DOI: 10.1111/eci.12408

ORIGINAL ARTICLE Procedure-related bleeding in elective percutaneous coronary interventions Gjin Ndrepepa*, Tilman Stephan*, Katrin Anette Fiedler*, Elena Guerra*, Sebastian Kufner* and Adnan Kastrati*,† * †

€t, Munich, Germany, Deutsches Herzzentrum, Technische Universita €t, Munich, Germany 1. Medizinische Klinik rechts der Isar, Technische Universita

ABSTRACT Background Prognostic impact of procedure-related bleeding in patients with stable coronary artery disease (CAD) undergoing elective percutaneous coronary intervention (PCI) remains incompletely investigated. The aim of this study was to investigate the association between peri-PCI bleeding and 1-year outcome of patients with stable CAD. Materials and methods The study included 9035 patients with stable CAD who underwent elective PCI. Bleeding within 30 days of PCI was defined using the Bleeding Academic Research Consortium (BARC) criteria. The primary outcome was 1-year mortality. Results Bleeding occurred in 844 patients (9
3%). Actionable bleeding (BARC class ≥ 2) occurred in 442 patients (4
9%). There were 210 deaths (2
3%) at 1 year following PCI: 41 deaths among patients with bleeding and 169 deaths among patients without bleeding [Kaplan–Meier estimates of mortality, 4
9% and 2
1%; odds ratio = 2
41, 95% confidence interval (CI) 1
73–3
36, P < 0
001]. The association between bleeding and mortality remained significant after adjustment for baseline risk variables (adjusted hazard ratio = 1
87, 95% CI 1
27–2
76, P = 0
002). Bleeding increased the discriminatory power of the model regarding prediction of 1-year mortality (absolute and relative integrated discrimination improvement, 0
006% and 16
3%, respectively, P = 0
001). Conclusions In patients with stable CAD undergoing elective PCI, the occurrence of bleeding within 30 days of the procedure was associated with increased risk of death at 1 year after PCI. These findings suggest that procedure-related bleeding may contribute to less than optimal results of PCI in terms of mortality reduction in patients with stable CAD. Keywords Bleeding, mortality, myocardial infarction, percutaneous coronary intervention, stable coronary artery disease. Eur J Clin Invest 2015; 45 (3): 263–273

Introduction Stable coronary artery disease (CAD) is the commonest form of ischaemic heart disease. Population-based studies have shown that 10–12% of women and 12–14% of men 65–84 years of age have angina [1]. The treatment of patients with stable CAD includes conservative and invasive options. The invasive treatment with percutaneous coronary intervention (PCI) represents a routine and safe procedure and has an established role in the treatment of patients with stable CAD [2]. However, due to the suboptimal results obtained in the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial [3] and other reasons, the use of PCI for the treatment of stable CAD has decreased [4,5]. Despite recent trends in performing PCI in patients with stable CAD, a

considerable number of patients with stable CAD still undergo PCI. A recent report from the CathPCI Registry showed that among 941 248 patients undergoing PCI, 17
6% had stable angina [6]. Evidence available suggests that peri-PCI complications may contribute to the suboptimal results of PCI in stable CAD [7]. Recent studies have shown that peri-PCI bleeding is as strong as myocardial infarction as a prognostic factor for short [8], and long-term mortality [9]. The prognostic role of peri-PCI bleeding has been studied in mixed groups of patients [9] or patients with acute coronary syndromes [8,10–12]. Evidence on the impact of peri-PCI bleeding on mortality in patients with stable CAD is rather limited [13]. So far, no study has investigated the prognostic role of bleeding defined by the Bleeding Academic Research Consortium (BARC) criteria [14] in patients with stable CAD undergoing elective PCI. We

European Journal of Clinical Investigation Vol 45

263

G. NDREPEPA ET AL.

www.ejci-online.com

undertook this study to assess the frequency, predisposing factors and impact of bleeding on the outcome of patients with stable CAD undergoing elective PCI.

Methods Study patients This study included 9035 patients with stable CAD who underwent diagnostic coronary angiography and PCI between June 2000 and May 2011 in the setting of the Intracoronary Stenting and Antithrombotic Regimen (ISAR) randomized trials. The ISAR trials investigated the efficacy and safety of various antithrombotic regimens in patients with CAD undergoing PCI. In brief, patients with ST-segment elevation myocardial infarction, hemodynamic instability, known malignancies, increased risk of bleeding (stroke within 3 months, active bleeding or bleeding diathesis, recent trauma or major surgery within the last month and suspected aortic dissection), oral anticoagulants (within 7 days), high blood pressure (≥ 180 mmHg), serum creatinine > 30 mg/L, known allergy to the study medication, reduced left ventricular ejection fraction (< 30%), pericarditis, subacute bacterial endocarditis or pregnancy (present or suspected) were excluded. Detailed criteria for the inclusion or exclusion of the

patients in these trials are reported in a prior work from our group [15]. Overall, there were 14 180 patients included in these trials. Of them, 5145 patients had non-ST-segment elevation acute coronary syndromes and were excluded from this analysis. The remaining 9035 patients presented with stable CAD and were included in this study (Fig. 1). Written informed consent was required for participation in the source studies, and each trial was approved by the institutional review board in the respective recruiting centres. The study conforms to the Declaration of Helsinki. Reporting of the study conforms to STROBE statement along with references to STROBE and the broader EQUATOR guidelines [16].

