Introduction

Non-steroidal anti-inflammatory drugs (NSAIDs) are a widely used group of pharmaceutical agents. In addition to being distributed by prescription, NSAIDs are distributed as over-the-counter products and are a component of many different drug formulations. Thus, many patients may unknowingly ingest NSAIDs, which can cause a variety of colonic abnormalities including colitis, ulcers, and strictures.¹

The mechanisms of damage caused by NSAIDs to the bowel mucosa involve the activities of prostaglandin-endoperoxide synthase 1 [PTGS1 or cyclooxygenase-1 (COX-1)] and PTGS2 (COX-2). Moreover, NSAIDs interact with phospholipids and uncouple mitochondrial oxidative phosphorylation, which initiates biochemical changes that impair the function of the gastrointestinal barrier. The resulting increase in intestinal permeability leads to low-grade inflammation. Furthermore, the NSAID’s inhibition of COX enzymes, along with luminal aggressors, results in erosions and ulcers, with the potential complications of bleeding, protein loss, stricture formation, and perforation.²

Inflammatory bowel disease (IBD), including ulcerative colitis (UC) and Crohn’s disease (CD), refers to chronic inflammatory disorders of the gastrointestinal tract (GIT) identified by episodes of relapse and remission.³ The two identified subtypes of the disease involve the GIT in different patterns.^3,4 IBD is thought to result from an inappropriate inflammatory response to gut microbial flora in genetically predisposed individuals.⁵

The NSAID mechanism of action has raised questions over whether these drugs can exacerbate IBD. These questions have been debated in many studies with divergent results.

The aim of this systematic review is to investigate evidence on the role NSAIDs play in the exacerbation of IBD.

Method

Study design: The review was conducted and reported in accordance with the recommendations in the Cochrane Handbook for Reviews of Interventions (http://www.cochrane.org) and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (Figure 1).⁶

Outcome measures: The primary outcome is disease exacerbation defined as a flare of disease activity after a period of remission.

The secondary outcome measure is the worsening of disease activity in patients with active IBD. An active IBD is defined as the following: an IBD usually runs a waxing and waning course. When there is severe inflammation, the disease is considered active, and the person experiences a flare of symptoms. When there is less or no inflammation, the person usually is without symptoms and the disease is said to be in remission.⁶

Eligibility criteria:

1. Patients with IBD, including CD and UC in addition to microscopic colitis and collagen colitis.

2. No age limits.

3. All known NSAIDs, including non-specific COX inhibitors (such as aspirin, paracetamol, and ibuprofen) and COX-2 inhibitors (such as celecoxib, etoricoxib, and parecoxib). Rofecoxib and valdecoxib, which were withdrawn from the market in 2004 and 2005, respectively, because they excessively increased the risk of heart attacks and strokes with long-term use, are also included in the review, when found, as they may have caused an inflammatory effect while prescribed to patients.

4. Duration of using NSAIDs.

5. Oral, intravenous, or other methods of drug intake.

6. All observational studies (case-control and cohort studies), interventional studies [blinded or non-blinded randomized controlled trials (RCTs)], and other reviews (narrative and systematic reviews).

Studies on humans published between 2000 and 2020 were included to ensure up-to-date data. No language limit was used.

Studies were identified by searching the electronic PubMed, EmBase, and Cochrane databases. The reference lists in relevant papers were also screened for any additional studies. Additional trials were identified through the World Health Organization search portal (www.who.int/trialsearch).

The search was conducted by two authors (MH, AE). The last search date was December 6, 2019.

The search thread used was as follows:

(((((((((((“anti-inflammatory agents, non-steroidal”[MeSH Terms] OR “nonsteroidal anti inflammatory”[Text Word]) OR “non steroidal anti inflammatory”[Text Word]) OR “non steroidal antiinflammatory”[Text Word]) OR “nonsteroidal antiinflammatory”[Text Word]) OR “NSAID”[Text Word]) OR “cyclooxygenase inhibitor*”[Text Word]) OR “cox inhibitor*”[Text Word]) OR “anti inflammatory analgesi*”[Text Word]) OR “anti inflammatory agent*”[Text Word]) AND (((((“Inflammatory Bowel Diseases”[MeSH Terms] OR “inflammatory bowel dis*”[Text Word]) OR “crohn*”[Text Word]) OR “colitis”[Text Word]) OR “irritable bowel dis*”[Text Word]) OR “irritable bowel syn*”[Text Word])) NOT ((((“Animals”[Mesh]) OR (mice[Text Word] OR rats[Text Word] OR rabbit*[Text Word])) NOT (((“Animals”[Mesh]) OR (mice[Text Word] OR rats[Text Word] OR rabbit*[Text Word])) AND (“Humans”[Mesh]))) AND ((((((((((“anti-inflammatory agents, non-steroidal”[MeSH Terms] OR “nonsteroidal anti inflammatory”[Text Word]) OR “non steroidal anti inflammatory”[Text Word]) OR “non steroidal antiinflammatory”[Text Word]) OR “nonsteroidal antiinflammatory”[Text Word]) OR “NSAID”[Text Word]) OR “cyclooxygenase inhibitor*”[Text Word]) OR “cox inhibitor*”[Text Word]) OR “anti inflammatory analgesi*”[Text Word]) OR “anti inflammatory agent*”[Text Word]) AND (((((“Inflammatory Bowel Diseases”[MeSH Terms] OR “inflammatory bowel dis*”[Text Word]) OR “crohn*”[Text Word]) OR “colitis”[Text Word]) OR “irritable bowel dis*”[Text Word]) OR “irritable bowel syn*”[Text Word])))) AND (((“Systematic Review”[Publication Type] OR “Systematic Reviews as Topic”[Mesh] OR systematic[sb] OR “Meta-Analysis as Topic”[Mesh] OR “Meta-Analysis” [Publication Type] OR metaanalys*[Title] OR meta-analys*[Title])) OR ((((((“Controlled Clinical Trial”[Publication Type] OR “Controlled Clinical Trials as Topic”[Mesh])) OR (((random*[Text Word] OR controlled[Text Word] OR crossover[Text Word] OR cross-over[Text Word] OR blind*[Text Word] OR mask*[Text Word])) AND (trial[Text Word] OR trials[Text Word] OR study[Text Word] OR studies[Text Word] OR analys*[Text Word] OR analyz*[Text Word]))) OR rct[Text Word]) OR (((singl*[Text Word] OR doubl*[Text Word] OR tripl*[Text Word])) AND (blind[Text Word] OR mask[Text Word]))) OR placebo[Text Word]))

