Abstract
BACKGROUND:
Clinical characteristics of patients with acute myocardial infarction after gastrointestinal bleeding are poorly characterized. We sought to evaluate the incidence, management and outcomes of myocardial infarction following hospitalization for gastrointestinal bleeding.
METHODS:
Patients admitted with a diagnosis of gastrointestinal bleeding with and without subsequent hospital readmissions for acute myocardial infarction within 90 days were identified in the 2014 U.S. Nationwide Readmission Database. Patients with myocardial infarction with and without a recent prior gastrointestinal bleed were compared to determine differences in management and in-hospital outcomes. Logistic regression models were used to estimate odds of invasive management and all-cause in-hospital mortality after covariate adjustment.
RESULTS:
A total of 644,622 patients with gastrointestinal bleeding were identified, of which 7523 (1.2%) were readmitted for myocardial infarction within 90 days. Compared to patients with myocardial infarction without recent gastrointestinal bleeding, patients with myocardial infarction within 90 days after gastrointestinal bleeding were older, more likely to be women, have kidney disease, presented with non-ST segment elevation myocarsdial infarction, and were less likely to undergo invasive management of acute myocardial infarction (28% vs 63%, P < .01). Prior gastrointestinal bleeding was associated with higher all-cause in-hospital myocardial infarction mortality (22% vs 9%, P < .01).
CONCLUSION:
In the first 3 months after hospitalization for gastrointestinal bleeding, 1 of every 83 patients was readmitted with acute myocardial infarction. Patients with myocardial infarction after gastrointestinal bleeding were less likely to undergo invasive management and coronary revascularization and had higher mortality than those without recent bleeding.
Keywords: Bleeding, Mortality, Myocardial infarction, Readmission
INTRODUCTION
Bleeding is a challenge in the management of cardiovascular disease. Acute blood loss often leads to interruption or discontinuation of clinically indicated antithrombotic therapies and may provoke a prothrombotic systemic response.1,2 Patients who survive acute bleeding are at increased risk for subsequent thrombotic complications, including acute myocardial infarction.3,4 When myocardial infarction occurs in the setting of antecedent bleeding, concerns for subsequent bleeding may impact the intensity and duration of guideline-directed antithrombotic therapy and decisions regarding coronary revascularization.5–7 However, data addressing the characteristics, management, and outcomes of patients with myocardial infarction following a recent bleeding event are limited. Studies focused on coronary revascularization in patients at high bleeding risk included relatively few individuals with a recent hospital admission for bleeding.8–10 Therefore, the objective of this study was to describe the baseline characteristics and risk factors associated with acute myocardial infarction occurring in the first 3 months after hospitalization for gastrointestinal bleeding and compare the management and outcomes of patients with myocardial infarction with and without recent gastrointestinal bleeding using a large national database.
METHODS
Study Population
Adults ≥18 years of age hospitalized in 2014 with a diagnosis of gastrointestinal bleeding were identified from the U.S. Nationwide Readmission Database (NRD). The NRD is an administrative database containing discharge-level administrative data from 22 states and 49.3% of all hospitalizations in the United States in 2014.11 Bleeding during hospitalization was identified by International Classification of Disease, Ninth Revision (ICD-9) diagnosis codes in any position, as previously described (Supplementary Appendix, available online).12 The year 2014 was the most recent full year in which ICD-9 codes were reported in the United States. We included all hospital discharges with gastrointestinal bleeding that occurred between January and September 2014 to ensure complete 90-day follow-up within the calendar year. Patients who died during the index admission for bleeding were excluded. Patients readmitted with acute myocardial infarction within 90 days of a hospitalization for gastrointestinal bleeding were identified using primary and non-primary ICD-9 diagnosis codes for ST-segment elevation myocardial infarction and non-ST-segment elevation myocardial infarction (410.01–410.61, 410.81, 410.91, and 410.71, respectively), as previously described.13 A cohort of patients hospitalized with acute myocardial infarction that was not preceded by gastrointestinal bleeding in the calendar year was selected as a comparator to evaluate the in-hospital management and outcomes of myocardial infarction.
Clinical Characteristics and In-Hospital Management
Demographics and clinical comorbidities were defined by the ICD-9 diagnosis codes, Clinical Classification Software (CCS) diagnosis codes, and U.S. Agency for Healthcare Research and Quality (AHRQ) Elixhauser comorbidities.14 Esophagogastroduodenoscopy and colonoscopy were defined by relevant ICD-9 and CCS procedure codes (Supplementary Appendix, available online). Management of acute myocardial infarction, including coronary thrombolysis, diagnostic invasive coronary angiography, percutaneous coronary intervention (PCI), and surgical revascularization with coronary artery bypass grafting (CABG) were identified by ICD-9 and CCS procedure codes (Supplementary Appendix, available online). Patients with acute myocardial infarction who underwent diagnostic coronary angiography, PCI, or CABG during the index admission, were considered to have been invasively managed. Patients who did not undergo invasive management were considered to have been conservatively managed.
Outcomes
Adverse outcomes included cardiac arrest (defined by ICD-9 diagnosis code 427.5), cardiogenic shock (defined by ICD-9 diagnosis code 785.51), acute ischemic stroke (defined by ICD-9 diagnosis codes 433.x1, 434.x1, 46, and 437.1), and acute respiratory failure (Supplementary Appendix, available online), as previously defined.15–17 All-cause, in-hospital mortality was determined from the NRD discharge disposition.
