Increased Risk of Acute Myocardial Infarction and Heart Failure in Patients With Inflammatory Bowel Diseases

Abstract

Background & Aims:

There are conflicting data as to whether inflammatory bowel diseases (IBD) increase risk for cardiovascular disease. We sought to examine the risk of acute myocardial infarction (AMI) and heart failure in patients with IBD.

Methods:

We identified patients diagnosed with IBD in Olmsted County, Minnesota from 1980 through 2010 (n=736). For each patient, 2 individuals without IBD (controls, n=1472) were randomly selected, matched for age, sex, and index date of disease diagnosis. Primary outcomes were AMI and heart failure. Cox proportional hazards analysis was used to estimate the risk of AMI and heart failure.

Results:

After adjustments for traditional cardiovascular disease risk factors, IBD associated independently with increased risk of AMI (adjusted hazard ratio [aHR], 2.82; 95% CI, 1.98— 4.04) and heart failure (aHR, 2.03; 95% CI, 1.36–3.03). The relative risk of AMI was significantly increased in patients with Crohn’s disease (aHR vs controls, 2.89; 95% CI, 1.65–5.13) or ulcerative colitis (aHR vs controls, 2.70; 1.69–4.35). The relative risk of AMI was increased among users of systemic corticosteroids (aHR vs controls, 5.08; 95% CI, 3.00–8.81) and non-users (aHR vs controls, 1.79; 1.08–2.98). The relative risk of heart failure was significantly increased among patients with ulcerative colitis (aHR, 2.06; 95% CI, 1.18–3.65) but not Crohn’s disease. The relative risk of heart failure was increased among users of systemic corticosteroids (aHR, 2.51; 95% CI, 1.93–4.57) but not non-users.

Conclusion:

In a population-based cohort study, we found that despite a lower prevalence of traditional risk factors for AMI and heart failure, patients with IBD are at increased risk for these cardiovascular disorders.

INTRODUCTION

Chronic systemic inflammation is a risk factor for atherosclerosis by causing endothelial dysfunction, promoting platelet aggregation and plaque formation. Previous studies have demonstrated an increased risk of cardiovascular diseases (CVD) in patients with chronic inflammatory diseases such as rheumatoid arthritis¹ and systemic lupus erythematous.² A recent meta-analysis highlighted the evidence for subclinical atherosclerosis in patients with inflammatory bowel disease (IBD).³ There are significant subclinical changes in arterial stiffness in IBD patients compared to controls.³

There are conflicting data as to whether IBD increases the risk of acute myocardial infarction (AMI) and heart failure (HF). Recently, a meta-analysis study showed that IBD was associated with a slight increase in the risk of ischemic heart disease (IHD) (odds ratio, 1.19; 95% confidence intervals [CI], 1.08–1.31).⁴ However, heterogeneity in this analysis was observed. For example, a population-based study from the United Kingdom reported no association between IBD and an increased risk of MI after adjusting for traditional risk factors.⁵ On the other hand, population-based studies from Canada and Denmark reported a modestly increased incidence of all types of IHD in IBD patients compared to the general population.^{6, 7} However, these studies did not adjust for traditional risk factors for CVD.

Despite their relatively large sample size and population-based design, prior studies had a short follow-up period and relied on diagnostic and billing codes which have inherent limitations, and are particularly vulnerable to misclassification. Furthermore, diagnostic codes differed across the studies. For example, a group of diagnostic codes for IHD using the International Classification of Diseases 9^th Revision comprises codes for angina pectoris, subacute and chronic IHD, old MI, AMI. Important confounders such as body mass index (BMI), smoking and familial coronary disease could not be identified from administrative data. Hence, the magnitude of the potential increase in the risk of AMI/HF on IBD remains unresolved. The aim of the present study was to address this gap in knowledge by examining the risk of AMI/HF in IBD patients compared with non-IBD patients in a population-based inception cohort.

METHODS

Rochester Epidemiology Project

The Rochester Epidemiology Project is a unique medical records linkage system developed in the 1950s. The health care for the residents of Olmsted County is primarily provided by two organizations: Mayo Medical Center, and Olmsted Medical Center. In any four-year period, over 95% of county residents are examined at either one of the two health care systems.⁸ Diagnoses generated from outpatients visits, hospitalizations, autopsy examinations and death certifications are contained in a single comprehensive medical records linkage system. Demographic data, diagnosis, procedures, laboratory, and drug prescriptions are electronically recorded in indexes. Trained abstractors are then able to review all the records for these patients to verify the information. Thus, this resource ensures virtually complete case ascertainment and follow-up in a well-defined geographic region.⁹

