Rheumatology

Rheumatol Int DOI 10.1007/s00296-014-3133-9

INTERNATIONAL

ORIGINAL ARTICLE - SAFETY AND PHARMACOVIGILLANCE

Assessing the likelihood of new‑onset inflammatory bowel disease following tumor necrosis factor‑alpha inhibitor therapy for rheumatoid arthritis and juvenile rheumatoid arthritis Asha Krishnan · Derrick J. Stobaugh · Parakkal Deepak 

Received: 9 July 2014 / Accepted: 9 September 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract  The association between inhibition of tumor necrosis factor-alpha (TNF-α) in patients with rheumatoid arthritis (RA) and juvenile rheumatoid arthritis (JRA) and the onset of inflammatory bowel disease (IBD) is unclear. We sought to evaluate this association by analyzing adverse events (AEs) reported to the Food and Drug Administration Adverse Event Reporting System (FAERS) with a standardized scoring tool for drug-induced AEs. A search of the FAERS for RA or JRA (January 2003–December 2011) reported with adalimumab, certolizumab pegol, etanercept, golimumab, or infliximab was performed. This dataset was then queried for cases indicating IBD. Full-length reports were accessed using the Freedom of Information Act and organized by age, sex, concomitant medications, co-morbidities, type of TNF-α inhibitor used, and diagnosis/treatment details. The Naranjo score was used to determine whether the drug-induced AEs were definite, probable, possible, or doubtful. There were 158 cases of IBD after TNF-α inhibitor exposure in RA or JRA patients. Use of the Naranjo score revealed that, in a majority of the cases (71.5 %), TNF-α inhibitor exposure was considered a ‘possible’ cause. A majority of the ‘probable cases’ in JRA were reported with etanercept (40 patients, 90.91 %). There were Electronic supplementary material  The online version of this article (doi:10.1007/s00296-014-3133-9) contains supplementary material, which is available to authorized users. A. Krishnan · D. J. Stobaugh  Research Institute, NorthShore University HealthSystem, 1001 University Place, Evanston, IL 60201, USA P. Deepak (*)  Division of Gastroenterology and Hepatology, Department of Gastroenterology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA e-mail: [email protected]

no ‘definite’ cases of anti-TNF-induced IBD. After applying the Naranjo scale, a weak association between newonset IBD and TNF-α inhibitor therapy in RA patients and a moderately strong association especially with etanercept exposure in JRA patients was observed. However, causality cannot be determined due to limitations of the FAERS and the Naranjo score. Keywords  Antibodies, monoclonal/adverse effects · Tumor necrosis factor-alpha/antagonists and inhibitors · Autoimmune diseases/chemically induced · Inflammatory bowel diseases/chemically induced

Introduction Tumor necrosis factor-alpha (TNF-α) is the central inflammatory cytokine identified in the pathogenesis of Crohn’s disease (CD) and ulcerative colitis (UC), the two main subtypes of inflammatory bowel disease (IBD) [1]. It is also responsible for the synovial inflammation in rheumatoid arthritis (RA) and juvenile rheumatoid arthritis (JRA) [2]. Treatment of these diseases often includes the use TNF-α inhibitors to block the cytokine’s activity. Two anti-TNF-α monoclonal antibodies, infliximab and adalimumab, and the TNF-α receptor fusion protein etanercept are the most commonly prescribed TNF-α inhibitors [1, 2]. A noticeable association between RA and IBD has been described, which is likely caused by a similar etiology [3]. Despite this predisposing factor, it has also been suggested that TNF-α inhibitors, namely etanercept, can further augment the risk of developing IBD in this select group of patients. New-onset CD, possibly triggered by etanercept, has been attributed to the drug’s pathways of action, though this is largely speculative [4]. This paradoxical relationship

