Abstract: Objectives. Patients with cirrhosis have increased bleeding risk due to coagulopathy and platelet sequestration, as well as inherent cardiovascular risk. We aim to assess the impact of cirrhosis on the revascularization rates and in-hospital outcomes in patients with acute myocardial infarction (AMI). Methods. We queried the National Inpatient Sample Database from 2010 to 2014 and identified hospitalizations with a primary diagnosis of AMI (n = 612,547); of these, a total of 3135 patients had a concomitant diagnosis of cirrhosis. We compared clinical outcomes between patients with cirrhosis and a propensity-score matched cohort without cirrhosis (n = 3086). Results. Patients with cirrhosis had a lower rate of ST-elevation MI (18.9% vs 26.7% in the cohort with no cirrhosis; P<.001), a lower rate of coronary angiography (51.4% vs 63.9% in the cohort with no cirrhosis; P<.001), and lower rates of revascularization by percutaneous coronary intervention (PCI) (28.7% vs 39.2% in the cohort with no cirrhosis; P<.001) or coronary artery bypass grafting (6.0% vs 12.9% in the cohort with no cirrhosis; P<.001). Gastrointestinal and postprocedural hemorrhage was more common in patients with cirrhosis (12.3% vs 7.1% in the cohort with no cirrhosis; P<.001), regardless of revascularization status, and cirrhosis patients also had a higher in-hospital mortality rate (8.7% vs 6.9% in the cohort with no cirrhosis; P<.01). PCI was independently associated with lower mortality in patients with cirrhosis (odds ratio, 0.57; 95% confidence interval, 0.33-0.98; P=.04). Conclusion. Patients with cirrhosis presenting with AMI were highly selected to undergo coronary angiography and subsequent revascularization, and had higher mortality than those without cirrhosis. However, PCI was independently associated with lower mortality in patients with cirrhosis, although to less effect than non-cirrhotics, perhaps due to higher bleeding rates.
J INVASIVE CARDIOL 2019;31(7):E162-E169.
Key words: acute myocardial infarction, bleeding, percutaneous coronary intervention, STEMI
Cirrhosis is the eighth leading cause of death in the United States and is escalating in prevalence.1 The etiology of chronic liver disease has been changing due to an increase in non-alcoholic fatty liver disease (NAFLD),2 which shares many risk factors with cardiovascular disease. NAFLD has been shown to be independently associated with higher rates of atherosclerosis and cardiovascular events,3-9 through several hypothesized mechanisms.8-10 Patients with cirrhosis have thrombocytopenia, decreased coagulation factor synthesis, and increased fibrinolysis, placing them at an inherent increased risk of bleeding, which in the setting of acute myocardial infarction (AMI) is well known to have a strong negative prognostic impact.11-13 These patients were not included in a risk-prediction model for bleeding in acute coronary syndromes, as those with bleeding diathesis were excluded from the study.14 The inherent bleeding risk in this population may dissuade health-care providers from considering coronary angiography (CAG) and subsequent revascularization in patients with cirrhosis presenting with AMI, thereby indirectly increasing mortality.15 The rate of invasive strategy utility and subsequent revascularization in patients with cirrhosis presenting with AMI is not known. Furthermore, it is uncertain whether the benefits of an invasive strategy are attenuated in cirrhosis patients due to increased risk of bleeding in comparison with the general population. In this context, we aimed to investigate the contemporary practice patterns and outcomes of AMI patients with cirrhosis vs those without cirrhosis.
This was a retrospective cohort of patients admitted to acute-care hospitals in the United States from 2010 to 2014. We reviewed the National Inpatient Sample (NIS) database from the Healthcare Cost and Utilization Project, developed by the Agency for Healthcare Research and Quality.16,17 This contains deidentified data on over 7 million hospital admissions, with a probability sample stratified by ownership, size, teaching status, urban/rural location, and geographic region. A representative from our institution’s Office of Research Integrity deemed that formal institutional board review was unnecessary, as the data are publicly available without any of the Health Insurance Portability and Accountability Act (HIPAA) patient identifiers. Discharge diagnoses and procedures were collected on each patient, using the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM).
