Assessment of extraintestinal manifestations in inflammatory bowel diseases: A systematic review and a proposed guide for clinical trials

Abstract

Background and aims

Extraintestinal manifestations are common in inflammatory bowel disease patients, although there are few data available on their diagnosis, management and follow-up. We systematically reviewed the literature evidence to evaluate tools and investigations used for the diagnosis and for the assessment of the treatment response in inflammatory bowel disease patients with extraintestinal manifestations.

Methods

We searched in PubMed, Embase and Web of Science from January 1999–December 2019 for all interventional and non-interventional studies published in English assessing diagnostic tools and investigations used in inflammatory bowel disease patients with extraintestinal manifestations.

Results

Forty-five studies (16 interventional and 29 non-interventional) were included in our systematic review, enrolling 7994 inflammatory bowel disease patients. The diagnostic assessment of extraintestinal manifestations was performed by dedicated specialists in a percentage of cases ranging from 60–100% depending on the specific condition. The clinical examination was the most frequent diagnostic strategy, accounting for 35 studies (77.8%). In patients with primary sclerosing cholangitis or rheumatological symptoms, biochemical and imaging tests were also performed. Anti-TNF agents were the most used biological drugs for the treatment of extraintestinal manifestations (20 studies, 44.4%), and the treatment response varied from 59.1% in axial spondyloarthritis to 88.9% in ocular manifestations. No benefit was detected in primary sclerosing cholangitis patients after treatment with biologics.

Conclusions

In the clinical management of inflammatory bowel disease patients with extraintestinal manifestations the collaboration of dedicated specialists for diagnostic investigations and follow-up is key to ensure the best of care approach. However, international guidelines are needed to homogenise and standardise the assessment of extraintestinal manifestations.

Introduction

Crohn's disease (CD) and ulcerative colitis (UC) are both phenotypes of inflammatory bowel disease (IBD).¹,² The most frequent IBD symptoms involve the gut tract, although a variety of extra-intestinal manifestations (EIMs) can be observed.¹,² EIMs have a broad spectrum of clinical expression: articular (axial and peripheral spondyloarthritis, dactylitis and enthesitis), ocular (uveitis, episcleritis), cutaneous (pyoderma gangrenosum, erythema nodosum and Sweet syndrome), hepatobiliary (primary sclerosing cholangitis) and immune and inflammatory associated diseases (psoriasis, atopic dermatitis and hidradenitis suppurativa).³ Prevalence of these manifestations ranges from 25–40% in IBD patients.^4–6 These manifestations have relevant consequences on patient’s quality of life and require complex and multidisciplinary management.⁷ Nowadays, a tumor necrosis factor antagonist (anti-TNF) is recognised as the best therapeutic option to manage EIMs of IBD patients,^8–10 while recent vedolizumab data are disappointing.¹¹ Other biologics (e.g. ustekinumab)¹² and small molecules (e.g. tofacitinib)¹³ may be efficacious in these patients and have been proposed as a possible therapeutic option. However, diagnosis of EIMs and accurate evaluation of drug efficacy on EIMs are challenging and not standardised in IBD trials. Hence, available data are difficult to interpret. For this reason, we performed a systematic literature review to identify the most appropriate tools and investigations to be used for diagnosis and assessment of treatment efficacy in IBD patients with EIMs.

Methods

Search strategy

The Cochrane Handbook¹⁴ and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)¹⁵ statement for reporting of systematic reviews were adopted as guidance for this work. We searched in PubMed, Embase and Web of Science databases to identify relevant studies from January 1999–December 2019. First, we screened all drug-related IBD studies (including anti-TNFα, vedolizumab, ustekinumab and tofacitinib). The following Medical Subject Heading (MeSH) terms alone or matched with the Boolean operators ‘AND’ or ‘OR’ were used: ‘Inflammatory Bowel Diseases’, ‘IBD’, ‘Colitis, Ulcerative’, ‘Crohn Disease’, ‘tumor necrosis factor alpha’, ‘certolizumab pegol’, ‘adalimumab’, ‘infliximab’, ‘golimumab’, ‘vedolizumab’, ‘ustekinumab’ and ‘tofacitinib’. Secondly, we screened all EIM-related IBD studies. The following MeSH terms alone or matched with the Boolean operators ‘AND’ or ‘OR’ were used: ‘Inflammatory Bowel Diseases’, ‘IBD’, ‘Colitis, Ulcerative’, ‘Crohn Disease’, ‘extraintestinal manifestations’, ‘extra intestinal manifestation’, ‘erythema nodosum’, ‘pyoderma gangrenosum’, ‘primary sclerosing cholangitis’, ‘uveitis’, ‘episcleritis’, ‘ocular manifestation’, ‘ocular’, ‘spondyloarthropathy’, ‘spondyloarthritis’, ‘arthritis’, ‘arthralgia’. Two reviewers (LG and FD) independently scrutinised titles and abstracts to identify eligible studies. Subsequently, full-text articles were examined for inclusion. Any disagreements between investigators were resolved through collegial discussion. Finally, we accurately checked the reference lists of the included studies for any additional relevant work.

Selection criteria

The search strategies included common search strings for disease-related and drug-related terms. A two-step process was applied: title and abstract assessment (step one), followed by full text analysis of relevant articles (step two). The inclusion criteria were: (a) confirmed diagnosis of IBD; (b) confirmed diagnosis of ≥1 EIM; (c) studies reporting EIM assessment; (d) interventional or non-interventional study; (e) study published in English. All editorials, notes, comments, letters or review articles were excluded.

Data extraction

After screening from eligible studies, two reviewers (LG and FDA) extracted data regarding the type of EIM, study characteristics, study design, IBD phenotype, time of follow-up, tools and investigations for EIM assessment, treatment for IBD, treatment for EIM and evaluation of treatment efficacy for EIMs. Finally, a multidisciplinary panel of experts (gastroenterologist, ophthalmologist, dermatologist and rheumatologist) was involved in the analysis of the collected data and in the subsequent definition of recommendations for the assessment of EIMs.

Quality of studies

Quality of non-randomised trials was measured through the Newcastle-Ottawa Scale (NOS)¹⁶ score, while in randomised clinical trials the Jadad score¹⁷ was used. NOS is made up of eight items: ‘representativeness of the exposed cohorts’, ‘selection of the non-exposed cohort’, ‘ascertainment of exposure’, ‘demonstration that the outcome of interest was not present at start of study’, ‘comparability of cohorts on the basis of the design or analysis’, ‘assessment of the outcome’, ‘follow-up is long enough for outcomes to occur’ and ‘adequacy of the follow-up’. The NOS score ranges from 0–9 and is based on the assignment of two stars for comparability of cohorts, while the other items can be assigned only one star. A NOS score ≥6 indicated high quality studies, while studies with NOS values ​​of 1–3 and 4–5 were defined as low and moderate quality studies respectively. The Jadad score ranges from 0–5 and assigns one point for each of the following parameters: randomised study, appropriate randomization, double blind study, appropriate double-blind study, description of withdrawals/dropouts. A study was of a good quality if the Jadad score was ≥3. Two authors (LG and FDA) independently graded the studies and disagreements were discussed with the other authors until their resolution.