Statistical Analysis
Continuous variables were reported as mean and standard errors and were compared using linear regression. Categorical variables were reported as frequency and percentages and compared using complex samples χ2 tests. Patients hospitalized for gastrointestinal bleeding were stratified by the occurrence of a readmission for acute myocardial infarction within 90 days of discharge. To identify characteristics associated with incident myocardial infarction, multivariable logistic regression models were used to estimate adjusted odds ratios (aOR) for 90-day readmission for myocardial infarction following hospitalization for gastrointestinal bleeding.
To evaluate the impact of recent gastrointestinal bleeding on myocardial infarction management and outcomes, we compared patients hospitalized with myocardial infarction within 90 days of a recent gastrointestinal bleeding hospitalization to comparator patients admitted with myocardial infarction without antecedent gastrointestinal bleeding during the calendar year. In the comparator cohort, we excluded all myocardial infarction hospitalizations that occurred from January 2014 to March 2014 to ensure that none had gastrointestinal bleeding within the 90-day window preceding acute myocardial infarction in the prior calendar year. Patients with gastrointestinal bleeding during the acute myocardial infarction hospitalization were also excluded from the comparator cohort because the sequence and timing of the bleeding and myocardial infarction within a single hospital admission could not be determined. To address the potential for bias from the exclusion of patients with gastrointestinal bleeding during acute myocardial infarction hospitalizations in the comparator cohort, we performed a sensitivity analysis excluding any patients in either group who had gastrointestinal bleeding during the acute myocardial infarction hospitalization. Multivariable logistic regression models adjusted for demographics and clinical covariates with a univariate P value ≤ .10 were used to estimate the odds of invasive management, coronary revascularization, and in-hospital outcomes. Among patients with acute myocardial infarction and recent gastrointestinal bleeding, the relationship between invasive management and in-hospital mortality was also assessed with multivariable logistic regression models adjusted for demographics and clinical comorbidities.
To confirm our primary findings in a larger cohort, we evaluated 30-day readmission for myocardial infarction after a gastrointestinal bleed hospitalization from January to November 2014. Similarly, we evaluated the management and outcomes of patients admitted with acute myocardial infarction within 30 days of discharge following hospitalization for gastrointestinal bleeding compared to patients admitted with acute myocardial infarction without a preceding gastrointestinal bleed in the calendar year. In a sensitivity analysis, we evaluated the differences in management and outcomes in patients with ST-segment elevation myocardial infarction with and without prior gastrointestinal bleeding. We also performed subgroup analyses to evaluate the differential impact of localized gastrointestinal bleeding versus diffuse or ambiguous sources of gastrointestinal bleeding (identified as both upper and lower gastrointestinal bleeding,] or when the bleeding source was unclear and could not be localized) on the subsequent management and outcomes of myocardial infarction.
Sampling weights were applied to all analyses to determine national incidence estimates, incorporating prespecified clustering and strata as per Agency for Healthcare Research and Quality guidance. In analyses evaluating the incidence of myocardial infarction readmission after gastrointestinal bleeding, sampling weights were applied based on the index hospitalization for gastrointestinal bleeding. In the analyses of acute myocardial infarction management with and without prior gastrointestinal bleeding, cases were weighted according to the acute myocardial infarction hospitalization. Differences in the reference sampling weights account for the slight differences in estimates in various analyses. All statistical analyses were performed using SPSS version 27 (IBM Corporation). The NRD is a publicly available, deidentified data set, and the study did not require local institutional board review.
RESULTS
Gastrointestinal Bleed with and Without Subsequent Acute Myocardial Infarction Readmission
A total of 644,622 hospital discharges for gastrointestinal bleeding were identified between January 2014 and September 2014. Among patients who survived to hospital discharge after their initial gastrointestinal bleed, readmission for myocardial infarction within 90 days occurred in 7523 (1.2%). The cumulative incidence of readmission for myocardial infarction in the 90 days after discharge from hospitalization for gastrointestinal bleeding is shown in Figure 1. When bleeding was localized to either the upper or lower gastrointestinal tract, 90-day readmission for myocardial infarction occurred in 1.0%, and in those with a diffuse or ambiguous bleeding source, 90-day readmission for myocardial infarction occurred in 1.6% (P < .01). Baseline characteristics of patients with and without 90-day readmission for myocardial infarction are shown in Table 1. Patients readmitted for myocardial infarction were older, had greater burden of cardiovascular risk factors and established cardiovascular disease, and were more likely to have had prior coronary revascularization. Patients who were readmitted with myocardial infarction were also more likely to have received red blood cell transfusion and more likely to have suffered a myocardial infarction during their index hospital admission for gastrointestinal bleeding. Characteristics of the index gastrointestinal bleed presentation in patients with and without 90-day readmission for myocardial infarction are shown in Table 1. Similar characteristics were observed in a sensitivity analysis of patients with and without 30-day readmission for myocardial infarction (Supplementary Table 1, available online).
Figure 1.
Cumulative incidence of readmission for acute myocardial infarction after hospitalization for gastrointestinal bleeding.
Table 1.