13



Rheumatol Int

Fig. 1  Flow diagram showing inclusion (and exclusions) criteria for RA and JRA cases of anti-TNF-induced IBD in the Food and Drug Administration Adverse Event Reporting System

has been investigated in several case reports, which has provided warrant for further analysis [5–7]. In a study by van Dijken et al. [8], RA patients on etanercept were shown to have an increased risk of developing IBD, although the authors were unable to distinguish between the increased risk due to RA and the increased risk from etanercept treatment, due to an inadequate control group. A similar study by Dallocchio et al. [9] showed an increase risk of developing IBD in JRA patients, although all patients in the study were on concomitant non-steroidal anti-inflammatory drugs (NSAIDs), a medication known to increase the risk of new-onset IBD [10]. With a broader focus toward general rheumatologic diseases, Toussirot et al. [11] suggested a link between the use of TNF-α inhibitors and the development of IBD. Although the findings of this study were provocative, it was again limited by excessive use of NSAIDs in those who developed IBD. We sought to determine whether treatment of RA or JRA with TNF-α inhibitors induces new-onset IBD using the Food and Drug Administration Adverse Event Reporting System (FAERS).

Methods The FAERS is a public database used for post-marketing safety surveillance of FDA-approved drugs [12]. The database contains a total of 2,852,661 medication-related adverse event reports reported over a 9-year period between January 2003 and December 2011, which were downloaded and analyzed using SPSS 20 (IBM Co. Armonk, NY, USA). The dataset was created by searching FAERS reports for Medical Dictionary for Regulatory Activities (MedDRA) terms indicating RA or JRA. These search terms were ‘rheumatoid arthritis,’ ‘rheumatoid lung,’ ‘rheumatoid nodule,’ ‘rheumatoid scleritis,’ ‘rheumatoid vasculitis,’ and ‘juvenile arthritis.’ The dataset was then queried for treatment with study’s medications of interest, TNF-α inhibitors approved for the treatment of RA or JRA (adalimumab, certolizumab pegol, etanercept, golimumab, and infliximab). Cases that

13

did not report a medication of interest as the primary suspect were excluded from the dataset (Fig. 1). These cases were then further queried for UC and CD using terms from MedDRA. The search terms used were ‘colitis ulcerative,’ ‘colitis ulcerative aggravated,’ ‘Crohn’s disease,’ ‘Crohn’s disease aggravated,’ ‘proctitis ulcerative,’ ‘proctitis ulcerative aggravated,’ ‘inflammatory bowel disease,’ and ‘inflammatory bowel disease NOS.’ Duplicate cases were removed based on individual safety report (ISR) and case number to reduce this error. Finally, all cases with an indication of preexisting IBD were removed from the dataset. The Freedom of Information Act was used to request full-text reports from the FDA for the selected cases. These were reviewed individually and analyzed to establish authenticity, identify duplication and eliminate repetition. Reports were perused to confirm the occurrence of IBD after the initiation of TNF-α inhibitors. The cases were then evaluated, and information was collected on age, gender, underlying disease state for TNF-α inhibitor usage (RA or JRA), dates of TNF-α inhibitor usage and the individual agent implicated in the report. The list of concomitant medications for each report was reviewed extensively, with a focus on medications previously described to induce IBD (NSAIDs, oral contraceptives, and antibiotics) [13]. Comorbidities commonly associated with IBD, such as asthma, hypertension, hypothyroidism, and psoriasis, were also noted [14–17]. All objective clinical and endoscopic features, genetic, and pathological findings detailed in the reports were used to review and confirm IBD diagnoses including the subtype (CD, UC or IBD). Information on treatment of the diagnosed IBD was also collected. Patient outcomes were also assessed. If the patient were in remission from IBD they were considered ‘recovered’, those with minimal symptoms but with some improvement in overall disease status were classified as ‘partially recovered’, while patients that did not improve or expired were classified as ‘did not recover’. Naranjo scale was used to analyze all the relevant reports to determine the likelihood that they were cases of TNF-α inhibitor induced IBD. This is a causality assessment score

