Background Inflammatory bowel disease (IBD) is a chronic inflammatory disorder. Systemic inflammation increases the risk of atherosclerosis and ischaemic heart disease (IHD).
Objective To examine the impact of IBD, including its duration and treatment, on the risk of IHD.
Methods In a nationwide population-based cohort of 4.6 million Danes aged ≥15 years, we compared people diagnosed with IBD during 1997–2009 (n=28 833) with IBD-free individuals. Subjects with IHD were identified in the National Patient Register. Using Poisson regression, we estimated the incidence rate ratios (IRRs) for IHD with 95% CI with adjustment for age, gender, socioeconomic status, calendar year and use of drugs for comorbidities.
Results A markedly increased risk of IHD was seen within the first year after IBD diagnosis (IRR=2.13 95% CI 1.91 to 2.38). During 1–13 years of follow-up after IBD diagnosis, the risk of IHD was 1.22 (95% CI 1.14 to 1.30). The risk of IHD was lower among patients with IBD using 5-aminosalicylic acids (IRR=1.16; 95% CI 1.06 to 1.26) than among non-users (IRR=1.36; 95% CI 1.22 to 1.51) (p=0.02), in particular among oral corticosteroid users, used as a proxy for disease severity. Likewise patients treated surgically or with thiopurines and tumour necrosis factor α antagonists tended to have reduced IRRs for IHD.
Conclusions The risk of IHD was highest in the first year after IBD diagnosis, possibly owing to ascertainment bias. The increased long-term risk of IHD in IBD may be related to chronic inflammation, and interventions reducing the inflammatory burden may attenuate this risk.
- Inflammatory Bowel Disease
- 5-Aminosalicylic Acid (5-ASA)
- Cardiovascular Disease
Statistics from Altmetric.com
Significance of this study
What is already known about this subject?
Systemic inflammation is associated with development of atherosclerosis and ischaemic heart disease (IHD).
The impact of inflammatory bowel disease on the risk of IHD continues to be debated.
What are the new findings?
Patients with inflammatory bowel disease (IBD) have a high risk of being diagnosed with IHD within the first year after IBD diagnosis, potentially reflecting ascertainment bias.
In the long term, a small, but still significantly excess risk of IHD persists.
Treatment with 5-aminosalicylic acids appears to decrease this risk.
Other interventions reducing the inflammatory burden may also decrease the risk.
How might it impact on clinical practice in the foreseeable future?
Our population-based cohort study suggests an increased long-term risk of IHD among patients with IBD which may be modified by treatment with 5-aminosalicylic acids and other interventions that reduce the inflammatory burden. This provides yet another reason for optimising control of inflammation in these patients.
The inflammatory bowel diseases (IBD) ulcerative colitis (UC) and Crohn's disease (CD) are chronic intestinal disorders characterised by local and systemic inflammation and an intermittent disease course with flares and periods of disease remission.1 ,2 The specific origin of IBD remains unknown, but it is generally accepted that development of both UC and CD is related to an uncontrolled inflammatory immune response in genetically predisposed individuals.3 Systemic inflammation increases the risk of cardiovascular disease (CVD) through an atherosclerosis-accelerating pathway.4 Few studies have investigated the risk of CVD among patients with IBD and results are conflicting, with a slight predominance of studies suggesting an increased risk of CVD among patients with IBD compared with non-IBD individuals.5–8 Focusing on the role of inflammation in development of CVD9 makes it of particular interest to study ischaemic heart disease (IHD), which is an inflammation–atherosclerosis-driven condition. Assessment of the time from diagnosis of IBD to diagnosis of IHD may also help to clarify whether the suggested increased risk of CVD in IBD is primarily a result of extensive clinical examination of patients owing to the diagnosis of IBD or actually reflects a long-term inflammation-driven association. Such time perspectives have to our knowledge not been investigated in an unselected, population-based cohort study. Likewise, no population-based study has assessed the impact of treatment with anti-inflammatory drugs, such as 5-aminosalicylates (5-ASA), thiopurines and tumour necrosis factor (TNF) α antagonist, or the impact of surgical procedures to lower the inflammatory burden, on the long-term risk of IHD in IBD.
