Gender-related differences in the incidence of bleeding and its relation to subsequent mortality in patients with ST-segment elevation myocardial infarction (STEMI) treated with fibrinolysis are not well understood. We studied patients with STEMI receiving fibrinolysis enrolled in 6 clinical trials. Outcomes included moderate or severe bleeding defined using Global Utilization of Strategies to Open Occluded Arteries (GUSTO) criteria and adjusted 1-year mortality (excluding deaths in first 24 hours). Moderate or severe bleeding was 1.9-fold higher in women compared to men (13.3% vs 7.1%, p <0.0001). Bleeding remained higher in women even after adjustment for baseline differences (odd ratios 1.52, 95% confidence interval [CI] 1.42 to 1.62). In fact, female gender was second most important prognostic factor (Wald chi-square 153.6) after older age (Wald chi-square 241.2) in the multivariable bleeding model. Overall 1-year mortality was similar in women and men after adjusting for prognostically important baseline differences (hazard ratio [HR] 1.06, 95% CI 0.97 to 1.17). However, after adjustment for baseline confounders and bleeding, female gender was associated with a lower risk of 1-year death. Thus, adjusted 1-year mortality was similar in women compared to men without bleeding (HR 1.08, 95% CI 0.97 to 1.19) but lower in women compared to men with bleeding (HR 0.85, 95% CI 0.73 to 0.98, p for interaction of gender by bleeding = 0.0016). The highest adjusted 1-year mortality was observed in men with bleeding (HR 2.42, 95% CI 2.20 to 2.66) followed by women with bleeding (HR 2.05, 95% CI 1.80 to 2.33) and women without bleeding (HR 1.08, 95% CI 0.97 to 1.19, referent men without bleeding). In conclusion, in patients with fibrinolytic-treated STEMI, women had a higher incidence but lower mortality with bleeding than men. These data highlight the importance of understanding factors associated with gender-related differences in bleeding and represent an opportunity for improving outcomes of women and men with fibrinolytic-treated STEMI.
Reperfusion with fibrinolytic therapy in patients with ST-segment elevation myocardial infarction (STEMI) has been shown to be associated with an increased risk of bleeding and hemorrhagic stroke in women compared to men. Similarly, women have higher mortality after fibrinolysis compared to men. However, factors associated with increased bleeding risk in women are poorly understood. More importantly, the relation of excess bleeding in women with subsequent outcomes has not been previously studied. The study goals were to determine differences in moderate or severe bleeding (henceforth referred to as “bleeding”) rates in men and women receiving fibrinolysis for STEMI, to investigate differences in clinical features potentially associated with differences in risk between women and men, and to evaluate relations among gender, bleeding, and 1-year mortality.
Methods
We studied subjects enrolled in 6 large randomized clinical trials—Global Utilization of Strategies to Open Occluded Arteries (GUSTO) I and III, Assessment of the Safety and Efficacy of a New Thrombolytic (ASSENT) 2 and 3, HERO-2, and patients with STEMI treated with fibrinolysis from GUSTO-IIb. In brief, these trials evaluated the efficacy of different fibrinolytic agents, the relative merits of fibrinolysis versus primary percutaneous coronary intervention, or adjunctive regimens of fibrinolytic therapy. For this retrospective observational analysis, patients were stratified into 4 groups: men without bleeding, men with bleeding, women without bleeding, and women with bleeding.
Data were collected prospectively in all trials. End points were predefined in all trials and independently adjudicated. Bleeding was classified according to GUSTO criteria. Severe or life-threatening bleeding was defined as intracerebral bleeding or hemorrhage or bleeding resulting in hemodynamic compromise needing treatment, and moderate bleeding was defined as bleeding requiring transfusion. All other overt bleeding was classified as minor.
We compared differences in clinical features, angiographic characteristics, and outcomes between men and women with and without bleeding. Univariable comparisons were made using Pearson chi-square test for categorical variables and Wilcoxon rank-sum test for continuous variables. Missing data were not defaulted to negative and denominators reflect cases reported.