Baseline and Presenting Characteristics of Patients Hospitalized with Gastrointestinal Bleeding Stratified by 90-Day Readmission for Acute Myocardial Infarction
| Variable | All GIB Hospitalizations (n = 644,622) | Readmission for AMI Within 90 Days (n = 7523) | No Readmission for AMI Within 90 Days (n = 637,099) | P Value |
|---|---|---|---|---|
| Demographics | ||||
| Age, mean (SE) | 67.2 (0.15) | 74.2 (0.27) | 67.1 (0.15) | <.01 |
| Female sex | 312,364 (48.5%) | 3490 (46.4%) | 308,874 (48.5%) | .03 |
| Cardiovascular Risk Factors and Diseases | ||||
| Tobacco use (current or prior) | 193,110 (30.0%) | 2090 (27.8%) | 191,020 (30.0%) | .01 |
| Hyperlipidemia | 223,895 (34.7%) | 3544 (47.1%) | 220,351 (34.6%) | <.01 |
| Hypertension | 414,526 (64.3%) | 5877 (78.1%) | 408,649 (64.1%) | <.01 |
| Diabetes mellitus | 190,089 (29.5%) | 3226 (42.9%) | 186,863 (29.3%) | <.01 |
| Atrial fibrillation/atrial flutter | 130,689 (20.3%) | 2108 (28.0%) | 128,581 (20.2%) | <.01 |
| Congestive heart failure | 103,286 (16.0%) | 2196 (29.2%) | 101,090 (15.9%) | <.01 |
| Valvular heart disease | 50,749 (7.9%) | 919 (12.2%) | 49,830 (7.8%) | <.01 |
| Peripheral artery disease | 59,654 (9.3%) | 1366 (18.2%) | 58,288 (9.1%) | <.01 |
| Coronary artery disease | 187,112 (27.3%) | 4203 (55.9%) | 182,909 (26.9%) | <.01 |
| Prior PCI | 43,596 (6.8%) | 1041 (13.8%) | 42,555 (6.7%) | <.01 |
| Prior CABG | 41,015 (6.4%) | 973 (12.9%) | 40,042 (6.3%) | <.01 |
| Other Comorbidities | ||||
| Alcohol use | 71,909 (11.2%) | 337 (4.5%) | 71,572 (11.2%) | <.01 |
| Drug abuse | 25,815 (4.0%) | 152 (2.0%) | 25,663 (4.0%) | <.01 |
| Obesity | 73,049 (11.3%) | 828 (11.0%) | 72,221 (11.3%) | .58 |
| Liver disease | 57,719 (9.0%) | 446 (5.9%) | 57,273 (9.0%) | <.01 |
| Chronic kidney disease | 131,129 (20.3%) | 2905 (38.6%) | 128,224 (20.1%) | <.01 |
| Malignancy | 45,426 (7.0%) | 544 (7.2%) | 44,882 (7.0%) | .72 |
| Anemia | 225,871 (32.9%) | 3077 (40.9%) | 222,794 (32.8%) | <.01 |
| Coagulopathy | 71,639 (11.1%) | 832 (11.1%) | 70,807 (11.1%) | .92 |
| Rheumatoid arthritis/collagen vascular disease | 21,367 (3.3%) | 302 (4.0%) | 21,065 (3.3%) | .04 |
| Chronic pulmonary disease | 138,431 (21.5%) | 2102 (27.9%) | 136,329 (21.4%) | <.01 |
| Hypothyroidism | 85,943 (13.3%) | 1212 (16.1%) | 84,731 (13.3%) | <.01 |
| Fluid and electrolyte disorders | 245,764 (38.1%) | 3145 (41.8%) | 242,619 (38.1%) | <.01 |
| Gastrointestinal Bleeding Source During Index GIB Admission | ||||
| Lower GIB | 109,567 (17.0%) | 1068 (14.2%) | 108,499 (17.0%) | <.01 |
| Upper GIB | 327,261 (50.8%) | 3220 (42.8%) | 324,041 (50.8%) | <.01 |
| Upper and lower GIB | 4507 (0.7%) | 64 (0.9%) | 4443 (0.7%) | .85 |
| Unknown source of GIB | 203,288 (31.5%) | 3171 (42.2%) | 200,117 (31.4%) | <.01 |
| Management of Gastrointestinal Bleeding During Index GIB Admission | ||||
| EGD during initial GIB admission | 303,208 (47.0%) | 3312 (44.0%) | 299,896 (47.1%) | <.01 |
| Colonoscopy during initial GIB admission | 119,437 (18.5%) | 1280 (17.0%) | 118,157 (18.5%) | .56 |
| RBC transfusion | 259,038 (40.2%) | 3681 (48.9%) | 255,357 (40.1%) | <.01 |
| Platelet transfusion | 16,127 (2.5%) | 213 (2.8%) | 15,914 (2.5%) | .31 |
| Acute Cardiopulmonary Disease During Index GIB Admission | ||||
| Respiratory failure/mechanical ventilation | 33,301 (5.2%) | 609 (8.1%) | 32,692 (5.1%) | <.01 |
| AMI during initial GIB hospitalization | 20,659 (3.2%) | 1519 (20.2%) | 19,140 (3.0%) | <.01 |
AMI = acute myocardial infarction; CABG = coronary artery bypass grafting; EGD = esophagogastroduodenoscopy; GIB = gastrointestinal bleeding; PCI = percutaneous coronary intervention; RBC = red blood cell; SE = standard error.
Patient characteristics independently associated with 90-day readmission for myocardial infarction after gastrointestinal bleeding are shown in Table 2. Recent myocardial infarction during the index gastrointestinal bleed hospitalization, coronary artery disease, older age, and comorbidities including chronic kidney disease, diabetes mellitus, heart failure, and peripheral artery disease were associated with 90-day readmission with acute myocardial infarction. Bleeding from an unknown gastrointestinal source was also independently associated with myocardial infarction read-mission. Similar clinical characteristics were also independently associated with 30-day readmission for myocardial infarction after gastrointestinal bleeding (Supplementary Table 2, available online).