We identified patients older than 18 years who were hospitalized with a primary diagnosis of AMI (ICD-9-CM code: 410.xx). From this, we identified patients with a diagnosis of cirrhosis (ICD-9-CM codes: 571.2, 571.5, and 571.6). Comorbidities and clinical outcomes were compared with those without cirrhosis using Pearson’s Chi-square test or the Mann-Whitney U-test for dichotomous and continuous variables, respectively. We then matched patients with cirrhosis by propensity score to those without cirrhosis (matched cohort), based on variables that were significant on univariate analysis with a match tolerance of 0.01. We excluded patients with prior liver transplantation (ICD-9-CM code: v42.7). Hemorrhage complications were defined by ICD-9-CM codes and are shown in Supplemental Table S1. We included diagnoses for gastrointestinal bleeding, and hemorrhage or hematoma complicating a procedure, henceforth referred to as “hemorrhage” unless otherwise specified. Any revascularization procedure included percutaneous coronary intervention (PCI) by balloon angioplasty or stent placement, systemic thrombolysis administration, or coronary artery bypass graft (CABG). We did not report those who received thrombolytics, as the number of patients was too small to report per Healthcare Cost and Utilization Project (HCUP) guidelines; however, they were included in the aggregate of any revascularization procedure. Comorbidities were described using the Elixhauser Cormorbidity Index in the Nationwide Inpatient Sample (NIS) and ICD-9-CM code. The burden of comorbidities in our sample was classified by the Charlson Comorbidity Index, which has been previously validated in cardiovascular studies.18,19 Cirrhosis patients with missing data were excluded from the multivariable analysis (n = 43; 1.4%). The primary outcome of interest was all-cause in-hospital mortality. Secondary outcomes were incidences of hemorrhage and revascularization by PCI or CABG.
We evaluated the association between cirrhosis and in-hospital mortality using a multivariable logistic regression model, adjusting for variables that were significant on univariate analysis of the propensity-score matched groups. To compare the impact of CAG, PCI, and hemorrhage on mortality between patients with and without cirrhosis, we ran the same logistic-regression model for the cohort without cirrhosis and then tested for interaction between the effect of each measure and in-hospital mortality. For this, we compared the natural logarithm of the estimated odds ratio (OR) for mortality in cirrhosis patients with non-cirrhosis patients by using the Student’s t-test, regarding these quantities as independent and approximately normally distributed on a logarithmic scale. Associations with a P-value <.05 were considered significant. Statistical analysis was performed with SPSS version 25.0 (IBM SPSS Statistics for Windows; IBM Corporation).
We identified 612,547 patients (admissions) with a primary diagnosis of AMI; of these, a total of 3135 had a concomitant diagnosis of cirrhosis and were matched by propensity score to patients without cirrhosis. The study flow diagram is shown in Figure 1. Table 1 exhibits baseline characteristics and comorbidities. Patients with cirrhosis were more likely to have a history of obesity, anemia, chronic lung and kidney disease, alcohol and drug abuse, hepatitis, and diabetes; they were less likely to have a history of dyslipidemia, coronary artery disease (CAD), congestive heart failure, and prior PCI. They were less likely to present with ST-segment elevation myocardial infarction (STEMI). Unmatched data are provided in Supplemental Table S2 with similar observations.
