Aspirin in the Aetiology of Crohn's Disease and Ulcerative Colitis

A European Prospective Cohort Study

S. S. M. Chan; R. Luben; M. M. Bergmann; H. Boeing; A. Olsen; A. Tjonneland; K. Overvad; R. Kaaks; H. Kennedy; K.-T. Khaw; E. Riboli; A. R. Hart


Aliment Pharmacol Ther. 2011;34(6):649-655. 

In This Article


Aspirin in the aetiology of IBD has not previously been investigated in a study using a prospective cohort design. Our main findings were: over a six times greater risk of developing CD in those taking regular aspirin, but none for UC. Therefore, regular aspirin use may have a role in the aetiology of CD, but not UC, although the reasons for this difference are unknown and require investigation. If this association is causal, then aspirin may be involved in the aetiology of approximately a fifth of patients with CD in this age group. Surprisingly an interaction between aspirin and smoking was observed in that smoking appeared to be protective for CD in those using regular aspirin. This finding is intriguing although the biological mechanism is unknown. Hypothetically, the pro-thrombotic and inflammatory effects of cigarette smoking leading to vasoconstriction, atherosclerosis and ischaemia might be diminished by aspirin's antiplatelet effects.

The relationship between nonsteroidal anti-inflammatory drugs, including aspirin and IBD has only previously been assessed in case–control studies. Unlike our study these reported the effects on IBD as a whole but not individually for CD or UC.[13–15,17–20] Three existing studies on NSAID and the development of new IBD reported positive associations. The first found that there was almost a two-fold nonsignificant increase in the development of new colitis associated with NSAID/salicylates usage in those with no prior history of IBD (OR = 1.75, 95% CI = 0.94–3.29) and that recent usage of NSAID/salicylates in patients with and without established IBD was associated with an increased risk of emergency hospital admission for IBD colitis (OR = 1.76, 95% CI = 1.02–3.03).[13] The second study reported that 25% of patients with a new onset of IBD, or an exacerbation, was associated with recent usage of NSAID (OR = 20.3, 95% CI = 2.6–159.7).[14] The authors commented that whilst those who developed de novo IBD following NSAID usage could have a NSAID specific enteropathy, this was unlikely given further relapses without taking these drugs. The third study reported a higher frequency of antecedent exposure to NSAID/salicylates in patients presenting with colitis for the first time compared with both hospital (OR = 6.2, 95% CI = 3.2–13.5) and community controls (OR = 9.1, 95% CI = 4.5–21.9).[15] Of these patients only 17% had disease remission on withdrawal of NSAID/salicylate therapy suggesting they had true IBD and not a NSAID enteropathy. A further four epidemiological studies have investigated NSAID usage and symptoms in patients with known IBD, with three showing positive associations[18–20] and one no effect.[17]

The main advantage of this prospective cohort study over previous retrospective case-control work was the minimisation of both selection and recall biases. Recruitment of the cohort before the development of symptoms and the recording of aspirin prior to diagnosis prevented recall bias. Studying cases and controls from the same baseline population ensured there was no selection bias. Our study is generalisable in that patients had a clinical distribution of disease comparable to that expected in a typical patient population. However, as the study recruited middle aged to elderly participants, these results are not as relevant to younger patients who will use aspirin less regularly. The use of aspirin in the CD control group was less than in the UC one. In this population, those who developed CD were predominantly women and younger than those diagnosed with UC, who were more likely to be older men. As cardiovascular disease is less common in younger women, than in older men, this may explain this lower prevalence of aspirin use in the age and gender matched controls in the CD study. In the latter, there were twice as many women as men, who were on average 3 years younger than men at recruitment in to the UC study. A further limitation was that the question of aspirin use varied between centres, namely on recording the time period for which the drug was taken. Importantly, most of the effect size was due to the findings from Denmark, which recorded aspirin use over 1 year, whereas the other centres enquired about use over shorter times. An analysis of dose-response would have been desirable to help infer a causal association for aspirin, but was not possible due to insufficient comprehensive data on dose from all centres. However, our findings suggest that it is regular aspirin usage over longer periods that may be relevant. Aspirin use was assessed in a single measure from questionnaires completed at baseline and although follow-up questionnaires were sent, we did not have access to this data. However, if aspirin use changes equally over time between those who subsequently become cases and those who remain well, this measurement error would result in an under-estimate of the true effect of aspirin. Another possibility for the association may be residual confounding if aspirin is a marker for another exposure which is the true aetiological factor. To exclude residual confounding other aetiological hypotheses will be explored in future work in this cohort and the use of aspirin adjusted for any associations to see if the drug effect is maintained. Finally, there is the possibility that existing inflammatory bowel disease is more likely to be diagnosed in participants on regular aspirin, as they are susceptible to gastrointestinal bleeding and require endoscopy. However, the exclusion of patients who developed IBD within 18 months of recruitment should have minimised this bias. Also this bias is unlikely as the association with aspirin was not seen for UC. As no effect was seen for paracetamol use in the Denmark centre, this suggests the association with aspirin was not a marker of a class effect for analgesics.

There are plausible biological mechanisms for the strong positive association between regular aspirin use and the development of incident Crohn's disease. Arachidonic acid (AA, a n-6-polyunsaturated fatty acid) is the substrate for COX enzyme which synthesises prostaglandins, thromboxanes and prostacyclins. These compounds are involved in the regulation of platelet aggregation, vasoconstriction and dilatation of vascular smooth muscle and have both pro-inflammatory and anti-inflammatory effects. Aspirin by inhibiting COX leads to a decreased synthesis of PGH2 and consequently less PGD2 and PGE2. These molecules have several anti-inflammatory properties,[21] and a decreased synthesis could hypothetically induce CD. PGE2 has pro-inflammatory effects, but also several anti-inflammatory properties including inhibiting the production of tumour necrosis factor (TNF)-α, interleukin (IL)-1 and other pro-inflammatory cytokines,[11,22–24] which are found in increased amounts in tissues affected with CD. Similarly, PGD2 has anti-inflammatory[25] and pro-resolution properties.[26] Patients with active IBD have alterations in their mesenteric vasculature and a decreased mesenteric perfusion.[27] COX-1 inhibition results in an inability to increase microcirculatory blood flow to enterocytes by inhibiting vasodilatation when cellular metabolic demands increase.[28] This leads to an ischaemic environment that may perpetuate a chronic inflammatory state. The mechanism is particularly relevant as those with CD and UC have lost nitric oxide (NO)-mediated vasodilatation.[29] With this loss, a novel vasodilating mechanism dependent on PGD2 appears to be important in increasing mucosal blood flow.

In summary, this prospective cohort study found that regular aspirin use was strongly associated with an increased risk of developing CD, for which there are plausible biological mechanisms and supportive previous epidemiological work. Future epidemiological studies in CD should measure regular aspirin use in their investigations of the aetiology of this chronic debilitating condition.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: