regression showed that platelet aggregation in the first quartile predicted increased risk for ARMYDA-BLEEDS-defined bleeding within 30 days, when adjusted for the other included variables. The ROC curve analysis showed a significant correlation between patients with and without ARMYDA-BLEEDSdefined bleeding according to platelet aggregation, with an area under the curve of 0.65 (95% CI 0.54 to 0.76, p ¼ 0.016). The optimal platelet aggregation cutoff was 74.8%, with 77% sensitivity and 52% specificity. The incidence of ARMYDABLEEDS-defined bleeding within 30 days was 7.3% (17 of 233 patients) in patients with 74.8% platelet aggregation and 2.1% (5 of 241 patients) in patients with >74.8% platelet aggregation (relative risk 3.5, 95% CI 1.3 to 9.4, p ¼ 0.007). Discussion This study indicates that clopidogrel-treated patients with low on-treatment platelet reactivity, according to a rapid point-of-care single-platelet function test at the time of intervention, have a significantly higher incidence of bleeding defined according to BARC1 and ARMYDABLEEDS2 <30 days after coronary angiography with and without PCI.
1352
The American Journal of Cardiology (www.ajconline.org)
The relation between BARC bleeding criteria and ontreatment platelet reactivity has not previously been studied in patients who undergo PCI.6 We decided to focus on type 2 BARC bleeding, as the US Food and Drug Administration supports bleeding definitions that reflect clinical outcome.7 Indeed, a relation between mortality and type 2 BARC bleeding has been shown in patients who undergo PCI.8 Also, a recent study by Matic et al9 indicated a hierarchical relation between BARC bleeding and mortality. The present study indicated that low platelet aggregation was nonlinearly correlated with type 2 BARC bleeding. Type 1 BARC bleeding is usually considered clinically unimportant, although these events have been associated with premature drug cessation, which may be associated with impaired clinical outcomes.10 Therefore, the association between the first platelet aggregation quartile and type 1 BARC bleeding in the present study may also be of clinical value. We used the NCDR CathPCI risk score5 to assess any differences in bleeding risk between platelet aggregation quartiles. This bleeding risk score was very similar among the different quartiles (Table 1) but was slightly higher in patients who experienced type 2 BARC bleeding. The multivariate Cox regression confirmed this relation, when adjusted for the other included variables (Table 4). However, the first platelet aggregation quartile was predictive for type 2 BARC bleeding, when adjusted for the NCDR CathPCI risk score. This may indicate that rapid platelet aggregation testing is of additional predictive value when assessing the risk for postprocedural coronary angiography bleeding. Nevertheless, the predictive value of the Plateletworks assay alone in the ROC curve analysis was relatively modest, with low sensitivity and specificity. The incidence of bleeding complications was 4.8% in the ARMYDA-BLEEDS study.2 Using the ARMYDA-BLEEDS combined end point as a bleeding definition, our study showed a similar incidence of 4.6%. Furthermore, the correlation between low platelet aggregation and bleeding described in the ARMYDA-BLEEDS study was also demonstrated in our study. The difference between studies in postprocedural coronary angiographic bleeding rates is highly dependent on the bleeding definition used.1 This was evident in the present study, in which the incidences of type 1 BARC bleeding and type 2 BARC bleeding were approximately 6 and 2 times, respectively, the incidence of ARMYDA-BLEEDS-defined bleeding. This may indicate a higher sensitivity of the BARC bleeding definitions compared with the ARMYDABLEEDS definition of bleeding. As shown in Figure 1 and Figure 2, most ARMYDABLEEDS-defined bleedings and type 2 BARC bleedings occurred within the first week after coronary angiography. This is an interesting finding, which, if it proves repeatable, may be of clinical value. The cause of the event plateau around day 5 to 6 is not clear, but it may indicate a relatively high proportion of procedural events, which occur shortly after the procedures. In the present study, the results obtained with the Plateletworks assay were not validated with another platelet function test, such as the reference standard method for platelet function testing, light transmittance aggregometry.
However, the Plateletworks assay has been shown to correlate well with light transmittance aggregometry in the prediction of adverse cardiovascular events.11,12 In contrast, a relation between platelet function testing with the Plateletworks assay and bleeding has not previously been shown,11 possibly because of less sensitive bleeding definitions. The present relatively small study found such a correlation, which adds valuable information on the usefulness of the Plateletworks assay as a method to identify patients receiving clopidogrel at increased risk for postprocedural coronary angiographic bleeding complications according to two relevant bleeding definitions. Acknowledgments: We thank Kristina Kilsand for data collection and excellent technical assistance and Associate Professor Ulrik Sartipy for valuable advice.