A multivariate logistic model was generated using prespecified factors shown to be associated with increased risk of bleeding. In addition, interaction of gender with treatment was entered into the model as a covariate. Baseline factors in addition to gender used to determine multivariable correlates of bleeding included age, race, medical history (MI, smoking, diabetes mellitus, coronary artery bypass grafting [CABG], stroke, and hypertension), presentation and findings on physical examination (Killip class, systolic and diastolic blood pressures, heart rate, and weight), infarct location, time to treatment, type of fibrinolytic therapy (fibrin specific vs nonspecific), and treatment and gender interaction. Spline transformations were generated to assess whether continuous variables were related to outcomes in nonlinear fashion. When nonlinearity was identified, appropriate transformations were applied. Odd ratios (ORs) and 95% confidence intervals (CIs) were generated to provide adjusted estimates of bleeding risks. To assess the association of gender and bleeding with mortality at 1 year, we generated hazard ratios (HRs) and 95% CIs after adjusting for baseline differences using a previously reported 30-day mortality model. For continuous variables (systolic blood pressure, pulse, height, and weight), linear splines were used. Age and time to treatment were modeled linearly. Interaction terms between age and gender and between gender and bleeding were also examined in the mortality model. Because date of bleeding was not known for most bleeds, to minimize the effect of survivor bias on the interpretation of results, we excluded all patients with deaths occurring within the first 24 hours of randomization from the 1-year mortality model. Sensitivity analyses were performed including all patients (not excluding deaths ≤24 hours) and further after excluding deaths ≤7 days to evaluate the effect of gender and bleeding on 1-year mortality. For all analyses, a 2-tailed p value <0.05 was considered statistically significant. All analyses were performed using SAS 8.2 (SAS Institute, Cary, North Carolina).
Results
Bleeding occurred in 5,279 of 73,994 men (7.1%) compared to 3,377 of 25,385 women (13.3%, p <0.0001). Moderate bleeding occurred in 4,363 of 73,934 men (5.9%) compared to 2,767 of 25,350 women (10.9%, p <0.0001). Intracranial hemorrhage occurred in 440 of 73,984 men (0.59%) and 264 of 25,381 women (1.04%, p <0.0001). The most common bleeding sites were CABG (28.6%), gastrointestinal (14.6%), and genitourinary (7.2%). Bleeding at the site of cardiac catheterization (reported in GUSTO III only) occurred in 35.0% of subjects (32.4% vs 38.5% in men vs women, p = 0.0249). Compared to those without bleeding, patients who had this event were more likely to be older and women and have more co-morbid conditions ( Table 1 ).
| Characteristics | Men Without Bleeding (n = 68,715) | Men With Bleeding (n = 5,279) | Women Without Bleeding (n = 22,008) | Women With Bleeding (n = 3,377) |
|---|---|---|---|---|
| Age (years) | 59 (50,68) | 65 (56,73) | 68 (59,75) | 71 (63,77) |
| Caucasian race | 51,247 (91.9%) | 4,153 (91.3%) | 15,850 (92.6) | 2,861 (92.0%) |
| Weight (kg) | 80 (72,90) | 78 (70,87) | 70 (60,79) | 65 (57,75) |
| Hypertension | 24,500 (35.7%) | 2,153 (40.9%) | 11,955 (54.4%) | 1,875 (55.7%) |
| Diabetes mellitus | 8,967 (13.1%) | 796 (15.1%) | 4,632 (21.1%) | 677 (20.1%) |