We aimed to investigate the risk of IHD and the subentity myocardial infarction (MI) in a nationwide Danish cohort study of patients with IBD and non-IBD individuals, while taking time elapsed since IBD diagnosis and treatment of IBD into consideration.
The source population for the study cohort consisted of about 5.3 million people identified through the Danish Civil Registration System, in which all Danes are registered by a unique 10-digit identification number.10 This number is given to citizens at birth or at immigration and serves as a unique identifier, enabling linkage between the Danish national registers. Patients with IBD were identified in the Danish National Patient Register (NPR), which contains individual-level healthcare information on inpatient and outpatient hospital contacts, diagnoses and surgical and other procedures performed in Danish hospitals since 1977 and in ambulatory outpatient settings since 1995.11 We identified patients with a first diagnosis of IBD after 1997 using the international classification of disease (ICD) 8th and 10th revision codes for CD (563.01 and code group K50) and UC (563.19, 569.04 and code group K51). When a pathology register was used as ‘gold standard’, the IBD diagnoses in the NPR were found to be correct and almost complete, with validity estimates for registered CD and UC of 97% and 90%, respectively.12 Patients who were recorded with a diagnosis of both UC and CD were included only in analyses of IBD combined. Prevalent IBD cases on 1 January 1997 were not included in analyses. Individuals with ICD-8 or ICD-10 codes indicating a history of IHD (4100–4149, I20–25), cardiomyopathies (425, I42–43.8), heart failure (42709, 42710, 42711, 42719, 42799, 42899, I50–99), atrial fibrillation (42793, 42794, I48), cerebral insult (43600–99, I63–I64) or IBD before entry (1 January 1997) were also excluded. This resulted in a final study population of 4 570 820 million people.
Treatment of IBD
Detailed individual-level information on medical treatment for IBD was extracted from the Danish Prescription Database, which contains information on all prescriptions redeemed from Danish pharmacies since 1995.13 As IBD drugs, we included azathioprine (the primary thiopurine used in Denmark; ATC code, L04AX01), 5-ASA/sulfasalazine (ATC code A07EC) and oral corticosteroids (ATC code, H02AB). Treatment with TNFα antagonists (L04AB01, L04B02, L04AB04, mainly infliximab and adalimumab) was identified by combining manually collected data, NPR procedure codes from inpatient and outpatient settings and the Danish Prescription Database. Surgical procedures for IBD covered colectomies, resections and other unspecified major intestinal operations as identified by codes from the Danish version of the Nordic medico-statistical committee classification (KJFA96, KJFA97, KJFA98, KJFB, KJFF (except KJFF41), KJFH, KJFK, KJFL, KJFW and KJGB).
Information on age and gender was obtained from the Danish Civil Registration System. From Statistics Denmark we obtained information on socioeconomic status at entry, as defined by labour marked status (employment with unknown, basic or no qualifications; employment with medium-level qualifications; employment with high-level qualifications; self-employed/co-working spouse; outside labour market or pensioned). Using the Danish Prescription Database we collected information on treatment for comorbidities related to the development of IHD: antidiabetic drugs (ATC code, A10), antithrombotic and antiarrhythmic agents, antihypertensive drugs and cholesterol-lowering drugs (ATC codes B01 and C0–C10).
The primary outcome was IHD, defined in NPR by ICD-10 codes I20–125. In a sensitivity analysis, we further assessed the specific risk of the IHD-subentity acute MI (ICD-10 codes, I21–I22), which in a validation study using a manually collected Danish clinical database (DANMONICA) as reference was found to have a sensitivity of 97% and a positive predictive value of 78%.14
All subjects were followed up from 1 January 1997 or their 15th year birthday, whichever was latest, until occurrence of outcome, migration, death, or end of the study, 31 December 2009, whichever came first. Individuals with IBD contributed to person-time as non-IBD individuals until the date of IBD diagnosis.
