Systematic Reviews
Copyright ©The Author(s) 2019.
World J Gastroenterol. Aug 21, 2019; 25(31): 4534-4554
Published online Aug 21, 2019. doi: 10.3748/wjg.v25.i31.4534
Table 1 Studies showing the impact of small bowel capsule endoscopy on disease reclassification and subsequent patient management during follow-up of patients suspected or diagnosed with inflammatory bowel disease
Ref.Study designSample sizePatient populationCompared modalityCE lesions considered diagnostic for CDPositive SB findingsImpact in patient management1
Reclassification of Crohn’s disease location
Chong et al[28], 2005Prospective, blinded43Group 1: Known CD (n = 22) Group 2: Suspected CD2 (n = 21)Push enteroscopy and enteroclysis≥ 1 erosion/ ulcerGroup 1: 17/22 (jejunum, n = 7) vs 3/22 at push enteroscopy, P < 0.001 and 4/21 at enteroclysis, P < 0.001 Group 2: 4/21 (jejunum, n = 2), no statistically significant difference vs other modalities30/43 (70%) Group 1: 16 (73%) Group 2: 14 (67%)
De Bona et al[38], 2006Prospective38Suspected CD2 Group 1: Ongoing symptoms (n = 12) Group 2: Ongoing symptoms and inflammatory biomarkers3 (n = 26)NADiagnostic if > 3 erosions/ ulcerations Suspicious if ≤ 3 and/or nodular patternDiagnostic: 13/38 (34.2%) (jejunum, n = 5) Suspicious: 2/38 (5.3%) Group 1: 1/12 (8.3%) Group 2: 14/26 (46.2%) P = 0.02215/38 (39.5%) i.e., 100% of patients with positive CE findings
Efthymiou et al[29], 2009Prospective, blinded55Group 1: Known CD (n = 29) Group 2: Suspected CD2 (n = 26)EnteroclysisDiffuse erythema, erosions, > 3 aphthoid ulcers, ulcers of different shape and stricturesGroup 1: 20/29 (jejunum, n = 8) vs 11/27 at enteroclysis4, incremental diagnostic yield= 33.4% (P = 0.035) Group 2: 16/26 (jejunum, n = 6), vs 6/20 at enteroclysis5, incremental diagnostic yield = 35.0% (P = 0.039)-
Tukey et al[78], 2009Retrospective105Suspected CD2NAAny ulcers39/105 (37%) Prevalence rate of CD diagnosis after a 12-mo follow-up = 13% Se 77%, Sp 89%, PPV 50%, NPV 96%-
Mehdizadeh et al[39], 2010Retrospective134Known CDNADiagnostic if > 3 ulcerations Suspicious if ≤ 3 ulcerationsDiagnostic: 52/134 (38.8%) Suspicious: 17/134 (12.7%) Jejunum lesions 53%, proximal ileum lesions 67%52/134 (38.8%) i.e., 100% of patients with positive CE findings
Lorenzo-Zúñiga et al[40], 2010Retrospective14Known CDNA≥ 7 mucosal breaks or ulcerations12/14 (86%) According to indications of CE: Abdominal pain = 3/3 Anemia = 5/5 Disease extent re-evaluation = 4/69/14 (64%) i.e., 100% of patients in whom CE was performed because of abdominal pain, 80% for anemia, 33% for disease extent re-evaluation
Petruzziello et al[31], 2010Prospective64Known CD of the distal ileum (n = 32) Control group (n = 32)SICUS> 3 aphthoid ulcers, deep ulcers, stricture(s)CD group: 16/32 (50%) with upper SB lesions vs 3/32 (9%) at SICUS, 30/32 (93%) with distal SB lesions vs 30/32 (93%) at SICUS Control group: 0/32 (0%)-
Dussault et al[79], 2013Retrospective71Known CDNAModerate: erythema and few aphthoid ulcers Severe: multiple and/or deep ulcers and/or stenosisModerate: 32/71 (45.1%) Severe: 12 (16.9%) According to indications of CE: Anemia = 4/6 Symptoms = 11/25 Disease re-evaluation = 28/3738/71 (53.5%) i.e., 75% of patients with severe lesions and 53% with moderate lesions
Kalla et al[80], 2013Retrospective315Known (n = 50) or suspected2 (n = 265) CDNA> 3 ulcers with erythema or edemaKnown CD: 33/50 (66%) (jejunum, n = 1 / diffuse, n = 16) Suspected CD: 45/265 (17%) (jejunum, n = 5 / diffuse, n = 7)Known CD: 73% Suspected CD: 90% of patients with positive CE findings
