Introduction

Inflammatory bowel disease (IBD), namely Crohn’s disease and ulcerative colitis, are conditions of chronic immune-mediated inflammation affecting the gastrointestinal tract. Individuals with IBD may also experience systemic inflammatory activity beyond the GI tract presenting as extra-intestinal manifestations or co-morbid immune mediated conditions. Although data are mixed, many studies indicate individuals with IBD have an increased risk of adverse cardiovascular (CV) outcomes not fully explained by traditional CV risk factors [1,2,3,4,5,6,7,8,9]. For instance, female and younger individuals with IBD may have higher risk of coronary heart disease (CHD) and stroke, in contrast to the traditional CV risk factors of male sex and older age [7, 10,11,12,13,14,15]. Endothelial dysfunction with accelerated atherosclerosis is proposed as a potential mechanism of adverse CV events in IBD [16,17,18] and is observed in other conditions of chronic inflammation, such as rheumatoid arthritis and systemic lupus erythematosus [19,20,21,22,23,24]. In addition, proinflammatory cytokines, such as C-reactive protein (CRP), tumor necrosis factor-alpha, and homocysteine are associated not only with IBD activity but also play a central role in the inflammatory cascade leading to atherosclerosis and thrombogenesis [25, 26].

Sex difference in the risk of coronary heart disease and stroke in IBD may be due to higher systemic inflammation in females [11, 27]. While endogenous estrogen is thought to be cardioprotective, decline in estrogen levels during the menopause transition is linked to decline in endothelial function [28, 29]. The menopause transition is also associated with increases in blood pressure, atherogenic lipids, visceral adiposity, and insulin resistance [28, 30,31,32]. It has been suggested that CV risk from a chronic inflammatory condition, such as IBD, may be less apparent in older populations due to the increasing prevalence of traditional CV risk factors with age. However, the risk of adverse CV outcomes among older women with IBD is unclear. Particularly in the setting of increased CV events observed with newer agents, such as Janus kinase (JAK) inhibitors, used in the treatment of IBD, identifying factors that elevate the CV risk for patients with IBD is increasingly important. Thus, this study aimed to assess the risk of incident cardiovascular disease (CVD) among postmenopausal women with IBD.

Methods

Study Population

The Women’s Health Initiative (WHI) is a long-term cohort study investigating clinical outcomes among postmenopausal women. Originally initiated in the early 1990s with the recruitment of women ages 50–79 years from 40 clinical centers nationwide, WHI has continued with extension studies that collect annual health updates and outcomes of participants. A total of 161,808 women were recruited and enrolled into the initial study to participate in either the clinical trials or the parallel observational study. The clinical trial component randomized women into four trials, which were overlapping for some participants: two hormone therapy (estrogen + progesterone vs. placebo and estrogen alone vs. placebo), dietary modification (low-fat diet vs. self-selected dietary behavior), and calcium with vitamin D supplementation (vs. placebo).

Extensive data from participants in both the clinical trials and observational study were collected by self-administered questionnaire, interview, and physical exam at the baseline visit and scheduled follow-ups. Further details of the study design have been previously described [33]. WHI data was analyzed at the WHI Clinical Coordinating Center. As secondary data analysis, this study was exempt from IRB review.

Exposure and Outcomes

All WHI participants completed a baseline questionnaire with information on medical history. In this study, IBD was defined as self-reported diagnosis by affirmative response to the question “Has a doctor told you that you have any of the following conditions? Ulcerative colitis or Crohn’s disease.” Details regarding IBD age of diagnosis and disease duration were not collected.

A major WHI clinical outcome of interest was CVD morbidity and mortality. The CVD outcomes of interest were coronary heart disease (CHD), ischemic stroke, pulmonary embolism or deep vein thrombosis (PE/DVT), or peripheral arterial disease (PAD). The outcomes of CHD, ischemic stroke, and PAD were adjudicated throughout the WHI study period. PE/DVT outcomes were by self-report [34].

Statistical Analysis

We performed a survival analysis of participants who were enrolled during the period 1993–1998 for the initial WHI cohort and were followed through Extension 1 of the study in 2010. Participants in both the WHI clinical trial and observational study were included. We excluded participants with missing data on self-reported IBD diagnosis at enrollment, missing model covariate data, no follow-up data, or a previous history of one of the CVD outcomes of interest.

We compared baseline characteristics of women with and without IBD as well as baseline characteristics between women who did and did not have a CVD outcome event, using t-test for continuous variables and chi-square test for categorical variables. We then assessed the risk of each CVD outcome between women with and without IBD using Cox proportional hazard models, stratified within the model by WHI component (clinical trial / observational study), hormone use (never, past, current; incorporating WHI hormone therapy trial component), WHI dietary modification trial arm (intervention, usual diet comparison group), and time-dependent WHI follow-up period (WHI, extension 1, extension 2). The proportional hazard assumption was assessed and confirmed by graphical methods and by fitting a model of the interaction between IBD and follow-up. Models were adjusted a priori for the following potential confounders: age; socio-demographics (ethnicity, race, education); comorbidities (treated hypertension, HTN; treated diabetes, DM; treated hypercholesterolemia, HLD; rheumatoid arthritis, RA; systemic lupus erythematosus, SLE; corticosteroids; body mass index, BMI); family history MI or stroke; and lifestyle factors (e.g. smoking status, alcohol use, physical activity, visit to a regular doctor in the past year, any health insurance). We performed a subgroup analysis to separately assess for interaction between IBD and age group (< 65 years vs. ≥ 65), RA/SLE, hormone therapy use, and smoking.