Results

A flow diagram reporting the study selection process is provided in Figure 1. A total of 2109 eligible studies were identified through the literature search on Pubmed (1000), EMBASE (748) and Web of Science (361). After removal of duplicates, 1257 articles were excluded after screening of titles and abstracts. An additional 343 studies were excluded after full-text analysis of the manuscripts since they did not meet the inclusion criteria. Eighteen works were identified through the reference lists of the studies. Finally, 45 studies⁴,^18–61 met the inclusion criteria and were evaluated. There were 15 interventional studies (14 open label (OL) trials^20–27,³⁰,³²,³³,³⁶,⁴⁹,⁵⁰ and one randomised controlled trial (RCT)⁴²) and 30 non interventional studies (12 retrospective cohort studies,²⁸,³¹,⁴⁵,⁴⁸,⁵²,⁵⁴,^56–61 10 case reports,¹⁸,¹⁹,³⁵,^38–41,⁴³,⁴⁴,⁴⁷ four prospective cohort studies⁴,²⁹,³⁷,⁵⁵ and four case control studies).³⁴,⁴⁶,⁵¹,⁵³ A Jadad score = 4 was assigned to the RCT.⁴² Among the studies evaluated through the NOS scale there were four low (8.9%),¹⁹,³⁵,³⁹,⁴³ 15 moderate (33.3%)¹⁸,²¹,²³,²⁵,³⁰,³³,³⁸,⁴⁰,⁴¹,⁴⁴,⁴⁵,^47–50 and 25 high quality studies (55.6%)⁴,²⁰,²²,²⁴,^26–29,³¹,³²,³⁴,³⁶,³⁷,⁴⁶,^51–61 (Supplementary Material Tables 1 and 2).

Figure 1.

Flow diagram reporting the study selection process.

Table 1.

Studies on rheumatological manifestations.