Definition of stable CAD, cardiovascular risk factors and bleeding Stable CAD was diagnosed in case of chest pain of cardiac origin that has not changed in character, frequency, intensity or duration over the last 2 months and if significant CAD was found in the diagnostic coronary angiography (angiographic documentation of coronary stenosis with ≥ 50% lumen obstruction in ≥ 1 of the 3 major coronary arteries). Main cardiovascular risk factors were defined as follows: arterial hypertension – presence of active treatment with

PaƟents with CAD undergoing PCI (n = 14180) PaƟents with ACS (n = 5145) PaƟents with stable CAD (n = 9035)

Bleeding within 30 days (n = 844)

Lost to follow-up (n = 16)

Complete 1-year follow-up (n = 828)

264

No bleeding within 30 days (n = 8191)

Lost to follow-up (n = 203)

Complete 1-year follow-up (n = 7988)

ª 2015 Stichting European Society for Clinical Investigation Journal Foundation

Figure 1 Study design flowchart. In patients with incomplete follow-up, the median follow-up (25th, 75th percentiles) was 6
8 (2
2–8
8) months. ACS, acute coronary syndrome; CAD, coronary artery disease; PCI, percutaneous coronary intervention.

BLEEDING IN CORONARY INTERVENTIONS

antihypertensive agents or documentation of a systolic blood pressure value of > 140 mmHg and/or diastolic blood pressure value > 90 mmHg on at least two separate occasions; hypercholesterolaemia – a documented total cholesterol ≥ 220 mg/dL or active treatment with cholesterol-lowering drugs; type 2 diabetes – active treatment with insulin or oral hypoglycaemic agents, or the documentation of an abnormal fasting blood glucose test (≥ 7
0 mmol/dL) or a blood glucose > 11
1 mmol/dL at any time or an abnormal 2-h postglucose load test (postload plasma glucose ≥ 11
1 mmol/L) according to the World Health Organization criteria. Current smokers were those with regular smoking in the prior 6 months. Patients’ weight and height were measured during the hospital stay and used to calculate the body mass index. Renal function was assessed by calculating the creatinine clearance using the Cockcroft–Gault equation [17]. Bleeding complications within 30 days of PCI were defined and scaled using the Bleeding Academic Research Consortium (BARC) criteria [14]. As the primary outcome of this investigation was mortality, bleeding events potentially belonging to BARC class 5 (fatal bleeding) were not analysed as a separate class but were distributed to other classes depending on the initial classification. According to location, bleeding was defined as access site bleeding (bleeding or haematoma arising directly from the arterial access site area or bleeding that spread from the access site to adjacent tissues including the retroperitoneal space) and nonaccess site bleeding (bleeding occurring in sites remote from the vascular access site). Only 1 bleeding event per patient was analysed. In case of more than 1 bleeding event per patient, the location of bleeding was defined as the site with the largest bleed (higher BARC class). Bleeding whose location remained unknown after imaging tests was classified as nonaccess site bleeding [15].

(9
7%) underwent balloon angioplasty only. Vascular access was achieved via the femoral artery. Vascular sheaths were removed, and manual compression was performed as soon as the activated partial thromboplastin time fell below 50 s. Post-PCI chronic antithrombotic therapy included aspirin (80–325 mg/day indefinitely) and clopidogrel (75–150 mg/day until discharge but for no longer than 3 days followed by 75 mg/day for ≥ 1 month in patients who received a baremetal stent or ≥ 6 months for patients who received a drugeluting stent). Other cardiac medications were prescribed at the discretion of the attending physician.

Study outcomes The length of follow-up was 1 year. The primary outcome was 1-year all-cause mortality. Nonfatal myocardial infarction was also assessed. Information on mortality was obtained from the hospital records, death certificates and telephone contact with relatives of the patient or referring physician, insurance companies or registration of address office. Diagnostic criteria for nonfatal myocardial infarction were development of new abnormal Q waves in ≥ 2 contiguous precordial leads or in ≥ 2 adjacent extremity leads or an elevation of the creatine kinasemyocardial band (CK-MB) activity of ≥ twofold the upper limit of normal. For patients within the first 48 h after PCI and those undergoing coronary artery bypass surgery, an elevation of CK-MB of ≥ threefold and ≥ 10-fold the upper limit of normal, respectively, was required for the diagnosis of myocardial infarction. Follow-up information was performed by personnel blinded to the clinical characteristics of the patients. Follow-up consisted of telephone interviews at 30 days, 6 months and 1 year after the index procedure. Patients with complaints or symptoms suspected to be of cardiac origin at those time points or at any time during the follow-up underwent a complete clinical and laboratory evaluation.

Antithrombotic/anticoagulant therapy and PCI procedure

Statistical analysis

All patients received 325–500 mg of aspirin and 600 mg loading dose of clopidogrel before PCI procedure. Peri-PCI anticoagulant therapy consisted of: a combination of glycoprotein IIb/IIIa inhibitor abciximab [an intravenous bolus of 0
25 mg/ kg of weight followed by an infusion of 0
125 lg/kg/min (a maximum of 10 lg/min) of the drug for 12 h] and unfractionated heparin (an intravenous bolus of 70 U/kg of weight; 1681 patients); unfractionated heparin as an intravenous bolus of 140 U/kg of weight (3547 patients); unfractionated heparin as an intravenous bolus of 100 U/kg of weight (1939 patients) and; bivalirudin as an intravenous bolus of 0
75 mg/kg of weight followed by an infusion of 1
75 mg/kg/h for the duration of the procedure (1868 patients). Coronary stenting was performed according to the standard practice in 8157 patients (90
3%). The remaining 878 patients

Data are presented as mean  standard deviation, median with 25th-75th percentiles, proportions (%) or Kaplan–Meier estimates (%). The normality of distribution of continuous data was tested with the Kolmogorov–Smirnov test. Continuous data were compared with the t-test or Kruskal–Wallis rank sum test, when appropriate. Categorical data were compared with the chi-square test. Multiple logistic regression analysis was performed to assess the associates of increased risk of bleeding. All variables of Table 1 but the type of antithrombotic therapy (to avoid co-linearity with bleeding) were entered into the model. The risk of the association with bleeding was estimated by calculating the adjusted odds ratio with the 95% confidence interval. The survival analysis was performed with the Kaplan– Meier method and the log-rank test. Homogeneity of odds ratios was assessed with the Breslow–Day test. The association