| Current smoker | 28,827 (42.3%) | 1,823 (34.9%) | 6,579 (32.2%) | 976 (29.2%) |
| Increased cholesterol | 15,742 (31.0%) | 1,451 (33.0%) | 5,891 (35.0%) | 1,129 (40.2%) |
| Previous myocardial infarction | 11,556 (16.8%) | 1,077 (20.4%) | 3,051 (13.9%) | 497 (14.7%) |
| Previous stroke | 860 (1.7%) | 118 (2.6%) | 415 (2.4%) | 102 (3.5%) |
| Previous heart failure | 1,250 (1.8%) | 159 (3.0%) | 823 (3.8%) | 147 (4.4%) |
| Previous coronary angioplasty | 3,073 (4.5%) | 263 (5.0%) | 607 (2.8%) | 126 (3.7%) |
| Previous coronary bypass | 2,654 (3.9%) | 316 (6.0%) | 455 (2.1%) | 101 (3.0%) |
| Presenting features | ||||
| Killip class ≥III | 1,203 (1.8%) | 150 (2.9%) | 729 (3.3%) | 125 (3.7%) |
| Heart rate (beats/min) | 74 (62,85) | 74 (62,87) | 75 (64,88) | 74 (63,88) |
| Systolic blood pressure (mm Hg) | 130 (117,149) | 130 (113,146) | 130 (115,150) | 129 (110,148) |
| Diastolic blood pressure (mm Hg) | 80 (70,90) | 80 (70,90) | 80 (69,90) | 74 (62,85) |
| Presenting pulse pressure (mm Hg) | 50 (40,60) | 50 (40,61) | 54 (42,66) | 54 (42,68) |
| Body mass index ≥30 kg/m 2 | 12,960 (20.3%) | 750 (15.3%) | 5,160 (25.8%) | 610 (19.9%) |
| Myocardial infarction location | ||||
| Anterior | 28,326 (41.6%) | 2,284 (43.6%) | 9,066 (41.6%) | 1,337 (39.9%) |
| Inferior | 37,254 (54.7%) | 2,770 (52.9%) | 11,682 (53.6%) | 1,902 (56.8%) |
| Other | 2,558 (3.8%) | 185 (3.5%) | 1,063 (4.9%) | 110 (3.3%) |
| Initial creatinine clearance (ml/min) | 83 (65,104) | 70 (54,88) | 63 (47,81) | 54 (41,70) |
| Baseline hemoglobin (g/dl) ⁎ | 15 (14,16) | 15 (13,16) | 14 (13,15) | 13 (12,14) |
| Electrocardiographic findings | ||||
| No evidence of myocardial infarction (GUSTO I, IIB, III) | 1,235 (3.4%) | 106 (3.0%) | 425 (3.7%) | 77 (3.2%) |
| Median sum of ST-segment elevations (mm) (GUSTO I, IIB) | 7.0 (4.0,12.0) | 8.0 (5.0,12.0) | 7.0 (4.0,12.0) | 8.0 (4.0,12.0) |
Medical treatments for the most part were similar in patients with and without bleeding ( Table 2 ). Prevalence of left main or 3-vessel coronary artery disease was greater for patients with bleeding ( Table 3 ). Compared to men with bleeding, women with this event were more likely to be older and nonsmokers and to have hypertension, diabetes, obesity (body mass index >30 kg/m 2 ), and previous congestive heart failure ( Table 1 ). Presenting pulse pressure was significantly higher, whereas initial hemoglobin and creatinine clearance were significantly lower, in women with bleeding compared to men who had this event. Invasive procedures were performed less frequently in women with bleeding compared to men with bleeding.
| Treatment | Men Without Bleeding (n = 68,715) | Men With Bleeding (n = 5,279) | Women Without Bleeding (n = 22,008) | Women With Bleeding (n = 3,377) |
|---|---|---|---|---|
| Drug therapies in hospital | ||||
| Aspirin | 50,962 (90.0%) | 4,615 (91.8%) | 15,533 (89.8%) | 2,958 (91.8%) |
| β Blockers | 44,492 (78.9%) | 3,813 (76.2%) | 12,816 (74.5%) | 2,393 (74.9%) |
| Angiotensin-converting enzyme inhibitors | 20,149 (36.3%) | 1,773 (35.8%) | 6,432 (37.9%) | 1,193 (37.8%) |
| Heparin | 24,598 (83.7%) | 1,548 (90.7%) | 7,655 (84.5%) | 1,017 (88.9%) |
| Fibrinolytic agents | ||||
| Fibrin specific (tissue plasminogen activator, reteplase, tenecteplase) | 40,868 (59.5%) | 3,155 (59.8%) | 12,543 (57.0%) | 2,083 (61.7%) |