Table 2.
Independent Predictors of 90-Day Readmission for Acute Myocardial Infarction Following Hospitalization for Gastrointestinal Bleeding
| Characteristic | aOR (95% CI)* | χ2 Statistic |
|---|---|---|
| AMI during initial GIB admission | 4.98 (4.44–5.60) | 743.6 |
| Coronary artery disease | 1.84 (1.67–2.03) | 155.4 |
| Age | 1.10 (1.08–1.12) | 125.1 |
| Chronic kidney disease | 1.40 (1.28–1.53) | 56.8 |
| Diabetes mellitus | 1.52 (1.34–1.72) | 44.1 |
| Congestive heart failure | 1.26 (1.15–1.37) | 27.2 |
| Peripheral artery disease | 1.26 (1.14–1.39) | 21.4 |
| Alcohol use | 0.68 (0.57–0.82) | 17.7 |
| Hypertension | 1.17 (1.07–1.29) | 11.1 |
| Anemia | 1.15 (1.06–1.25) | 11.1 |
| Source of GIB: | 9.0 | |
| Upper GIB | (Reference) | |
| Lower GIB | 0.93 (0.82–1.06) | |
| Upper and lower GIB | 1.36 (0.89–2.09) | |
| Unknown bleeding source | 1.20 (1.11–1.31) | |
| Mechanical ventilation during initial GIB admission | 1.21 (1.06–1.37) | 8.4 |
| Chronic pulmonary disease | 1.14 (1.04–1.25) | 7.8 |
| Prior PCI | 1.18 (1.05–1.32) | 7.7 |
| Atrial fibrillation/atrial flutter | 0.88 (0.80–0.97) | 6.7 |
| RBC transfusion | 1.10 (1.01–1.19) | 5.3 |
AMI = acute myocardial infarction; aOR = adjusted odds ratio; CI = confidence interval; GIB = gastrointestinal bleeding; PCI = percutaneous coronary intervention; RBC = red blood cell.
Model adjusted for age, sex, and all baseline characteristics that differed between groups by prespecified criteria (P ≤ .10).
Acute Myocardial Infarction with and Without Antecedent Hospitalization for Gastrointestinal Bleeding
A total of 531,019 admissions for acute myocardial infarction that were not preceded by gastrointestinal bleeding during the calendar year were compared to 7444 readmissions for myocardial infarction that occurred within 90 days of discharge from a hospitalization for gastrointestinal bleeding. Compared to patients with myocardial infarction and without recent gastrointestinal bleeding, patients with myocardial infarction after gastrointestinal bleeding were older, more likely to be women, and had a higher burden of cardiovascular and noncardiovascular comorbidities (Table 3). Similar characteristics were observed in sensitivity analyses of patients who were readmitted for myocardial infarction within 30 days of gastrointestinal bleeding hospitalization (Supplementary Table 3, available online).
Table 3.
Baseline Characteristics of Patients with Acute Myocardial Infarction Stratified by Presence of Prior Gastrointestinal Bleeding
| Variable | All AMI Hospitalizations (n = 538,463) | AMI with Prior GIB (≤90 Days) (n = 7444) | AMI Not Preceded by GIB in the Calendar Year (n = 531,019) | P Value |
|---|---|---|---|---|
| Demographics | ||||
| Age (mean, SE) | 68.5 (0.10) | 74.4 (0.27) | 68.4 (0.10) | <.01 |
| Female sex | 221,698 (41.2%) | 3468 (46.6%) | 218,230 (41.1%) | <.01 |
| AMI Subtype | ||||
| STEMI | 142,866 (26.5%) | 1337 (18.0%) | 141,529 (26.7%) | <.01 |
| Non-STEMI | 395,597 (73.5%) | 6107 (82.0%) | 389,490 (73.3%) | <.01 |
| Cardiovascular Risk Factors and Diseases | ||||
| Tobacco use (current or prior) | 213,294 (39.6%) | 1987 (26.7%) | 211,307 (39.8%) | <.01 |
| Hyperlipidemia | 318,431 (59.1%) | 3587 (48.2%) | 314,844 (59.3%) | <.01 |
| Hypertension | 404,894 (75.2%) | 5639 (75.8%) | 399,255 (75.2%) | .49 |
| Diabetes mellitus | 199,680 (37.1%) | 3148 (42.3%) | 196,532 (37.0%) | <.01 |
| Atrial fibrillation/atrial flutter | 111,320 (20.7%) | 2469 (33.2%) | 108,851 (20.5%) | <.01 |
| Congestive heart failure | 53,341 (9.9%) | 1680 (22.6%) | 51,661 (9.7%) | <.01 |
| Valvular heart disease | 18,922 (3.5%) | 587 (7.9%) | 18,335 (3.5%) | <.01 |
| Peripheral artery disease | 70,291 (13.1%) | 1485 (19.9%) | 68,806 (13.0%) | <.01 |
| Coronary artery disease | 399,555 (74.2%) | 4792 (64.4%) | 394,763 (74.3%) | <.01 |
| Prior PCI | 74,190 (13.8%) | 1116 (15.0%) | 73,074 (13.8%) | .05 |
| Prior CABG | 43,406 (8.1%) | 914 (12.3%) | 42,492 (8.0%) | <.01 |
| Other Comorbidities | ||||
| Alcohol use | 19,836 (3.7%) | 301 (4.0%) | 19,535 (3.7%) | .31 |
| Drug abuse | 17,130 (3.2%) | 185 (2.5%) | 16,945 (3.2%) | .03 |
| Obesity | 89,322 (16.6%) | 807 (10.8%) | 88,515 (16.7%) | <.01 |
| Liver disease | 11,666 (2.1%) | 434 (5.8%) | 10,732 (2.0%) | <.01 |
| Chronic kidney disease | 122,069 (22.7%) | 3213 (43.2%) | 118,856 (22.4%) | <.01 |
| Malignancy | 21,447 (4.0%) | 680 (9.1%) | 20,767 (3.9%) | <.01 |
| Anemia | 103,952 (19.3%) | 3391 (45.6%) | 100,561 (18.9%) | <.01 |
| Coagulopathy | 41,511 (7.7%) | 881 (11.8%) | 40,630 (7.7%) | <.01 |
| Rheumatoid arthritis/collagen vascular disease | 15,522 (2.9%) | 279 (3.7%) | 15,243 (2.9%) | <.01 |