Study	Type of EIM	IBD patients	Study design	Follow-up	Tools/method for diagnosis	Treatment for EIM	Efficacy assessment of EIM treatment
Van den Bosch et al., 200018	Ankylosing spondylitis (n = 2)	CD (n = 2)	Case report	No data	Clinical evaluation + ESSG criteria (rheumatologist)	IFX 5 mg/kg (n = 2), 1 to 2 infusions	Clinical assessment + CRP; 2 complete response
	Peripheral spondyloarthritis (n = 3)	CD (n = 3)			Clinical evaluation + ESSG criteria (rheumatologist)	IFX 5 mg/kg (n = 3), 2 infusions	Clinical assessment + CRP; 3 complete response
Ellman et al., 200119	Sacroiliac joint (n = 1), lower back pain (n = 2)	CD (n = 2) UC (n = 1)	Case report	No data	Clinical evaluation + X-ray (rheumatologist)	IFX 5 mg/kg (n = 4), ≥4 infusions	Clinical assessment; 2 complete and 2 partial response
	Peripheral spondyloarthritis(n = 4)	CD (n = 3) UC (n = 1)
Herfarth et al., 200220	Arthralgia (n = 43)	CD (n = 43)	OL prospective trial, multicentre	12 weeks	Clinical evaluation: painful, no swollen or joint effusion (rheumatologist)	IFX 5 mg/kg (n = 43)	Clinical assessment: 4-point scale (severe, moderate, mild or none); 61% improved by at least 1 point
Reinisch et al., 200321	Peripheral spondyloarthritis (n = 13)	IBD (n = 13)	OL prospective trial, single centre	20 days	Clinical evaluation: swollen or joint effusion + HAQ (gastroenterologist)	Rofecoxib 12.5 to 25 mg/day (n = 11)	Clinical assessment + HAQ; 36% of response
	Arthralgia (n = 19)	IBD (n = 19)			Clinical evaluation: joint pain + HAQ (gastroenterologist)	Rofecoxib 12.5 to 25 mg/day (n = 18)	Clinical assessment + HAQ; 50% of response
Generini et al., 200422	Axial pain (n = 36)	CD (n = 36)	OL prospective trial, single centre	3–18 months	Clinical evaluation + ESSG criteria (rheumatologist)	IFX 3 or 5 mg/kg (n = 24)	Clinical assessment, BASDAI, VAS; Significant improvement of BASDAI and VAS
Kaufman et al., 200423	Ankylosing spondylitis (n = 11), sacroiliitis (n = 3)	CD (n = 11)	OL prospective trial, single centre	No data	Clinical evaluation + X-ray (gastroenterologist or rheumatologist)	IFX 5 mg/kg (n = 11)	Clinical assessment; 7 complete or partial response
	Peripheral spondyloarthritis (n = 3)	CD (n = 3)			Clinical evaluation (gastroenterologist or rheumatologist)	IFX 5 mg/kg (n = 3)	Clinical assessment; 1 complete and 2 no response
	Arthralgia (n = 11)	CD (n = 11)			Clinical evaluation (gastroenterologist or rheumatologist)	IFX 5 mg/kg (n = 11)	Clinical assessment; 7 complete or partial response
Rispo et al., 200524	Sacroiliitis (n = 5), ankylosing spondylitis (n = 5)	CD (n = 10)	OL prospective trial, single centre	22 months	Clinical evaluation + ESSG criteria + bone scintigraphy and X-ray (New York criteria, BASRI) (rheumatologist)	IFX 5 mg/kg (10)	Clinical assessment, ASAS20/ASAS40, BASDAI; 80% ASAS20 and 60% ASAS40
	Arthralgia (n = 6)	CD (n = 6)			Clinical evaluation (rheumatologist)	IFX 5 mg/kg (n = 6)	Clinical assessment; 6 complete response
Conti et al., 200525	Buttock pain (n = 11), spinal pain (n = 9), enthesopathy (n = 8)	CD (n = 3) UC (n = 4) IBD-U (n = 12)	OL prospective trial, single centre	No data	Clinical evaluation + ESSG criteria + X-ray and MRI (rheumatologist)	No data	No data
Vavricka et al., 20114	Ankylosing spondylitis (n = 39)	CD (n = 33) UC (n = 6)	Prospective cohort	34 months	Clinical exam + questionnaire at inclusion + X-ray (rheumatologist)	No exhaustive data	No data
	Peripheral spondyloarthritis (n = 272)	CD (n = 193) UC (n = 79)			Clinical exam + questionnaire at inclusion: painful/ swollen/redness joints (rheumatologist)	No exhaustive data	No data
Löfberg et al., 201226	Sacroiliitis (n = 34), ankylosing spondylitis (n = 16)	CD (n = 50)	OL prospective trial, multicentre	20 weeks	Assess by investigator at baseline	ADA 40 mg (n = 50)	Assessed by investigator at each visit and Week 20; 33 cases at Week 20
	Peripheral spondyloarthritis (n = 82)	CD (n = 82)			Assessed by investigator at baseline	ADA 40 mg (n = 82)	Assessed by investigator at each visit and Week 20; 20 cases at Week 20
	Arthralgia (n = 445)	CD (n = 445)			Assess by investigator at baseline	ADA 40 mg (n = 445)	Assessed by investigator at each visit and Week 20; 252 cases at Week 20
Barreiro-de-Acosta et al., 201227	Sacroiliitis (n = 1), ankylosing spondylitis (n = 7)	CD (n = 8)	OL prospective trial, single centre	6 months	Clinical evaluation + radiology (rheumatologist)	ADA 40 mg (n = 8)	Clinical assessment (number, size, location); 6 complete or partial response, 2 no response
	Peripheral spondyloarthritis (n = 31)	CD (n = 31)			Clinical evaluation: swollen or joint effusion (rheumatologist)	ADA 40 mg (n = 31)	Clinical assessment; 19 complete or partial and 12 no response
Karmiris et al., 201628	Sacroiliitis (n = 89), ankylosing spondylitis (n = 39)	CD (n = 104) UC (n = 21)	Retrospective multicentre study	No data	Clinical evaluation + X-ray or MRI changes (rheumatologist)	No data	No data
	Peripheral spondyloarthritis (n = 221)	CD (n = 155) UC (n = 66)			Painful/swollen joints at clinical evaluation (rheumatologist)	No data	No data
	Arthralgia (n = 311)	CD (n = 218) UC (n = 93)			Self-reporting + motility restriction + clinical evaluation (rheumatologist)	No data	No data
Landi et al., 201629	Ankylosing spondylitis (n = 45)	IBD (n = 45)	Observational cross-sectional multicentre study	No data	Clinical evaluation + ESSG criteria + New York modified criteria + BASDAI/BASRI/ BASFI/BASMI (rheumatologist)	No data	No data
Luchetti et al., 201730	Ankylosing spondylitis (n = 16), other (n = 13)	CD (n = 16) UC (n = 13)	OL prospective trial, multicentre	12 months	Clinical evaluation + BASDAI/ ASDAS/ BASFI/BASMI + ASAS criteria + MRI/X-ray (New York criteria) (rheumatologist)	ADA 40 mg (n = 20)	Clinical evaluation + BASDAI/ASDAS/BASFI/BASMI
Vavricka et al., 201731	Axial arthropathy (n = 45)	CD (n = 35) UC (n = 9) IBD-U (n = 1)	Retrospective multicentre study	4 years	Clinical evaluation (rheumatologist)	IFX, ADA or CZP (n = 34)	Clinical assessment; 59.1 to 61.8% response rate
	Peripheral spondyloarthritis (n = 160)	CD (n = 131) UC (n = 24) IBD-U (n = 5)			Clinical evaluation (rheumatologist)	IFX, ADA or CZP (n = 158)	Clinical assessment; 73.4 to 81.2% response rate
Tadbiri et al., 201732	Axial arthropathy (n = 12)	IBD (n = 47)	OL prospective trial, multicentre	54 weeks	Clinical evaluation at baseline (gastroenterologist)	VDZ (n = 47)	Clinical assessment; 21 complete and 10 partial response
	Peripheral spondyloarthritis (n = 41)
Orlando et al., 201733	Axial arthropathy (n = 2)	IBD (n = 14)	OL prospective trial, single centre	2.6 months (median)	Clinical evaluation	VDZ (n = 14)	Clinical assessment; 6 complete or partial response
	Peripheral spondyloarthritis (n = 14)
Yamamoto-Furusho et al., 201834	Sacroiliitis (n = 7), ankylosing spondylitis (n = 8), both (n = 1)	UC (n = 16)	Case control single centre study	No data	Clinical evaluation (rheumatologist)	No data	No data
	Peripheral spondyloarthritis (n = 9)	UC (n = 9)			Clinical evaluation (rheumatologist)	No data	No data
	Arthralgia (n = 138)	UC (n = 138)			Clinical evaluation (rheumatologist)	No data	No data
Fleisher et al., 201835	Sacroiliitis + ankylosing spondylitis (n = 1)	CD (n = 1)	Case report	5 months 4–24 months	Clinical evaluation	VDZ (n = 1)	Clinical assessment; 1 complete response
	Peripheral spondyloarthritis (n = 3)	CD (n = 1) UC (n = 2)			Clinical evaluation	VDZ (n = 3)	Clinical assessment; 3 complete response but 2 recurrence
Macaluso et al., 201836	Axial arthropathy (n = 15)	IBD (n = 43)	OL prospective trial, multicentre	22 weeks	Clinical evaluation + ASAS criteria (rheumatologist)	VDZ (n = 43)	Clinical assessment; 17 complete or partial response
	Peripheral spondyloarthritis (n = 39)
Feagan et al., 201937	Peripheral spondyloarthritis (n = 793)	CD (n = 708) UC (n = 85)	Post hoc analyses of RCT	No data	Screening at baseline	VDZ (n = 590)	Assessed by investigator at each visit
	Arthralgia (n = 793)

ADA: adalimumab; ASAS20: 20% improvement in the ASAS; ASAS40: 40% improvement in the ASAS; ASDAS: ankylosing spondylitis disease activity score; BASDAI: Bath Ankylosing Spondylitis Disease Activity Index; BASFI: Bath Ankylosing Spondylitis Functional Index; BASMI: Bath Ankylosing Spondylitis Metrology Index; BASRI: Bath Ankylosing Spondylitis Radiology Index; CD: Crohn's disease; CRP: C-reactive protein; CZP: certolizumab pegol; EIM: extra-intestinal manifestation; ESSG: European Spondylarthropathy Study Group; HAQ: health assessment questionnaire; IBD: inflammatory bowel disease; IBD-U: inflammatory bowel disease unclassified; IFX: infliximab; MRI: magnetic resonance imaging; OL: open label; RCT: randomised controlled trial; SPA: spondyloarthropathy; UC: ulcerative colitis; VAS: visual analog scale; VDZ: vedolizumab.