Disclosures The authors have no conflicts of interest to disclose. 1. Mehran R, Rao SV, Bhatt DL, Gibson CM, Caixeta A, Eikelboom J, Kaul S, Wiviott SD, Menon V, Nikolsky E, Serebruany V, Valgimigli M, Vranckx P, Taggart D, Sabik JF, Cutlip DE, Krucoff MW, Ohman EM, Steg PG, White H. Standardized bleeding definitions for cardiovascular clinical trials: a consensus report from the Bleeding Academic Research Consortium. Circulation 2011;123:2736e2747. 2. Patti G, Pasceri V, Vizzi V, Ricottini E, Di Sciascio G. Usefulness of platelet response to clopidogrel by point-of-care testing to predict bleeding outcomes in patients undergoing percutaneous coronary intervention (from the Antiplatelet Therapy for Reduction of Myocardial Damage During Angioplasty-Bleeding Study). Am J Cardiol 2011;107:995e1000. 3. Smith SC Jr, Feldman TE, Hirshfeld JW Jr, Jacobs AK, Kern MJ, King SB 3rd, Morrison DA, O’Neill WW, Schaff HV, Whitlow PL, Williams DO, Antman EM, Adams CD, Anderson JL, Faxon DP, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Nishimura R, Ornato JP, Page RL, Riegel B. ACC/AHA/SCAI 2005 guideline update for percutaneous coronary interventionesummary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/SCAI Writing Committee to Update the 2001 Guidelines for Percutaneous Coronary Intervention). Catheter Cardiovasc Interv 2006;67:87e112. 4. Kushner FG, Hand M, Smith SC Jr, King SB 3rd, Anderson JL, Antman EM, Bailey SR, Bates ER, Blankenship JC, Casey DE Jr, Green LA, Jacobs AK, Hochman JS, Krumholz HM, Morrison DA, Ornato JP, Pearle DL, Peterson ED, Sloan MA, Whitlow PL, Williams DO. 2009 focused updates: ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction (updating the 2004 guideline and 2007 focused update) and ACC/ AHA/SCAI guidelines on percutaneous coronary intervention (updating the 2005 guideline and 2007 focused update): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Catheter Cardiovasc Interv 2009;74:E25eE68. 5. Rao SV, McCoy LA, Spertus JA, Krone RJ, Singh M, Fitzgerald S, Peterson ED. An updated bleeding model to predict the risk of postprocedure bleeding among patients undergoing percutaneous coronary intervention: a report using an expanded bleeding definition from the National Cardiovascular Data Registry CathPCI Registry. JACC Cardiovasc Interv 2013;6:897e904. 6. Tantry US, Bonello L, Aradi D, Price MJ, Jeong YH, Angiolillo DJ, Stone GW, Curzen N, Geisler T, Ten Berg J, Kirtane A, Siller-Matula J, Mahla E, Becker RC, Bhatt DL, Waksman R, Rao SV, Alexopoulos D, Marcucci R, Reny JL, Trenk D, Sibbing D, Gurbel PA; Working Group on OnTreatment Platelet Reactivity. Consensus and update on the definition of on-treatment platelet reactivity to adenosine diphosphate associated with ischemia and bleeding. J Am Coll Cardiol 2013;62:2261e2273.
Coronary Artery Disease/Bleeding After Coronary Angiography 7. Hicks KA, Stockbridge NL, Targum SL, Temple RJ. Bleeding Academic Research Consortium consensus report: the Food and Drug Administration perspective. Circulation 2011:2664e2665. 8. Ndrepepa G, Schuster T, Hadamitzky M, Byrne RA, Mehilli J, Neumann FJ, Richardt G, Schulz S, Laugwitz KL, Massberg S, Schomig A, Kastrati A. Validation of the Bleeding Academic Research Consortium definition of bleeding in patients with coronary artery disease undergoing percutaneous coronary intervention. Circulation 2012;125: 1424e1431. 9. Matic DM, Milasinovic DG, Asanin MR, Mrdovic IB, Marinkovic JM, Kocev NI, Marjanovic MM, Antonijevic NM, Vukcevic VD, Savic LZ, Zivkovic MN, Mehmedbegovic ZH, Dedovic VM, Stankovic GR. Prognostic implications of bleeding measured by Bleeding Academic Research Consortium (BARC) categorisation in patients undergoing primary percutaneous coronary intervention. Heart 2014;100:146e152.
1353
10. Roy P, Bonello L, Torguson R, de Labriolle A, Lemesle G, Slottow TL, Steinberg DH, Kaneshige K, Xue Z, Satler LF, Kent KM, Suddath WO, Pichard AD, Lindsay J, Waksman R. Impact of “nuisance” bleeding on clopidogrel compliance in patients undergoing intracoronary drugeluting stent implantation. Am J Cardiol 2008;102:1614e1617. 11. Breet NJ, van Werkum JW, Bouman HJ, Kelder JC, Ruven HJ, Bal ET, Deneer VH, Harmsze AM, van der Heyden JA, Rensing BJ, Suttorp MJ, Hackeng CM, ten Berg JM. Comparison of platelet function tests in predicting clinical outcome in patients undergoing coronary stent implantation. JAMA 2010;303:754e762. 12. van Werkum JW, Kleibeuker M, Postma S, Bouman HJ, Elsenberg E, ten Berg JM, Hackeng CM. A comparison between the Plateletworks (TM)-assay and light transmittance aggregometry for monitoring the inhibitory effects of clopidogrel. Int J Cardiol 2010;140: 123e126.