Using Poisson regression, we calculated the incidence rate ratios (IRRs) with 95% CI for risk of IHD and MI, given IBD. The IRRs for IHD were adjusted for age, gender, socioeconomic status and calendar year (termed ‘partly adjusted’) and in a second analysis also for the above-mentioned drugs related to relevant comorbidities (termed ‘fully adjusted’). All subsequent IRRs were based on this fully adjusted model. All variables used for adjustment were treated as time dependent except for sex and socioeconomic status. In analyses of IBD medication use, this was defined as ever use of the drug in question, and non-users of the drug in question were used as reference. Patients could have received other IBD drugs. In further analyses, we stratified the 5-ASA group according to use of oral steroids as a marker of disease severity. We also stratified the TNFα antagonists group according to the use of azathioprine to assess the separate effect of the two on the risk of IHD. Major intestinal surgery was treated as a dichotomous variable (yes/no). Homogeneity between IRRs was tested by likelihood ratio. All analyses were performed using SAS V.9.3 (SAS institute, Cary, North Carolina, USA).
The study was register based and followed the regulations and instructions set up by the Danish Data Protection Agency.
The cohort consisted of 4 570 820 people, of whom 28 833 were diagnosed with IBD (CD, n=7521; UC, n=19 990). During up to 13 years of follow-up (51 704 947 person-years), we identified 245 019 incident cases of IHD (97 501 MIs), of which 1175 occurred among patients with IBD and 863 of these more than 1 year after IBD diagnosis. Of the 1175 IHD cases in patients with IBD, 436 were MIs, of which 330 occurred more than 1 year after IBD diagnosis. Demographic and clinical characteristics of patients with IBD and non-IBD individuals are listed in table 1.
Risk of IHD
Patients with IBD had a significantly increased risk of IHD (IRR=1.59; 95% CI 1.50 to 1.69) compared with non-IBD individuals. The risk of IHD was particularly high in the first 3 months after IBD diagnosis (IRR=4.57; 95% CI 3.89 to 5.36; figure 1) and in the first year (IRR=2.13 95% CI 1.91 to 2.38). Disregarding the first year after diagnosis of IBD, the IRR was 1.40 (95% CI 1.31 to 1.50). This estimate decreased further after adjustment for comorbidity-related medication (IRR=1.22; 95% CI 1.14 to 1.30). In all subsequent analyses of IHD, we excluded follow-up within 1 year after diagnosis of IBD and included adjustment for comorbidity-related drugs. The risk for IHD was higher among women (IRR=1.33; 95% CI 1.21 to 1.46) than among men (IRR=1.14; 95% CI 1.04 to 1.26) (p=0.03) and this tendency was seen in patients with UC and in those with CD (table 2).
Stratifying the IBD-population according to age at IBD diagnosis, we found an overall increased risk of IHD in all age groups, which was most pronounced in patients diagnosed with IBD at young age (table 2). The same trend was seen when assessing patients with UC and CD separately (table 2).
Treatment of IBD
The IRR of IHD more than 1 year after IBD diagnosis was 1.16 (95% CI 1.06 to 1.26) among patients with IBD using 5-ASA and 1.36 (95% CI 1.22 to 1.51) among patients with IBD not using 5-ASA as compared with IBD-free never users of 5-ASA (table 2; p=0.02). Restricting the analysis to long-term users of 5-ASA (defined as three or more redeemed prescriptions) showed an even lower risk of IHD in this subgroup of patients (IRR=1.08; 95% CI 0.98 to 1.19). When stratifying analyses according to treatment with oral corticosteroids, as a proxy for underlying disease severity, we observed that the impact of ever use of 5-ASA on the risk of IHD was significantly different between patients with IBD previously treated or not treated with oral corticosteroids (p<0.01). Thus in patients with IBD previously receiving oral corticosteroids, the risk of IHD was 1.16 (95% CI 1.04 to 1.29) among 5-ASA users versus 1.71 (95% CI 1.46 to 2.01) among 5-ASA non-users (p=0.01), whereas among patients with IBD never treated with oral corticosteroids the risk of IHD was similar among 5-ASA users (IRR=1.16; 95% CI 1.01 to 1.33) and non-users (IRR=1.18; 95% CI 1.02 to 1.35) (p=0.78).
Patients requiring oral corticosteroids had a significantly higher risk of IHD (IRR=1.37; 95% CI 1.25 to 1.50) than patients never receiving oral corticosteroids (IRR=1.23; 95% CI 1.12 to 1.36) (p<0.01), using non-IBD patients with no use of oral corticosteroids as reference (table 2). This tendency persisted when UC and CD were analysed separately, but only reached statistical significance in patients with UC.