Rheumatol Int

used to determine the likelihood of a medication induced AE [18]. Answers to a series of 10 questions provide a total score that is used to make this classification. Each question provides a score that ranges from −2 to +2. It has been previously validated, with high within-rater and inter-rater retest reliability in addition to consensual, content and concurrent validity [18]. This has been used previously by our group to analyze the association between exposure to TNF-α inhibitors and the development of neurological adverse events [19]. For questions with possible ambiguity; a priori were developed to prevent subjectivity. Question 1 that deals with previous conclusive reports of IBD reported after treatment of RA or JRA with TNF-α inhibitors was scored as +1. The second question addresses temporal association of IBD to TNF-α inhibitor exposure. Here, all reports were scored as a +2 after assessing for the occurrence of IBD after the initiation of treatment with a TNF-α inhibitor and not being a preexisting condition that worsened or flared with TNF-α inhibitor exposure. All endoscopies, genetic tests, CT scans, and stool cultures performed were used as the ‘objective evidence’ to confirm IBD in Question 10. Each case was scored, and a compilation of these scores from the ten questions was computed, and then classified as ‘definite’ (>8 points), ‘probable’ (5–8 points), ‘possible’ (1–4 points), or ‘doubtful’ (0 points) to determine the probability that treatment of RA/JRA with TNF-α inhibitors induces IBD. Additional analysis was performed to assess for effect of reporting biases resulting in excess reporting of newonset IBD with TNF-α inhibitors or with a particular TNFα inhibitor compared to others in the same class of medications. This was done by calculating odds ratios for the reporting of cases with TNF-α inhibitors compared to control drugs (mycophenolate mofetil, leflunomide, methotrexate, and thiopurines (azathioprine and 6-mercaptopurine)) and control reactions (hernia, tinnitus, vertigo, deafness, and syncope) predefined to have no association with either study or control drugs.

while the average RA patient was much older, 51 (±15) years. There was a female predominance in the dataset, with 102 females (64.6 %), and the majority of patients were reported to be non-smokers (86.7 %). In 71 cases (44.9 %), the patients developed CD, UC developed in 58 cases (36.7 %), and the subtype of IBD was not specified in 29 cases (18.4 %) (Table 1). Comorbidities and concomitant medications Table  2 shows that there were ten patients with hypertension and six patients with hypothyroidism, common IBD comorbidities. Airway diseases, such as asthma and chronic bronchitis were comorbidities in three patients and one patient, respectively. Psoriasis, another risk factor, was seen in five patients, and diabetes mellitus was seen in six patients. Many patients were on drugs which may cause IBD, with NSAIDs (34 patients) being the most common. Less common drugs included concomitant estrogen (six patients) and antibiotics (two patients). Diagnosis Few rechallenges were performed in these cases, as seen in Table 3. Of the two rechallenges performed, there was one positive result in an RA patient. A total of 61 cases (37.4 %) had confirmed IBD based on a colonoscopy, biopsy, genetic test or CT scan, with a positive result occurring in 57 % of those who had a test performed. Treatment

Results

After IBD symptoms appeared, 83 patients (52.5 %) discontinued their current TNF-α inhibitor treatment in favor of another TNF-α inhibitor or a different class of medication (Table 4). Thirty-five patients (22.2 %) continued taking the TNF-α inhibitor post-action while the action taken with the TNF-α inhibitor was unknown in 40 patients (25.3 %). Treatment for the new-onset IBD included 5-aminosalicylates (26 patients), corticosteroids (19 patients), or antibiotics (10 patients).

Clinical and demographic characteristics

Naranjo scale

There were a total of 158 FAERS cases from January 2003 to December 2011 that matched this study’s inclusion and exclusion criteria (Fig. 1). As seen in Table 1, a majority of the JRA patients who developed IBD were on etanercept (50 patients, 90.9 %), with a similar pattern in those with RA (53 patients, 51.5 %). No cases of RA or JRA patients with new-onset IBD were reported with certolizumab pegol and only one case was reported with golimumab in a RA patient. The average age of JRA patients was 17 (±9) years

After applying the Naranjo scale, there were no ‘definite’ cases of TNF-α inhibitor IBD (Table 5). One hundred and thirteen (71.5 %) cases with a Naranjo score between 1 and 4 were considered ‘possible’ adverse events. The remaining 45 ‘probable’ reports (28.5 %) had a score between 5 and 8. In patients with RA, the majority of cases were scored as ‘possible’ (102, 99 %), while the majority were ‘probable’ in JRA (44, 80 %). Furthermore, the ‘possible’ cases varied across all TNF-α inhibitors, especially in those with RA. A