Study population

The institutional review boards of both Mayo Clinic and Olmsted Medical Center approved the protocol. This was a retrospective longitudinal cohort study of Olmsted County. The IBD cohort was comprised of all Olmsted County residents with first diagnosis of IBD between 1980 and 2010 according to well-defined criteria as previously described.^{10, 11} The index date was based on the first IBD diagnosis. For each IBD patient, two individuals with non-IBD diagnosis (controls) were randomly selected after matching for age, sex and index date of IBD diagnosis. Patients with IBD and controls who had a history of AMI/HF before the index date were excluded. The records of IBD patients and controls were reviewed longitudinally from the index date until first AMI/HF event, death, relocation out of the county, or the end of the study (30 June 2016), whichever came first.

Ascertainment of AMI and HF events

All medical records of individual patients in the IBD and control cohorts were reviewed in detail to ascertain AMI/HF events using standardized criteria. Patients who were diagnosed with AMI/HF from outside Olmsted County were included only if their medical documents were available for review. According to the third universal definition of MI¹², a diagnosis of AMI was defined as meeting any one of the following criteria: 1) detection of a rise and/or fall of cardiac biomarker and with at least one of the following—a) symptoms of ischemia, b) new or presumed new significant ST-segment-T wave changes or new left bundle branch block (LBBB), c) development of pathological Q waves on the electrocardiogram (ECG), d) imaging evidence of new loss of viable myocardium or new regional wall motion abnormality, or e) identification of intracoronary thrombus at cardiac catheterization or autopsy; or 2) cardiac death with symptoms suggestive of AMI and presumed new ischemic ECG changes or new LBBB but death occured before cardiac biomarkers were obtained or before cardiac biomarker would be increased. For cardiac biomarkers, creatine kinase (CK) and CK-MB were used until 2000, and troponin was measured thereafter.

According to the Framingham Heart Study criteria¹³, a diagnosis of HF was defined as meeting either 2 major or 1 major and 2 minor criteria. Major criteria included paroxysmal nocturnal dyspnea or orthopnea, neck vein distention, pulmonary rales, cardiomegaly, acute pulmonary edema, S3 gallop, increased venous pressure >16 cm of water, circulation time ≥25 seconds, or hepatojugular reflux. Minor criteria included ankle edema, night cough, dyspnea on exertion, hepatomegaly, pleural effusion, decreased vital capacity (1 in 3 from maximum), or tachycardia (≥120 rate/min). Minor criteria could not be attributable to another condition.

Data collection

Data on Montreal classification for IBD were collected at 3 months after IBD diagnosis. Data on IBD treatment and traditional CVD risk (i.e., age, sex, hypertension, diabetes mellitus, hyperlipidemia, familial coronary disease, current smoking and BMI) were collected at the first AMI/HF event, death, relocation out of the county, or the end of the study, whichever came first. Data on IBD treatment were collected as follows: 1) ever used systemic corticosteroids; 2) ever used immunosuppressives (i.e., 6-mercaptopurine, azathioprine, methotrexate or cyclosporine); 3) ever used biologics (i.e., infliximab, adalimumab, certolizumab pegol, golimumab, natalizumab, vedolizumab, or ustekinumab); or 4) ever underwent IBD-related intraabdominal surgery. Hypertension was defined by the seventh report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure (JNC 7)¹⁴ or by having treatment with antihypertensive drugs. Diabetes mellitus was defined by the American Diabetes Association¹⁵ or by having treatment with hypoglycemic drugs. Hyperlipidemia was defined by the National Cholesterol Education Program (NCEP)¹⁶ or by having treatment with lipid-lowering drugs. Familial coronary disease was defined as subject report of a first-degree relative with coronary disease. Smoking data was retrieved from the patient provided information questionnaire which is routinely used for every episode of care for every patient in the medical systems. Current smoking was defined as the response of “yes” to a question of “Do you currently smoke?”

Statistical analysis

Incidence rates of AMI/HF for the IBD and control cohorts were obtained by the number of first AMI/HF events divided by the person-years of follow-up in each cohort. We calculated an incidence rate ratio (IRR) with 95%CI to determine if the incidence rate of AMI/HF for the IBD cohort was higher or lower than in the control cohort. The cumulative incidence of AMI/HF in the IBD cohort was estimated using the 1- Kaplan-Meier method and was compared to that of the controls using the log-rank test.