| Chronic pulmonary disease | 122,037 (22.7%) | 2188 (29.4%) | 119,849 (22.6%) | <.01 |
| Hypothyroidism | 66,895 (12.4%) | 1174 (15.8%) | 65,721 (12.4%) | <.01 |
| Fluid and electrolyte disorders | 171,673 (31.9%) | 3454 (46.4%) | 168,219 (31.7%) | <.01 |
AMI = acute myocardial infarction; aOR = adjusted odds ratio; CABG = coronary artery bypass grafting; CI = confidence interval; GIB = gastrointestinal bleeding; PCI = percutaneous coronary intervention; SE = standard error; STEMI = ST-segment-elevation myocardial infarction.
In-hospital management of acute myocardial infarction differed considerably in patients with and without a preceding hospitalization for gastrointestinal bleeding. Patients with myocardial infarction patients and gastrointestinal bleeding in the preceding 90 days were less likely to undergo invasive management of the myocardial infarction (27.7% vs 63.1%, P < .01) compared to patients without recent gastrointestinal bleeding. PCI and CABG were also less common in patients with a recent hospitalization for gastrointestinal bleeding, as shown in Table 4. Similar associations were observed in sensitivity analyses restricted to patients with ST-segment elevation myocardial infarction with and without a preceding gastrointestinal bleed and in analyses of patients with any myocardial infarction within 30 days of hospitalization for gastrointestinal bleeding (Supplementary Tables 4 and 5, available online). The location of the gastrointestinal bleeding source during the preceding hospitalization was associated with myocardial infarction management, with lower rates of invasive management and coronary revascularization in patients with a diffuse or ambiguous gastrointestinal bleeding source (Supplementary Table 6, available online). In an analysis of patients with acute myocardial infarction who were hospitalized for gastrointestinal bleeding within the prior 90 days, invasive management of acute myocardial infarction was associated with lower risk of in-hospital mortality (11.2% vs 26.5%, P < .01; aOR 0.28; 95% confidence interval 0.26–0.29) than conservative management.
Table 4.
Management of Patients with Acute Myocardial Infarction Stratified by Presence of Prior Gastrointestinal Bleeding
| Intervention | AMI with Prior GIB (≤90 Days) (n = 7444) | AMI Not Preceded by GIB in the Calendar Year (n = 531,019) | P Value | aOR Model 1* | aOR Model 2† |
|---|---|---|---|---|---|
| Thrombolysis | 51 (0.7%) | 6560 (1.2%) | .03 | 0.61 (0.35–1.07) | 0.65 (0.38–1.12) |
| Invasive management | 2059 (27.7%) | 334,994 (63.1%) | <.01 | 0.26 (0.24–0.28) | 0.40 (0.37–0.44) |
| Coronary revascularization | 1194 (16.0%) | 248,861 (46.9%) | <.01 | 0.26 (0.23–0.28) | 0.42 (0.37–0.47) |
| PCI | 1039 (14.0%) | 211,277 (39.8%) | <.01 | 0.30 (0.26–0.33) | 0.53 (0.46–0.60) |
| CABG | 168 (2.3%) | 41,248 (7.8%) | <.01 | 0.31 (0.24–0.39) | 0.33 (0.25–0.42) |
AMI = acute myocardial infarction; aOR = adjusted odds ratio; CABG = coronary artery bypass grafting; CI = confidence interval; GIB = gastrointestinal bleeding; PCI = percutaneous coronary intervention.
Model 1: Adjusted for age, sex.
Model 2: Adjusted for age, sex, and all baseline characteristics that differed between groups by pre-specified criteria (P ≤ .10).
All-cause mortality during myocardial infarction hospitalization was substantially higher in patients with versus without an antecedent hospitalization for gastrointestinal bleeding in the prior 90-days (22.3% vs 8.9%, P < .01; aOR 1.86; 95% confidence interval 1.68–2.06). Recent gastrointestinal bleeding was also associated with increased odds of cardiac arrest and acute respiratory failure but not cardiogenic shock or ischemic stroke (Figure 2). Similar findings were observed in a sensitivity analysis of patients with and without gastrointestinal bleeding within 30 days prior to hospitalization for myocardial infarction (Supplementary Figure) as well as in a sensitivity analysis of patients with ST-segment elevation myocardial infarction (Supplementary Table 7, available online). Associations between prior gastrointestinal bleeding and in-hospital mortality during hospitalization for myocardial infarction were also observed in sensitivity analyses excluding any patients with gastrointestinal bleeding reported during the myocardial infarction hospitalization (Supplementary Tables 8 and 9, available online). Among patients with a prior gastrointestinal bleed hospitalized for myocardial infarction, all-cause mortality was higher in patients with a diffuse or ambiguous gastrointestinal bleeding source compared to those with bleeding localized to the upper or lower gastrointestinal tract (Supplementary Table 10, available online).