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2014.07.068
without PCI at the Karolinska University Hospital from October 2006 to May 2011. The indications for coronary angiography were acute coronary syndromes, stable angina, or chest pain with high suspicion of coronary origin. Sixteen patients were excluded because of treatment with glycoprotein IIb/IIIa inhibitors before platelet function testing, and 1 patient was excluded because of treatment with prasugrel, leaving a study cohort of 474 patients. All patients not previously receiving clopidogrel and/or aspirin treatment received a loading dose of clopidogrel (150 to 800 mg) in addition to aspirin (300- to 500-mg loading dose, followed by 75 mg/day) before coronary angiography. If PCI was performed, a daily maintenance dose of clopidogrel 75 mg was postprocedurally recommended in addition to aspirin for 1 year in patients receiving drug-eluting stents, whereas 3 months of dualantiplatelet treatment was recommended to patients receiving bare-metal stents. Patients already receiving clopidogrel treatment for >5 days before coronary angiography did not receive additional loading dose but continued with their daily maintenance dose (75 mg once daily). Ten patients were receiving warfarin treatment, which was discontinued 7 days before coronary angiography. All interventions were performed according to international guidelines.3,4 The femoral approach was used in all but 32 interventions, in which the radial approach was used, and unfractionated heparin was given in weight-adjusted doses www.ajconline.org
1348
The American Journal of Cardiology (www.ajconline.org)
Table 1 Demographic characteristics Characteristic Median [range] platelet aggregation Age (years) Women BMI (kg/m2) Hypertension† Diabetes mellitus Current smokers Prior myocardial infarction Prior percutaneous coronary intervention Prior coronary bypass NCDRÒ CathPCI Bleeding Risk5 Total score (points) Low risk (<25) Medium risk (26e65) High risk (>65) Medication Clopidogrel LD (mg) 150 300 450 600 800 Time (hours) from LD to coronary angiography Time (hours) from last dose to coronary angiography Clopidogrel maintenance treatment (months) Lipid-lowering drugs Proton pump inhibitors Fondaparinux Bivalirudin
1st Quartile (n ¼ 118)
2nd Quartile (n ¼ 119)
3rd Quartile (n ¼ 117)
4th Quartile (n ¼ 120)
29% [0e46%] 63.9 8.9 28 (24%) 26.3 5.0 65 (55%) 23 (19%) 19 (16%) 33 (28%) 25 (21%) 20 (17%)
63% [47e75%] 65.0 11.2 31 (26%) 26.9 3.9 62 (52%) 35 (29%) 24 (20%) 36 (30%) 20 (17%) 18 (15%)
84% [75e91%] 65.9 11.8 23 (20%) 27.3 5.4 58 (50%) 29 (25%) 27 (23%) 36 (31%) 29 (25%) 16 (14%)
96% [91e100%] 65.5 10.9 23 (19%) 28.7 4.4 73 (61%) 32 (27%) 21 (18%) 36 (30%) 26 (22%) 15 (13%)
0.20 0.39 <0.001 0.37 0.19 0.77 0.73 0.93 0.44
63.3 20.4 — 66 (56%) 52 (44%)
65.8 20.1 1 (1%) 64 (54%) 54 (45%)
64.3 21.3 5 (4%) 63 (54%) 49 (42%)
65.0 21.7 3 (3%) 64 (53%) 53 (44%)
0.63 0.25 0.69 0.99
— 51 (43%) — 22 (19%) — 47 [24/97] 4.5 [2.5/6.0] 3 [1/3] 102 (86%) 16 (14%) 43 (36%) —
1 (1%) 59 (50%) 1 (1%) 22 (18%) 1 (1%) 69 [35/122] 5.0 [3.0/6.0] 3 [0/3] 102 (86%) 23 (19%) 47 (39%) —
1 60 2 22
(1%) (51%) (2%) (19%) — 45 [22/100] 5.5 [3.5/6.5] 3 [0/3] 96 (82%) 21 (18%) 49 (42%) 2 (2%)
1 64 2 28
(1%) (53%) (17%) (23%) — 27 [16/52] 5.0 [3.0/7.0] 3 [1/3] 102 (85%) 16 (13%) 52 (43%) 1 (1%)
p-Value*
0.32 0.12 0.50 0.38 N/A <0.001 0.107 0.75 0.96 0.28 1.00
Data are expressed by number (percentage) for categorical variables, and as mean SD or median [25th/75th percentile] for continuous variables. Quartiles were established for the percentage of platelet aggregation measured by adenosine diphosphate-induced single-platelet aggregation. LD ¼ loading dose. * Comparison between quartile 1 and quartile 4. † Defined as documented and treated hypertension.