Thiopurines and TNFα antagonists
Risk of IHD was slightly lower among users than non-users of azathioprine or TNFα antagonists, although it did not reach statistical significance (table 2). Stratification on TNFα antagonist use showed the risk of IHD to be lowest among TNFα antagonist users with (IRR=0.58; 95% CI 0.22 to 1.56) or without use of azathioprine (IRR=0.68; 95% CI 0.10 to 4.82). Among non-users of TNFα antagonists, on the other hand, the risk of IHD tended to be higher than among users of TNFα antagonists and this was slightly attenuated by use of azathioprine (IRR=1.14; 95% CI 0.94 to 1.39) versus no use of azathioprine (IRR=1.25; 95% CI 1.17 to 1.35) (p=0.23).
Both among patients with IBD combined (p=0.01) and for patients with CD separately (p<0.01), there was a lower risk for IHD in patients who had undergone major intestinal surgery than in patients who had not undergone surgery (table 2).
In a sensitivity analysis, we substituted the outcome IHD with the subentity MI, as this diagnosis has undergone thorough validation in the NPR.14 As for IHD, patients with IBD were at increased risk of MI both in the first year after diagnosis and adjusted for comorbidity-related drugs (IRR=1.88; 95% CI 1.56 to 2.28) and >1 year after IBD diagnosis (IRR=1.17; 95% CI 1.05 to 1.30). The risk of MI was similar in patients with CD (IRR=1.18; 95% CI 0.92 to 1.51) and in those with UC (IRR=1.14; 95% CI 1.00 to 1.29).
In this nationwide Danish cohort study with 4.6 million individuals including 28 833 patients with IBD, the risk of IHD was increased twofold in the first year after IBD diagnosis and approximately 20% increased 1–13 years after diagnosis when taking age, gender, calendar year, socioeconomic status and use of glucose-lowering drugs, cardiovascular drugs and treatment for hypercholesterolaemia into account. Risk estimates were similar in patients with CD and UC. We found a significantly lower risk for IHD among users than among non-users of 5-ASA, even after adjustment for corticosteroids as a proxy for disease severity, which itself is associated with increased risk of IHD. We also observed a tendency towards lower risk of IHD in patients with IBD receiving thiopurines and TNFα antagonists, and in patients with CD, who had undergone surgery.
The major strength of our study was the population-based design covering a whole nation with free access to healthcare and comprehensive registration of citizens in national registers. The IBD and MI diagnoses in the Danish NPR have formerly been validated and found to be correct and almost complete, as described above.12 ,14 In addition to valid information on exposure and outcomes, we were able to account for socioeconomic status and prior treatment with drugs prescribed for conditions such as hypertension, hyperlipidaemia and diabetes mellitus, which are known risk factors for IHD.
A potential limitation was that individual-level data on smoking were not available in national registers. However, no differential effect of CD and UC on risk of IHD was seen despite well-known differences in smoking habits.15 Furthermore, adjustment for socioeconomic status may to some extent have captured the effect of smoking. Still, it should be kept in mind that our measure of socioeconomic status does not preclude the existence of residual confounding by other socioeconomic aspects not captured by this measure.
Another potential limitation was the lack of data on body mass index. However, since Scandinavian patients with IBD,16 and in particular patients with CD, tend to have a reduced rather than an excess body mass index, this limitation was judged to be minor.
Lastly, it could be seen as a limitation to the study that use of 5-ASA and other IBD drugs might not only reflect need of treatment but also general healthcare use, which may, in turn, be associated with prevention of IHD. However, the free and easy access to healthcare for all citizens in Denmark, and the adjustment for drugs related to risk factors for IHD, suggest that these factors did not have a great influence on the reported results.
This study is one of few and is the largest population-based study on IHD in IBD. We observed a 1.6-fold increased risk of IHD including the first year after IBD diagnosis, which declined to 1.40 after exclusion of the first year and to 1.22 after adjustment for relevant comorbidities. In accordance with this, a previous population-based IBD study from Canada, which included the first year after IBD diagnosis and did not take relevant comorbidities into account, reported an IRR of 1.26 for IHD in IBD.5 A smaller single-centre hospital study from the USA reported an HR of 4.08 (95% CI 2.49 to 6.70) for coronary heart events among patients with IBD,7 which might be due to assessment of a select subset of more severely ill patients or due to comparison with a selected reference population. In contrast, a previous study from the UK General Practice Research Database6 found no association between IBD and MI, probably owing to selective assessment of the less severely ill members of the IBD population.