Flamant et al[81], 2013Retrospective108Known CD (32 L1, 25 L2, 51 L3)NADiffuse erythema and edema, linear/ circumferential ulcerations, ≥ 3 aphthous ulcers, or stenosis68/108 (63%) (jejunum, n = 60 of whom n = 18 i.e., 17% only in the jejunum) Restricted colonic location of the disease associated with a significantly decreased risk of jejunal lesions by 80% (OR = 0.21, P = 0.002)- Jejunal lesions=sole independent factor associated with increased risk of clinical relapse (HR = 1.99, P = 0.02)
Cotter et al[82], 2014Retrospective50Known CDNAModerate: Lewis score ≥ 135 Severe: Lewis score > 790Moderate: 33/50 (66%) Severe: 11/50 (22%)Proportion of patients on thiopurines and/or biologics increasing from 2/50 (4%) to 15/50 (30%) after CE, P = 0.023
Urgesi et al[41], 2015Retrospective492Suspected CD on obscure gastrointestinal bleedingNAMucosal fissure, ulcers of different shape, cobblestoning mucosa, aphthous ulcers, stricture(s), erythema/edema, loss of villi94/492 (19.1%) (jejunum, n = 31)64/94 -68%) i.e., 100% of confirmed CD
Greener et al[10], 2016Prospective79Known CDMRELewis score ≥ 135Proximal disease location detected by CE in 51% of patients vs 26% by MRE (P < 0.01) (isolated proximal lesions, n = 9)-
Chao et al[83], 2018Retrospective197Suspected CD in elderly patients2NALewis score > 7908/197 (4.1%)4/197 (2.0%) i.e., 50% of patients with positive CD findings
Carter et al[30], 2018Prospective, blinded50Suspected CD2Intestinal ultrasoundLewis score ≥ 135Similar diagnostic yield: 19/50 (38%) for SBCE and intestinal ultrasound, correlation r = 0.532, P < 0.001-
Sorrentino and Nguyen[32], 2018Retrospective43Known CD (20 never had surgery, 23 in the post-operative setting)Ileocolono-scopy and MRE/CTE and CRP/FLAny ulcerations or multiple erosionsSurgery-naïve group: 13/20 (65%) vs 8/20 (40%) at ileocolonoscopy vs 9/206 (45%) at imaging vs 12/207 (60%) at biomarkers Post-operative group: 20/23 (87%) vs 16/238 (70%) at ileocolonoscopy vs 0/239 (0%) at imaging vs 13/23 (57%) at biomarkersSurgery-naïve group: 6/20 (30%) Post-operative group: 12/23 (52%)
Hansel et al[84], 2018Prospective50Known CD with normal imagingNADiffuse erythema and edema, linear or circumferential ulceration(s), ≥ 3 aphthous ulcers, or stenosis14/50 (28%) with proximal SB lesions (duodenum, jejunum)17/50 (34%)
González-Suárez et al[33], 2018Retrospective47Known CD (n = 32) or suspected (n = 15) CDMRELewis score ≥ 13536/47 (76.6%) vs 21/47 (44.7%) at MRE, P = 0.001, of which jejunal lesions: 15/47 (31.9%) vs 3/47 (6.4%), P = 0.02-
Xavier et al[34], 2018Retrospective71Perianal CD (n = 17) and non-perianal CD (n = 54)NAVillous edema, erosions, ulcers or stenosisPerianal CD: 94.1% vs 66.6% in non-perianal CD (P = 0.03), with more frequently a Lewis Score ≥ 135: 94.1% vs 64.8% (P = 0.03), and higher Lewis scores in the first and second tertiles but not in the third tertile-
Reclassification of inflammatory bowel disease type
Maunoury et al[35], 2007Prospective30IBD-U with negative ASCA/ANCA and normal SBFTNA≥ 3 ulcerations5/30 (16.7%) (jejunum, n = 4)-
Lopes et al[36], 2010Prospective18IBD-U (n = 14) or IC (n = 4) with negative ASCA/ANCANADiagnostic if ≥ 4 erosions/ulcers and/or stricture(s) Suspicious if < 4 and/or focal villi denudationDiagnostic: 7/18 (38.9%) Suspicious: 9/18 (50.0%) Jejunum and proximal ileum lesions: 8/18 (44.4%)0 (0%)
Long et al[37], 2011Retrospective124CD (n = 86) or IC (n = 15) or pouchitis (n = 23)NAErythema, few aphthae/ulcers, multiple aphthae/ulcers, stenosisCD: 67/86 (77.9%) IC: 7/15 (46.7%) Pouchitis: 15/23 (65.2%)Medication: CD: 34/86 (39.5%) IC: 6/15 (40.0%) Pouchitis: 13/23 (56.5%) Surgery: CD: 11/86 (12.8%) IC: 6/15 (40.0%) Pouchitis: 1/23 (4.4%)