13



Rheumatol Int

Table 1  Demographics of JRA and RA patients that developed IBD in the Food and Drug Administration Adverse Event Reporting System (January 2003–December 2011) JRA

RA

Adalimumab (n = 2)

Etanercept (n = 50)

Infliximab (n = 3)

Adalimumab (n = 24)

Etanercept (n = 53)

Golimumab (n = 1)

Infliximab (n = 25)

15

16 ± 9

24 ± 15

48 ± 14

51 ± 17

69

55 ± 13

2 (100)

18 (36)

2 (67)

17 (71)

42 (79)

1 (100)

20 (80)

–

14 (28)

–

5 (21)

9 (17)

–

5 (20)

–

18 (36)

1 (33)

2 (8)

2 (4)

–

–

 CD, n (%)

2 (100)

22 (44)

1 (33)

13 (54)

23 (43)

–

10 (40)

 UC, n (%)

–

6 (12)

1 (33)

11 (46)

26 (49)

1 (100)

13 (52)

 IBD, n (%) Outcome

–

22 (44)

1 (33)

–

4 (8)

–

2 (8)

 Recovered, n (%)

–

15 (30)

–

5 (21)

7 (13)

1 (100)

6 (24)

 Partially  recovered, n (%)

–

4 (8)

1 (33)

11 (46)

21 (40)

–

4 (16)

 Did not recover, n (%)

1 (50)

9 (18)

–

1 (4)

6 (11)

–

7 (28)

1 (50)

22 (44)

2 (67)

7 (30)

19 (36)

–

8 (32)

 No, n (%)

1 (50)

49 (98)

3 (100)

16 (67)

46 (87)

1 (100)

21 (84)

 Previous, n (%)

1 (50)

–

–

5 (21)

1 (2)

–

–

 Yes, n (%)

–

–

–

3 (13)

3 (6)

–

3 (12)

 Unknown, n (%)

–

1 (2)

–

–

3 (6)

–

1 (4)

Age (in years)  Average (±SD) Gender  Female, n (%)  Male, n (%)  Not Specified IBD subtype

 Unknown, n (%) Smoker

CD Crohn’s disease, IBD inflammatory bowel disease, JRA juvenile rheumatoid arthritis, RA rheumatoid arthritis, UC ulcerative colitis

majority of the ‘probable cases’ in JRA were reported with etanercept (40 patients, 90.91 %). Reporting bias Additional analysis performed to assess for reporting biases did not demonstrate excessive reporting of cases compared to control reactions with any of the TNF-α inhibitors compared to control drugs in patients with RA or JRA in the FAERS (supplementary table 1).

Discussion Analysis of the FAERS revealed 158 cases of IBD after TNF-α inhibitor exposure in RA or JRA patients. In a majority of the cases (71.5 %), TNF-α inhibitor exposure was considered a ‘possible’ cause of IBD per the Naranjo scale. There were no ‘definite’ cases of anti-TNF-induced IBD. Similar results were demonstrated in a French retrospective study by Dallocchio et al. [9] In this study, the relationship between etanercept therapy and IBD development was

13

also considered ‘possible’ using replies to a questionnaire regarding the medical history, demographics, clinical findings, and patient outcome of eight JRA patients [9]. This strong association between IBD and etanercept has also been found in this study as well as other recent studies. The higher number of probable cases of etanercept induced IBD found in this study may be due to the differences in the binding affinity and structure of the various TNF-α inhibitors, with etanercept showing a lack of efficacy in the treatment of IBD [20]. Infliximab is a chimeric monoclonal antibody has been shown to neutralize both soluble and membrane bound TNF-α (expressed in macrophages and activated T cells in inflamed human gut) with activation of apoptosis of T cells through a caspase-dependent pathway [21]. In contrast, etanercept is a recombinant TNF receptor/immunoglobulin G fusion protein that neutralizes soluble TNF-α that does not activate T cell apoptosis due to the lower binding affinity and inability to cross-fix membrane bound TNF due to its monomeric structure [21]. Similar results have been shown in vitro with adalimumab compared to etanercept [22]. Such differences in the ability to induce apoptosis may result in the development of new-onset IBD after treatment for RA or JRA. Additional