Cox proportional hazard regression was used to assess the association of IBD with the risk of AMI/HF expressed in HR with 95%CI. IBD and the traditional CVD risk factors were tested by univariate analysis. Age, sex, and any variables with a p <0.1 in the univariate analysis were then included in the multivariate analysis. Separate analyses for IBD subtypes (CD and UC), sex (male and female), age at IBD diagnosis (<40 years and ≥40 years), disease location (for CD, ileal/ileocolon and colon and for UC, proctitis/left-sided colitis and extensive colitis) and disease behavior (for CD, non-stricturing/non-fistulizing and stricture/fistula) were performed. Subgroup analyses for use of systemic corticosteroids (ever or never), immunosuppressives (ever or never), biologics (ever or never), and IBD-related intraabdominal surgery (ever or never) were performed. We defined requiring systemic corticosteroids and IBD-related intraabdominal surgery as markers of disease severity. An alpha-level of 0.05 was considered as statistically significant. For assessment of the association of IBD with the risk of AMI/HF, at least a total of 100 AMI/HF events were required to detect hazard ratios (HR) of ≥1.8, at least 75 events were required to detect HR of ≥2 and at least 50 events were required to detect HR of ≥2.3, with a power of 80% at a two-sided significance level of 0.05. Statistical analyses were performed by using JMP 10 statistical software package (SAS Institute Inc. Cary, NC, USA).

RESULTS

Demographic characteristics

A total of study population comprised 736 IBD patients (339 CD, 397 UC) and 1,472 controls who were followed for a total of 11,398 person-years and 17,880 person-years, respectively. Table 1 shows demographic characteristics of the IBD and control cohorts. Supplement Table 1 shows Montreal classification and IBD treatment.

Table 1:

Demographic characteristics of the IBD and non-IBD cohorts, Olmsted County, MN, 1980–2010.

Characteristics	All IBD			Crohn’s disease			Ulcerative colitis
	Cases (n=736)	Controls (n=1472)	p- value	Cases (n=339)	Controls (n=678)	p- value	Cases (n=397)	Controls (n=794)	p- value
Median age at index date, years (IQR)	34.1 (24.0 – 47.8)	34.1 (24.0 – 47.8)	1.00	31.2 (21.8 – 47.4)	31.2 (21.8 – 47.4)	1.00	35.6 (25.0 – 48.1)	35.6 (25.0 – 48.1)	1.00
Male, n (%)	405 (55%)	810 (55%)	1.00	177 (52%)	354 (52%)	1.00	228 (57%)	456 (57%)	1.00
Median duration of follow-up, years (IQR)	14.3 (8.3 – 22.3)	10.0 (4.4 – 18.6)	<0.01	13.8 (8.3 – 21.0)	10.1 (4.4–17.8)	<0.01	14.4 (8.3 – 22.7)	9.9 (4.6– 19.3)	<0.01
Familial coronary disease, n (%)	165 (22.5%)	366/1321 (27.7%)	0.01	76 (22.4%)	162/609 (26.6%)	0.15	89 (22.5%)	204/712 (28.7%)	0.03
Diabetes mellitus, n (%)	52 (7.1%)	136 (9.2%)	0.08	18 (5.3%)	60 (8.8%)	0.05	34 (8.6%)	76 (9.6%)	0.57
Hypertension, n (%)	201 (27.3%)	399 (27.1%)	0.92	82 (24.2%)	171 (25.2%)	0.72	119 (30.0%)	228 (28.7%)	0.65
Hyperlipidemia, n (%)	163 (22.2%)	427 (29%)	<0.01	72 (21.2%)	185 (27.3%)	0.04	91 (22.9%)	242 (30.5%)	<0.01
Current smoking, n (%)	92 (12.5%)	228/1327 (17.2%)	<0.01	55 (16.2%)	96/610 (15.7%)	0.84	37 (9.3%)	132/717 (18.4%)	<0.01
Mean BMI, kg/m2, ± S.D.	28.2 ± 6.5	27.7 ± 6.3	0.14	28.0 ± 6.6	27.9 ± 6.6	0.78	28.3 ± 6.4	27.5 ± 6.0	0.07

IBD, inflammatory bowel disease; CD, Crohn’s disease; UC, ulcerative colitis; IQR, interquartile range; BMI, body mass index; SD, standard deviation

Risk of AMI

A total of 134 first AMI events occurred during the study period. The first AMI event developed in 75 IBD patients and 59 controls, which corresponded to IRR of AMI was 1.99 (95%CI, 1.42–2.80). The cumulative probability of developing first AMI from time of IBD diagnosis was 3.2% at 5 years (95%CI, 2.6%−4.0%), 6.4% at 10 years (5.1%−7.9%), and 12.3% at 20 years (10.0%−15.2%). The cumulative incidence of AMI was significantly greater among IBD patients than controls (p<0.001) (Figure 1). Table 2 shows the univariate and multivariate analyses for the association of potential factors and the development of AMI. In multivariate analysis, IBD was significantly associated with developing AMI (adjusted hazard ratio [aHR], 2.82; 95%CI, 1.98–4.04).