European Journal of Clinical Investigation Vol 45

265

G. NDREPEPA ET AL.

www.ejci-online.com

Table 1 Baseline characteristics according to bleeding status Bleeding Yes (n = 844)

Characteristic

69
6  9
9

Age (years) Women

66
4  9
9

289 (34
2)

1813 (22
1)

26
7  4
1

2

Body mass index (kg/m ) Type 2 diabetes

No (n = 8191)

27
6  4
1

P value < 0
001 < 0
001 < 0
001

290 (34
4)

2622 (32
0)

0
164

64 (7
6)

661 (8
1)

0
620

Arterial hypertension

687 (81
4)

6395 (78
1)

0
025

Hypercholesterolaemia (≥ 220 mg/dL)

573 (67
9)

5663 (69
1)

0
456

Current smoker

112 (13
3)

1232 (15
0)

0
168

Prior myocardial infarction

239 (28
3)

2755 (33
6)

0
002

90 (10
7)

945 (11
5)

0
448 0
763

On insulin therapy

Prior coronary artery bypass surgery Serum creatinine (mg/dL)

0
95 [0
80; 1
20]

1
00 [0
80; 1
10]

Creatinine clearance (mL/min)

72
1 [54
2; 93
3]

83
1 [64
5; 104
1]

1
6 [0
5; 5
0]

C-reactive protein (mg/L)

1
3 [0
0; 5
0]

< 0
001 0
042

Extent of coronary artery disease 1-vessel disease

152 (18
0)

1566 (19
1)

2-vessel disease

224 (26
5)

2326 (28
4)

3-vessel disease

468 (55
5)

4299 (52
5)

Multivessel disease

692 (82
0)

6625 (80
9)

0
434

Left ventricular ejection fraction (%)

60
0 [50
0; 65
0]

60
0 [51
0; 65
0]

0
084

221
0 [186
0; 265
0]

216
0 [183
0; 255
0]

0
005

9

Platelet count (910 /L) Thrombocytopenia (< 2 9 10 cells/mm ) 4

3

0
258

4 (0
5)

14 (0
2)

0
060

Abciximab plus heparin (70 U/kg)

165 (19
5)

1516 (18
5)

0
090

Unfractionated heparin (100 U/kg)

188 (22
3)

1751 (21
4)

Unfractionated heparin (140 U/kg)

345 (40
9)

3202 (39
1)

Bivalirudin

146 (17
3)

1722 (21
0)

Periprocedural anticoagulant therapy

Data are mean  standard deviation, median with 25th-75th percentiles (mentioned in square bracket) or counts (%).

between bleeding and the risk of mortality or nonfatal myocardial infarction was tested in the multivariable Cox proportional hazards model. All variables of Table 1 but the type of antithrombotic therapy were entered into the Cox model. The impact of the study type on the outcome was accounted for by including a random effect for the study into the model. The proportional hazards assumption was checked with the method by Grambsch and Therneau [18]. Discriminatory power of the multivariable (Cox) model regarding mortality before (with baseline variables only) and after inclusion of bleeding

266

(baseline variables plus bleeding) was assessed by calculating the integrated discrimination improvement (IDI) [19]. All analyses were performed using the R package. A two-sided P value < 0
05 was considered to indicate statistical significance.

Results Baseline data The study included 9035 patients with stable CAD who underwent elective PCI. Based on the occurrence of bleeding

ª 2015 Stichting European Society for Clinical Investigation Journal Foundation

BLEEDING IN CORONARY INTERVENTIONS

within the first 30 days of PCI, patients were divided into two groups: the group with bleeding (n = 844 patients; 9
3%) and the group without bleeding (8191 patients; 90
7%). Baseline data of the patients are shown in Table 1. Compared with patients without bleeding, patients with bleeding were of older age and were more likely to be women and to have lower body mass index, higher frequency of arterial hypertension, lower creatinine clearance, higher C-reactive protein level and higher platelet count. Coronary stents were implanted in 8157 patients (90
3%): 751 patients with bleeding and 7406 patients without bleeding (89
0% vs. 90
4%; P = 0
180). Drug-eluting stents were used in 455 patients with bleeding and 4571 patients without bleeding (53
9% vs. 55
8%; P = 0
291). Bleeding events in various ISAR trials are shown in Table 2.

plus 70 U/kg of weight of unfractionated heparin; OR = 0
79 (0
63–0
99), P = 0
040 for bivalirudin vs. 100 U/kg of weight of unfractionated heparin; and OR = 0
78 (0
64–0
96), P = 0
020 for bivalirudin vs. 140 U/kg of weight of unfractionated heparin]. The multiple logistic regression analysis was used to assess the association of various clinical and demographic variables with bleeding (see methods for variables entered into the model). Of all variables, we adjusted for, the patient’s age, being women, body mass index, arterial hypertension, prior myocardial infarction, creatinine clearance and reduced left ventricular function were independently associated with the increased risk of bleeding. The direction and the strength of association and the risk estimates for the association between various variables and bleeding are shown in Table 4.

Severity and associates of bleeding Scaling of bleeding severity was performed using the BARC criteria. Of the 844 patients with bleeding, class 1 bleeding occurred in 402 patients (47
6%), class 2 in 118 patients (14%), class 3a in 227 patients (26
9%), class 3b in 84 patients (10
0%), class 3c in 7 patients (0
8%) and class 4 in 6 patients (0
7%). Bleeding class ≥ 2 was occurred in 442 patients (4
9%). Access and nonaccess site bleeding occurred in 535 (5
9%) and 309 patients (3
4%), respectively. Baseline data of patients with access site and nonaccess site bleeding are shown in Table 3. Analysis by type of antithrombotic drug used showed that bleeding occurred in 146 patients (7
8%) treated with bivalirudin, 165 patients (9
8%) treated with abciximab plus 70 U/kg of weight of unfractionated heparin, 188 patients (9
7%) treated with 100 U/kg of weight of unfractionated heparin (9
7%) and 345 patients (9
7%) treated with 140 U/kg of weight of unfractionated heparin [odds ratio (OR) = 0
77, 95% confidence interval (CI) 0
61–0
98, P = 0
035 for bivalirudin vs. abciximab