(50 to 100 IE/kg). The sheath size was 6Fr. A vascular closure device (Angio-Seal; St. Jude Medical, St. Paul, Minnesota) was used in 247 patients. A compression assist device (Femostop; St. Jude Medical) was used in the rest of the cohort and in patients with vascular closure devices when required for hemostasis. Use of periprocedural antiplatelet agents other than clopidogrel and aspirin, for example, glycoprotein IIb/IIIa inhibitors, was at the discretion of the interventionist. At the start of each coronary angiographic procedure, a 4-ml blood sample was drawn from the arterial line. Assessment of adenosine diphosphateeinduced platelet aggregation was performed by using single-platelet counting with the Plateletworks assay (Helena Laboratories, Beaumont, Texas). The test was always performed <10 minutes after blood sampling. The baseline platelet count was obtained by the addition of 1 ml whole blood to the first Plateletworks tube, primed with the synthetic anticoagulant ethylenediaminetetraacetic acid. One milliliter of whole blood was then added to the second Plateletworks tube, containing citrate and adenosine diphosphate (20 mmol), inducing platelet aggregation. For each tube, the platelet count was then measured with a cell counter (ABX Micros 60; Horiba ABX Diagnostics, Holliston, Massachusetts).
Because platelet aggregates exceed normal platelet size, it is possible for the cell counter to discriminate between aggregated and nonaggregated platelets on the basis of size. The difference in platelet count between the 2 samples was used as a measurement of platelet aggregation. A research nurse, who was well familiar with the testing equipment and had received training from the manufacturer, conducted all blood sampling and subsequent platelet function testing. The physicians responsible for the patients during their hospital stays were not aware of the platelet function test results. Written informed consent was obtained from all patients. The regional human research ethics committee in Stockholm, Sweden, approved the study. Data on in-hospital bleeding events were prospectively acquired, including location and extent, laboratory data, imaging data, medications, and treatment. The data on outof-hospital bleeding events that did not require direct visits to health care professionals were registered at later routine follow-up visits. The study database and patients’ medical records were reexamined for every bleeding event by 2 researchers blinded to platelet aggregation to classify them according to the bleeding definitions. The primary outcome was the 30-day incidence of bleeding complications after coronary angiography in relation to quartile distribution of
Coronary Artery Disease/Bleeding After Coronary Angiography
1349
Table 2 Procedural characteristics Characteristic Median [range] platelet aggregation Procedural characteristics Femoral approach Radial approach Vascular closing device PCI performed GP IIb/IIIa inhibitors Laboratory data Platelet count, 109/L Hemoglobin (g/L) Hematocrit Creatinine clearance (ml/min)† Diagnosis at discharge STEMI NSTEMI/unstable angina Stable angina Unspecific chest pain Otherz
1st Quartile (n ¼ 118)
2nd Quartile (n ¼ 119)
3rd Quartile (n ¼ 117)
4th Quartile (n ¼ 120)
29% [0e46%]
63% [47e75%]
84% [75e91%]
96% [91e100%]
114 4 58 70 14
(97%) (3%) (49%) (59%) (12%)
220.9 59.5 138.0 16.4 0.39 0.04 86.6 25.9 2 68 36 5 6
(2%) (58%) (31%) (4%) (5%)
112 7 57 56 9
(94%) (6%) (48%) (47%) (8%)
220.4 51.4 136.8 16.4 0.38 0.05 83.1 29.2 2 66 31 12 6
(2%) (55%) (26%) (10%) (5%)
108 9 64 64 16
(92%) (8%) (55%) (55%) (14%)
219.1 69.3 138.6 16.1 0.38 0.05 87.5 35.0 5 67 31 6 8
(4%) (57%) (26%) (5%) (7%)
110 10 68 75 14
(92%) (8%) (57%) (63%) (12%)
221.1 53.8 140.8 16.5 0.40 0.04 93.3 31.7 11 68 34 3 4
(9%) (57%) (38%) (3%) (3%)
p-Value*
0.11 0.11 0.25 0.62 0.96 0.98 0.19 0.16 0.11 0.01 0.88 0.71 0.50 0.54
Data are expressed by number (percentage) for categorical variables, or as mean SD for continuous variables. Quartiles were established for the percentage of platelet aggregation measured by adenosine diphosphate-induced single-platelet aggregation. GP ¼ glycoprotein; NSTEMI ¼ non ST-elevation myocardial infarction; PCI ¼ percutaneous coronary intervention; STEMI ¼ ST-elevation myocardial infarction. * Comparison between quartile 1 and quartile 4. † Estimated with Cockcroft-Gault equation. z Includes heart failure, aortic stenosis, arterio-ventricular block, and tachycardia.