In-hospital outcomes are shown in Table 2. Patients with cirrhosis were less likely to undergo CAG (51.4% vs 63.9% in the non-cirrhosis group; P<.001) or revascularization procedure (34.2% vs 50.8% in the non-cirrhosis group; P<.001) by either PCI (28.7% vs 39.2% in the non-cirrhosis group; P<.001) or CABG (6.0% vs 12.9% in the non-cirrhosis group; P<.001). This lower utilization of an invasive approach was seen in both STEMI (55.2% vs 65.7% in the non-cirrhosis group; P<.001) and non-STEMI (NSTEMI) cohorts (22.5% vs 29.5% in the non-cirrhosis group; P<.001). Patients with cirrhosis had a higher mortality rate in the overall cohort (8.7% vs 6.9% in the non-cirrhosis group; P<.01), but this difference was not significant in the subgroup undergoing revascularization (5.7% vs 3.5% in the non-cirrhosis group; P=.32). Among those who underwent stent placement, the rate of drug-eluting stent (DES) use was lower in patients with cirrhosis (13.9% vs 25.1% in the non-cirrhosis group; P<.001). Mortality rate was not significantly different between groups in those undergoing DES placement (3.7% in the cirrhosis group vs 2.5% in the non-cirrhosis group; P=.14); however, the rate of hemorrhage was significantly higher in the cirrhosis group (7.8% vs 6.2% in the non-cirrhosis group; P<.001). While bare-metal stent (BMS) implantation was associated with a higher rate of mortality and hemorrhage overall, there was no significant difference between cirrhosis and non-cirrhosis groups (mortality, 5.6% in the cirrhosis group vs 8.1% in the non-cirrhosis group [P=.07]; hemorrhage, 11.3% in the cirrhosis group vs 9.7% in the non-cirrhosis group [P=.16]).
Patients with cirrhosis had a higher rate of hemorrhage (12.3% vs 7.1% in the non-cirrhosis group; P<.001) and in the subgroups undergoing CAG (10.1% vs 6.5% in the non-cirrhosis group; P<.001), PCI (10.2% vs 7.1% in the non-cirrhosis group; P=.01), or CABG (14.8% vs 8.5% in the non-cirrhosis group; P=.02). The increased hemorrhage risk was associated with increased blood transfusion requirement (16.0% in the cirrhosis group vs 13.9% in the non-cirrhosis group; P=.02). Among patients who had hemorrhagic events, patients with cirrhosis incurred higher mortality (12.0% vs 11.9% in the non-cirrhosis group; P=.97). Compared with non-cirrhotics, there were higher rates of hemorrhage seen with CAG and PCI in those with cirrhosis (Figure 2). Patients with cirrhosis were also more likely to have an acute kidney injury, but were less likely to have in-hospital complications such as cardiac arrest, shock, stroke, pneumonia, and endotracheal intubation. There was no difference in mortality between alcoholic and non-alcoholic cirrhosis (9.7% vs 8.4%, respectively; P=.25) and neither were independent predictors of mortality (OR, 3.49; 95% CI, 0.68-17.86 [P=.13] and OR, 4.21; 95% CI, 0.85-20.84 [P=.08], respectively).
On multivariate analysis of patients with cirrhosis, we adjusted for gender; race; insurance; baseline comorbidities, including diabetes, dyslipidemia, chronic kidney and lung disease, obesity, anemia, alcohol and drug abuse, depression, coagulopathy, hepatitis, congestive heart failure, prior CAD, prior PCI, and alcoholic vs non-alcoholic cirrhosis; hospital characteristics, including teaching status, size, municipality, and transfer status; and outcome variables, including type of MI (STEMI vs NSTEMI), cardiac arrest, shock, stroke, pneumonia, intubation, acute kidney injury, hemorrhage, blood transfusion, intra-aortic balloon pump placement, and type of revascularization procedure (CAG, PCI, CABG). Independent predictors for higher mortality in patients with cirrhosis included STEMI, chronic kidney disease, coagulopathy, cardiac arrest, shock, acute kidney injury, and intubation with mechanical ventilation (significant and select values are shown in Figure 3).
Other findings include PCI being independently associated with lower mortality (OR, 0.57; 95% CI, 0.33-0.98; P=.04), although this was more impactful in the non-cirrhosis cohort (OR, 0.27; 95% CI, 0.16-0.46; P<.001). This difference in effect of PCI on in-hospital mortality was significant between the two groups (P-value for OR interaction = .04). However, CAG had an association with lower mortality in both the cirrhosis and non-cirrhosis cohorts (OR, 0.33; 95% CI, 0.21-0.51 [P<.001] and OR, 0.63; 95% CI, 0.41-0.98 [P=.04], respectively) with a difference in impact between cirrhosis vs non-cirrhosis only approaching significance (P-value for OR interaction = .07).