One-year incidence of mortality and myocardial infarction Overall, there were 210 deaths within the first year after PCI (2
3%). Of them, 41 deaths occurred among patients with bleeding and 169 deaths occurred among patients without bleeding [Kaplan–Meier (KM) estimates of 1-year mortality 4
9% and 2
1%, respectively; OR = 2
41 (1
73–3
36), P < 0
001]. Within the first 30 days after PCI, there were nine deaths among patients with bleeding and 14 deaths among patients without bleeding [KM estimates 1
1% and 0
2%; OR = 6
28 (3
03–13
05), P < 0
001]. Time-to-event curves, obtained by KM analysis for 30-day and 1-year mortality, are shown in Fig. 2. According to the bleeding site, there were 20 deaths among 535 patients with access site bleeding and 21 deaths among 309 patients with nonaccess site bleeding [KM estimates of 1-year mortality 3
8% and 6
9%, respectively, OR = 1
84 (1
15–2
95), P = 0
009 for access site bleeding vs. no bleeding; OR = 3
46

Table 2 Number of patients and bleeding events in the primary trials included in this analysis Bleeding events Trial

Number of patients (n = 9035)

BARC (n = 844)

TIMI major* (n = 68)

TIMI minor* (n = 164)

ISAR-REACT

2159 (23
9)

191 (22
6)

20 (0
9)

48 (2
2)

ISAR-REACT 3

3734 (41
3)

355 (42
1)

28 (0
8)

66 (1
8)

ISAR-REACT 3a

1939 (21
5)

188 (22
3)

9 (0
5)

16 (0
8)

ISAR-SMART

502 (5
6)

48 (5
7)

4 (0
8)

17 (3
4)

ISAR-SWEET

701 (7
7)

62 (7
3)

7 (1
0)

17 (2
4)

BARC, Bleeding Academic Research Consortium; ISAR-REACT, The Intracoronary Stenting and Antithrombotic Regimen–Rapid Early Action for Coronary Treatment; ISAR-SMART, The Intracoronary Stenting to Abrogate Restenosis in Small Arteries; ISAR-SWEET, The Intracoronary Stenting and Antithrombotic Regimen: Is Abciximab a Superior Way to Eliminate Elevated Thrombotic Risk in Diabetics; TIMI, Thrombolysis in Myocardial Infarction. Data are number of patients/events (%). *Bleeding events according to TIMI group criteria were prospectively evaluated in the primary trials included in this analysis.

European Journal of Clinical Investigation Vol 45

267

G. NDREPEPA ET AL.

www.ejci-online.com

Table 3 Baseline characteristics according to bleeding location Bleeding Access site (n = 535)

Characteristic

69
2  9
9

Age (years) Women

205 (38
3) 26
7  4
2

Body mass index (kg/m2) Type 2 diabetes

Nonaccess site (n = 309)

P value

70
3  9
9

0
117

84 (27
2)

0
001

26
6  4
0

0
416

183 (34
2)

107 (34
6)

0
901

33 (6
2)

31 (10
0)

0
041

Arterial hypertension

426 (79
6)

261 (84
5)

0
081

Hypercholesterolaemia (≥ 220 mg/dL)

368 (68
8)

205 (66
3)

0
461

71 (13
3)

41 (13
3)

0
999

146 (27
3)

93 (30
1)

0
383

54 (10
1)

36 (11
6)

0
480

On insulin therapy

Current smoker Prior myocardial infarction Prior coronary artery bypass surgery Serum creatinine (mg/dL)

0
94 [0
80; 1
12]

0
99 [0
80; 1
20]

0
305

Creatinine clearance (mL/min)

71
0 [53
9; 93
3]

73
2 [55
7; 90
9]

0
888

1
6 [0
0; 5
0]

1
6 [0
4; 5
0]

0
403

1-vessel disease

107 (20
0)

45 (14
4)

0
140

2-vessel disease

138 (25
8)

86 (27
8)

3-vessel disease

290 (54
2)

178 (57
6)

Multivessel disease

428 (80
0)

264 (85
4)

0
047

Left ventricular ejection fraction (%)

60
0 [52
0; 65
0]

59
0 [49
0; 64
0]

0
068

225
0 [190
0; 267
0]

214
5 [179
0; 257
2]

0
026

C-reactive protein (mg/L) Extent of coronary artery disease

9

Platelet count (910 /L) Thrombocytopenia (< 2 9 10 cells/mm ) 4

3

0 (0)

4 (1
3)

0
008

Abciximab plus heparin (70 U/kg)

113 (21
1)

52 (16
8)

0
001

Unfractionated heparin (100 U/kg)

227 (42
4)

118 (38
2)

Unfractionated heparin (140 U/kg)

123 (23
0)

65 (21
0)

72 (13
5)

74 (24
0)

Periprocedural anticoagulant therapy

Bivalirudin Time-to-bleeding interval (days)*

0
63  1
14

2
66  5
11

< 0
001

Data are mean  standard deviation, median with 25th-75th percentiles (mentioned in square bracket) or counts (%). *Measured from the beginning of the procedure.

(2
17–5
53), P < 0
001 for nonaccess site bleeding vs. no bleeding and OR = 1
86 (1
02–3
40), P = 0
043 for nonaccess site bleeding vs. access site bleeding]. Time-to-event curves in patients with access site and nonaccess site bleeding are shown in Fig. 3. One-year mortality according to bleeding severity is shown in Table 5. There was a progressive increase in mortality with the increase in the severity of bleeding (P for trend < 0
001).