on-treatment platelet reactivity measured by Plateletworks. Bleeding was defined according the criteria of BARC1 and of the ARMYDA-BLEEDS study (>10-cm hematoma, pseudoaneurysm, arteriovenous fistula, or major Thrombolysis In Myocardial Infarction [TIMI] bleeding criteria).2 The National Cardiovascular Data Registry (NCDR) CathPCI Registry model was used for retrospective evaluation of preprocedural risk for bleeding. This predictive risk score is based on the variables ST-segment elevation myocardial infarction, age, body mass index, previous PCI, chronic kidney disease, shock, cardiac arrest <24 hours, gender, hemoglobin level, and PCI status.5 After returning to a referring hospital, 1 patient died from momentary iatrogenic lung bleeding during a chest drainage procedure, which aimed to evacuate a pleural effusion. This bleeding event was not regarded as an outcome for this study, because according to the autopsy protocol, it was a direct result of a surgical trauma. A power calculation was done on the basis of the ARMYDA-BLEEDS study, which showed 10.1% and 1.3% incidence rates of bleeding in the first and fourth platelet aggregation quartiles, respectively. For the present study, assuming that the bleeding event rate would be similar, it would require 85 patients in each platelet aggregation quartile (a total of 340 patients) to show a statistical difference between the first and the fourth quartiles (p <0.05, 80% power). Patients were divided into quartiles according to adenosine diphosphateeinduced platelet aggregation. Normal distribution for continuous variables was tested with the Shapiro-Wilk test. Continuous variables are described as mean SD or medians with interquartile ranges (IQRs) and
were compared with Student’s t test when normally distributed and with the Mann-Whitney U test when not normally distributed. Categorical variables are described as percentages and were compared with either the chi-square test or Fisher’s exact test. The optimal cutoff for each bleeding definition was assessed with receiver-operating characteristic (ROC) curve. To identify possible variables associated with bleeding, an initial univariate Cox regression analysis was performed. Variables with p values <0.10 in the univariate analysis were considered for the stepwise forward and backward manual multivariate Cox regression analysis. The bleeding events in the different quartiles over time were visualized with Kaplan-Meier curves. Two-sided p values <0.05 were considered significant. All statistical analyses were performed with SPSS version 21.0 (SPSS, Inc., Chicago, Illinois). Results Detailed clinical and procedural patient data are listed in Table 1 and Table 2, including comparisons between the lowest and the highest platelet aggregation quartiles. As listed in Table 1, the different clopidogrel loading doses were evenly distributed among quartiles. Median time from clopidogrel loading dose to coronary angiography was higher in quartile 1 compared with quartile 4. At 30-day follow-up, data on bleeding were available for all patients, except for 1 patient who had moved abroad (0.2%). Three patients had gastrointestinal bleedings, 2 had bladder or urethral bleedings, 1 had an intracranial bleeding, 1 had an eye bleeding, and the remaining bleeding complications were entry-site bleedings. If a patient had >1 bleeding event
1350
The American Journal of Cardiology (www.ajconline.org)
Table 3 Quartile distribution of bleeding events
Median [range] of platelet aggregation BARC bleeding Type 1 Type 2 Type 3a 3b 3c Type 4 Type 5a 5b Type 1 Type 2 ARMYDA-BLEEDS defined bleeding Major TIMI bleeding Entry site complications >10 cm hematoma Pseudoaneurysm AV-fistula Combined end pointz
Total (n ¼ 474)
1st Quartile (n ¼ 118)
2nd Quartile (n ¼ 119)
3rd Quartile (n ¼ 117)
4th Quartile (n ¼ 120)
75% [0e100%]
29% [0e46%]
63% [47e75%]
84% [75e91%]
96% [91e100%]
76 39 5 5 1 1
(16%) (8%) (1%) (1%) (1%) (1%) — — 127 (27%) 51 (11%)
28 14 3 2
(24%) (12%) (3%) (2%) — 1 (1%) — — 48 (41%) 20 (17%)
20 12 1 1
(17%) (10%) (1%) (1%) — — — — 34 (29%) 14 (12%)
p-Value*
(14%) (4%) (1%) (2%) (1%) — — — 25 (21%) 9 (8%)†
12 (10%) 8 (7%) — — — — — — 20 (17%) 8 (7%)
<0.01 0.17 0.12 0.25 N/A 0.50 N/A N/A <0.001 0.014
16 5 1 2 1
6 (1%)
4 (3%)
1 (1%)
1 (1%)
—
0.06
14 (3%) 6 (1%) — 22 (5%)
5 (4%) 1 (1%) — 10 (8%)
4 (3%) 3 (3%) — 7 (6%)
3 (3%) — — 3 (3%)x
2 (2%) 2 (2%) — 2 (2%)
0.28 1.0 N/A 0.016
Data are expressed by number (percentage) for categorical variables, or as mean SD for continuous variables. Quartiles were established for the percentage of platelet aggregation measured by adenosine diphosphate-induced single-platelet aggregation. AV ¼ arterio-venous; TIMI ¼ Thrombolysis in Myocardial Infarction. * Comparison between quartile 1 and quartile 4. † p ¼ 0.031 (quartile 1 vs quartile 3). z >10 cm hematoma, pseudo-aneurysm, AV-fistula, or major TIMI bleeding (4 patients had 2 events each). x p ¼ 0.048 (quartile 1 vs quartile 3).
Figure 1. Frequency of type 2 BARC bleeding over time according to platelet aggregation quartiles.