In this retrospective study evaluating the outcomes of patients with cirrhosis who were admitted for AMI in comparison with a propensity-score matched cohort without cirrhosis, we noted a lower utilization of an invasive approach and subsequent percutaneous or surgical revascularization, a higher rate of combined gastrointestinal and postprocedural hemorrhage, and a higher mortality rate for patients with cirrhosis. However, revascularization with PCI was independently associated with a lower mortality rate in select patients with cirrhosis.
One of the key findings of our analysis is that patients with cirrhosis had a substantially lower rate of any invasive approach, including subsequent percutaneous or surgical revascularization, regardless of having a STEMI or NSTEMI diagnosis. This lower adoption of an invasive strategy in cirrhosis patients with MI (51.4%), even when presenting with STEMI (55.2%), denotes a significant difference in contemporary practice patterns in such patients. This may be due to a higher comorbidity burden (such as active bleeding, coagulopathy or advanced kidney disease) precluding an invasive strategy, but may also be related to a risk-aversion tendency among health-care providers due to concern for subsequent bleeding. Interestingly, cirrhosis patients were less likely to present with STEMI. Although this poses the question of a separate pathophysiology, there may be a proportional shift to higher NSTEMI diagnoses due to a “type II” mechanism.
As noted in previous studies, the presence of classical risk factors and higher prevalence of diabetes3 may be related to the increasing prevalence of NAFLD, which shares many of the same risk factors as cardiovascular disease.4,6,20 We also found a higher prevalence of chronic kidney disease in cirrhosis patients, which may be due to intrinsic kidney disease, diabetic nephropathy, diuretic use for edema (secondary to low oncotic pressure), or hepato-renal syndrome.
A higher incidence of bleeding events in patients with cirrhosis in our study is in accord with prior reports.15,21,22 Even diagnostic CAG can be high risk; pre-liver transplant recipients receiving CAG had a higher rate of significant bleeding (14.8% vs 3.7%; P=.01).22 Our study confirmed unfavorable short-term outcomes in patients with cirrhosis presenting as AMI, which is likely related to higher comorbid burden and a higher incidence of bleeding events, but may also be due to lower utilization of an invasive approach. A study using data from the NIS comparing in-hospital mortality between patients with and without cirrhosis admitted for STEMI found that despite decreasing mortality over a decade in both cohorts, mortality remained significantly higher in patients with cirrhosis.15
We found an independent association between PCI and lower mortality rate in cirrhosis patients. Although in the general population, both primary PCI for STEMI and an early invasive strategy for high-risk NSTEMI have been shown to improve outcomes,23–25 there are limited data on the benefits of revascularization specifically in cirrhosis patients. Prior landmark randomized control trials evaluating revascularization strategies for AMI commonly excluded patients with bleeding disorders, thrombocytopenia, or clinically significant liver failure.26-28 Clinicians are often concerned that increased bleeding risk in cirrhotic patients may negate the benefits of revascularization. This lower mortality rate occurred with a significant interaction of OR between PCI and cirrhosis (vs non-cirrhosis) in terms of decreased odds for death (0.57 in the cirrhosis cohort and 0.27 in the non-cirrhosis cohort; P-value for interaction = .04). This finding may suggest that the absolute negative impact of bleeding is deleterious to the beneficial effect of PCI, but that there still remains at least a short-term benefit to percutaneous revascularization. However, residual selection bias and survivor treatment bias likely remain, inflating the survival benefit of PCI, given that the 73% risk reduction noted with PCI in the non-cirrhosis cohort is greater than the risk reduction demonstrated in prior randomized trials. Therefore, this likely reflects a substantial therapeutic benefit of an invasive strategy in a highly selected low-risk population of patients with cirrhosis.
The discrepancy in significance of hemorrhage based on stent type may be due to the presence of bleeding (or higher risk of bleeding) at the time of stent placement, leading to operator preference for BMS over DES. The mortality difference likely relates to the baseline risk of the patient, given that DES is the preferred stent during PCI and a BMS is used when reservations exist about bleeding or medication compliance, as was demonstrated by the similarity in outcomes of BMS implantation between groups. As such, the higher rate of hemorrhage in the overall cirrhosis cohort vs those undergoing CAG or PCI (Figure 2) suggests that bleeding may have prohibited a revascularization procedure, as opposed to this finding indicating less bleeding after revascularization. Unfortunately, we are not able to temporally determine whether bleeding or revascularization came first during the admission using the NIS database. However, the higher rate of hemorrhage may explain the attenuated relative-risk reduction in in-hospital mortality seen with CAG and PCI in patients with cirrhosis vs those without.