| Combination (streptokinase + tissue plasminogen activator) | 6,884 (10.0%) | 886 (16.8%) | 1,991 (9.1%) | 577 (17.1%) |
| Time from symptom onset to treatment (hours) | 2.8 (2.0, 4.0) | 2.8 (1.9, 4.0) | 3.2 (2.3, 4.3) | 3.0 (2.1, 4.3) |
| Glycoprotein IIb/IIIa antagonist | 1,597 (5.9%) | 163 (10.1%) | 431 (5.1%) | 87 (8.5%) |
| Drug therapies at discharge | ||||
| Aspirin | 10,975 (88.7%) | 640 (74.4%) | 3,512 (85.4%) | 452 (73.4%) |
| β Blockers | 23,668 (64.2%) | 1,759 (48.8%) | 6,552 (58.9%) | 1,085 (48.4%) |
| Angiotensin-converting enzyme inhibitor | 9,157 (24.8%) | 872 (23.4%) | 2,992 (26.9%) | 601 (26.0%) |
| Procedures | ||||
| Cardiac catheterization | 29,871 (43.5%) | 3,531 (67.0%) | 7,988 (36.2%) | 2,043 (60.6%) |
| Coronary angioplasty | 13,834 (20.2%) | 1,185 (22.5%) | 3,849 (17.5%) | 886 (26.3%) |
| Coronary bypass | 2,927 (4.3%) | 1,790 (34.0%) | 622 (2.8%) | 792 (23.5%) |
| Angiographic Findings | Men Without Bleeding (n = 1,434) | Men With Bleeding (n = 451) | Women Without Bleeding (n = 319) | Women With Bleeding (n = 224) |
|---|---|---|---|---|
| Time to catheterization from study drug initiation (hours) | 2.9 (1.6,24.0) | 1.9 (1.5,4.0) | 3.0 (1.8,24.8) | 1.9 (1.5,3.2) |
| Left ventricular ejection fraction (%) | 59.0 (48.8,67.9) | 57 (46.6,65.6) | 61.3 (50.6,69.9) | 59.4 (48.2,69.5) |
| Thrombosis In Myocardial Infarction flow grade | ||||
| 0 | 253 (18.4%) | 92 (21.0%) | 46 (15.8%) | 46 (20.6%) |
| 1 | 96 (7.0%) | 36 (8.2%) | 15 (5.1%) | 18 (8.3%) |
| 2 | 409 (29.8%) | 134 (30.6%) | 89 (30.5%) | 72 (33.0%) |
| 3 | 615 (44.8%) | 176 (40.2%) | 142 (48.6%) | 83 (38.1%) |
| Patency of infarct-related artery ⁎ | 1,024 (74.6%) | 310 (70.9%) | 231 (79.1%) | 155 (71.1%) |
| Number of narrowed coronary arteries | ||||
| 0 | 118 (8.6%) | 23 (5.2%) | 35 (11.9%) | 6 (2.8%) |
| 1 | 749 (54.5%) | 206 (46.9%) | 163 (55.6%) | 109 (50.7%) |
| 2 | 313 (22.8%) | 129 (29.4%) | 57 (19.5%) | 62 (28.8%) |
| 3 | 194 (14.1%) | 81 (18.5%) | 38 (13.0%) | 38 (17.7%) |
| Left main coronary artery disease >50% stenosis | 24 (1.7%) | 16 (3.6%) | 3 (1.0%) | 7 (3.3%) |
Rates of most in-hospital adverse events were higher in patients with bleeding compared to those without bleeding ( Table 4 ). In-hospital complications were higher for women compared to men with bleeding.
| Events | Men Without Bleeding (n = 68,715) | Men With Bleeding (n = 5,279) | Women Without Bleeding (n = 22,008) | Women With Bleeding (n = 3,377) |
|---|---|---|---|---|
| In-hospital events | ||||
| Recurrent ischemia or angina | 1,131 (6.7%) | 75 (9.6%) | 355 (7.3%) | 58 (11.7%) |
| Reinfarction | 2,243 (3.3%) | 415 (7.9%) | 918 (4.2%) | 347 (10.3%) |
| Heart failure or pulmonary edema | 6,176 (11.0%) | 1,251 (25.2%) | 2,775 (16.2%) | 1,017 (32.1%) |
| Shock | 2,645 (3.9%) | 667 (12.7%) | 1,541 (7.0%) | 509 (15.1%) |
| Stroke | 359 (0.5%) | 526 (10.4%) | 178 (0.8%) | 330 (10.2%) |
| Death | 3,152 (4.6%) | 781 (14.8%) | 2,369 (10.8%) | 577 (17.1%) |
| Death or reinfarction | 5,063 (7.4%) | 1,080 (20.5%) | 3,046 (13.8%) | 843 (25.0%) |
| 30-day death | 3,462 (5.1%) | 793 (15.0%) | 2,506 (11.4%) | 598 (17.8%) |
| Death at 1 year | 7.10% | 18.32% | 13.10% | 21.26% |
| Death at 1 year in patients with moderate bleeding | 12.76% | 16.47% | ||
| Death at 1 year in patients with severe bleeding | 35.49% | 36.41% | ||
| Death at 30 days to 1 year | 2.45% | 5.27% | 3.26% | 5.90% |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