within 30 days, the most extensive bleeding was included in the analysis. The incidence of any BARC bleeding (type 1) was 26.8% (127 of 474 patients; Table 3). Patients who experienced any BARC bleeding within 30 days had a higher median NCDR CathPCI risk score compared with patients without bleeding (70 [IQR 55 to 80] vs 60 [IQR 50 to 75], p ¼ 0.019). Median platelet aggregation was lower in patients
with any BARC bleeding compared with patients without bleeding (51.0% [IQR 32.8% to 84.0%] vs 79.1% [IQR 52.5% to 92.0%], p ¼ 0.001). Patients in the first quartile had a higher incidence of any BARC bleeding compared with the fourth quartile and the third quartile, as listed in Table 3. The incidence of type 2 BARC bleeding was 10.8% (51 of 474 patients; Table 3). Patients who experienced type 2 BARC bleeding within 30 days had a higher median NCDR CathPCI risk score compared with patients with type 0 or type 1 bleeding (70 [IQR 60 to 85] vs 65 [IQR 50 to 75], p ¼ 0.035). Median platelet aggregation was lower in patients with type 2 BARC bleeding compared with patients with type 0 or type 1 bleeding (59.9% [IQR 31.5% to 83.8%] vs 76.9% [IQR 48.0% to 91.1%], p ¼ 0.005). Patients in the first platelet aggregation quartile had a higher frequency of type 2 BARC bleeding within 30 days compared with the fourth quartile and the third quartile, as listed in Table 3. There was no significant difference in type 2 BARC bleeding between clopidogrel-naive patients and those already on a clopidogrel maintenance dose (10.7% vs 10.9%, p ¼ 0.932). The frequency of type 2 BARC bleeding complications over time is depicted in Figure 1. As listed in Table 4, multivariate Cox regression showed that platelet aggregation in the first quartile predicted increased bleeding risk for type 2 BARC bleeding within 30 days, when adjusted for NCDR risk score, concurrent antiplatelet treatment, and anticoagulant treatment. Moreover, the NCDR risk score significantly predicted type 2 BARC bleeding complications in the multivariate Cox regression model. The ROC curve analysis showed a significant correlation with type 2 BARC bleeding and platelet aggregation, with
Coronary Artery Disease/Bleeding After Coronary Angiography
1351
Table 4 Cox regression analysis Variable
Univariate Analysis OR (95% CI)
BARC type 2 bleeding Quartile 1 (platelet aggregation) NCDR risk score (10-point increase)* Fondaparinux Periprocedural GP IIb/IIIa inhibitors Diabetes Bilivarudin Angioseal ARMYDA-BLEEDS defined bleeding Quartile 1 (platelet aggregation) NCDR risk score (10-point increase)* Fondaparinux Periprocedural GP IIb/IIIa inhibitors Diabetes Bilivarudin Angioseal
Multivariate Analysis p-Value
OR (95% CI)
p-Value
1.99 1.17 1.23 1.26 0.54 0.05 1.11
(1.14e3.50) (1.03e1.33) (0.71e2.14) (0.57e2.79) (0.26e1.15) (0.0e2.01 106) (0.64e1.93)
0.016 0.014 0.457 0.572 0.112 0.698 0.71
2.08 1.18 1.04 1.19
(1.19e3.66) (1.03e1.35) (0.59e1.83) (0.54e2.65) — — —
0.011 0.014 0.902 0.666 — — —
2.55 1.01 1.02 1.77 0.47 0.05 0.63
(1.10e5.89) (0.83e1.24) (0.44e2.40) (0.60e5.22) (0.14e1.58) (0.0e7.1 108) (0.27e1.46)
0.034 0.900 0.956 0.303 0.220 0.801 0.28
2.54 1.02 1.02 1.74
(1.10e5.90) (0.82e1.26) (0.42e2.50) (0.59e5.14) — — —
0.029 0.868 0.964 0.318 — — —
* The NCDR Cath PCI Bleeding Risk Score5 is based on the following variables: STEMI, age, BMI, Previous PCI, chronic kidney disease, shock, cardiac arrest <24 h, gender, hemoglobin levels, and PCI status.
Figure 2. Frequency of ARMYDA-BLEEDS-defined bleeding over time according to platelet aggregation quartiles.
an area under the curve of 0.62 (95% confidence interval [CI] 0.54 to 0.70, p ¼ 0.005). The optimal platelet aggregation cutoff was 76.7%, with 71% sensitivity and 51% specificity. The incidence of type 2 BARC bleeding within 30 days was 14.7% (36 of 245 patients) in patients with 76.7% platelet aggregation and 6.6% (15 of 229 patients) in patients with >76.7% platelet aggregation (relative risk 2.2, 95% CI 1.3 to 4.0, p ¼ 0.004). As listed in Table 3, the incidence of ARMYDABLEEDS-defined bleeding was 4.6% (22 of 474 patients). Patients who experienced bleeding according to the ARMYDA-BLEEDS-defined bleeding end point within 30 days had an equal median NCDR CathPCI risk score compared with patients without bleeding (65 [IQR 60 to 76.25] vs 65 [IQR 50 to 80], p ¼ 1.00). Median platelet
aggregation was lower in patients with ARMYDABLEEDS-defined bleeding compared with patients without bleeding (50.1% [IQR 31.8% to 75.3%] vs 76.4% [IQR 47.0% to 91.0%], p ¼ 0.016). Patients in the first platelet aggregation quartile had a higher frequency of ARMYDABLEEDS-defined bleeding within 30 days compared with the third quartile and the fourth quartile, as listed in Table 3. There was no significant difference in ARMYDA-BLEEDSdefined bleeding between clopidogrel-naive patients and those already receiving clopidogrel maintenance doses (4.2% vs 5.8%, p ¼ 0.429). The frequency of ARMYDA-BLEEDS-defined bleeding events over time is presented in Figure 2. As listed in Table 4, multivariate Cox regression showed that platelet aggregation in the first quartile predicted increased risk for ARMYDA-BLEEDS-defined bleeding within 30 days, when adjusted for the other included variables. The ROC curve analysis showed a significant correlation between patients with and without ARMYDA-BLEEDSdefined bleeding according to platelet aggregation, with an area under the curve of 0.65 (95% CI 0.54 to 0.76, p ¼ 0.016). The optimal platelet aggregation cutoff was 74.8%, with 77% sensitivity and 52% specificity. The incidence of ARMYDABLEEDS-defined bleeding within 30 days was 7.3% (17 of 233 patients) in patients with 74.8% platelet aggregation and 2.1% (5 of 241 patients) in patients with >74.8% platelet aggregation (relative risk 3.5, 95% CI 1.3 to 9.4, p ¼ 0.007). Discussion This study indicates that clopidogrel-treated patients with low on-treatment platelet reactivity, according to a rapid point-of-care single-platelet function test at the time of intervention, have a significantly higher incidence of bleeding defined according to BARC1 and ARMYDABLEEDS2 <30 days after coronary angiography with and without PCI.