Rheumatol Int Table 2  Comorbidities and concomitant medications of JRA and RA patients that developed IBD in the Food and Drug Administration Adverse Event Reporting System (January 2003–December 2011) JRA

Comorbiditiesa  Hypertension  Diabetes mellitus  Hypothyroidism  Psoriasis  Asthma  Polyarthritis  Ankylosing spondylitis  Chronic bronchitis  Hyperthyroidism  Overlap syndrome Concomitant medicationsa  NSAID  Opioid  Estrogen  Antibiotic

RA

Adalimumab (n = 2)

Etanercept (n = 50)

Infliximab (n = 3)

Adalimumab (n = 24)

Etanercept (n = 53)

Golimumab (n = 1)

Infliximab (n = 25)

– – – – – – – – – –

– – – 1 – – – – – 1

– – – – – – – – – –

2 2 1 – – 1 – 1 1 –

6 3 2 3 2 2 1 – – –

1 – 1 – – – – – – –

1 1 2 1 1 – – – – –

1 1 –

5 – –

– – –

9 7 –

10 6 2

– – –

9 5 4

–

–

–

1

–

–

1

JRA juvenile rheumatoid arthritis, NSAID non-steroidal anti-inflammatory drug, RA rheumatoid arthritis a

  Many patients had multiple comorbidities and concomitant medications; only relevant comorbidities and concomitant medications are shown

Table 3  Diagnostic tests performed on JRA and RA patients that developed IBD in the Food and Drug Administration Adverse Event Reporting System (January 2003–December 2011) JRA

Rechallenge  Total  Positive Endoscopy  Total  Positive Biopsy  Total  Positive Genetic test  Total  Positive Other testsa  Confirming IBD

RA

Adalimumab (n = 2)

Etanercept (n = 50)

Infliximab (n = 3)

Adalimumab (n = 24)

Etanercept (n = 53)

Golimumab (n = 1)

Infliximab (n = 25)

– –

1 –

– –

– –

1 1

– –

– –

– –

20 20

2 2

11 10

20 15

– –

12 10

– –

18 18

1 1

4 2

8 5

– –

9 6

1 –

17 1

1 –

– –

1 –

– –

– –

–

1

–

–

–

–

1

CD Crohn’s disease, CT computerized tomography, IBD inflammatory bowel disease, JRA juvenile rheumatoid arthritis, RA rheumatoid arthritis a

  Other tests confirming IBD: CT confirmed CD, CT revealed terminal ileitis

13



Rheumatol Int

Table 4  Treatment of JRA and RA patients that developed IBD in the Food and Drug Administration Adverse Event Reporting System (January 2003–December 2011) JRA TNF used Adalimumab (n = 2) Continued TNF-α inhibitor  No 1  Yes –  Unknown 1 Treatment offered for IBDa  5-ASA –  Prednisone –  Antibiotics –  Infliximab –  Adalimumab –  Surgical resection –  Azathioprine –  Unspecified IBD – medication  Cyclosporin –  Etanercept –  Leukocytapheresis –  Unknown

2

RA TNF used Etanercept (n = 50)

Infliximab (n = 3)

Adalimumab (n = 24)

Etanercept (n = 53)

Golimumab (n = 1)

Infliximab (n = 25)

31 2 17

1 – 2

11 8 5

27 19 7

– 1 –

12 5 8

8 3 1 6 4 – 1 2

– 1 – 1 – – 1 –

5 3 2 – – 3 1 –

6 7 4 3 4 2 1 –

– – – – – – – –

7 5 3 – 1 – – –

– – –

– 1 –

1 – –

– – –

– – –

– – 1

34

1

15

35

1

13

5-ASA 5-aminosalicylic acid, IBD inflammatory bowel disease, JRA juvenile rheumatoid arthritis, RA rheumatoid arthritis a