Figure 1: Cumulative probability of developing first acute myocardial infraction from time of IBD diagnosis between IBD cohort and control cohort.

Table 2:

Univariate and multivariate analysis for risk of AMI and HF in IBD patients, Olmsted County, MN, 1980–2010.

Acute Myocardial Infarction	Univariate Analysis		Multivariate Analysis†
	Unadjusted HR (95% CI)	p-value	Adjusted HR (95% CI)	p-value
Inflammatory bowel disease	1.98 (1.41 – 2.79)	<0.01	2.82 (1.98 – 4.04)	<0.01
Male sex	2.35 (1.61 – 3.50)	<0.01	2.47 (1.68 – 3.71)	<0.01
Age (per 1-year increase)	1.08 (1.07 – 1.09)	<0.01	1.08 (1.07 – 1.09)	<0.01
Familial coronary disease	2.06 (1.46 – 2.91)	<0.01	1.31 (0.92 – 1.86)	0.13
Diabetes Mellitus	2.79 (1.86 – 4.09)	<0.01	1.16 (0.75 – 1.75)	0.50
Hypertension	5.01 (3.47 – 7.39)	<0.01	1.39 (0.90 – 2.18)	0.14
Hyperlipidemia	3.09 (2.18 – 4.41)	<0.01	1.72 (1.18 – 2.55)	<0.01
Current smoking	1.28 (0.79 – 1.98)	0.30	-	-
BMI (per 1-m2/kg increase)	1.00 (0.97 – 1.03)	0.84	-	-

Heart Failure	Univariate Analysis		Multivariate Analysis†
	Unadjusted HR (95% CI)	p-value	Adjusted HR (95% CI)	p-value
Inflammatory bowel disease	1.55 (1.05 – 2.27)	0.03	2.03 (1.36 – 3.03)	<0.01
Male sex	1.05 (0.71 – 1.55)	0.82	1.11 (0.74 – 1.67)	0.60
Age (per 1-year increase)	1.11 (1.09 – 1.12)	<0.01	1.11 (1.09 – 1.13)	<0.01
Familial coronary disease	2.17 (1.46 – 3.21)	<0.01	1.32 (0.89 – 1.97)	0.17
Diabetes Mellitus	3.03 (1.93 – 4.62)	<0.01	1.24 (0.77 – 1.96)	0.37
Hypertension	5.40 (3.54 – 8.47)	<0.01	1.30(0.79 – 2.19)	0.31
Dyslipidemia	2.20 (1.50 – 3.25)	<0.01	1.06 (0.70 – 1.61)	0.79
Current smoking	0.79 (0.41 – 1.38)	0.43	-	-
BMI (per 1-m2/kg increase)	0.48 (0.08 – 2.46)	0.39	-	-

AMI, acute myocardial infarction; HF, heart failure; IBD, inflammatory bowel disease; HR, hazard ratio; CI, confidence intervals; BMI, body mass index

^†Age, sex, and any variables with a p < 0.1 in the univariate analysis were included in the multivariate analysis.

Stratification by IBD subtype, fifty-five first AMI events occurred in the CD cohort and their controls and seventy-nine first AMI events occurred in the UC cohort and their controls. The incidence and relative risk of developing AMI was significantly elevated for both CD and UC (Table 3). According to the Montreal classification, an increased risk of AMI in CD patients was observed for ileal/ileocolonic localization and for non-stricturing and non-fistulizing behavior. An increased risk of AMI in UC patients was observed for both proctitis/left-sided colitis and extensive colitis. Regardless of age at IBD diagnosis and sex, an increased risk of AMI in IBD persisted significantly (Table 4).

Table 3:

Incidence and risk of AMI and HF in IBD patients compared to their matched controls, Olmsted County, MN, 1980–2010.