We showed that a substantial part of the overall increased risk of IHD was due to a marked excess risk in the first year after diagnosis of IBD, hence to some extent possibly reflecting ascertainment bias (ie, that patients undergoing an extensive physical examination due to symptoms of IBD are more likely also to be diagnosed with other diseases such as IHD). Nevertheless, a 22% increased risk of IHD persisted during up to 13 years of follow-up of patients, suggesting a long-term effect of intermittent or chronic systemic IBD-related inflammation on risk of IHD. In line with this, patients with rheumatoid arthritis from the same population have formerly been shown to have an increased risk of stroke,17 whereas patients with psoriasis were at increased risk of MI and cardiovascular death.18 Our observation is further substantiated by the fact that systemic levels of C-reactive protein, a marker of inflammatory activity in these chronic disorders, is known to be associated with risk of IHD.4 ,19
An important measure and predictor of both current and later disease activity in IBD—and hence of systemic inflammation—is the use of oral corticosteroids.2 ,20 We observed a higher risk of IHD in patients with a history of receiving oral steroids at some point during the disease course as compared with patients who did not need these drugs. This was also the case after adjustment for medical treatment of hypertension, raised blood glucose levels and dyslipidaemia, which may be seen as complications to treatment with corticosteroids.21 On the other hand, patients receiving 5-ASA, which might have aspirin like properties,22 had a decreased risk of IHD in comparison with patients never receiving 5-ASA. This was the case both among patients with CD and UC and when restricted to long-term use (more than three redeemed prescriptions). Stratification according to the use of oral corticosteroids as a proxy for disease activity further showed that the protective effect of 5-ASA was only seen among patients in need of treatment with oral corticosteroids. It has formerly been suggested that 5-ASA may have a protective effect on development of several CVDs,23 and our findings needs confirmation in other IBD populations.
In line with our findings for 5-ASA, we observed a tendency towards a lower risk of IHD among patients treated with anti-inflammatory drugs, most pronounced among users of both thiopurines and TNFα antagonists, as compared with non-users of these drugs. This is interesting considering that use of these drugs might also reflect disease severity, and their use would be expected to be associated with an increased risk of IHD. In accordance with the observed trends, others have found a decreased risk for CVD in patients with rheumatoid arthritis treated with TNFα antagonists.24 ,25 Lastly, patients with CD who had undergone major intestinal surgery had a lower risk of IHD during follow-up than patients who had not undergone surgery. To our knowledge this has not been seen previously and together with our observation of medication use, it suggests that lowering of the inflammatory burden protects against development of IHD among patients with IBD.
Finally, we found that women with IBD had a slightly higher risk for IHD than men, which is in accordance with previous observations.5 ,8 Also, patients diagnosed with IBD at young age were at higher risk of IHD than patients diagnosed at later ages, which may reflect that chronic inflammation in young people adds relatively more to the underlying—and for those ages minor—expected risk of IHD.
In conclusion, this nationwide Danish cohort study showed a markedly increased risk of IHD and the subentity MI in the first year after IBD diagnosis, which may to some extent reflect ascertainment bias. However, a statistically significant 22% increased long-term risk persisted, suggesting an impact of IBD-related intermittent or chronic systemic inflammation on risk of IHD. Patients treated with 5-ASA had a lower risk of IHD than non-users of 5-ASA—in particular, patients with corticosteroid-requiring disease. A tendency towards a lower risk of IHD was also seen among patients treated with thiopurines and TNFα antagonists and among patients with CD who had undergone surgery. Our new findings suggest that lowering the inflammatory burden reduces the risk of IHD in IBD.
Correction notice This article has been corrected since it was published Online First. The paragraph under the heading ‘Thiopurines and TNFα antagonists’ has been amended.
Contributors The study was conceived and designed by CR, MFR, EL and TJ. Statistical analyses were performed by SB and JW. All authors were involved in interpretation of the data. CR drafted the manuscript, which was critically revised by all authors, who also approved the final version of the manuscript.
Funding TJ was supported by a female research leader grant (No 09-066323) from the Danish Council of Independent Research.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.