Table 2 Studies evaluating the use of small bowel capsule endoscopy in the assessment of mucosal healing in patients diagnosed with Crohn’s disease
Ref.Study designSample sizeTreatment evaluatedPrior biologic exposureInterval between the CECE findings considered for assessing mucosal healingPositive SB findings before treatmentPositive SB findings after treatmentP-value
Efthymiou et al[85], 2008Prospective, blinded40CS (60%) Mesalamine (70%) Azathioprine (10%) Infliximab (5%) Metronidazo-le (20%)-After achievement of clinical remission: 4 wk (75%)Number of apthous ulcers, mean ± SE26.0 ± 7.512.7 ± 2.30.07
6 wk (15%)Number of large ulcers, mean ± SE8.3 ± 1.45.0 ± 0.80.01
8 wk (10%)Percentage of the SBTT in which any endoscopic lesion was visible, mean ± SE22.0 ± 3.117.8 ± 2.50.08
Tsibouris et al[86], 2013Prospective, blinded1021--≥ 15 d after CDAI dropped < 150: 2-3 mo; (26.5%) 3-6 mo; (19.6%) 6-12 mo; (53.9%)CECDAI score, mean ± SD14 ± 64 ± 2-
Niv et al[44], 2014Prospective, blinded19Copaxone (68.4%) 5-ASA (52.6%) Antibiotics (15.8%) CS (5.3%) IS (10.5%) Vitamins (26.3%) Others (36.8%)-12 wkLewis score, mean ± SD1730 ± 1780No correlation between changes in CDAI/IBDQ and Lewis score2-
Hall et al[42], 2014Prospective, blinded43Adalimumab (84%) 160 mg W0, 80 mg W2, then 40 mg /2 wk or Azathioprine (16%) 2-2.5 mg/kgNaïve 38/43 Exposed 5/4352 wk3Complete MH = Absence of ulcers, n (%) Normalizatio-n of CECDAI score < 3.5, n (%) Change in CECDAI score, n (%)- - CECDAI < 3.5: 4/43 (9%) 3.5 ≤ CECDAI < 5.8: 13/43 (30%) CECDAI ≥ 5.8: 26/43 (61%)Complete MH: 12/28 (43%) CECDAI < 3.5: 2/28 (7%) CECDAI ≤ 3.5 < 5.8: 6/28 (21%) CECDAI ≥ 5.8: 8/28 (29%)< 0.0001
Kopylov et al[45], 2015Prospective52None (15.4%) 5-ASA (9.6%) Thiopurine (36.6%) Anti-TNF (26.9%) Anti-TNF+IS (11.5%)-NA Included patients were all in clinical remission (CDAI < 150) and had only one CE.MH = Lewis score < 135NAMH: 8/52 (15.4%) 135 ≤ Lewis < 790: 33/52 (63.5%) Lewis score ≥ 790: 11/52 (21.2%)-
Shafran et al[43], 2016Prospective, open-label15Certolizumab pegol 400 mg W0, W2, W4 then /4 wkNaïve 3/15 Exposed 12/15424 wk in respondersLewis score, mean1663226-
Aggarwal et al[46], 2017Prospective, blinded43None (14%) 5-ASA (60%) CS (12%) IS (74%) Anti-TNF (21%)-NA Included patients were all in clinical remission (CDAI < 150) and had only one CE.MH = Lewis score < 135NAMH: 17/43 (40%) 135 ≤ Lewis < 790: 19/43 (44%) Lewis score ≥ 790: 7/43 (16%) Significant correlation between Lewis score and fecal calprotectin (r = 0.82, P < 0.0001)-
Mitselos et al[47], 2018Retrospective30None (37%) 5-ASA (17%) Budesonide (10%) Azathioprine (10%) Anti-TNF (20%) Anti-TNF+IS (6%)-NA Included patients had only one CE (60% in both clinical and biochemical remission)MH = Lewis score < 1355NAMH: 6/15 (40%) Weak correlation between CDAI and Lewis score (r = 0.32, P = 0.088) and between CRP and Lewis score (r = 0.52, P = 0.004)-
Nakamura et al[87], 2018Prospective, blinded92None (27%) 5-ASA (18%) Thiopurines (17%) Infliximab (20%); Adalimumab (10%) Elemental diet (5%) CS (3%)Naïve 38/92Exposed 54/926 mo in the active group (40/92) Non-active patients ended the study at baseline (52/92)Lewis score, mean MH = Lewis score of 0 Active CD: Lewis score > 135458233 MH: 2/296 Improvement of LS in all 7 patients who received biologics, and in 8/11 (73%) of asymptomatic patients receiving additional medication0.0004
Table 3 Studies assessing the use of small bowel capsule endoscopy in the monitoring of patients with Crohn’s disease in the post-operative setting