Acute Myocardial Infarction	Incidence Rate (per 1,000 person-years, 95% CI)		Incidence Rate Ratio (95% CI)	Hazard Ratio (95% CI)
	Cases	Controls		Unadjusted	Adjusted†
Inflammatory Bowel Disease	6.58 (5.21 – 8.20)	3.30 (2.54 – 4.23)	1.99 (1.42 – 2.80)	1.98 (1.41 – 2.79)	2.82 (1.98 – 4.04)
Crohn’s Disease	5.78 (3.98 – 8.15)	3.10 (2.05 – 4.50)	1.89 (1.15 – 3.12)	1.85 (1.09 – 3.16)	2.89 (1.65 – 5.13)
Ulcerative Colitis	7.25 (5.35 – 9.60)	3.47 (2.44 – 4.79)	2.09 (1.34 – 3.26)	2.09 (1.34 – 3.28)	2.70 (1.69 – 4.35)

Heart Failure	Incidence Rate (per 1,000 person-years, 95% CI)		Incidence Rate Ratio (95% CI)	Hazard Ratio (95% CI)
	Cases	Controls		Unadjusted	Adjusted†
Inflammatory Bowel Disease	4.55 (3.44 – 5.90)	2.91 (2.20 – 3.78)	1.56 (1.07 – 2.29)	1.55 (1.05 – 2.27)	2.03 (1.36 – 3.03)
Crohn’s Disease	4.78 (3.17 – 6.94)	3.22 (2.16 – 4.65)	1.48 (0.86 – 2.57)	1.48 (0.85 – 2.57)	1.73 (0.98 – 3.07)
Ulcerative Colitis	4.36 (2.96 – 6.21)	2.65 (1.77 – 3.82)	1.65 (0.97 – 2.81)	1.60 (0.94 – 2.74)	2.06 (1.18 – 3.65)

AMI, acute myocardial infarction; HF, heart failure; IBD, inflammatory bowel disease; CD, Crohn’s disease; UC, ulcerative colitis; CI, confidence intervals

^†Age, sex, and any variables with a p < 0.1 in the univariate analysis were included in the multivariate analysis.

Table 4:

Subgroup analyses of incidence and risk of AMI and HF in IBD patients compared to their matched controls, Olmsted County, MN, 1980–2010.

Acute Myocardial Infarction	Incidence rate (per 1,000 person-years, 95% CI)		Incidence rate ratio (95% CI)	Hazard ratio (95% CI)
	IBD Cases	Controls		Unadjusted	Adjusted†
Sex
• Male	8.06 (6.05 – 10.54)	5.05 (3.78 – 6.62)	1.60 (1.08 – 2.37)	1.59 (1.07 – 2.36)	2.46 (1.62 – 3.73)
• Female	4.81 (3.19 – 6.99)	1.22 (0.63 – 2.17)	3.94 (1.89 – 8.20)	3.92 (1.94 – 8.56)	4.77 (2.34 – 10.49)
Age at IBD diagnosis
• <40 years	1.70 (0.95 – 2.83)	0.55 (0.23 – 1.14)	3.08 (1.17 – 8.09)	3.36 (1.36 – 9.44)	4.64 (1.74 – 14.00)
• ≥40 years	16.50 (12.77 – 21.01)	7.54 (5.71 – 9.78)	2.19 (1.52 – 3.16)	2.20 (1.52 – 3.18)	2.43 (1.67 – 3.55)
Systemic corticosteroids
• Ever used	7.64 (5.62 – 10.15)	2.52 (1.64 – 3.72)	3.03 (1.83 – 5.01)	3.03 (1.85 – 5.11)	5.08 (3.00 – 8.81)
• Never used	5.50 (3.81 – 7.70)	4.10 (2.92 – 5.62)	1.34 (0.83 – 2.16)	1.31 (0.81 – 2.12)	1.79 (1.08 – 2.98)
Immunosuppressives
• Ever used	4.32 (2.61 – 6.76)	0.82 (0.31 – 1.80)	5.26 (1.94 – 14.25)	5.20 (2.06 – 5.84)	8.35 (3.11 – 26.54)
• Never used	7.77 (5.96 – 9.97)	4.58 (3.48 – 5.93)	1.70 (1.17 – 2.46)	1.69 (1.16 – 2.45)	2.37 (1.61 – 3.49)
Biologics
• Ever used	4.11 (2.11 – 7.29)	1.09 (0.36 – 2.59)	3.78 (1.18 – 12.04)	3.79 (1.27 −13.83)	2.78 (0.88 – 10.47)
• Never used	7.25 (5.65 – 9.18)	3.87 (2.95 – 5.00)	1.87 (1.31 – 2.68)	1.86 (1.30 – 2.67)	2.67 (1.84 – 3.89)
Intraabdominal surgery
• Ever surgery	6.29 (4.17 – 9.14)	2.46 (1.44 – 3.96)	2.56 (1.35 – 4.85)	2.55 (1.36 −4.96)	4.10 (2.10 – 8.31)
• Never surgery	6.74 (5.06 – 8.80)	3.73 (2.75 – 4.96)	1.80 (1.20 – 2.70)	1.78 (1.19 – 2.68)	2.57 (1.68 – 3.96)
CD location
• Ileal or ileocolon	5.89 (3.71 – 8.91)	2.27 (1.24 – 3.85)	2.59 (1.27 – 5.30)	2.53 (1.25 – 5.33)	3.28 (1.54 – 7.26)
• Colon	5.59 (2.87 – 9.92)	4.66 (2.61 – 7.74)	1.20 (0.53 – 2.74)	1.19 (0.51 – 2.70)	2.00 (0.81 – 4.90)
CD behavior
• Inflammatory	6.08 (4.03 – 8.83)	3.79 (2.49 – 5.55)	1.60 (0.92 – 2.81)	1.58 (0.90 – 2.78)	2.87 (1.57 – 5.29)
• Stricture or fistula	4.66 (1.77 – 10.21)	0.57 (0.05 – 2.67)	8.14 (0.95 – 69.64)	7.81 (1.26–149.7)	12.61(0.94–442.9)
UC location
• Proctitis or left- sided colitis	5.82 (3.71 – 8.73)	3.34 (2.08 – 5.11)	1.74 (0.94 – 3.24)	1.73 (0.93 – 3.24)	2.34 (1.24 – 4.46)
• Extensive colitis	9.22 (6.06 – 13.49)	3.64 (2.13 – 5.86)	2.53 (1.33 – 4.82)	2.62 (1.38 – 5.13)	3.05 (1.58 – 6.09)