  Many patients had multiple types of treatment

Table 5  Distribution of Naranjo score among JRA and RA patients that developed IBD in the Food and Drug Administration Adverse Event Reporting System (January 2003–December 2011) Possible ADR

Probable ADR

 Adalimumab, n (%)

1 (50)

1 (50)

 Etanercept, n (%)

10 (20)

40 (80)

–

3 (100)

 Adalimumab, n (%)

24 (100)

–

 Etanercept, n (%)

52 (98)

1 (2)

 Golimumab, n (%)

1 (100)

–

 Infliximab, n (%)

25 (100)

–

Juvenile rheumatoid arthritis

 Infliximab, n (%) Rheumatoid arthritis

ADR adverse drug reaction, IBD inflammatory bowel disease

mechanisms for this paradoxical phenomenon may include increased peripheral T cell reactivity both to microbial antigens and to self-antigens that has been demonstrated in RA patients after treatment with etanercept [23]. These findings may be due to this possible adverse drug relationship, though it may also be caused by the common, genetic pathways found between IBD and RA.

13

A notable association between IBD and RA or JRA has already been found in large-scale studies assessing risk factors and comorbidities in these diseases [3, 14, 16]. There have been common loci (IL2/IL21, TNFRSF14, and IRF5) found between both UC and RA as well as the IL2RA locus found in both CD and RA. Furthermore, the REL locus, the PRDM1/ATG5 locus, and FCGR2A locus have been associated with all three of these diseases [24]. Increased susceptibility to UC or CD may be due to this genetic link, causing its occurrence in some RA patients. In this study, the majority of the cases only met the criteria to be considered ‘possible’ adverse events, calling into question the role of concomitant medications and underlying diseases as alternative causes. Antibiotics have been found to alter the intestinal microbiota, potentially modulating immune tolerance and increasing the risk of IBD. Such a risk has also been found with the mucosal injury linked to NSAIDs. Additionally, the proinflammatory effects of estrogen have also been suspected to cause mucosal lesions in patients on oral contraceptives [13]. Such environmental exposures could result in the development of de novo IBD. The high prevalence of psoriasis and asthma as underlying diseases in this study confirms the findings of previous literature [14, 17]. This increased risk of IBD is suspected

Rheumatol Int

due to genetic findings that show loci associated with IBD in psoriasis and asthma [24]. Moreover, the bowel and bronchial tree have similar, embryological origins that are suspected to contribute to the pathogenesis of both IBD and asthma [14]. Though the underlying mechanisms are still unknown, an increased risk due to a suspected similar etiology has also been associated with diabetes mellitus [3]. These shared origins of IBD and other diseases could increase the risk of RA patients with such comorbidities, leading to the development of the disease. Though this relationship is not definite, it raises questions about the safety of TNF-α inhibitor therapy in patients with RA or JRA, an issue previously addressed in a retrospective study by Toussirot et al. [11]. In this study, the clinical and personal information of 16 cases of de novo IBD in rheumatologic disease patients was acquired by means of a questionnaire. All patients experienced a favorable outcome when discontinuing their current anti-TNF therapy, switching to another TNF-α inhibitor or another treatment method. These findings suggest that such paradoxical cases should be carefully evaluated for a previous history and family history of IBD because use of TNF-α inhibitors in RA or JRA patients with such a history is ill-advised. A different TNF-α inhibitor or another treatment should be considered whether intestinal symptoms do develop or worsen while on this treatment method [7]. This study was affected by inherent limitations to the FAERS due to both under and over reporting. These cases are spontaneously reported by consumers, healthcare professionals, and lawyers, and are thus susceptible to such reporter bias [25]. We have attempted to compensate for misdiagnoses and report duplications by reviewing the fulllength reports. Additionally, the FAERS does not require definite causation for filing a report [26]. However, we have attempted to quantify the association of these AEs after exposure to TNF-α inhibitors by applying the Naranjo score, a validated probability scoring tool specific to druginduced AEs [18]. Though this scoring system was used, it also has limitations, limitations which may have contributed to the strong association found between JRA and IBD in this study. The Naranjo score takes into account any alternative causes of the AE and, since RA has been found to associated with IBD (and scored as an alternative cause) while JRA has not, more JRA cases have scored in the ‘probable’ category compared to RA [3]. Other imitations of the dataset included lack of information on duration of exposure to TNF-α inhibitors as well as action taken with these medications in 25.3 % in relation to the onset of IBD. This may have underestimated the strength of association between new-onset IBD and exposure to TNF-α inhibitors in RA and JRA patients. Though we have objective verification for 37.4 % of the cases, we could not verify whether