Figure 2.
In-hospital outcomes of AMI stratified by presence of prior gastrointestinal bleeding. AMI = acute myocardial infarction; aOR = adjusted odds ratio; GIB = gastrointestinal bleeding. aOR adjusted for age, sex, and baseline characteristics that differed between groups by prespecified criteria (P ≤ .10). Parentheses indicate 95% confidence intervals.
DISCUSSION
In a nationwide analysis, we report that 1 of every 83 patients hospitalized with gastrointestinal bleeding was readmitted with myocardial infarction within 3 months. Patients with myocardial infarction after gastrointestinal bleeding were older and had a higher burden of cardiovascular risk factors and disease, and many of these clinical characteristics were independently associated with readmission for myocardial infarction. Among patients with myocardial infarction, those with a recent gastrointestinal bleed were less likely to undergo invasive management of myocardial infarction, coronary thrombolysis, or coronary revascularization with PCI or CABG, and these patients had a nearly 2-fold greater adjusted odds of mortality than patients without recent bleeding. Nearly one-quarter of patients rehospitalized with myocardial infarction within 90 days of a gastrointestinal bleed died during the myocardial infarction admission.
Risks of bleeding in patients with cardiovascular disease are well established, and bleeding in the setting of myocardial infarction is strongly associated with mortality.18 Although the clinical characteristics and impact of gastrointestinal bleeding following myocardial infarction and PCI are well described, data describing risks of acute myocardial infarction following gastrointestinal bleeding are limited.19–21
The present analysis identifies several characteristics associated with readmission for acute myocardial infarction within 90 days of hospitalization for gastrointestinal bleeding. Chronic kidney disease, age, traditional risk factors for cardiovascular disease, and established ischemic cardiovascular disease were all independent predictors of readmission for acute myocardial infarction after gastrointestinal bleeding. Acute myocardial infarction during the index admission for gastrointestinal bleeding was also strongly associated with readmission for recurrent myocardial infarction.
Management of patients with coronary artery disease and gastrointestinal bleeding can be challenging because antithrombotic therapy indicated for the secondary prevention of cardiovascular disease can lead to increased risk of recurrent bleeding.22,23 Due to the concern for recurrent bleeding in such patients, clinically indicated antithrombotic therapy is often withheld at discharge in patients hospitalized with gastrointestinal bleeding.20 We report that patients hospitalized with gastrointestinal bleeding are at substantial risk for readmission for myocardial infarction. Patients with the highest risk of cardiovascular events, including older adults, those with established coronary artery disease, or a myocardial infarction during the index admission, may benefit from multidisciplinary collaboration between cardiologists and gastroenterologists to facilitate early resumption of statins and antithrombotic therapy.
Compared to patients with an upper gastrointestinal bleed, bleeding from an unknown gastrointestinal source was independently associated with readmission for acute myocardial infarction. This may reflect provider hesitance to prescribe antiplatelet therapy when the source of bleeding cannot be localized or controlled. Although there are few data to guide antiplatelet therapy according to the source of gastrointestinal bleeding, a recent American College of Cardiology expert consensus statement advocates for delaying resumption of anticoagulation in patients with bleeding of an unidentified source as opposed to bleeding of a known location.24
Among patients hospitalized with acute myocardial infarction, we observed that recent gastrointestinal bleeding was associated with a lower likelihood of an invasive management strategy and coronary revascularization compared to patients with acute myocardial infarction without recent gastrointestinal bleeding. Providers may be hesitant to pursue invasive management of myocardial infarction due to the need for dual antiplatelet therapy after PCI and perceived risks of postprocedural bleeding. This observation is consistent with prior studies that report lower rates of PCI in myocardial infarction patients who are at high risk of bleeding.25 Unfortunately, in the current study, conservative management of myocardial infarction in patients with prior gastrointestinal bleeding was associated with 2-fold higher in-hospital mortality than invasive management of myocardial infarction. Although recent studies have examined strategies using abbreviated courses of dual antiplatelet therapy after PCI for myocardial infarction, additional investigation is needed to establish best practices in high-risk patients with recent bleeding.26–28 Abbreviated antithrombotic regimens, contemporary ultra-thin strut drug eluting stents, and drug eluting balloon angioplasty in lieu of stent implantation may be considered to lower thrombotic risks in patients with gastrointestinal bleeding in whom dual antiplatelet therapy may not be tolerated, but many of these strategies are largely untested in this population and require further study.