Heart Failure	Incidence rate (per 1,000 person-years, 95% CI)		Incidence rate ratio (95% CI)	Hazard ratio (95% CI)
	IBD Cases	Controls		Unadjusted	Adjusted†
Sex
• Male	4.34 (2.94 – 6.18)	3.19 (2.21 – 4.47)	1.36 (0.82 – 2.26)	1.33 (0.79 – 2.22)	-
• Female	4.81 (3.19 – 6.99)	2.57 (1.64 – 3.86)	1.87 (1.05 – 3.34)	1.87 (1.05 – 3.37)	2.17 (1.17 – 4.05)
Age at IBD diagnosis
• <40 years	0.52 (0.17 – 1.23)	0.37 (0.12 – 0.88)	0.80 (0.23 – 2.74)	1.32 (0.31– 5.58)	-
• ≥40 years	12.46 (9.33 – 16.33)	6.82 (5.09 – 8.96)	1.83 (1.23 – 2.72)	1.80 (1.21 – 2.69)	1.92 (1.28 – 2.86)
Systemic corticosteroids
• Ever used	4.71 (3.20 – 6.71)	2.64 (1.73 – 3.86)	1.79 (1.04 – 3.08)	1.77 (1.03 – 3.08)	2.51 (1.93 – 4.57)
• Never used	4.38 (2.90 – 6.36)	3.19 (2.17 – 4.55)	1.37 (0.80 – 2.35)	1.36 (0.79 – 2.34)	-
Immunosuppressives
• Ever used	2.24 (1.11 – 4.10)	0.82 (0.31 – 1.80)	2.73 (0.91 – 8.13)	2.69 (0.93 – 8.76)	2.76 (0.88 – 9.63)
• Never used	5.76 (4.24 – 7.65)	3.98 (2.96 – 5.24)	1.45 (0.96 – 2.18)	1.44 (0.95 – 2.18)	1.96 (1.28 – 3.01)
Biologies
• Ever used	1.61 (0.54 – 3.82)	0.54 (0.11 – 1.75)	2.95 (0.54 – 16.11)	3.00 (0.58 – 21.64)	-
• Never used	5.35 (4.00 – 7.01)	3.52 (2.64 – 4.60)	1.52 (1.02 – 2.25)	1.51 (1.02 – 2.24)	1.98 (1.31 – 2.99)
Intraabdominal surgery
• Ever surgery	4.42 (2.72 – 6.84)	2.48 (1.45 – 3.98)	1.79 (0.90 – 3.54) 1.47	1.74 (0.88 – 3.50) 1.49	1.96 (0.94 – 4.20) 2.06
• Never surgery	4.61 (3.27 – 6.34)	3.13 (2.24 – 4.26)	1.47 (0.93 – 4.26)	1.49 (0.93 – 2.37)	2.06 (1.26 – 3.35)
CD location
• Ileal or ileocolon	4.35 (2.54 – 6.99)	3.02 (1.80 – 4.79)	1.44 (0.71 – 2.91)	1.41 (0.69 – 2.86)	-
• Colon	5.61 (2.88 – 9.96)	3.60 (1.85 – 6.39)	1.56 (0.65 – 3.74)	1.59 (0.65 – 3.88)	-
CD behavior
• Inflammatory	4.83 (3.04 – 7.31)	3.47 (2.24 – 5.16)	1.39 (0.76 – 2.55)	1.39 (0.75 – 2.55)	-
• Stricture or fistula	4.61 (1.75 – 10.10)	2.31 (0.77 – 5.50)	1.99 (0.53 – 7.41)	1.99 (0.53 – 8.05)	-
UC location
• Proctitis or left- sided colitis	2.95 (1.56 – 5.10)	2.98 (1.80 – 4.66)	0.99 (0.46 – 2.11)	0.97 (0.44 – 2.05)	-
• Extensive colitis	6.32 (3.83 – 9.90)	2.19 (1.08 – 3.99)	2.89 (1.29 – 6.49)	2.89 (1.32 – 6.81)	3.37 (1.32 – 9.80)