Ref.Study designSample sizeIndication for surgeryRisk factors for PORPost-operative prophylac-tic treatmentCompared modalityInterval between surgery and endosco-pic re-assess-mentCE findings considered for defining PORRate of clinical recurrence, n (%)Rate of endosco-pic recurren-ce, s (%)
Bourreille et al[49], 2006Prospective, blinded32Resistance to medical treatment (19%) Stenosis (37%) Fistula/abs-cess (44%)-None (28%) 5-ASA (22%) CS (3%) IS (9%) Others (44%)Ileocolono-scopyMedian (IQR): 6 mo (4-7)Rutgeerts score ≥ i,1-Colonosco-py: 19/311 (61%) Se 90%, Sp 100% WCE: 21/31 (68%) Se 76%, Sp 91% SB lesions up to 72%
Biancone et al[50], 2007Prospective, blinded22Resistance to medical treatment (9%) Stenosis (64%) Fistula/abs-cess (14%) Other (13%)Smoking (32%) Penetrating phenotype (23%)Mesalamine (100%)Ileocolono-scopy (gold standard), SICUS1 yearUlcers, strictures, or stenosis in the neoterminal ileum and/or anastomosis0 (0%)Ileocolonosc-opy: 16/172 (94%) SICUS: 17/172 (1 FP) (100%) WCE: 16/172 (94%) Se 93%, Sp 67%
Pons Beltrán et al[51], 2007Prospective, blinded24Resistance to medical treatment (21%) Stenosis (63%) Other (16%)Smoking (50%) Penetrating phenotype (38%)None (100%)Ileocolono-scopyMedian (range): 254 d (118-439)Rutgeerts score ≥ i,20 (0%)Ileocolonos-copy: 6/24 (25%) WCE: 15/24 (63%) Jejunal lesions (54%)
Kono et al[52], 2014Prospective, blinded19-Smoking (11%) Penetrating phenotype (58%) Prior resection (68%)5-ASA (39%) Anti-TNF (61%)Ileocolono-scopy at 6-8 momean ± SD: 17.3 ± 5.6 d then 216.9 ± 23.6 dLewis score ≥ 135, n (%) and Mean (range)0 (0%)Week 2-3: 14/183 (78%) 428.3 (8-4264) 6-8 mo: 9/134 (69%) vs 3/6 (50%) at colonoscopy 196.1 (8-450) 5/13 (38%) with LS higher by ≥ 100 than shortly after surgery
Hausmann et al[53], 20175Prospective, blinded22-Penetrating phenotype (18%) Prior resection (50%)None (76%) Azathiopri-ne (6%) Adalimum-ab (18%)Ileocolono-scopy at 4-8 moMean (range): 57.5 (34 – 83) d then 220 (159 – 322) dModified Rutgeerts score ≥ i,2-Week 4-8: 3/166 (19%) 4-8 mo: 7 6/12 (50%) vs 5/15 (33%) at colonoscopy
Han et al[54], 2018Retrospec-tive, blinded83Resistance to medical treatment (24.3%) Stenosis (75.7%)None (100%)None (100%) before date 1 After date 1 if POR: None (53.1%) Azathiopri-ne (21.6%) Infliximab (25.3%)Group 1 (37/83): ileocolono-scopy + CE (date 1) then repeat colonoscopy (date 2) Group 2 (46/83): ileocolono-scopy (date 1 and 2)Date 1: 3-7 mo; after surgery Date 2: 1 year after date 1Rutgeerts score ≥ i,2 in the terminal ileum or > 5 aphthous lesions in proximal SB or ulcers in proximal SBGroup 1: 1/37 (2.7%) Group 2: 10/46 (21.7%) P = 0.019Date 1: Group 1: 13/37 (35.1%) at IC vs 24/37 (64.9%) at CE Group 2: 15/46 (32.6%) at IC (P = 0.809) Date 2: Group 1: 8/37 (21.6%) Group 2: 20/46 (43.5%) (P = 0.036)
Kusaka et al[55], 2018Prospective25-Smoking (22%) Penetrating phenotype (7%) Prior resection (48%)5-ASA (96%) Elemental diet (30%) IS (19%) Anti-TNF (59%)8-< 3 moLewis score ≥ 1355/25 (20%)21/25 (84%) mean ± SD: 751.3 ± 984.0 Clinical recurrence rate significantly higher in the group with highest third tertile score (distal SB)
Table 4 Advantages and limitations of small bowel capsule endoscopy in Crohn’s disease
AdvantagesLimitations
Less invasive than conventional endoscopyRisk of capsule retention in stricturing CD
No need for sedationNo therapeutic or biopsy capability
High diagnostic yield comparable to other endoscopic or imaging modalitiesSB evaluation may be incomplete due to:
Uncontrolled air insufflation
Retention or delayed transition
Limited battery life
Impossible to maneuver
Direct mucosal evaluationLonger procedure time compared to other modalities
Patient-friendlyAnalysis is time-consuming for the physician