Limitations
There are several limitations to these analyses. First, comorbidities and outcomes were identified using administrative coding, which may be subject to inaccurate reporting or coding errors. However, previous studies demonstrate that ICD-9 codes for gastrointestinal diseases and myocardial infarction have a high sensitivity and specificity for adjudicated diagnoses.19,29 Second, although mechanism of myocardial infarction could not be determined, sensitivity analyses evaluating the management and outcomes of ST segment elevation myocardial infarction with and without antecedent gastrointestinal bleeding demonstrated results consistent with those from the overall cohort and demonstrate the robustness of our findings. Third, within each hospital admission, the sequence of diagnoses could not be determined. In some cases, gastrointestinal bleeding may have preceded myocardial infarction, and vice versa. As such, all cases with gastrointestinal bleeding occurring during the same hospital admission as myocardial infarction were excluded from the comparator cohort without prior gastrointestinal bleeding, which may have biased study outcomes. However, sensitivity analyses excluding concurrent gastrointestinal bleeding during myocardial infarction hospitalization from both cohorts corroborated results from the primary analysis. Fourth, the interval between the onset of ischemic symptoms and hospital presentation was not recorded, and the proportion of patients with a late presentation of myocardial infarction is unknown. Fifth, discrete data regarding the extent and severity of coronary artery disease, left ventricular ejection fractions, hemoglobin concentrations, troponin concentrations and other laboratory data were not recorded in the NRD. Sixth, the NRD does not contain the prehospital, in-hospital, or discharge medication use, and consequently, the proportion of patients prescribed antithrombotic therapy is unknown. Seventh, the NRD does not contain data on outpatient diagnostics, outof-hospital death, or long-term follow-up beyond the calendar year. Finally, despite hierarchical multivariable adjustment we cannot exclude the possibility of residual confounding.
CONCLUSION
This is the first study to describe the incidence, predictors, management, and outcomes of readmission for myocardial infarction following hospitalization for gastrointestinal bleeding. In the first 3 months after hospitalization for gastrointestinal bleeding, 1 of every 83 patients was readmitted with myocardial infarction. Myocardial infarction patients with versus without recent gastrointestinal bleeding were less likely to undergo invasive management and coronary revascularization and had increased odds of acute respiratory failure and in-hospital death. Efforts are necessary to prevent acute myocardial infarction after gastrointestinal bleeding, and strategies to reduce morbidity and mortality of thrombosis in patients at high risk of bleeding are needed.
Supplementary Material
CLINICAL SIGNIFICANCE.
Approximately 1 out of 83 patients hospitalized for gastrointestinal bleeding are readmitted with acute myocardial infarction within 3 months of discharge.
These patients are less likely to undergo invasive management of the acute myocardial infarction compared with patients without a recent gastrointestinal bleed.
Patients readmitted for acute myocardial infarction within 3 months of a hospitalization for gastrointestinal bleeding have in-hospital mortality that exceeds 20%.
Funding:
NRS is supported, in part, by the National Heart, Lung, and Blood Institute of the National Institutes of Health under Award Number K23HL150315. JSB is funded, in part, by the National Heart and Lung Blood Institute of the National Institute of Health (R01HL139909 and R35HL144993).
Footnotes
Conflicts of Interest: None.
SUPPLEMENTARY DATA
Supplementary data to this article can be found online at https://doi.org/10.1016/j.amjmed.2022.03.030.
References
- 1.Stephan B, Schenk JF, Pindur G, et al. Influence of bleeding on haemorheology and haemostasis in surgery. Clin Hemorheol Microcirc 2010;46(2–3):183–8. [DOI] [PubMed] [Google Scholar]
- 2.Campbell CL, Steinhubl SR, Hooper WC, et al. Bleeding events are associated with an increase in markers of inflammation in acute coronary syndromes: an ACUITY trial substudy. J Thromb Thrombolysis 2011;31(2):139–45. [DOI] [PubMed] [Google Scholar]
- 3.Held C, Hylek EM, Alexander JH, et al. Clinical outcomes and management associated with major bleeding in patients with atrial fibrillation treated with apixaban or warfarin: insights from the ARISTOTLE trial. Eur Heart J 2015;36(20):1264–72. [DOI] [PubMed] [Google Scholar]
- 4.Wu CJ, Lin HJ, Weng SF, et al. Acute myocardial infarction risk in patients with coronary artery disease doubled after upper gastrointestinal tract bleeding: a nationwide nested case-control study. PLoS One 2015;10(11):e0142000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Levine GN, Bates ER, Bittl JA, et al. 2016 ACC/AHA guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines. J Am Coll Cardiol 2016;68(10):1082–115. [DOI] [PubMed] [Google Scholar]
- 6.Urban P, Mehran R, Colleran R, et al. Defining high bleeding risk in patients undergoing percutaneous coronary intervention: a consensus document from the Academic Research Consortium for High Bleeding Risk. Eur Heart J 2019;40(31):2632–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Roy P, Bonello L, Torguson R, et al. Impact of “nuisance” bleeding on clopidogrel compliance in patients undergoing intracoronary drug-eluting stent implantation. Am J Cardiol 2008;102(12):1614–7. [DOI] [PubMed] [Google Scholar]