the diagnosis of RA/JRA was correct or these could have been cases of enteropathic arthritis, a disease that has an association with IBD [27]. Additionally, subclinical gut inflammation has been reported in patients with spondyloarthropathy including juvenile rheumatoid arthritis [28]. Cases reported especially in patients with JRA could represent such cases where the IBD eventually manifested. Finally, no cases of RA or JRA have been reported after the treatment of IBD with TNF-α inhibitors [29]. Hence, it is unlikely that we misidentified such cases in the study dataset. In summary, there may be only a weak association between new-onset IBD and TNF-α inhibitor therapy in RA patients. After applying the Naranjo scale, a majority of the cases scored as ‘probable’ especially with etanercept exposure in JRA patients, indicating a moderately strong association. However, causality cannot be determined due to limitations of the FAERS and the Naranjo score. Acknowledgments  We thank Harold Stepper at the Food and Drug Administration for support in sending case reports obtained through the Freedom of Information Act. Conflict of interest  None of the authors report any conflict of interest with the subject matter of this manuscript.

References 1. Sandborn WJ, Targan SR (2002) Biologic therapy of inflammatory bowel disease. Gastroenterology 122:1592–1608 2. Olsen NJ, Stein CM (2004) New drugs for rheumatoid arthritis. N Engl J Med 350:2167–2179 3. García Rodríguez LA, González-Pérez A, Johansson S, Wallander MA (2005) Risk factors for inflammatory bowel disease in the general population. Aliment Pharmacol Ther 22:309–315 4. Haraoui B, Krelenbaum M (2009) Emergence of Crohn’s disease during treatment with the anti-tumor necrosis factor agent etanercept for ankylosing spondylitis: possible mechanisms of action. Semin Arthritis Rheum 39:176–181 5. Wiegering V, Morbach H, Dick A, Girschick HJ (2010) Crohn’s disease during etanercept therapy in juvenile idiopathic arthritis: a case report and review of the literature. Rheumatol Int 30:801–804 6. Ruemmele FM, Prieur AM, Talbotec C, Goulet O, Schmitz J (2004) Development of Crohn disease during anti-TNF-alpha therapy in a child with juvenile idiopathic arthritis. J Pediatr Gastroenterol Nutr 39:203–206 7. Oh J, Arkfeld DG, Horwitz DA (2005) Development of Crohn’s disease in a patient taking etanercept. J Rheumatol 32:752–753 8. van Dijken TD, Vastert SJ, Gerloni VM, Pontikaki I, Linnemann K, Girschick H et al (2011) Development of inflammatory bowel disease in patients with juvenile idiopathic arthritis treated with etanercept. J Rheumatol 38:1441–1446 9. Dallocchio A, Canioni D, Ruemmele F, Duquesne A, Scoazec JY, Bouvier R et al (2010) Occurrence of inflammatory bowel disease during treatment of juvenile idiopathic arthritis with etanercept: a French retrospective study. Rheumatology (Oxford) 49:1694–1698