Study limitations. There are several limitations of our study inherent to the nature of utilizing a large administrative database. Mainly, the NIS database consists of time-limited administrative data that are related to a specific hospitalization. The possibility of incomplete diagnoses, misclassified diagnoses or procedures, or omitted documentation exists. Specifically, important patient-level information could not be retrieved from this database, which includes specific values of vitals, laboratory data, anticoagulation and antiplatelet status, the severity of cirrhosis, use of bleeding avoidance strategies relating to cardiac catheterization (such as radial access, small size catheters, and closure devices), type of NSTEMI (type I vs type II), long-term postdischarge outcomes, rationale for revascularization decisions, and ejection fraction. This also includes the exact etiology of non-alcoholic cirrhosis, given the narrow breadth of ICD codes for such diseases. The inherent bleeding risk in cirrhosis patients while taking dual-antiplatelet therapy may portend a negative outcome over time, and thus requires long-term follow-up data.
Patients with cirrhosis who are admitted to the hospital for AMI have higher rates of mortality and bleeding than those without cirrhosis. They are less likely to undergo invasive treatment and subsequent percutaneous or surgical revascularization, even when presenting with STEMI. PCI was an independent predictor of decreased mortality in this population, although to less effect than in non-cirrhotics. Given the unbalanced distribution of revascularization procedures in this population, further prospective data in this population are needed.
1. Murray CJL, Atkinson C, Bhalla K, et al. The state of US health, 1990-2010: burden of diseases, injuries, and risk factors. JAMA. 2013;310:591-608.
2. Younossi Z, Stepanova M, Afendy M, et al. Changes in the prevalence of the most common causes of chronic liver diseases in the United States from 1988 to 2008. Clin Gastroenterol Hepatol. 2011;9:524-530.
3. Targher G, Bertolini L, Padovani R, et al. Prevalence of nonalcoholic fatty liver disease and its association with cardiovascular disease among type 2 diabetic patients. Diabetes Care. 2007;30:1212-1218.
4. Hamaguchi M, Kojima T, Takeda N, et al. Nonalcoholic fatty liver disease is a novel predictor of cardiovascular disease. World J Gastroenterol. 2007;13:1579-1584.
5. Mirbagheri SA, Rashidi A, Abdi S, Saedi D, Abouzari M. Liver: an alarm for the heart? Liver Int. 2007;27:891-894.
6. Chhabra R, O’Keefe JH, Patil H, et al. Association of coronary artery calcification with hepatic steatosis in asymptomatic individuals. Mayo Clin Proc. 2013;88:1259-1265.
7. Fracanzani AL, Tiraboschi S, Pisano G, et al. Progression of carotid vascular damage and cardiovascular events in non-alcoholic fatty liver disease patients compared to the general population during 10 years of follow-up. Atherosclerosis. 2016;246:208-213.
8. Zeb I, Li D, Budoff MJ, et al. Nonalcoholic fatty liver disease and incident cardiac events: the multi-ethnic study of atherosclerosis. J Am Coll Cardiol. 2016;67:1965-1966.
9. Ekstedt M, Hagström H, Nasr P, et al. Fibrosis stage is the strongest predictor for disease-specific mortality in NAFLD after up to 33 years of follow-up. Hepatology. 2015;61:1547-1554.
10. Kim D, Choi S-Y, Park EH, et al. Nonalcoholic fatty liver disease is associated with coronary artery calcification. Hepatology. 2012;56:605-613.
11. Rao SV, Eikelboom JA, Granger CB, Harrington RA, Califf RM, Bassand JP. Bleeding and blood transfusion issues in patients with non-ST-segment elevation acute coronary syndromes. Eur Heart J. 2007;28:1193-1204.