1352
The American Journal of Cardiology (www.ajconline.org)
The relation between BARC bleeding criteria and ontreatment platelet reactivity has not previously been studied in patients who undergo PCI.6 We decided to focus on type 2 BARC bleeding, as the US Food and Drug Administration supports bleeding definitions that reflect clinical outcome.7 Indeed, a relation between mortality and type 2 BARC bleeding has been shown in patients who undergo PCI.8 Also, a recent study by Matic et al9 indicated a hierarchical relation between BARC bleeding and mortality. The present study indicated that low platelet aggregation was nonlinearly correlated with type 2 BARC bleeding. Type 1 BARC bleeding is usually considered clinically unimportant, although these events have been associated with premature drug cessation, which may be associated with impaired clinical outcomes.10 Therefore, the association between the first platelet aggregation quartile and type 1 BARC bleeding in the present study may also be of clinical value. We used the NCDR CathPCI risk score5 to assess any differences in bleeding risk between platelet aggregation quartiles. This bleeding risk score was very similar among the different quartiles (Table 1) but was slightly higher in patients who experienced type 2 BARC bleeding. The multivariate Cox regression confirmed this relation, when adjusted for the other included variables (Table 4). However, the first platelet aggregation quartile was predictive for type 2 BARC bleeding, when adjusted for the NCDR CathPCI risk score. This may indicate that rapid platelet aggregation testing is of additional predictive value when assessing the risk for postprocedural coronary angiography bleeding. Nevertheless, the predictive value of the Plateletworks assay alone in the ROC curve analysis was relatively modest, with low sensitivity and specificity. The incidence of bleeding complications was 4.8% in the ARMYDA-BLEEDS study.2 Using the ARMYDA-BLEEDS combined end point as a bleeding definition, our study showed a similar incidence of 4.6%. Furthermore, the correlation between low platelet aggregation and bleeding described in the ARMYDA-BLEEDS study was also demonstrated in our study. The difference between studies in postprocedural coronary angiographic bleeding rates is highly dependent on the bleeding definition used.1 This was evident in the present study, in which the incidences of type 1 BARC bleeding and type 2 BARC bleeding were approximately 6 and 2 times, respectively, the incidence of ARMYDA-BLEEDS-defined bleeding. This may indicate a higher sensitivity of the BARC bleeding definitions compared with the ARMYDABLEEDS definition of bleeding. As shown in Figure 1 and Figure 2, most ARMYDABLEEDS-defined bleedings and type 2 BARC bleedings occurred within the first week after coronary angiography. This is an interesting finding, which, if it proves repeatable, may be of clinical value. The cause of the event plateau around day 5 to 6 is not clear, but it may indicate a relatively high proportion of procedural events, which occur shortly after the procedures. In the present study, the results obtained with the Plateletworks assay were not validated with another platelet function test, such as the reference standard method for platelet function testing, light transmittance aggregometry.
However, the Plateletworks assay has been shown to correlate well with light transmittance aggregometry in the prediction of adverse cardiovascular events.11,12 In contrast, a relation between platelet function testing with the Plateletworks assay and bleeding has not previously been shown,11 possibly because of less sensitive bleeding definitions. The present relatively small study found such a correlation, which adds valuable information on the usefulness of the Plateletworks assay as a method to identify patients receiving clopidogrel at increased risk for postprocedural coronary angiographic bleeding complications according to two relevant bleeding definitions. Acknowledgments: We thank Kristina Kilsand for data collection and excellent technical assistance and Associate Professor Ulrik Sartipy for valuable advice.