268

Nonfatal myocardial infarction occurred in 349 patients (3
9%). Eighty-three cases of myocardial infarction occurred among patients with bleeding and 266 cases among patients without bleeding [KM estimates 9
9% and 3
3%, respectively, OR = 3
16 (2
50–3
99), P < 0
001]. Time-to-event curves are shown in Fig. 4. The majority of myocardial infarctions occurred within a shorttime interval after PCI (initial steep part of the KM curves in Fig. 4). To better characterize the occurrence of myocardial

ª 2015 Stichting European Society for Clinical Investigation Journal Foundation

BLEEDING IN CORONARY INTERVENTIONS

Table 4 Results of multiple logistic regression analysis applied to assess associates of bleeding Characteristic

Chi-square*

Age (for a 10-year increase) Women

Body mass index (for 5 kg/m increase)

P value

9
62

1
20 (1
07–1
34)

0
002

17
55

1
48 (1
23–1
78)

< 0
001

1
20

1
10 (0
93–1
30)

0
274

Type 2 diabetes 2

Adjusted OR (95% CI)

10
27

0
83 (0
74–0
93)

0
001

Arterial hypertension

6
11

1
32 (1
06–1
64)

0
013

Hypercholesterolaemia

0
95

0
92 (0
78–1
09)

0
331

Current smoker

0
0

1
18 (0
78–1
27)

0
994

Prior myocardial infarction

5
33

0
80 (0
67–0
97)

0
021

Prior coronary artery bypass graft surgery

0
59

0
90 (0
70–1
17)

0
444

Creatinine clearance (for 30 mL/min decrease)

4
06

1
14 (1
01–1
29)

0
044

C-reactive protein (for 5 mg/L increase)

0
01

1
00 (0
97–1
03)

0
923

Multivessel disease (vs. single-vessel disease)

0
03

1
02 (0
83–1
25)

0
872

Platelet count (for 50 9 10 /L increase)

2
63

1
05 (0
99–1
12)

0
105

Left ventricular ejection fraction (for 10% decrease)

4
70

1
08 (1
01–1
16)

0
030

9

CI, confidence interval; OR, odds ratio. *The Wald chi-square values show the strength of association of variables with the risk for bleeding.

infarction within the time interval immediately after PCI, a 48-h analysis was performed. Thus, within 48 h following PCI, 74 nonfatal infarctions (event rate 8
8% or 89
1% of all infarctions) occurred among patients with bleeding and 198 infarctions (event rate 2
4% or 74
4% of all infarctions) occurred among patients without bleeding [OR = 3
87 (2
94–5
12), P < 0
001]. The occurrence of bleeding almost quadrupled the risk of nonfatal myocardial infarction within the first 48 h of PCI. Time-to-event curves of myocardial infarction over the 48-h interval obtained by KM analysis are shown in Fig. 4 (right top corner). There was no difference in the risk of myocardial infarction between patients with or without bleeding beyond 48 h (1
1% vs. 0
8%; P = 0
477). There were 51 nonfatal infarctions among patients with access site bleeding and 32 nonfatal infarctions among patients with nonaccess site bleeding [KM estimates of 1-year nonfatal myocardial infarction 9
6% and 11
4%, respectively, OR = 3
14 (2
29–4
29), P < 0
001 for access site vs. no bleeding; OR = 3
44 (2
34–5
06), P < 0
001 for nonaccess site vs. no bleeding; and OR = 1
10 (0
71–1
71), P = 0
659 for nonaccess site vs. access site bleeding]. One-year occurrence of myocardial infarction according to bleeding severity is shown in Table 5. There was a progressive increase in the frequency of myocardial infraction with the increase in the severity of bleeding (P for trend < 0
001). The risk of 1-year mortality associated with bleeding (BARC class ≥ 2) or myocardial infarction within 30 days of PCI was compared. Of the 442 patients with BARC class bleeding ≥ 2, 31

patients (7
0%) died over the 1-year follow-up [OR = 3
58 (2
41– 5
31), P < 0
001 compared with patients without bleeding]. Of the 299 patients with nonfatal myocardial infarction occurring within 30 days of PCI, 23 patients (7
7%) died over the 1-year follow-up [OR = 3
81 (2
43–5
97), P < 0
001 compared with patients without myocardial infarction (187 deaths/8736 patients; 2
1%)]. Thus, bleeding and myocardial infarction within 30 days of PCI appear to be associated with a similar risk of 1-year mortality (Breslow–Day test P = 0
478).

Bleeding outcome association after adjustment in multivariable model The association between bleeding and the risk of 1-year mortality or myocardial infarction was tested in the multivariable Cox proportional hazards model (see methods for variables entered into the model). Bleeding was an independent associate of increased risk for both outcomes. Other associates of mortality or myocardial infarction, the risk estimates and the direction of association(s) are shown in Table 6. The discriminatory power of the multivariable model for prediction of the risk of 1-year mortality or myocardial infarction was assessed by calculating the integrated discrimination improvement (IDI) of the models without bleeding (with baseline variables only) and after inclusion of bleeding. The inclusion of bleeding in the multivariable model increased the discriminatory power of the model for prediction of 1-year mortality (absolute and relative IDI, 0
006% and 16
3%, respectively,

European Journal of Clinical Investigation Vol 45

269

G. NDREPEPA ET AL.

www.ejci-online.com

Table 5 One-year mortality or nonfatal myocardial infarction according to bleeding severity

Figure 2 Kaplan–Meier curves of 1-year mortality. Numbers in brackets represent Kaplan–Meier estimates of mortality. The inset (Kaplan–Meier curves) on the right upper corner represents time-to-event curves within the first 30 days of PCI on a larger scale. OR, odds ratio; CI, confidence interval.

Figure 3 Kaplan–Meier curves of 1-year mortality among patients with access site (AS) and nonaccess site (NAS) bleeding. Dashed line represents the Kaplan–Meier curve of mortality for the whole group of patients with bleeding. Numbers in brackets represent Kaplan–Meier estimates of 1-year mortality. OR, odds ratio; CI, confidence interval.

P = 0
001) and nonfatal myocardial infarction (absolute and relative IDI 0
014% and 28
0%, respectively, P < 0
001).