Table 2.

Studies on dermatological manifestations.

Study	Type of EIM	IBD patients	Study design	Follow-up	Tools/method for diagnosis	Treatment for EIM	Efficacy assessment of EIM treatment
Tan et al., 200138	Pyoderma gangrenosum (n = 2)	CD (n = 2)	Case report	3–12 months	Clinical evaluation (dermatologist)	IFX 5 mg/kg (n = 2)	Clinical assessment (number, size, location); 1 complete and 1 partial response
Ljung et al., 200239	Pyoderma gangrenosum (n = 8)	CD (n = 8)	Case report	No data	Clinical evaluation (dermatologist)	IFX 5 mg/kg (n = 8)	Clinical assessment (number, size, location); 3 complete, 3 partial (>50% reduction symptoms) and 2 no response
Regueiro et al., 200340	Pyoderma gangrenosum (n = 13)	CD (n = 12) UC (n = 1)	Case report	1 day–4 years	Clinical evaluation (dermatologist)	IFX 5 mg/kg (n = 13), 1 to 24 infusions+ CS (n = 9), AZA/6-MP (n = 11)	Clinical assessment (number, size, location); 13 complete response, 3 patients treated solely with IFX for PG
Sapienza et al., 200441	Pyoderma gangrenosum (n = 4)	CD (n = 4)	Case report	9–20 months	Clinical evaluation (dermatologist)	IFX 5 mg/kg (n = 4), 1 infusion (n = 3), 3 infusions (n = 1)	Clinical assessment (number, size, location); 4 complete response
Kaufman et al., 200423	Pyoderma gangrenosum (n = 4)	CD (n = 3) UC (n = 1)	OL prospective trial, single centre	No data	Clinical evaluation (gastroenterologist or rheumatologist)	IFX 5 mg/kg (n = 4), 2 to 4 infusions	Clinical assessment (number, size, location); 1 complete and 3 partial response
Rispo et al., 200524	Pyoderma gangrenosum (n = 1)	CD (n = 1)	OL prospective trial, single centre	22 months	Clinical evaluation + skin biopsy (dermatologist)	IFX 5 mg/kg	Clinical assessment (number, size, location); Complete response
	Erythema nodosum (n = 2)	CD (n = 2)			Clinical evaluation + skin biopsy (dermatologist)	IFX 5 mg/kg (n = 2)	Clinical assessment (number, size, location); 2 complete response
Brooklyn et al., 200642	Pyoderma gangrenosum (n = 30)	CD (n = 13) UC (n = 6) No IBD (n = 11)	Multicentre RCT	6 weeks	Clinical evaluation + skin biopsy if clinical doubt (dermatologist)	IFX 5 mg/kg (n = 29), 1 infusion	Clinical assessment (number, size, location); 6 complete, 14 partial and 9 no response
Baglieri et al., 201043	Pyoderma gangrenosum (n = 1)	UC (n = 1)	Case report	14 weeks	Clinical evaluation + skin biopsy (dermatologist)	IFX 5 mg/kg, 4 infusions + AZA 50 mg/day	Clinical assessment (number, size, location); Nearly complete response
Vavricka et al., 20114	Pyoderma gangrenosum (n = 17)	CD (n = 9) UC (n = 8)	Prospective cohort	34 months	Clinical exam + questionnaire at inclusion (dermatologist)	No exhaustive data	No data
	Erythema nodosum (n = 48)	CD (n = 36) UC(n = 12)			Clinical exam + questionnaire at inclusion (dermatologist)	No exhaustive data	No data
Hayashi et al., 201244	Pyoderma gangrenosum (n = 1)	IBD-U(n = 1)	Case report	22 weeks	Clinical evaluation + skin biopsy (dermatologist)	IFX 5 mg/kg, 5 infusions+ AZA 50 mg/day	Clinical assessment (number, size, location); Complete response
Löfberg et al., 201226	Pyoderma gangrenosum (n = 4)	CD (n = 4)	OL prospective trial, multicentre	20 weeks	Assess by investigator at baseline	ADA 40 mg (n = 4)	Assessed by investigator at each visit and Week 20; 2 cases at Week 20
	Erythema nodosum (n = 23)	CD (n = 23)			Assess by investigator at baseline	ADA 40 mg (n = 23)	Assessed by investigator at each visit and Week 20; 4 cases at Week 20
Barreiro-de-Acosta et al., 201227	Pyoderma gangrenosum (n = 2)	CD (n = 2)	OL prospective trial, single centre	6 months	Clinical evaluation (dermatologist)	ADA 40 mg (n = 2)	Clinical assessment (number, size, location); 2 complete or partial response
Argüelles-Arias et al., 201345	Pyoderma gangrenosum (n = 67)	CD (n = 41) UC (n = 25) IBD-U (n = 1)	Retrospective multicentre study	No data	Clinical evaluation (gastroenterologist)	CS (n = 51), IFX (24), ADA (n = 7), AZA (n = 6), TC (n = 3), Cyclosporin (n = 10)	Clinical assessment (number, size, location)
Weizman et al., 201446	Pyoderma gangrenosum (n = 92)	CD (n = 57) UC (n = 35)	Retrospective multicentre case control study	No data	Clinical evaluation + skin biopsy if clinical doubt (dermatologist)	No data	No data
	Erythema nodosum (n = 103)	CD (n = 69) UC (n = 34)			Clinical evaluation (dermatologist)	No data	No data
Karmiris et al., 201628	Pyoderma gangrenosum (n = 15)	CD (n = 7) UC (n = 8)	Retrospective multicentre study	No data	Clinical evaluation + skin biopsy if clinical doubt (dermatologist)	No data	No data
	Erythema nodosum (n = 99)	CD (n = 75) UC (n = 24)			Clinical evaluation + skin biopsy if clinical doubt (dermatologist)	No data	No data
Vavricka et al., 201731	Pyoderma gangrenosum (n = 11)	CD (n = 9) UC (n = 2)	Retrospective multicentre study	4 years	Clinical evaluation (dermatologist)	IFX, ADA or CZP (n = 11)	Clinical assessment (number, size, location); 54.5% response rate
	Erythema nodosum (n = 25)	CD (n = 18) UC (n = 7)			Clinical evaluation (dermatologist)	IFX, ADA or CZP (n = 10)	Clinical assessment (number, size, location); 80% response rate (n = 10)
Fleisher et al., 201835	Pyoderma gangrenosum (n = 2)	CD (n = 1) UC (n = 1)	Case report	12–20 monthsNo data	Clinical evaluation	VDZ (n = 2), 6 infusions (n = 1) + MTX (n = 1)	Clinical assessment (number, size, location); 1 complete and 1 complete response but recurrence
	Erythema nodosum (n = 1)	CD (n = 1)			Clinical evaluation	VDZ, 3 infusions	Clinical assessment (number, size, location); Complete response
Perricone et al., 201847	Pyoderma gangrenosum (n = 1)	UC (n = 1)	Case report	3 months	Clinical evaluation + skin biopsy (dermatologist)	IFX, 4 infusions	Clinical assessment (number, size, location); Complete response
Feagan et al., 201937	Pyoderma gangrenosum (n = 10)	CD (n = 9) UC (n = 1)	Post hoc analyses of RCT	No data	Screening at baseline	VDZ (n = 9)	Assessed by investigator at each visit
	Erythema nodosum (n = 57)	CD (n = 51) UC (n = 6)			Screening at baseline	VDZ (n = 41)	Assessed by investigator at each visit
De Risi-Pugliese et al., 201948	Pyoderma gangrenosum (n = 4)	CD (n = 4)	Retrospective multicentre study	12–47 months	Clinical evaluation + skin biopsy (dermatologist)	Ustekinumab (n = 4) + CS (n = 1), Topical TC (n = 1), dCS (n = 1)	Clinical assessment (number, size, location); 3 complete and 1 partial response (>50% reduction symptoms)