Disclosures The authors have no conflicts of interest to disclose. 1. Mehran R, Rao SV, Bhatt DL, Gibson CM, Caixeta A, Eikelboom J, Kaul S, Wiviott SD, Menon V, Nikolsky E, Serebruany V, Valgimigli M, Vranckx P, Taggart D, Sabik JF, Cutlip DE, Krucoff MW, Ohman EM, Steg PG, White H. Standardized bleeding definitions for cardiovascular clinical trials: a consensus report from the Bleeding Academic Research Consortium. Circulation 2011;123:2736e2747. 2. Patti G, Pasceri V, Vizzi V, Ricottini E, Di Sciascio G. Usefulness of platelet response to clopidogrel by point-of-care testing to predict bleeding outcomes in patients undergoing percutaneous coronary intervention (from the Antiplatelet Therapy for Reduction of Myocardial Damage During Angioplasty-Bleeding Study). Am J Cardiol 2011;107:995e1000. 3. Smith SC Jr, Feldman TE, Hirshfeld JW Jr, Jacobs AK, Kern MJ, King SB 3rd, Morrison DA, O’Neill WW, Schaff HV, Whitlow PL, Williams DO, Antman EM, Adams CD, Anderson JL, Faxon DP, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Nishimura R, Ornato JP, Page RL, Riegel B. ACC/AHA/SCAI 2005 guideline update for percutaneous coronary interventionesummary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/SCAI Writing Committee to Update the 2001 Guidelines for Percutaneous Coronary Intervention). Catheter Cardiovasc Interv 2006;67:87e112. 4. Kushner FG, Hand M, Smith SC Jr, King SB 3rd, Anderson JL, Antman EM, Bailey SR, Bates ER, Blankenship JC, Casey DE Jr, Green LA, Jacobs AK, Hochman JS, Krumholz HM, Morrison DA, Ornato JP, Pearle DL, Peterson ED, Sloan MA, Whitlow PL, Williams DO. 2009 focused updates: ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction (updating the 2004 guideline and 2007 focused update) and ACC/ AHA/SCAI guidelines on percutaneous coronary intervention (updating the 2005 guideline and 2007 focused update): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Catheter Cardiovasc Interv 2009;74:E25eE68. 5. Rao SV, McCoy LA, Spertus JA, Krone RJ, Singh M, Fitzgerald S, Peterson ED. An updated bleeding model to predict the risk of postprocedure bleeding among patients undergoing percutaneous coronary intervention: a report using an expanded bleeding definition from the National Cardiovascular Data Registry CathPCI Registry. JACC Cardiovasc Interv 2013;6:897e904. 6. Tantry US, Bonello L, Aradi D, Price MJ, Jeong YH, Angiolillo DJ, Stone GW, Curzen N, Geisler T, Ten Berg J, Kirtane A, Siller-Matula J, Mahla E, Becker RC, Bhatt DL, Waksman R, Rao SV, Alexopoulos D, Marcucci R, Reny JL, Trenk D, Sibbing D, Gurbel PA; Working Group on OnTreatment Platelet Reactivity. Consensus and update on the definition of on-treatment platelet reactivity to adenosine diphosphate associated with ischemia and bleeding. J Am Coll Cardiol 2013;62:2261e2273.
Coronary Artery Disease/Bleeding After Coronary Angiography 7. Hicks KA, Stockbridge NL, Targum SL, Temple RJ. Bleeding Academic Research Consortium consensus report: the Food and Drug Administration perspective. Circulation 2011:2664e2665. 8. Ndrepepa G, Schuster T, Hadamitzky M, Byrne RA, Mehilli J, Neumann FJ, Richardt G, Schulz S, Laugwitz KL, Massberg S, Schomig A, Kastrati A. Validation of the Bleeding Academic Research Consortium definition of bleeding in patients with coronary artery disease undergoing percutaneous coronary intervention. Circulation 2012;125: 1424e1431. 9. Matic DM, Milasinovic DG, Asanin MR, Mrdovic IB, Marinkovic JM, Kocev NI, Marjanovic MM, Antonijevic NM, Vukcevic VD, Savic LZ, Zivkovic MN, Mehmedbegovic ZH, Dedovic VM, Stankovic GR. Prognostic implications of bleeding measured by Bleeding Academic Research Consortium (BARC) categorisation in patients undergoing primary percutaneous coronary intervention. Heart 2014;100:146e152.
1353
10. Roy P, Bonello L, Torguson R, de Labriolle A, Lemesle G, Slottow TL, Steinberg DH, Kaneshige K, Xue Z, Satler LF, Kent KM, Suddath WO, Pichard AD, Lindsay J, Waksman R. Impact of “nuisance” bleeding on clopidogrel compliance in patients undergoing intracoronary drugeluting stent implantation. Am J Cardiol 2008;102:1614e1617. 11. Breet NJ, van Werkum JW, Bouman HJ, Kelder JC, Ruven HJ, Bal ET, Deneer VH, Harmsze AM, van der Heyden JA, Rensing BJ, Suttorp MJ, Hackeng CM, ten Berg JM. Comparison of platelet function tests in predicting clinical outcome in patients undergoing coronary stent implantation. JAMA 2010;303:754e762. 12. van Werkum JW, Kleibeuker M, Postma S, Bouman HJ, Elsenberg E, ten Berg JM, Hackeng CM. A comparison between the Plateletworks (TM)-assay and light transmittance aggregometry for monitoring the inhibitory effects of clopidogrel. Int J Cardiol 2010;140: 123e126.