Remission endpoints in ulcerative colitis: A systematic review

Table 4 Validation studies of ulcerative colitis endoscopic index of severity and ul-cerative colitis disease activity index

	Ref.	Patient number	Outcomes
UCEIS
Validity	Corte et al[18]	89	Correlation between UCEIS and outcomes The UCEIS score was directly proportional to requirement of rescue therapy UCEIS ≥ 5 was significantly linked to requiring colectomy 18/54 (33%) patients with UCEIS ≥ 5 compared to 3/33 (9%) with UCEIS ≤ 4 No definition of remission
	Fernandes et al[19]	108	Prediction of outcomes in acute severe colitis UCEIS was applied to score of the rectum and sigmoid, seg-UCEIS Seg-UCEIS predicted to develop steroid-refractory disease and the likelihood of colectomy (seg-UCEIS = 14 had a 17 times higher risk of steroid-refractory disease and a 25 times higher risk of requiring colectomy) Every 1 point increase in the UCEIS or Seg-UCEIS increased the need of colectomy by 2.78 and 1.79 respectively Mayo score did not predict these No definition of remission
	Arai et al[20]	285	Reflection of true UC activity and remission The recurrence rate was directly proportional to the UCEIS score (5.0% for UCEIS = 0, 22.4% for UCEIS = 1, 27.0% for UCEIS = 2, 35.7% for UCEIS = 3, 75% for UCEIS = 4-5) The absence of bleeding and mucosal damage were independent factors for continued clinical remission UCEIS ranged from 0 to 5 when clinical remission, Mayo ≤ 1 UCEIS ≤ 1 for clinical remission, which showed sensitivity of 68% and specificity 57% The expected duration of recurrence is also prolonged when UCEIS ≤ 1
	Kucharski et al[21]	49	Assessment of 9 endoscopic indices correlate well with (1) clinical indices; and (2) histological Geboes Index[22] The UCEIS showed the strongest correlation with the Geboes Index (the coefficient: 0.434 to 0.629) Recommends the UCEIS for the best overall correlations with both clinical and histological indices
Responsiveness	Ikeya et al[23]	41	The ability to detect to change after Tacrolimus remission induction treatment for moderate to severe UC Although Mayo endoscopic score is easy to use, it does not distinguish depth of ulcers unlike UCEIS Despite UCEIS score improved from 7 to 4, Mayo endoscopic score remained at 3 (severe) An improvement of UCEIS ≥ 3 showed close correlation with clinical remission, colectomy-free and relapse free rates Proposed remission (score 0-1), mild (2-4), moderate (5-6), severe (7-8) UCEIS 1 in remission is only from vascular pattern
	Menasci et al[24]	80	Comparison of the global UCEIS score from 5 segments and a traditional method of UCEIS score The regular method of the UCEIS is to score the most inflamed segment of the bowel This was compared with the sum of the score of five colonic segments A very good correlation (Spearman’s r = 0.86, P < 0.0001) for disease with UCEIS score ≤ 5 Less correlation (r = 0.48, P < 0.01) for disease with UCEIS > 5
Reliability	Travis et al[15]		Investigation of intra- and inter-observer consistency assessment 25 readers evaluated 28 videos including 4 duplicates to assess intra-reader reliability The intra and inter-reader reliability ratios for the UCEIS were 0.96 and 0.88 respectively The USCEI revealed a strong correlation with overall assessment of severity without being influenced by knowledge of clinical information No definition of remission
	Feagan et al[25]	281	The effect of centralized review of images on inter-observer variations Patients with UCDAI ≥ 2 were randomised to evaluate the efficacy of delayed mesalamine treatment (4.8 g/d for 10 wk) UCEIS was used as a part of inter-observer agreement study and showed interclass correlation coefficient of 0.83 amongst 7 central readers, which is superior to UCDAI
	Travis et al[26]		Clinical information influences UCEIS score 40 readers evaluated 28 of 44 videos No discrepancy between blinded and unblended readers Intra- and inter-reader variability demonstrated moderate to substantial agreement (κ = 0.47 to 0.74 and κ = 0.40 to 0.50 respectively) UCEIS correlated well with patient-reported symptoms - rectal bleeding, stool frequency and patient functional assessment (rank correlation = 0.76 to 0.82)
UCDAI
Validity	Higgins et al[27]	66	Finding endpoints in disease activity indices for remission and improvement in UC UCDAI < 2.5 for remission, which had a sensitivity and specificity of 0.82 and 0.89 Remission in this study was defined by patients
	Poole et al[28]	126	Establish the relationship between the UCDAI and patient reported EQ-5D The UCDAI with or without endoscopy assessment demonstrated a good correlation with EQ-5D Endoscopy assessment may not link with the disease activity
	Kucharski et al[21]	49	Assessment of 9 endoscopic indices correlate well with (1) clinical indices; and (2) histological Geboes Index (22) The UCDAI showed strong correlations with all 9 endoscopic indices (the coefficient in a range of 0.712 to 0.790) The UCDAI showed the highest correlation amongst clinical activity indices with the Geoboes Index (the Spearman’s coefficient 0.478) Compared to UCEIS, the UCDAI is less correlated with the Geboes Index
Reliability	Feagan et al[25]	281	The effect of centralized review of images on inter-observer variations Patients with UCDAI ≥ 2 were randomised to evaluate the efficacy of delayed mesalamine treatment (4.8 g/d for 10 wk) 31% of patients with UCDAI ≥ 2 enrolled in the RCT initially were considered ineligible by the central readers Inter-observer agreement amongst 7 central readers was good (interclass correlation coefficient: 0.78)

UC: Ulcerative colitis; UCEIS: Ulcerative colitis endoscopic index of severity; Seg-UCEIS: The sum of the rectal and sigmoid segmental UCEIS score; UCDAI: Ulcerative colitis disease activity index; EQ-5D: EuroQoI Five Dimensions Questionnaire; RCT: Randomised control trial.