Statistical significance for all analyses was defined as p < 0.05. All statistical analyses were performed using SAS for Windows 9.4.

Results

Of the total 161,808 women enrolled in WHI, there were 134,022 who met inclusion criteria, of whom 1367 (1.0%) had self-reported IBD at baseline. As seen in Table 1, the mean age among women with and without IBD was similar (63.4 vs. 63.5 years). The majority of women in both groups were of non-Hispanic ethnicity (95.1% vs. 94.9%) and White race (87.9% vs. 85.6%). Overall hormone therapy use, defined by either self-report at enrollment or assignment into the hormone therapy trial, was not significantly different among women with and without IBD (p = 0.21). With respect to co-morbid CVD risk factors, there were no significant differences in HTN (28.4% vs. 27.8%) or DM (3.4 vs. 3.7%), whereas more women with than without IBD had HLD (17.0% vs. 12.5%), RA (8.3% vs. 4.7%), and SLE (SLE, 6.5% vs. 0.4%); p < 0.001 for all comparisons. More women with than without IBD also had a family history of MI (52.5% vs. 48.9%, p = 0.02) and stroke (38.0% vs. 35.9%, p = 0.04) and were either past or current smokers (never smokers: 43.7% vs. 51.0%, p < 0.001). Of note, less women with IBD were on regular aspirin, defined as ≥ 80mg for thirty days or more, than those without IBD (15.0% vs. 18.8%, p < 0.001) (Table 1). From baseline (1993–1998) through the end of the study period (2010), very few women reported use of IBD biologics that were just coming to market (i.e. infliximab, adalimumab, certolizumab).

Table 1 Baseline characteristics of women with and without IBD, defined as self-report of colitis at enrollment (n = 134,022)
Full size table

In the overall study population, traditional CV risk factors were more common among women who had an adverse CVD event compared to women who did not have an event (HTN—no event 26.4% vs. event 36.6–49.7%; DM—no event 3.3% vs. event 5.3–12.3%, p < 0.001 for all comparisons). In addition, RA and SLE were more common in women who experienced CHD, ischemic stroke, and PE/DVT (no event—RA 4.5%, SLE 0.4% vs. event RA 6.5–7.3%, SLE 0.7–0.8%, p < 0.001 for all comparisons). Finally, smoking and less physical activity were also more common among women who had an event compared to those who did not (never smokers—no event 51.3% vs. event 25.4–50.8%; mean metabolic equivalent hours—no event 12.8 vs. event 9.6–11.3%, p < 0.05).

The annual rate of CVD events was approximately 0.36% for CHD, 0.25% for PE/DVT, and 0.04% for PAD among all women regardless of IBD status. The annual rate of ischemic stroke, however, was 0.09% higher for women with IBD than women without IBD (Table 2). On survival analysis, women with IBD had overall lower stroke-free survival than those without IBD during the study period (Fig. 1).

Table 2 Hazard ratios of self-report IBD on outcome events
Full size table
Fig. 1
figure 1

Survival-free of ischemic stroke between women with and without IBD, Women’s Health Initiative, 1993–2010

Full size image

After adjusting for age and other potential confounders (ethnicity, race, education, HTN, DM, HLD, family history of MI or stroke, RA, SLE, corticosteroids, BMI, smoking, alcohol, physical activity, visit to doctor in the past year, any insurance), no significant difference was observed between women with IBD compared to those without IBD for risk of CHD (HR 0.96, 95% CI 0.73–1.24), PE/DVT (HR 1.11, 95% CI 0.81–1.52), or PAD (HR 0.64, 95% CI 0.28–1.42) (Table 2). After adjusting for age, the risk of ischemic stroke was significantly higher (HR 1.41, 95% CI 1.06–1.88) in women with IBD compared to those without IBD. On further adjustment for the abovementioned socio-demographic characteristics, comorbidity, family history, and lifestyle factors, the higher risk of ischemic stroke among women with IBD was attenuated and no longer statistically significant (HR 1.31, 95% CI 0.98–1.76) (Table 2).

In a subgroup analysis among women who had ischemic stroke, no significant interaction was observed between IBD and age (< 65 vs. ≥ 65 years), comorbid RA/SLE, hormone therapy use, or smoking on the risk of ischemic stroke. (Table 3).