12. Rao SV, Jollis JG, Harrington RA, et al. Relationship of blood transfusion and clinical outcomes in patients with acute coronary syndromes. JAMA. 2004;292:1555.
13. Rao SV, O’Grady K, Pieper KS, et al. Impact of bleeding severity on clinical outcomes among patients with acute coronary syndromes. Am J Cardiol. 2005;96:1200-1206.
14. Mehran R, Pocock SJ, Nikolsky E, et al. A risk score to predict bleeding in patients with acute coronary syndromes. J Am Coll Cardiol. 2010;55:2556-2566.
15. Abougergi MS, Karagozian R, Grace ND, Saltzman JR, Qamar AA. ST elevation myocardial infarction mortality among patients with liver cirrhosis: a nationwide analysis across a decade. J Clin Gastroenterol. 2015;49:778-783.
16. HCUP Databases. Healthcare Cost and Utilization Project (HCUP). Agency for Healthcare Research and Quality, Rockville MD. http://www.hcup-us.ahrq.gov/nisoverview.jsp. Published 2017. Accessed January 1, 2017.
17. Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Eff Clin Pract. 2002;5:143-151.
18. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373-383.
19. Theuns DA, Schaer BA, Soliman OI, et al. The prognosis of implantable defibrillator patients treated with cardiac resynchronization therapy: comorbidity burden as predictor of mortality. Europace. 2011;13:62-69.
20. Targher G, Bertolini L, Padovani R, et al. Prevalence of non-alcoholic fatty liver disease and its association with cardiovascular disease in patients with type 1 diabetes. J Hepatol. 2010;53:713-718.
21. Marui A, Kimura T, Tanaka S, et al. Coronary revascularization in patients with liver cirrhosis. Ann Thorac Surg. 2011;91:1393-1399.
22. Sharma M, Yong C, Majure D, et al. Safety of cardiac catheterization in patients with end-stage liver disease awaiting liver transplantation. Am J Cardiol. 2009;103:742-746.
23. Keeley EC, Boura JA, Grines CL. Primary angioplasty versus intravenous thrombolytic therapy for acute myocardial infarction: a quantitative review of 23 randomised trials. Lancet. 2003;361:13-20.
24. O’Donoghue M, Boden WE, Braunwald E, et al. Early invasive vs conservative treatment strategies in women and men with unstable angina and non-ST-segment elevation myocardial infarction: a meta-analysis. JAMA. 2008;300:71-80.
25. Cannon CP, Weintraub WS, Demopoulos LA, et al. Comparison of early invasive and conservative strategies in patients with unstable coronary syndromes treated with the glycoprotein IIb/IIIa inhibitor tirofiban. N Engl J Med. 2001;344:1879-1887.
26. Bonnefoy E, Lapostolle F, Leizorovicz A, et al. Primary angioplasty versus prehospital fibrinolysis in acute myocardial infarction: a randomised study. Lancet. 2002;360:825-829.
27. Cannon CP, Weintraub WS, Demopoulos LA, Robertson DH, Gormley GJ, Braunwald E. Invasive versus conservative strategies in unstable angina and non-Q-wave myocardial infarction following treatment with tirofiban: rationale and study design of the international TACTICS-TIMI 18 trial. Am J Cardiol. 1998;82:731-736.
28. TIMI study group. Comparison of invasive and conservative strategies after treatment with intravenous tissue plasminogen activator in acute myocardial infarction. N Engl J Med. 1989;320:618-627.
From the 1Gill Heart and Vascular Institute; and 2Department of Internal Medicine, University of Kentucky, Lexington, Kentucky.
Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Abdel-Latif is supported by the University of Kentucky COBRE Early Career Program (P20 GM103527) and NIH Grant R01 HL124266. The remaining authors report no conflicts of interest regarding the content herein.
Manuscript submitted January 13, 2019 and accepted January 25, 2019.
Address for correspondence: Dustin Hillerson, MD, University of Kentucky College of Health Sciences, 900 S. Limestone, CTW 326, Lexington KY 40536. Email: firstname.lastname@example.org