13

10. Felder JB, Korelitz BI, Rajapakse R, Schwarz S, Horatagis AP, Gleim G (2000) Effects of nonsteroidal antiinflammatory drugs on inflammatory bowel disease: a case-control study. Am J Gastroenterol 95:1949–1954 11. Toussirot É, Houvenagel É, Goëb V, Fouache D, Martin A, Le Dantec P et al (2012) Development of inflammatory bowel disease during anti-TNF-α therapy for inflammatory rheumatic disease. A nationwide series. Joint Bone Spine 79:457–463 12. The Adverse Event Reporting System (AERS) Older quar terly data files [homepage on the Internet]. U.S. Food and Drug Administration [updated 2013 Aug 15; cited 2014 June 20]. http://www.fda.gov/ 13. Dubeau MF, Iacucci M, Beck PL, Moran GW, Kaplan GG, Ghosh S et al (2013) Drug-induced inflammatory bowel disease and IBD-like conditions. Inflamm Bowel Dis 19:445–456 14. Bernstein CN, Wajda A, Blanchard JF (2005) The clustering of other chronic inflammatory diseases in inflammatory bowel disease: a population-based study. Gastroenterology 129:827–836 15. Ha C, Magowan S, Accortt NA, Chen J, Stone CD (2009) Risk of arterial thrombotic events in inflammatory bowel disease. Am J Gastroenterol 104:1445–1451 16. Kappelman MD, Galanko JA, Porter CQ, Sandler RS (2011) Association of paediatric inflammatory bowel disease with other immune-mediated diseases. Arch Dis Child 96:1042–1046 17. Naldi L, Mercuri SR (2010) Epidemiology of comorbidities in psoriasis. Dermatol Ther 23:114–118 18. Naranjo CA, Busto U, Sellers EM, Sandor P, Ruiz I, Roberts EA et al (1981) A method for estimating the probability of adverse drug reactions. Clin Pharmacol Ther 30:239–245 19. Deepak P, Stobaugh DJ, Sherid M, Sifuentes H, Ehrenpreis ED (2013) Neurological events with tumour necrosis factor alpha inhibitors reported to the Food and Drug Administration Adverse Event Reporting System. Aliment Pharmacol Ther 38:388–396

13

Rheumatol Int 20. Sandborn WJ, Hanauer SB, Katz S, Safdi M, Wolf DG, Baerg RD et al (2001) Etanercept for active Crohn’s disease: a randomized, double-blind, placebo-controlled trial. Gastroenterology 121(5):1088–1094 21. Van den Brande JM, Braat H, van den Brink GR, Versteeg HH, Bauer CA, Hoedemaeker I et al (2003) Infliximab but not etanercept induces apoptosis in lamina propria T-lymphocytes from patients with Crohn’s disease. Gastroenterology 124(7):1774–1785 22. Shen C, Assche GV, Colpaert S, Maerten P, Geboes K, Rutgeerts P et al (2005) Adalimumab induces apoptosis of human monocytes: a comparative study with infliximab and etanercept. Aliment Pharmacol Ther 21(3):251–258 23. Berg L, Lampa J, Rogberg S, van Vollenhoven R, Klareskog L (2001) Increased peripheral T cell reactivity to microbial antigens and collagen type II in rheumatoid arthritis after treatment with soluble TNFalpha receptors. Ann Rheum Dis 60(2):133–139 24. Lees CW, Barrett JC, Parkes M, Satsangi J (2011) New IBD genetics: common pathways with other diseases. Gut 60:1739–1753 25. Gale EA (2009) Collateral damage: the conundrum of drug safety. Diabetologia 52:1975–1982 26. FDA Adverse Event Reporting System (FAERS) (formerly AERS) [homepage on the Internet]. U.S. Food and Drug Administration [updated 2012 Sep 10; cited 2014 June 20]. http://www.fda.gov/ 27. Holden W, Orchard T, Wordsworth P (2003) Enteropathic arthritis. Rheum Dis Clin North Am 29:513–530 28. Leirisalo-Repo M, Turunen U, Stenman S, Helenius P, Seppälä K (1994) High frequency of silent inflammatory bowel disease in spondylarthropathy. Arthritis Rheum 37:23–31 29. Ramos-Casals M, Brito-Zerón P, Soto MJ, Cuadrado MJ, Khamashta MA (2008) Autoimmune diseases induced by TNF-targeted therapies. Best Pract Res Clin Rheumatol 22:847–861