AMI, acute myocardial infarction; HF, heart failure; IBD, inflammatory bowel disease; CI, confidence intervals

^†Age, sex, and any variables with a p < 0.1 in the univariate analysis were included in the multivariate analysis.

IBD Treatment and Risk of AMI

An increased risk of AMI in systemic corticosteroid users was noted (aHR, 5.08; 95%CI, 3.00–8.81), while the magnitude of the risk in non-users was lower (aHR, 1.79; 1.08–2.98). Both patients who had undergone IBD-related surgery and those never undergoing IBD-related surgery had a significantly elevated risk of AMI. The relative risk of AMI for immunosuppressives and biologics are demonstrated in Table 4.

Risk of HF

A total of 105 first HF events (caused by ischemic cardiomyopathy (60%), diastolic dysfunction (23%) and dilated cardiomyopathy (17%)) occurred during the study period. Of those, 52 patients developed both first AMI and HF events. The first HF event developed in 53 IBD patients and 52 controls which corresponded to IRR of HF was 1.56 (95%CI, 1.07– 2.29)(Table 3). The cumulative probability of developing first HF from time of IBD diagnosis was 2.2% at 5 years (95% CI, 1.7%−2.9%), 4.4% at 10 years (3.3%−5.7%), and 8.5% at 20 years (6.6%−11.0%). The cumulative incidence of HF was significantly greater among IBD patients than controls (p=0.02) (Figure 2). In multivariate analysis, IBD was significantly associated with time to developing HF (aHR, 2.03; 95%CI, 1.36–3.03) (Table 2).

Figure 2: Cumulative probability of developing first heart failure from time of IBD diagnosis between IBD cohort and control cohort.

When stratified by IBD subtype, fifty-one HF events occurred in the CD cohort and their respective controls and fifty-four HF events occurred in the UC cohort and controls. Although not statistically significant, both the CD and UC cohorts had a higher incidence rate of HF compared to controls. In multivariate analysis, UC patients had a significantly increased relative risk of HF, but not CD (Table 3). In subgroup analysis, an increased risk of HF was observed in UC patients with extensive colitis, females and IBD patients diagnosed at age of ≥40 years (Table 4).

IBD Treatment and Risk of HF

The risk of HF was significantly elevated among users of systemic corticosteroids (aHR, 2.51; 95%CI, 1.93–4.57), but not among non-users. An increased risk of HF was observed for patients never undergoing IBD-related surgery. The relative risk of HF for immunosuppressives and biologics are demonstrated in Table 4.

DISCUSSION

In this population-based study, IBD patients were almost three times as likely as the matched control population to develop AMI, and about twice as likely to develop HF. Among systemic corticosteroid users, the risk for AMI increased to five-fold as compared to the controls, whereas their risk was 2.5-fold elevated for HF, suggesting an association between disease activity and the developing AMI/HF. Regardless of IBD subtypes, IBD patients remained at increased risk for AMI. For HF, subgroups that appeared to be at higher risk than matched controls among UC patients.