6-MP: 6-mercaptopurine; ADA: adalimumab; AZA: azathioprine; CD: Crohn's disease; CS: corticosteroid; CZP: certolizumab pegol; dCS: dermo-corticosteroid; EIM: extra-intestinal manifestation; IBD: inflammatory bowel disease; IBD-U: inflammatory bowel disease unclassified; IFX: infliximab; MTX: methotrexate; OL: open label; RCT: randomised controlled trial; TC: tacrolimus; UC: ulcerative colitis; VDZ: vedolizumab.

Rheumatological manifestations

Axial spondyloarthritis

Eighteen studies reported axial signs of spondyloarthritis. A total of 468 patients were included: 311 CD (76.1%), 70 UC (20.1%), 13 IBD-unclassified (IBD-U) (3.8%) and 74 unspecified patients. The diagnostic evaluation was performed by rheumatologists in 14 studies (78%).⁴,¹⁸,¹⁹,^22–25,^27–31,³⁴,³⁶ The European Spondylarthropathy Study Group (ESSG) or the Assessment of SpondyloArthritis international Society (ASAS) criteria were used in seven studies (39%), which include careful evaluation of symptoms and medical history, clinical examination, radiographic assessment and biochemical tests (human leukocyte antigen (HLA) B27, C-reactive protein (CRP)).¹⁸,²²,²⁴,²⁵,²⁹,³⁰,³⁶ The other studies did not include clinical evaluation details. X-ray or magnetic resonance imaging (MRI) were performed in nine cases (50%).⁴,¹⁹,^23–25,^27–30 Concerning treatment, infliximab was used in six studies (33%),¹⁸,¹⁹,^22–24,³¹ adalimumab in four (22%),²⁶,²⁷,³⁰,³¹ vedolizumab in four (22%)³²,³³,³⁵,³⁶ and certolizumab in one (5.5%).³¹ The response to treatment was investigated in 13 articles¹⁸,¹⁹,^22–24,²⁶,²⁷,^30–33,³⁵,³⁶ and the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) was used in three studies (16.7%),²²,²⁴,³⁰ while the Ankylosing Spondylitis Disease Activity Score (ASDAS) was used in one case (5.5%).³⁰ The axial spondyloarthritis showed a good response to anti-TNF agents. Interestingly, in a retrospective multicentre study by Vavricka et al.³¹ including 45 patients, the response rate to anti-TNF ranged from 59.1–61.8% (Table 1). Vedolizumab results varied depending on the study. In the work by Tadbiri et al.,³² 21 complete (44.7%) and 10 partial responses (21.3%) were detected in 47 patients with axial and/or peripheral spondyloarthritis. Orlando et al.³³ described six partial or complete responses (43%) in 14 patients with peripheral and/or axial spondyloarthritis. Finally, Fleisher et al.³⁵ reported one complete response in one treated patient while, in the study of Macaluso et al.,³⁶ 17 cases of partial or complete response were found in 43 patients (39.5%).

Peripheral spondyloarthritis

Fifteen studies reported signs of peripheral spondyloarthritis. A total of 1688 patients were studied: 1310 CD (77.6%), 266 UC (15.8%), five IBD-U (0.3%) and 107 unspecified patients (6.3%). Rheumatologists performed the diagnostic assessment in nine studies (60%).⁴,¹⁸,¹⁹,²³,²⁷,²⁸,³¹,³⁴,³⁶ ESSG or ASAS criteria were used in two studies (13.3%)¹⁸,³⁶ and the Health Assessment Questionnaire (HAQ) was used in one case (6.6%).²¹ Clinical examination consisted in the count of painful and swollen joints in four articles (26.7%).⁴,²¹,²⁷,²⁸ The other studies did not include the details of the clinical evaluation. Surprisingly, X-ray or MRI were performed in only one case.¹⁹ Concerning treatment, vedolizumab was used in five studies (33.3%),³²,³³,^35–37 infliximab in four (26.7%),¹⁸,¹⁹,²³,³¹ adalimumab in three (20%)²⁶,²⁷,³¹ and certolizumab in one (6.7%).³¹ The response to treatment was evaluated through clinical examination in 12 trials.¹⁸,¹⁹,²¹,²³,²⁶,²⁷,^31–33,^35–37 A good response to anti-TNF agents was achieved (Table 1). Indeed, the response rate ranged from 73.4–81.2% in the retrospective multicentre study by Vavricka et al.,³¹ which included 160 patients. Among vedolizumab-treated patients, Fleisher et al.³⁵ reported one complete response (33.3%) and two partial responses (66.6%), while Macaluso et al.³⁶ found partial or complete responses in 17/43 patients (39.5%).

Arthralgia

Eight studies assessed arthralgia. A total of 1766 patients were studied: 1431 CD (81%), 316 UC (17.9%) and 19 unspecified patients (1.1%). The diagnostic screening was carried out by rheumatologists in five studies (62.5%).²⁰,²³,²⁴,²⁸,³⁴ Clinical examination (evaluation of painful joints without swelling and effusions) was performed in three trials (37.5%).²⁰,²¹,²⁸ The other studies did not provide specific information on the diagnostic approach. Infliximab was the most-used drug (three studies, 37.5%),²⁰,²³,²⁴ while adalimumab and vedolizumab were administered in one study each (12.5%).²⁶,³⁷ The response to treatment was evaluated in six articles (75%) and improvement was assessed by clinical examination²⁰,²¹,²³,²⁴,²⁶,³⁷ (Table 1).