Table 3 Hazard ratios of self-report IBD on ischemic stroke, stratified by baseline subgroups
Full size table

Discussion

In this large cohort study of postmenopausal women, those with IBD had a 41% age-adjusted increased risk of ischemic stroke compared to women without IBD. After adjusting for socio-economic factors as well as co-morbid rheumatologic disorders (RA, SLE) and traditional CV risk factors, the excess risk of ischemic stroke among women with IBD was no longer statistically significant. Importantly, this higher risk of ischemic stroke among women with IBD did not differ based on age (< 65 years vs. ≥ 65 years), comorbid RA/SLE, hormone therapy use, or smoking status. We found no significant difference in the risk of the CVD outcomes of CHD, PE/DVT, or PAD between women with IBD compared to those without IBD.

In a previous meta-analysis of observational studies that reported incident cases of stroke and ischemic heart disease among an adult IBD and non-IBD population, those with IBD had an approximately 20% increased risk of stroke (28% for women) [5]. Both ischemic and hemorrhagic stroke as well as transient ischemic attack were included in the meta-analysis, which may account for difference in estimated risk compared to the current study in which only ischemic stroke was included. The results of the meta-analysis suggested a 26% increased risk of ischemic heart disease (inclusive of acute coronary syndrome, history of MI, and/or angina) among women of all ages with IBD, while we did not find a significant difference in risk of CHD in our population. Reasons for this difference in findings are unclear but may be due to the specific focus on postmenopausal women in the current study.

The landmark WHI hormone therapy clinical trial was designed to assess the safety and efficacy of hormone therapy for the primary prevention of CHD in postmenopausal women. Data from the trial, however, demonstrated excess risk of CHD, stroke, VTE, and breast cancer among hormone therapy use participants. As a result of these findings, the use of hormone therapy among postmenopausal women has declined substantially [35] so that the prevalence of hormone therapy use in our study population is likely greater than that of the present day IBD population. However, there was no significant difference in hormone therapy use among women with and without IBD in our analysis. In addition, there was no interaction between IBD and hormone therapy use for the risk of stroke, suggesting the excess age-adjusted risk of stroke observed among IBD women was not associated with exposure to hormone therapy.

Notably, women with IBD in our study had higher rates of HLD but no difference in other traditional CVD risk factors, such as DM and HTN, compared to women without IBD. It is also interesting to note that most women regardless of IBD status did not report regular aspirin use, but even less so among women with IBD. One of the largest trials examining aspirin primary prevention in women showed a reduction in ischemic stroke among those receiving aspirin on alternate days compared to those receiving placebo [36]. However, the role of aspirin as primary prophylaxis against CVD is controversial [37] and currently is only recommended by the US Preventive Services Task Force for individuals 40–59 years with increased CVD risk factors [38]. In IBD, there has been concern for precipitating flares of disease activity with use of non-steroidal anti-inflammatory drugs (NSAIDs) [39], although recent data suggest lack of adverse effect from aspirin on IBD outcomes [40, 41]. Specific data on the role of aspirin for CVD risk factor reduction in IBD are lacking.

This study has several strengths including a well-characterized cohort of women in the WHI studies with data on socio-economic and traditional CV risk factors, a robust process of adjudication of CVD events, and a substantial follow-up period. Our study is also one of the first to specifically examine adverse CVD outcomes among IBD women after the menopause transition, a period of known increased CVD risk for women. We do acknowledge there are some limitations. The majority of women included in the study were of non-Hispanic White race and ethnicity, limiting the generalizability of study findings among an increasingly racial and ethnically diverse IBD population. IBD diagnosis was self-reported at baseline, so misclassification bias is a potential limitation. We assessed for diagnosis of IBD after baseline enrollment (1993–1998) by expanding the definition of IBD to include use of IBD medications during follow-up in 2010. However, this did not alter the results. Of note, very few women in the study reported taking an IBD biologic medication, which is unsurprising since initial WHI enrollment occurred in the pre-biologic era. Infliximab was only first approved for use in Crohn’s disease in 1998, which is when initial WHI enrollment closed. While WHI data included the rheumatologic conditions RA and SLE, data on psoriasis was not available. Given that psoriasis is also more common in IBD and a known CVD risk factor, absence of this data is a limitation. Data on IBD activity and subtype (Crohn’s disease vs. ulcerative colitis) were also not available in the data set, and the CVD outcomes reported here may differ based on level of underlying inflammation [3, 7, 13]. Measurement of CRP as a biomarker of IBD activity and a marker of atherosclerotic disease was missing for the majority of women. Future studies should examine the influence of IBD activity on adverse CVD outcomes in postmenopausal women, particularly now in the biologic era.

In summary, we found that IBD is associated with an age-adjusted excess risk of ischemic stroke among postmenopausal women that is no longer statistically significant after accounting for traditional CVD risk factors. Future studies should investigate underlying mechanisms and potential interventions for this increased risk. As newer medications used in the treatment of IBD, such as JAK inhibitors, are associated with increased CV events, understanding other factors that may elevate the CV risk for patients with IBD is increasingly important. Additionally, with an aging population, the number of postmenopausal women with IBD will almost certainly continue to rise, and findings from this study highlight the importance of pursuing modifiable risk factor reduction among women with IBD.