Discussion The main findings of present study may be summarized as follows: (i) In patients with stable CAD undergoing elective PCI, bleeding within the first 30 days following the procedure

270

BARC bleeding class

Mortality (n = 210)

Myocardial infarction (n = 349)

0 (n = 8191)

169 (2
1)*

266 (3
2)

1 (n = 402)

10 (2
5)

27 (6
7)

2 (n = 118)

5 (4
2)

13 (11
0)

3a (n = 227)

13 (5
7)

24 (10
6)

3b (n = 84)

9 (10
7)

3c (n = 7)

4 (57
1)

0 (0)

4 (n = 6)

0 (0)

5 (83
3)

14 (16
7)

BARC, Bleeding Academic Research Consortium. Data are number of events (%). *Percentages in parentheses are row percentages and represent proportions of patients (column 1) with bleeding events that died or had nonfatal myocardial infarction.

Figure 4 Kaplan–Meier curves of 1-year myocardial infarction. The inset (Kaplan–Meier curves) on the right upper corner represents time-to-event curves within the first 48 h of PCI on a larger scale. Numbers in brackets represent Kaplan–Meier estimates of 1-year mortality. OR, odds ratio; CI, confidence interval.

occurred in 9
3% of the patients. Actionable bleeding (BARC class ≥ 2) occurred in 4
9% of the patients. (ii) The independent correlates of increased risk of bleeding ranked from highest to lowest strength of association were as follows: female sex, low body mass index, advanced age, arterial hypertension, prior myocardial infarction, reduced left ventricular function and reduced estimated creatinine clearance. (iii) The occurrence of bleeding within 30 days of procedure was associated with an increased risk of death or nonfatal myocardial infarction up to 1 year after PCI. Bleeding particularly increased the risk of

ª 2015 Stichting European Society for Clinical Investigation Journal Foundation

BLEEDING IN CORONARY INTERVENTIONS

Table 6 Results of Cox proportional hazards model applied to assess associates of one-year mortality or myocardial infarction Mortality

Myocardial infarction Adjusted HR (95% CI)

P value

0
002

3
29 (2
51–4
31)

< 0
001

1
25 (1
01–1
54)

0
044

0
88 (0
75–1
03)

0
113

0
76 (0
52–1
10)

0
145

1
04 (0
78–1
37)

0
804

2
36 (1
73–3
21)

< 0
001

1
44 (1
13–1
83)

0
003

BMI (for 5 kg/m increase)

0
83 (0
67–1
03)

0
093

1
05 (0
89–1
23)

0
577

Arterial hypertension

1
22 (0
84–1
79)

0
306

0
75 (0
57–0
99)

0
045

Hypercholesterolaemia

0
78 (0
57–1
07)

0
119

1
10 (0
85–1
41)

0
471

Current smoker

1
18 (0
76–1
83)

0
461

1
13 (0
82–1
57)

0
452

Prior myocardial infarction

0
91 (0
66–1
26)

0
579

1
01 (0
78–1
31)

0
934

Prior CABG surgery

0
65 (0
39–1
07)

0
089

1
00 (0
70–1
44)

0
984

CC (for 30 mL/min decrease)

1
37 (1
08–1
74)

0
011

1
32 (1
10–1
57)

0
002

CRP (for 5 mg/L increase)

1
05 (1
03–1
07)

< 0
001

1
01 (0
98–1
04)

0
350

1
65 (0
92–2
50)

0
103

1
63 (1
15–2
32)

0
007

Platelet count (for 50 9 10 /L increase)

1
16 (1
05–1
28)

0
003

1
06 (0
97–1
16)

0
217

LVEF (for 10% decrease)

1
46 (1
31–1
63)

< 0
001

0
97 (0
87–1
07)

0
529

Characteristic

Adjusted HR (95% CI)

Bleeding (vs. no bleeding)

1
87 (1
27–2
76)

Age (for a 10-year increase) Women Type 2 diabetes 2

Multivessel disease 9

P value

BMI,body mass index; CABG, coronary artery bypass graft; CC, creatinine clearance; CI, confidence interval; CRP, C-reactive protein; HR, hazard ratio; LVEF, left ventricular ejection fraction.

PCI-related myocardial infarction. The association between bleeding and mortality was stronger for nonaccess site than for access site bleeding. (iv) Bleeding and myocardial infarction within the first 30 days after PCI showed a comparable strength regarding the association with 1-year mortality. Patients with stable CAD undergoing elective PCI are considered to be at low risk of procedure-related complications. In a recent study of patients undergoing elective coronary stenting, the rate of bleeding defined by the GUSTO, ACUITY and TIMI group criteria was 2
3%, 1
9% and 2
1%, respectively [13]. A higher frequency of bleeding as found in the present study may be explained by the use of BARC criteria which are more sensitive than other criteria in detecting the bleeding events [20]. By demonstrating a close association between bleeding and mortality, the present study showed that occurrence of bleeding may shift the risk-benefit relationship of the PCI procedure towards more risk and lead to less than expected benefits of PCI in patients with stable CAD. As the risk-benefit relationship of PCI may be more balanced in patients with stable CAD compared with acute coronary syndromes, this finding may offer some explanation for the suboptimal results of PCI in stable CAD [3]. Most of the evidence available with regard to the prognostic impact of peri-PCI bleeding comes from studies

of patients with acute coronary syndrome. These studies have shown that procedure-related bleeding is associated with a worse prognosis compared with patients without bleeding. As elective PCI is perceived to be associated with low risk [2], the prognostic impact of procedure-related bleeding has drawn little attention. A prior study by Fleming et al. [13]. showed that bleeding occurring in patients undergoing elective stenting was associated with increased risk of combined endpoint of allcause death or myocardial infarction at 12 months after procedure. The present study corroborates and further extends these findings. We found that bleeding almost doubled the risk of 1-year mortality and more than tripled the risk of myocardial infarction. Analysis of mortality by bleeding location showed that nonaccess site bleeding was associated with almost double the risk of mortality compared with access site bleeding. Mechanisms of the deleterious impact of bleeding on the risk for mortality may not differ from those offered to explain the association of bleeding with mortality in patients with acute coronary syndromes [8,10]. The association of several cardiovascular risk factors with increased risk of bleeding shows that bleeding and mortality have similar risk factors. As a consequence, bleeding per se accumulates cardiovascular risk which mediates the increased risk of subsequent mortality [21]. In addition a direct impact of bleeding on mortality has been