Dermatological manifestations

Pyoderma gangrenosum (PG)

PG was investigated in 20 studies, including a total of 289 patients: 186 CD (64.4%), 90 UC (31.1%), two IBD-U (0.7%) and 11 without IBD (3.8%). Dermatologists were involved in diagnosis in 15 studies (75%)⁴,²⁴,²⁷,²⁸,³¹,^38–44,^46–48 and clinical evaluation (e.g. number, size and location) was the most adopted method (18/20, 90%).⁴,²³,²⁴,²⁷,²⁸,³¹,³⁵,^38–48 A skin biopsy was performed in eight studies (40%).²⁴,²⁸,^42–44,^46–48 Infliximab was the most-used treatment (12 studies, 60%),²³,²⁴,³¹,^38–45,⁴⁷ followed by adalimumab (four studies, 20%),²⁶,²⁷,³¹,⁴⁵ vedolizumab (two studies, 10%),³⁵,³⁷ certolizumab (one study, 5%)³¹ and ustekinumab (one study, 5%).⁴⁸ The response to treatment was clinically evaluated in 17 cases,²³,²⁴,²⁶,²⁷,³¹,³⁵,^37–45,⁴⁷,⁴⁸ assessing number, size and location of PG. A complete response was defined as a total resolution of the skin lesion in 12 studies (60%),²³,²⁴,²⁷,³⁵,^38–42,⁴⁴,⁴⁷,⁴⁹ while the partial response was defined in two studies as an improvement >50% of the lesion.³⁹,⁴⁸ In 12 studies a high PG response rate to TNF inhibitors was found²³,²⁴,²⁷,³¹,^38–44,⁴⁷ with 32 complete responses and 26 partial responses (Table 2). In the randomised controlled trial by Brooklyn et al.,⁴² six complete responses, 14 partial responses and nine non-responses were found among 29 patients treated with infliximab. Vavricka et al.³¹ reported a response rate of 54.5%. De Risi-Pugliese et al.⁴⁸ showed good response to ustekinumab with three complete responses and one partial response. Fleisher et al.³⁵ found one complete response and one partial response in vedolizumab-treated patients.

Erythema nodosum (EN)

Eight studies investigated EN, enrolling a total of 358 patients: 275 CD (76.8%) and 83 UC (23.2%). The diagnosis was obtained by dermatologists in five studies (62.5%)⁴,²⁴,²⁸,³¹,⁴⁶ and clinical evaluation of number, size and location of EN was the most frequent approach (six studies, 75%).⁴,²⁴,²⁸,³¹,³⁵,⁴⁶ A skin biopsy was performed in two trials.²⁴,²⁸ EN was treated with infliximab (two studies, 25%),²⁴,³¹ adalimumab (two studies, 25%),²⁶,³¹ vedolizumab (two studies, 25%)³⁵,³⁷ and certolizumab (one study, 12.5%).³¹ The response to treatment was clinically evaluated (number, size and location of EN) in five trials.²⁴,²⁶,³¹,³⁵,³⁷ In the study by Vavricka et al.,³¹ 8/10 patients (80%) had a clinical response after therapy with anti-TNF agents, while in the experience of Rispo et al.,²⁴ 2/2 infliximab-treated patients (100%) achieved complete response (Table 2).

Ocular manifestations

Eleven studies reported ocular EIMs. Multiple symptoms were observed but uveitis and episcleritis were the two most common symptoms (88.5%). A total of 396 patients were involved: 270 CD (68.2%) and 126 UC (31.8%). Ophthalmologists were frequently implicated in the diagnosis (nine studies, 82%).⁴,²⁴,²⁷,²⁸,³¹,³⁵,^49–51 Slit lamp examination, visual acuity measurement and tonometry were used for EIM assessment in two studies.⁴⁹,⁵⁰ The three-mirror Goldmann lens, fluorescein angiography and Break-Up Time (BUT) test were used by Felekis et al.⁵⁰ The other studies did not provide specific information on diagnostic assessment. The most frequent treatments included adalimumab (three studies, 27.3%),²⁶,²⁷,³¹ infliximab (two studies, 18.2%),²⁴,³¹ vedolizumab (two studies, 18.2%)³⁵,³⁷ and certolizumab (one study, 9.1%).³¹ The response to treatment was clinically evaluated in eight articles.²⁴,²⁶,²⁷,³¹,³⁵,³⁷,⁴⁹,⁵⁰ Vavricka et al.³¹ reported a response rate ranging from 72–88.9% in 25 patients with uveitis treated with anti-TNF drugs (Table 3).

Table 3.

Studies on ocular manifestations.

Study	Type of EIM	IBD patients	Study design	Follow-up	Tools/method for diagnosis	Treatment for EIM	Efficacy assessment of EIM treatment
Rispo et al., 200524	Uveitis (n = 2) Episcleritis (n = 1)	CD (n = 3)	OL prospective trial, single centre	22 months	Clinical evaluation (ophthalmologist)	IFX 5 mg/kg	Clinical assessment; 3 complete response
Yilmaz et al., 200749	Conjunctivitis (n = 10) Blepharitis (n = 8) Uveitis (n = 6) Cataract (n = 6) Episcleritis (n = 4)	CD (n = 12) UC (22)	OL prospective trial, single centre	12 months	Slit lamp biomicroscopy, tonometry, indirect ophthalmoscopy, and visual acuity (ophthalmologist)	No data	Monthly clinical assessment
Felekis et al., 200950	Episcleritis (n = 2) Iridocyclitis (n = 3) Conjunctivitis (n = 1) Dry eye (n = 13) Choroiditis (n = 1) Optic neuritis (n = 1) Retinal vasculitis (n = 1)	CD (n = 23) UC (n = 37)	OL prospective trial, single centre	10 years (median)	Slit lamp, tonometry, visual acuity, 3-mirror Goldmann lens, Fluorescein angiography, BUT test (ophthalmologist)	CS (n = 1), CS drop (n = 2), NSAID drop (n = 2)	every 6 monthsclinical assessment
Vavricka et al., 20114	Uveitis (n = 50)	CD (n = 36) UC (n = 14)	Prospective cohort	34 months	Clinical exam + questionnaire at inclusion (ophthalmologist)	No exhaustive data	No data
Löfberg et al., 201226	Iritis (n = 7) Uveitis (n = 3)	CD (n = 10)	OL prospective trial, multicentre	20 weeks	Assess by investigator at baseline	ADA 40 mg (n = 10)	Assessed by investigator at each visit and Week 20; 5 cases at Week 20
Barreiro-de-Acosta et al., 201227	Uveitis (n = 1)	CD (n = 1)	OL prospective trial, single centre	6 months	Clinical evaluation (ophthalmologist)	ADA 40 mg (n = 1)	Clinical assessment; 1 complete or partial response
Taleban et al., 201651	Episcleritis/scleritis/uveitis (n = 124)	CD (n = 90) UC/IBD-U (n = 34)	Retrospective multicentre case control study	No data	Clinical evaluation (ophthalmologist)	No data	No data
Karmiris et al., 201628	Episcleritis (n = 16) Uveitis (n = 38)	CD (n = 44) UC (n = 10)	Retrospective multicentre study	No data	Clinical evaluation (ophthalmologist)	No data	No data
Vavricka et al., 201731	Uveitis (n = 33)	CD (n = 29) UC (n = 4)	Retrospective multicentre study	4 years	Clinical evaluation (ophthalmologist)	IFX, ADA or CZP (n = 25)	Clinical assessment; 72 to 88.9% response rate
Fleisher et al., 201835	Uveitis (n = 1)	UC (n = 1)	Case report	1 year	Clinical evaluation (ophthalmologist)	VDZ (n = 1)+ CS (n = 1)	Clinical assessment; 1 complete response
Feagan et al., 201937	Iritis/uveitis (n = 26)	CD (n = 22) UC (n = 4)	Post hoc analyses of RCT	No data	Screening at baseline	VDZ (n = 21)	Assessed by investigator at each visit