- 8.Valgimigli M, Patialiakas A, Thury A, et al. Zotarolimus-eluting versus bare-metal stents in uncertain drug-eluting stent candidates. J Am Coll Cardiol 2015;65(8):805–15. [DOI] [PubMed] [Google Scholar]
- 9.Urban P, Meredith IT, Abizaid A, et al. Polymer-free drug-coated coronary stents in patients at high bleeding risk. N Engl J Med 2015;373 (21):2038–47. [DOI] [PubMed] [Google Scholar]
- 10.Mehran R, Chandrasekhar J, Urban P, et al. Sex-based outcomes in patients with a high bleeding risk after percutaneous coronary intervention and 1-month dual antiplatelet therapy: a secondary analysis of the LEADERS FREE randomized clinical trial. JAMA Cardiol 2020;5 (8):939–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Eff Clin Pract 2002;5(3):143–51. [PubMed] [Google Scholar]
- 12.Selak V, Kerr A, Poppe K, et al. Annual risk of major bleeding among persons without cardiovascular disease not receiving antiplatelet therapy. JAMA 2018;319(24):2507–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Smilowitz NR, Gupta N, Guo Y, Berger JS, Bangalore S. Perioperative acute myocardial infarction associated with non-cardiac surgery. Eur Heart J 2017;38(31):2409–17. [DOI] [PubMed] [Google Scholar]
- 14.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care 1998;36(1):8–27. [DOI] [PubMed] [Google Scholar]
- 15.Smilowitz NR, Gupta N, Ramakrishna H, Guo Y, Berger JS, Bangalore S. perioperative major adverse cardiovascular and cerebrovascular events associated with noncardiac surgery. JAMA Cardiol 2017;2 (2):181–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Brinjikji W, Rabinstein AA, Kallmes DF, Cloft HJ. Patient outcomes with endovascular embolectomy therapy for acute ischemic stroke: a study of the national inpatient sample: 2006 to 2008. Stroke 2011;42 (6):1648–52. [DOI] [PubMed] [Google Scholar]
- 17.Vallabhajosyula S, Kashani K, Dunlay SM, et al. Acute respiratory failure and mechanical ventilation in cardiogenic shock complicating acute myocardial infarction in the USA, 2000–2014. Ann Intensive Care 2019;9(1):96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Mehran R, Pocock S, Nikolsky E, et al. Impact of bleeding on mortality after percutaneous coronary intervention results from a patient-level pooled analysis of the REPLACE-2 (randomized evaluation of PCI linking angiomax to reduced clinical events), ACUITY (acute catheterization and urgent intervention triage strategy), and HORIZONS-AMI (harmonizing outcomes with revascularization and stents in acute myocardial infarction) trials. JACC Cardiovasc Interv 2011;4 (6):654–64. [DOI] [PubMed] [Google Scholar]
- 19.Gupta K, Khan A, Kumar M, Sawalha K, Abozenah M, Singhania R. Readmissions rates after myocardial infarction for gastrointestinal bleeding: a national perspective. Dig Dis Sci 2021;66(3):751–9. [DOI] [PubMed] [Google Scholar]
- 20.Nikolsky E, Stone GW, Kirtane AJ, et al. Gastrointestinal bleeding in patients with acute coronary syndromes: incidence, predictors, and clinical implications: analysis from the ACUITY (Acute Catheterization and Urgent Intervention Triage Strategy) trial. J Am Coll Cardiol 2009;54(14):1293–302. [DOI] [PubMed] [Google Scholar]
- 21.Moukarbel GV, Signorovitch JE, Pfeffer MA, et al. Gastrointestinal bleeding in high risk survivors of myocardial infarction: the VALIANT Trial. Eur Heart J 2009;30(18):2226–32. [DOI] [PubMed] [Google Scholar]
- 22.Chan FK, Leung Ki EL, Wong GL, et al. Risks of bleeding recurrence and cardiovascular events with continued aspirin use after lower gastrointestinal hemorrhage. Gastroenterology 2016;151(2):271–7. [DOI] [PubMed] [Google Scholar]
- 23.Sung JJ, Lau JY, Ching JY, et al. Continuation of low-dose aspirin therapy in peptic ulcer bleeding: a randomized trial. Ann Intern Med 2010;152(1):1–9. [DOI] [PubMed] [Google Scholar]
- 24.Tomaselli GF, Mahaffey KW, Cuker A, et al. 2020 ACC Expert Consensus Decision Pathway on Management of Bleeding in Patients on Oral Anticoagulants: A Report of the American College of Cardiology solution set oversight committee. J Am Coll Cardiol 2020;76(5):594–622. [DOI] [PubMed] [Google Scholar]
- 25.Rubinfeld GD, Smilowitz NR, Berger JS, Newman JD. Association of thrombocytopenia, revascularization, and in-hospital outcomes in patients with acute myocardial infarction. Am J Med 2019;132 (8):942–948.e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Vranckx P, Valgimigli M, Juni P, et al. Ticagrelor plus aspirin for 1 month, followed by ticagrelor monotherapy for 23 months vs aspirin plus clopidogrel or ticagrelor for 12 months, followed by aspirin monotherapy for 12 months after implantation of a drug-eluting stent: a multicentre, open-label, randomised superiority trial. Lancet 2018;392(10151):940–9. [DOI] [PubMed] [Google Scholar]
- 27.Mehran R, Baber U, Sharma SK, et al. Ticagrelor with or without aspirin in high-risk patients after PCI. N Engl J Med 2019;381(21):2032–42. [DOI] [PubMed] [Google Scholar]
- 28.Watanabe H, Domei T, Morimoto T, et al. Effect of 1-month dual antiplatelet therapy followed by clopidogrel vs 12-month dual antiplatelet therapy on cardiovascular and bleeding events in patients receiving PCI: The STOPDAPT-2 randomized clinical trial. JAMA 2019;321 (24):2414–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Freedberg DE, Kim LS, Yang YX. The risks and benefits of long-term use of proton pump inhibitors: expert review and best practice advice from the American Gastroenterological Association. Gastroenterology 2017;152(4):706–15. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.