European Journal of Clinical Investigation Vol 45

271

G. NDREPEPA ET AL.

postulated considering the hemodynamic deterioration and pro-thrombotic state it generates [22]. The stronger association between nonaccess site bleeding and mortality compared with access site bleeding may be not specific to patients with stable CAD, but it may be explained by a worse cardiovascular risk profile and more severe bleeding in patients with nonaccess site bleeding [15]. The present study also found an association between bleeding and myocardial infarction and that the majority of myocardial infarctions were procedure related. Although the exact reasons for the observed association between bleeding and procedure-related myocardial infarction remain unclear, three mechanisms may be postulated: first, bleeding may cause hemodynamic deterioration through hypovolaemia and anaemia leading to reduced oxygen-carrying capacity of blood and enhanced endogenous neuro-humoral response leading to increased circulating levels of catecholamines which may exacerbate tissue ischaemia; second, bleeding may cause a prothrombotic state through multiple mechanisms; and third, measures taken to treat bleeding such as blood transfusion, antiplatelet drug discontinuation and invasive measures (invasive monitoring, reparative surgery or endoscopic procedures) may further accentuate prothrombotic state or increase stress on already vulnerable myocardium [22,23]. The present study also found that bleeding and myocardial infarction within 30 days of PCI seem to comparably increase the risk for 1-year mortality. We recognize three limitations of the present analysis. First, as bleeding criteria were retrospectively applied, we acknowledge the possibility of having missed some mild forms of bleeding (BARC 1 class). However, patients were obtained from randomized studies, and detailed information on bleeding including mild forms was available for all patients. Moreover, studies that have performed retrospective [20] and prospective [24] validation of the BARC criteria concur with regard to prognostic value of bleeding defined by these criteria. Second, due to the lack of high-sensitivity cardiac troponin assays (not available in patients recruited before 2009), the diagnosis of peri-procedural myocardial infarction did not involve this biomarker as recommended [25]. However, if a cardiac troponin assay is not available, CK-MB is considered as the best alternative to diagnose myocardial infarction [25]. Third, these data are obtained in patients undergoing PCI via femoral artery approach and ought not be extrapolated to patients undergoing PCI via radial artery route. The present findings may have implications. In patients with stable CAD considered suitable candidates for treatment with PCI and who have risk factors for bleeding, bleeding avoidance strategies such as the use of radial artery for vascular access and antithrombotic regimens with favourable bleeding profile may be recommended. The use of radial artery for vascular access

272

www.ejci-online.com

may be particularly beneficial in these patients. It has been reported that patients at highest risk of bleeding complications gain the greatest benefit from the use of transradial approach during PCI [26]. Furthermore, bivalirudin was associated with the reduced risk of peri-PCI bleeding particularly compared with unfractionated heparin plus glycoprotein IIb/IIIa inhibitors [27], and the drug may be recommended in patients with stable CAD at high risk for bleeding. However, the cost-effectiveness issues related to the use of bivalirudin as peri-PCI anticoagulant strategy may be more prominent in patients stable CAD compared with other higher risk CAD presentations. In conclusion, in patients with stable CAD undergoing elective PCI, occurrence of bleeding within 30 days of the procedure was associated with an increased risk of death or nonfatal myocardial infarction up to 1 year after PCI. These findings may offer some explanation for the suboptimal results of PCI in patients with stable CAD and may serve to increase the awareness on the deleterious effects of peri-PCI bleeding even in low-risk patients with stable CAD. Results from this post hoc exploratory analysis should be interpreted with caution but warrant additional investigation.

Author contribution Drs Ndrepepa and Kastrati contributed to conception and design of the study. Drs Ndrepepa, Stephan, Fiedler, Guerra and Kufner contributed to acquisition, analysis or interpretation of the data. Drs Ndrepepa and Kastrati contributed to writing and critiquing of drafts of the manuscript. Drs Ndrepepa, Stephan, Fiedler, Guerra, Kufner and Kastrati contributed to approval of the final manuscript for submission.

Funding None. Conflicts of interest None. Address Deutsches Herzzentrum M€ unchen, Technische Universit€ at, Lazarettstrasse 36, 80636 Munich, Germany. (G. Ndrepepa, T. Stephan, K. A. Fiedler, E. Guerra, S. Kufner, A. Kastrati); 1. Medizinische Klinik rechts der Isar, Technische Universit€ at, Ismaninger Strasse 22, 81675 Munich, Germany (A. Kastrati). Correspondence to: Gjin Ndrepepa, Deutsches Herzzentrum M€ unchen, Lazarettstrasse 36, 80636 Munich, Germany. Tel.: +49-89-12181535; fax: +49-89-12184053; e-mail: [email protected] Received 8 August 2014; accepted 19 January 2015