All studies identified by the search were screened for inclusion, primarily based on title and abstract. Eligible studies were retrieved in full text. Three authors (MH, MA, SDA) performed the inclusion/exclusion phase of the study. Any disagreement was resolved by discussion among the three authors or involvement of a senior author (AE).

A Rayyan intelligent systematic review was used for the inclusion/exclusion phase.⁷

This web-based application allows a blinded inclusion/exclusion of studies to be conducted and then disagreements to be resolved.

Data extraction: Three authors (M.H., M.A., and S.D.A.)independently extracted data based on the pre-defined study protocol’s inclusion criteria. Differences were resolved by consulting senior author (A.E.H.).

Data on patients, study methodology, study quality [case-control and cohort studies, interventional studies (blinded or non-blinded RCTs), narrative and systematic reviews], trial setting (single or multicenter, secondary or tertiary center/department, country of origin), duration of follow-up, outcomes assessed, definition of assessed outcome measures, intervention characteristics (type, dose, duration, mode of administration), and outcome measures were extracted.

No data synthesis was performed due to the heterogeneity of the included studies and inherent qualitative differences among studies.

Risk of bias in individual studies: The quality of bias control in the included studies was assessed by three authors independently of each other. The Cochrane risk-of-bias tool can be used for randomized trials and the Newcastle-Ottawa Scale to assess bias in observational studies. To assess bias (MH, MA, and SDA) in the included randomized trials, we used the Cochrane risk-of-bias tool for RCTs (RoB 2.0),⁸ which focuses on random sequence generation (selection bias), allocation concealment (selection bias), the blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), and selective reporting (reporting bias). The Newcastle-Ottawa Scale is used to assess the quality of non-randomized studies included in a systematic review.⁹ Each study was assigned a number of stars based on the selection of patients (maximum 4 stars), the comparability of cohorts (maximum 2 stars), and the ascertainment of the outcome (maximum 3 stars). The lower the number of stars is, the greater the risk of bias.

Results

In total, 1,352 articles were selected, with 11 included based on the inclusion criteria, 1,325 excluded based on the exclusion criteria, 6 disputed, and 10 that might be included. The final review included 15 articles after reading the full research text. These 15 studies were undertaken between 2000 and 2020: 1 double-blind placebo-controlled study, 1 prospective randomized placebo-controlled pilot study, 1 prospective open-label trial, 1 prospective open-label monocentric trial, 1 retrospective case-control trial, 1 prospective case-control trial, 7 systematic reviews, and 2 meta-analyses published before December 2019.^10-24 The number of participants per study ranged from 11 to 217 for the prospective and retrospective case-control, randomized, and placebo-controlled studies. The number of studies included in the meta-analyses varied between 2 and 21. The study characteristics are summarized in Table 1, 2.

Quality assessment results

Studies involving the relationship between NSAID use and IBD disease exacerbation, disease exacerbation after the IBD remission period, and the worsening of disease activity in patients with active IBD are described in Table 1,​ 2. The number of studies included in the meta-analyses varied between 2 and 21. The number of participants per study ranged from 11 to 217 for the prospective and retrospective case-control, randomized, and placebo-controlled studies. Using RoB 2.0, we assessed the risk of bias in the randomized trials (Table 3), and quality was assessed using the Newcastle-Ottawa Scale due to inherent qualitative differences between studies (Table 4).

The publication year of the studies, subtype of inflammatory disease, type and dose of NSAIDs used in treatment, duration of NSAID drug use, and results were reviewed.