ADA: adalimumab; CD: Crohn's disease; CS: corticosteroid; CZP: certolizumab pegol; EIM: extra-intestinal manifestation; IBD: inflammatory bowel disease; IBD-U: inflammatory bowel disease unclassified; IFX: infliximab; NSAID: non-steroidal anti-inflammatory drug; OL: open label; RCT: randomised controlled trial; UC: ulcerative colitis; VDZ: vedolizumab.

Primary sclerosing cholangitis (PSC)

Thirteen studies included screening for PSC. A total of 3844 patients were recruited: 706 CD (18.3%), 3024 UC (78.7%) and 114 IBD-U (3%). Hepato-gastroenterologists were involved in the diagnostic approach of all studies (100%).⁴,²⁸,³¹,^52–61 The diagnosis was obtained through clinical examination (three studies, 23%) which included careful examination of symptoms and medical history,⁴,³¹,⁶⁰ elevation of alkaline phosphatase (ALP) (11 studies, 84.6%)²⁸,^52–61 and bile duct imaging (MRI or endoscopic retrograde cholangio-pancreatography (ERCP), 11 studies, 84.6%).²⁸,^52–61 A liver biopsy was performed in 10 studies (77%).²⁸,^52–59,⁶¹ The biological drugs used for treatment were infliximab (two studies, 15.4%),³¹,⁵⁷ adalimumab (two studies, 15.4%),³¹,⁵⁷ vedolizumab (two studies, 15.4%)⁶⁰,⁶¹ and certolizumab (1, 7.7%).³¹ Ursodeoxycholic acid (UDCA) was also administered in six trials (46.1%).⁵²,^54–57,⁶¹ The response to treatment was evaluated in only four articles through clinical assessment and biological decrease of ALP levels.³¹,⁵⁷,⁶⁰,⁶¹ In all studies, biologics had no impact on the PSC course (Table 4).

Table 4.

Studies on primary sclerosing cholangitis.

Study	IBD patients	Study design	Follow-up	Tools/method for diagnosis	Treatment for EIM	Efficacy assessment of EIM treatment
Pardi et al., 200352	UC (n = 52)	Retrospective study of single centre RCT	166–189 person-years	ALP level + ERCP + biopsy	UDCA (n = 29)	No data
Moayyeri et al., 200553	UC (n = 19)	Case control single centre study	12.2 years (median)	ALP level + ERCP + biopsy	No data	No data
Claessen et al., 200954	CD (n = 3) C (n = 24)	Retrospective multicentre study	2.3–2.7 years (median)	ALP level + ERCP or biopsy	UDCA (n = 15)	No data
Vavricka et al., 20114	CD (n = 4) UC (n = 13)	Prospective cohort	34 months	Clinical exam + questionnaire at inclusion	No exhaustive data	No data
Eaton et al., 201155	UC (n = 25)	Nested cohort study of multicentre RCT	4.0 years (median)	ALP level + MRI or ERCP + biopsy	UDCA 28–30 mg/kg/day	No data
Navaneethan et al., 201256	CD (n = 41)	Retrospective multicentre study	17.3 years (median)	ALP level + MRI or ERCP + biopsy if clinical doubt	5-ASA (n = 30), AZA (n = 3), MTX (n = 1), CS (n = 17) + UDCA (n = 36)	No data
Karmiris et al., 201628	CD (n = 5) UC (n = 3)	Retrospective multicentre study	No data	ALP level + MRI and/or biopsy	No data	No data
Franceschet et al., 201657	CD (n = 10) UC (n = 38) IBD-U (n = 1)	Retrospective single centre study	103.8 months (median)	ALP and γGT level + MRI or ERCP + biopsy (for small duct PSC)	5-ASA (n = 42), AZA (n = 5), CS (n = 19), ADA (n = 2), IFX (n = 1) + UDCA (n = 49)	ALP and γGT level; No significant change of ALP and γGT level
Weismüller et al., 201758	CD (n = 595) UC (n = 2761), IBD-U (113)	Retrospective multicentre study	No data	ALP level + MRI or ERCP + biopsy	No data	No data
Vavricka et al., 201731	CD (n = 6) UC (n = 4)	Retrospective multicentre study	4 years	Diagnosis by gastroenterologist	IFX, ADA or CZP	Clinical and biological assessment
Park et al., 201859	UC (n = 18)	Retrospective multicentre study	No data	ALP level + MRI or ERCP + biopsy	No data	No data
Christensen et al., 201860	CD (n = 16) UC (n = 18)	Retrospective multicentre study	9 months (median)	Clinical + ALP level + MRI or ERCP	VDZ (n = 34)	ALP level; No significant change of ALP level
Caron et al., 201961	CD (n = 26) UC (n = 49)	Retrospective multicentre study	6–54 weeks	ALP level + MRI + biopsy	VDZ (n = 75) + CS (n = 38), IS (n = 23), UDCA (n = 65)	ALP level; No significant change of ALP level

5-ASA: 5-aminosalicylic acid; ADA: adalimumab; ALP: alkaline phosphatase; AZA: azathioprine; CD: Crohn's disease; CS: corticosteroid; CZP: certolizumab pegol; EIM: extra-intestinal manifestation; ERCP: endoscopic retrograde cholangio-pancreatography; γGT: gamma-glutamyltransferase; IBD: inflammatory bowel disease; IBD-U: inflammatory bowel disease unclassified; IFX: infliximab; IS: immunosuppressant; MRI: magnetic resonance imaging; MTX: methotrexate; OL: open label; RCT: randomised controlled trial; UC: ulcerative colitis; UDCA: ursodeoxycholic acid; VDZ: vedolizumab.