In the general population, traditional CVD risk factors contribute to AMI/HF. Nevertheless, we demonstrated that IBD patients had a lower prevalence of these traditional CVD risk factors. Hence, the increased risk of AMI/HF in IBD patients might be due to chronic systemic inflammation-related atherosclerosis. The incidence and risk of CVD in IBD patients has been demonstrated in previous studies. In large population-based studies from the United States and from the United Kingdom, confirmed an increased risk of IHD in IBD patients, but not an increased risk of AMI.^{5, 17} A hospital-based study from Miami reported an increased risk of unstable angina or AMI (aHR, 2.69; 95%CI, 1.37–5.30) with approximately 4-years duration of follow-up, but not AMI alone.¹⁸ With regard to the risk of developing HF for IBD patients, there has only been a small amount of evidence in the literature. Recently, a study from Denmark showed that IBD patients had 37% increased risk of hospitalized HF compared with the reference population.¹⁹

In our study, we chose AMI and HF as representative of CVD because these are the hard endpoints of CVD, and their diagnostic criteria included objective evidence which could be verified in the medical record. For instance, ECG, cardiac biomarkers, cardiac imaging findings or histopathology were required for AMI diagnosis, and physical examination findings were required for HF diagnosis. With our medical record linkage system, we were able to identify AMI/HF according to standard criteria from population-based cohort with a median duration of 14 years follow-up.

We demonstrated not only a strong association between IBD and the development of AMI/HF, but also an association between higher severity of IBD and higher risk to develop AMI/HF. Systemic corticosteroid users had a five-fold elevated risk of AMI compared to their respective controls, versus a 1.8-fold higher risk of AMI in non-users. This is in accordance with Danish study, where there was a greater risk of IHD in IBD for corticosteroid users than those in non-users.⁷ In addition, we demonstrated a four-fold higher risk of AMI for IBD patients undergoing surgery as compared to their controls, compared to an approximately 2.5-fold higher relative risk of AMI for IBD patients never undergoing surgery.

Furthermore, our findings suggest an association between the extent of inflammation and the risk of development of AMI/HF. CD patients with ileal/ileocecal localization and UC patients with extensive colitis were at three-fold higher risk of AMI as compared to their respective controls. Even though an increased risk of AMI in non-stricturing/non-fistulizing CD patients was observed, there was no statistically significant association for stricturing/fistulizing CD. However, this should be interpreted with caution. There was a small number of stricturing/fistulizing CD patients at the time of IBD diagnosis. Furthermore, the later progression from inflammatory to fistulizing/stricturing behavior during the study period may have contributed to the results.

Regarding to subgroups of IBD, both IBD subtypes, sexes and all ages at IBD diagnosis were at risk for AMI. Although both IBD subtypes, sexes and all ages at IBD diagnosis had a higher absolute risk of developing HF relatively to their controls, the relative risks were significantly increased in UC patients, females and IBD patients who were 40 years or older at IBD diagnosis. Nonetheless, these results should be interpreted with caution due to the small number of HF cases in these subgroups.

It is interesting to note that most traditional CVD risk factors were significantly associated with AMI/HF in the univariate analysis; however, they were no longer statistically significant in the multivariate analyses. A combination of the effect of unbalanced samples and the influence of missing data for each of the traditional risk factors may have contributed to the results after adjusting other variables in the model. By study design, the non-IBD patients were matched for age, sex, and index date of IBD diagnosis. With a median age of 34 years at the study enrollment, less than 10% of our study population had diabetes mellitus and only one-third had other traditional CVD risk factors.

There were some limitations in this study. First, we did not evaluate the risk of cerebrovascular disease and peripheral vascular disease. Atherosclerosis is the most common pathogenesis of these diseases; however, other factors, such as cardiac emboli, can play a role in the pathogenesis. Second, we did not classify disease activity according to disease activity indexes. Because disease course of IBD is variable from disease in remission to severe disease over time, adjustment for disease activity would have been very difficult. A third limitation is the retrospective nature of the study; any data not recorded in the medical records would have been missed. Fourth, Olmsted County had less ethnic diversity than the U.S. population (85% White).⁹ The generalization of these results to the populations with more diverse ethnicity may not be possible. Lastly, the relatively small number of AMI/HF events in our subgroups may have contributed to the lack of statistical significance.

In conclusion, IBD was associated with an increased risk of AMI and HF in this population-based cohort compared to matched controls, despite a lower prevalence of traditional coronary risk factors in IBD patients. These findings support the role of chronic inflammation in IBD-associated CVDs. This confirmation of the risks of AMI/HF in IBD patients should prompt physicians to be vigilant for the development of these disorders. Further studies regarding whether controlling systemic inflammation could ultimately prevent these AMI/HF events are needed.

Abbreviations:

aHR

adjusted hazard ratio

AMI

acute myocardial infarction

BMI

body mass index

CD

Crohn’s disease

CI

confident interval

CK

creatine kinase

CVD

cardiovascular disease

ECG

electrocardiogram

HF

heart failure

HR

hazard ratio

IBD

inflammatory bowel disease

IHD

ischemic heart disease

IRR

incidence rate ratio

LBBB

left bundle branch block

UC

ulcerative colitis