ª 2015 Stichting European Society for Clinical Investigation Journal Foundation

BLEEDING IN CORONARY INTERVENTIONS

References 1 National Institutes of Health NH, Lung, and Blood Institute. Morbidity & Mortality: 2012 Chart Book on Cardiovascular, Lung, and Blood Diseases. Bethesda, MD: National Heart, Lung, and Blood Institute; 2012. 2 Montalescot G, Sechtem U, Achenbach S, Andreotti F, Arden C, Budaj A et al. 2013 ESC guidelines on the management of stable coronary artery disease: the Task Force on the management of stable coronary artery disease of the European Society of Cardiology. Eur Heart J 2013;34:2949–3003. 3 Boden WE, O’Rourke RA, Teo KK, Hartigan PM, Maron DJ, Kostuk WJ et al. Optimal medical therapy with or without PCI for stable coronary disease. N Engl J Med 2007;356:1503–16. 4 Kim LK, Feldman DN, Swaminathan RV, Minutello RM, Chanin J, Yang DC et al. Rate of percutaneous coronary intervention for the management of acute coronary syndromes and stable coronary artery disease in the United States (2007-2011). Am J Cardiol 2014;114:1003–10. 5 Mohan AV, Fazel R, Huang PH, Shen YC, Howard D. Changes in geographic variation in the use of percutaneous coronary intervention for stable ischemic heart disease after publication of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial. Circ Cardiovasc Qual Outcomes 2014;7:125–30. 6 Dehmer GJ, Weaver D, Roe MT, Milford-Beland S, Fitzgerald S, Hermann A et al. A contemporary view of diagnostic cardiac catheterization and percutaneous coronary intervention in the United States: a report from the CathPCI Registry of the National Cardiovascular Data Registry, 2010 through June 2011. J Am Coll Cardiol 2012;60:2017–31. 7 Hochman JS, Steg PG. Does preventive PCI work? N Engl J Med 2007;356:1572–4. 8 Manoukian SV, Feit F, Mehran R, Voeltz MD, Ebrahimi R, Hamon M et al. Impact of major bleeding on 30-day mortality and clinical outcomes in patients with acute coronary syndromes: an analysis from the ACUITY Trial. J Am Coll Cardiol 2007;49:1362–8. 9 Ndrepepa G, Berger PB, Mehilli J, Seyfarth M, Neumann FJ, Schomig A et al. Periprocedural bleeding and 1-year outcome after percutaneous coronary interventions: appropriateness of including bleeding as a component of a quadruple end point. J Am Coll Cardiol 2008;51:690–7. 10 Mehran R, Pocock SJ, Stone GW, Clayton TC, Dangas GD, Feit F et al. Associations of major bleeding and myocardial infarction with the incidence and timing of mortality in patients presenting with non-ST-elevation acute coronary syndromes: a risk model from the ACUITY trial. Eur Heart J 2009;30:1457–66. 11 Kikkert WJ, van Geloven N, van der Laan MH, Vis MM, Baan J Jr, Koch KT et al. The prognostic value of bleeding academic research consortium (BARC)-defined bleeding complications in ST-segment elevation myocardial infarction: a comparison with the TIMI (Thrombolysis In Myocardial Infarction), GUSTO (Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries), and ISTH (International Society on Thrombosis and Haemostasis) bleeding classifications. J Am Coll Cardiol 2014;63:1866–75. 12 Matic DM, Milasinovic DG, Asanin MR, Mrdovic IB, Marinkovic JM, Kocev NI et al. Prognostic implications of bleeding measured by

13

14

15

16

17 18 19

20

21

22

23

24

25

26

27

Bleeding Academic Research Consortium (BARC) categorisation in patients undergoing primary percutaneous coronary intervention. Heart 2014;100:146–52. Fleming LM, Novack V, Novack L, Cohen SA, Negoita M, Cutlip DE. Frequency and impact of bleeding in elective coronary stent clinical trials–utility of three commonly used definitions. Catheter Cardiovasc Interv 2012;80:E23–9. Mehran R, Rao SV, Bhatt DL, Gibson CM, Caixeta A, Eikelboom J et al. Standardized bleeding definitions for cardiovascular clinical trials: a consensus report from the Bleeding Academic Research Consortium. Circulation 2011;123:2736–47. Ndrepepa G, Neumann FJ, Richardt G, Schulz S, Tolg R, Stoyanov KM et al. Prognostic value of access and non-access sites bleeding after percutaneous coronary intervention. Circ Cardiovasc Interv 2013;6:354–61. Simera I, Moher D, Hoey J, Schulz KF, Altman DG. A catalogue of reporting guidelines for health research. Eur J Clin Invest 2010;40: 35–53. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31–41. Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 1994;81:515–26. Pencina MJ, D’Agostino RB Sr., D’Agostino RB Jr., Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med 2008;27:157–72; discussion 207-12. Ndrepepa G, Schuster T, Hadamitzky M, Byrne RA, Mehilli J, Neumann FJ et al. Validation of the Bleeding Academic Research Consortium definition of bleeding in patients with coronary artery disease undergoing percutaneous coronary intervention. Circulation 2012;125:1424–31. Spencer FA, Moscucci M, Granger CB, Gore JM, Goldberg RJ, Steg PG et al. Does comorbidity account for the excess mortality in patients with major bleeding in acute myocardial infarction? Circulation 2007;116:2793–801. Doyle BJ, Rihal CS, Gastineau DA, Holmes DR Jr. Bleeding, blood transfusion, and increased mortality after percutaneous coronary intervention: implications for contemporary practice. J Am Coll Cardiol 2009;53:2019–27. Ndrepepa G, Kastrati A. Bleeding complications in patients undergoing percutaneous coronary interventions: current status and perspective. Coron Artery Dis 2014;25:247–57. Vranckx P, Leonardi S, Tebaldi M, Biscaglia S, Parrinello G, Rao SV et al. Prospective validation of the Bleeding Academic Research Consortium classification in the all-comer prodigy trial. Eur Heart J 2014;35:2524–9. Thygesen K, Alpert JS, Jaffe AS, Simoons ML, Chaitman BR, White HD et al. Third universal definition of myocardial infarction. J Am Coll Cardiol 2012;60:1581–98. Mamas MA, Anderson SG, Carr M, Ratib K, Buchan I, Sirker A et al. Baseline bleeding risk and arterial access site practice in relation to procedural outcomes after percutaneous coronary intervention. J Am Coll Cardiol 2014;64:1554–64. Cavender MA, Sabatine MS. Bivalirudin versus heparin in patients planned for percutaneous coronary intervention: a meta-analysis of randomised controlled trials. Lancet 2014;384:599–606.

European Journal of Clinical Investigation Vol 45

273