Summary of the study findings: The results of the included studies are summarized in Table 1, 2.^10-24 Three of the observational studies included in this review investigated the impact of rofecoxib in flares of IBD, reporting no flares.^12-14 Similarly, a study reported that etoricoxib was safe in cases of IBD.¹⁰ Two studies determined that celecoxib was unrelated to IBD flares and could be used for the management of inflammatory symptoms where indicated,^11,14 as detailed in Table 1.^10-15

Five out of 9 review studies (systematic reviews and meta-analyses) documented that NSAIDs induce relapse in IBD as a result of unknown mechanisms and induce colitis in previously asymptomatic patients. However, some studies reinforced the safety of selective COX-2 inhibitors.^20,21 The evidence synthesis in these reviews was weak, and studies with a greater sample size were recommended.^22-24

Discussion

Active infection such as amoeba, parasite, bacterial, and viral infections in blood smear and/or stool cultures, presence of cytomegalo-virus pp65 and clostridium difficult toxin a and b antigens, use of NSAIDs, drug compliance, and type of current treatment (corticosteroid, salicylates, immunosuppressive drugs, and anti-tumor necrosis factor) were considered causes of exacerbation.^12,25,26

Regarding the administration of NSAIDs, according to consensus guidelines from the British Society of Gastroenterology, individuals with UC (including those with extensive disease) should be given a mix of oral and enema 5-ASA, and those who do not respond well to oral 5-ASA should also receive topical medication.²⁷ Even in patients with pancolitis, oral and topical 5-ASA therapy is preferable to monotherapy. Despite the clear advantages of enema therapy, patients continue to find the administration and maintenance of enemas difficult, and support and education in this area are urgently required.²⁶

The most common indications of the use of NSAIDs in IBD are extraintestinal manifestations, such as IBD-associated arthralgia, ankylosing spondylitis, sacroiliitis, and arthritis. However, the use of these drugs in the management of extraintestinal symptoms may lead to the exacerbation of the disease itself.²⁸ Hence, it is important to consider the side effects of such medications for extraintestinal manifestations because of their potential role in disease exacerbation. According to the American College of Gastroenterology, the use of NSAIDs is a possible trigger for disease exacerbation in patients with diagnosed IBD.²⁹ Similarly, Evans et al.³⁰ noted that NSAIDs play a role in relapse in patients with IBD. Therefore, patients with IBD are encouraged to avoid using NSAIDs because of concerns relating to their potential adverse effects on disease activity. In addition, Takeuchi et al.³¹ reported that non-selective NSAID intake is associated with the frequent and early clinical recurrence of IBD, as measured using the Harvey-Bradshaw Clinical Disease Activity Index.

Regarding the possible mechanism involved in the pathophysiology of gastrointestinal damage, several mechanisms have been proposed.² COX-1 and COX-2, when used concomitantly, cause damage to gastric mucosa by reducing blood flow and increasing the tendency of leukocytes to adhere to the blood vessels of the GIT, thus decreasing GIT defense.³² Although this is not the only manner in which NSAIDs can harm the gastrointestinal mucosa, the inhibition of prostaglandin synthesis is crucial in the development of mucosal injury.^33,34

Moreover, NSAIDs cause gastrointestinal damage by interacting with cellular phospholipids and oxidative phosphorylation.³⁵ These drugs frequently uncouple mitochondrial oxidative phosphorylation processes, leading to changes associated with a weakened gastrointestinal barrier. These biochemical changes are important in the pathophysiology of the disease, causing intestinal permeability to rise, which then causes low-grade inflammation. Erosion and ulceration are the outcomes of the NSAID suppression of COX enzymes in conjunction with luminal aggressors, with the possibility of perforation, hemorrhage, stricture development, and protein loss as sequelae.^2,34,36

The aforementioned processes might be used to account for the biological plausibility of IBD exacerbation with NSAID use. Because the major therapeutic objective of medicinal interventions for IBD is intestinal mucosal repair, non-selective COX inhibitors may cause GIT mucosal injury, which could delay healing. Similarly, NSAID use may lead to frequent relapses, as revealed by Forrest et al.¹⁶. However, the possible safety of selective COX-2 inhibitors may be explained by their lower interaction with the gastrointestinal barrier.

Study Limitations

This study has several limitations, such as the heterogeneity of the included studies. Even well-designed studies on NSAID use and IBD exacerbation risk have significant limitations in defining outcomes. Although some studies have defined the exacerbation of the disease as a subjective criterion, such as emergency admission to the hospital, there are also more objective studies using the “disease activity index.”

Conclusion

The published data remain contradictory and confusing. No consistent association between NSAID use and the risk of CD and UC exacerbation has been established, and it remains uncertain whether NSAIDs are key to inducing IBD relapse.

Peer-review: Externally peer-reviewed.

Acknowledgement: OpenSourceResearch collaboration is an international independent organization with special focus on implementing information technologies and artificial intelligence in clinical research. More about the organization and its projects can be found on its website: OSRC.network

Authorship Contributions

Concept: A.E.H., Design: M.H., A.E.H., Data Collection or Processing: M.H., M.A., S.D.A., Analysis or Interpretation: S.A., Literature Search: M.H., M.A., S.D.A., S.A., A.E.H., Writing: M.H., M.A., S.D.A., S.A., A.E.H.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.