Recommendations for the assessment of EIMs

Proposed recommendations for the assessment of EIMs in upcoming IBD trials are reported in Figure 2. Based on our systematic literature review, recourse to specialist expertise is essential for both diagnosis and follow-up of each EIM. In patients with articular symptoms, the diagnostic approach should include clinical evaluation, biochemical tests (CRP and HLA B-27), and imaging (MRI/X-ray), while the response to therapy should be evaluated through clinical data and scores commonly used in clinical practice (BASDAI and ASDAS for axial spondyloarthritis,²²,²⁴,³⁰,⁶²,⁶³ BASDAI, number of painful or swollen joints and HAQ for peripheral spondyloarthritis, Disease Activity in PSoriatic Arthritis (DAPSA) score for psoriatic arthritis).²¹,⁶²,⁶³ In cases of dermatologic manifestations the diagnosis should be clinical and a skin biopsy should be performed only in doubtful diagnoses.²⁴,²⁸,^42–44,^46–48 Follow-up should be based on clinical assessment and in patients with PG, response to therapy could be defined as a reduction >50% of symptoms.³⁹,⁴⁸ Ocular manifestations should be clinically investigated with the support of slit lamps and visual acuity tests. Similarly, ophthalmological follow-up should include clinical assessment, slit lamp and visual acuity tests.⁴⁹,⁵⁰ Further examinations (e.g. tonometry, fluorescein angiography and BUT test) should be adopted according to the symptoms reported by the patient.⁵⁰ Finally, clinical, biochemical (ALP level), and bile duct imaging by MRI (currently preferred imagery) or ERCP should guide PSC assessment. Liver biopsy should not be systematically performed, but it should be reserved for doubtful situations, small duct PSC or overlap syndrome.⁴,²⁸,^52–59,⁶¹ In addition, clinical examination, ALP levels and MRI or ERCP should be used for the hepatological follow-up of PSC patients.³¹,⁵⁷,⁶⁰,⁶¹

Figure 2.

Proposed recommendations.

ALP: alkaline phosphatase; ASAS: Assessment of SpondyloArthritis international Society; BASDAI: Bath Ankylosing Spondylitis Disease Activity Index; CRP: C-reactive protein; DAPSA: Disease Activity in PSoriatic Arthritis; EIM: extraintestinal manifestation; ERCP: endoscopic retrograde cholangio-pancreatography; ESSG: European Spondylarthropathy Study Group; HAQ: Health Assessment Questionnaire; MRI: magnetic resonance imaging; PSC: primary sclerosing cholangitis.

Discussion

We investigated the diagnostic assessment of rheumatological, dermatological, ophthalmological and gastroenterological manifestations in approximately 8000 IBD patients. EIMs were diagnosed by dedicated specialists in percentages ranging from 60% for peripheral spondyloarthritis to 100% for PSC. The diagnostic approach was mainly based on clinical evaluation⁴,^18–25,^27–51,⁶⁰ even if a considerable percentage of patients underwent biochemical and imaging tests.⁴,¹⁹,^23–25,^27–30,^42–44,^46–50,^52–61 TNF inhibitors were the main biologic drugs for the treatment of EIMs and the response to therapy was primarily assessed through clinical data. Importantly, anti-TNF agents were confirmed to be effective in patients with EIMs, achieving a clinical response in over 50% of cases (except for PSC).^18–20,^22–24,²⁶,²⁷,³⁰,³¹,^38–45,⁴⁷,⁵⁷ These data are in line with a previous study, demonstrating the efficacy of anti-TNFs for musculoskeletal, cutaneous and ocular manifestations of IBD.⁸ Vedolizumab efficacy was evaluated in only eight articles (mainly case reports), preventing definitive conclusions from being drawn. However, in a recent systematic review no strong evidence supported the efficacy of vedolizumab for pre-existing EIMs, although its use could play a role in reducing the occurrence of new events.¹¹ Only one study reported the efficacy of ustekinumab⁴⁸ and no article on tofacitinib met our inclusion criteria, suggesting the lack of data in this field and the need for further studies. In 2016, a European Crohn’s and Colitis Organization consensus⁶⁴ provided guidelines for diagnosis of IBD patients with EIMs, but no mention was made regarding follow-up and evaluation of the response to therapy. Based on the results of our research, we have proposed recommendations for evaluation and follow-up of each EIM. In the absence of clear guidelines, the approach to be used should be based on the most frequent management and all patients with EIMs should be referred to the dedicated specialist.

Regarding strengths, to our knowledge, this is the first systematic review specifically designed to evaluate the diagnostic approach of EIMs in IBD patients. Secondly, the high number of studies and patients included are another relevant strength of our work. Thirdly, we adopted strict inclusion criteria, and study selection and data extraction were performed through the combined work of two authors who operated independently, reducing the risk of errors. Fourthly, more than 50% of the included studies were high quality studies according to globally accepted assessment scores such as the NOS and Jadad scores. On the other hand, several limitations must also be reported. First of all, there was a lack of standardised approach for EIM evaluation. In fact, the clinical characteristics leading to the diagnosis were rarely specified and the diagnostic accuracy could be questionable depending on the experience of the individual specialist. In addition, the analysed studies were heterogeneous as they included a wide variety of designs (e.g. randomised and open label trials, prospective and retrospective cohort studies, and case reports) and there was no commonly accepted definition to assess the responses to different therapies, making data interpretation and comparison difficult. Finally, details of the clinical assessment of EIMs were not reported in the majority of the selected studies, probably due to their retrospective design preventing them from specifying if the evaluation of EIMs was performed before or after the IBD diagnosis and defining if this assessment had an impact on disease outcomes.

In conclusion, EIMs are challenging conditions and require the collaboration of dedicated specialists, from the dermatologist to the ophthalmologist, from the rheumatologist to the hepatologist according to the type of symptom, to guarantee the best management of these patients. In this context, it is essential that international expert organizations provide clear statements to homogenise and standardise the evaluation of IBD patients with EIMs. This consensus meeting will also clarify the role of imaging, such as ultrasound, in assessing EIMs.