Is Endoscopic Remission in Ulcerative Colitis Still Good Enough?

Abstract

Histology is a treatment target for investigational agents. Histologic activity predicts relapses and increased risk of colorectal neoplasia. Recent studies demonstrated that the proportion of patients achieving histologic improvement is low. Research is needed before providers treat to histologic remission.

The Selecting Therapeutic Targets in Inflammatory Bowel Disease (STRIDE) program, initiated by the International Organization for the Study of Inflammatory Bowel Diseases (IOIBD), recommended the following treatment targets for patients with ulcerative colitis (UC):

1. patient-reported remission, defined as resolution of symptoms such as rectal bleeding, diarrhea, or other altered bowel habits;

2. endoscopic remission, defined as a Mayo endoscopic subscore (MES) of 0 or 1.¹

STRIDE recommended assessment of these end points 3–6 months after initiating treatment and considered histologic remission to be only an adjunctive target of treatment. However, in 2016, the Food and Drug Administration (FDA) recommended that endoscopic remission no longer be assessed by visual appearance of the mucosa alone.² What is the current evidence for use of histology as a target of treatment?

Before considering the evidence surrounding use of histologic activity as a target of treatment, one needs to consider how histologic activity is measured. Definitions for histologic remission range from complete resolution of any inflammation to resolution of active inflammation.^{2, 3} There are multiple scoring systems to assess histologic activity including the Geboes Score, Nancy Index, and Robarts Histopathologic Index; however, none are universally validated or considered to be the gold standard.³

Prior studies have shown evidence of discordance between endoscopic and histologic activity. One such study followed 103 patients in symptomatic remission on maintenance therapy.⁴ These patients underwent surveillance colonoscopies during which time endoscopic activity was scored using the MES and histologic activity using the Geboes score. Fifty-four percent of patients demonstrated some degree of inflammation on biopsy, and 37% of patients met Geboes criteria for abnormal histologic inflammation (score > 3.1). The overall correlation between the MES and Geboes score was good (ρ = 0.65). However, 11% of patients had histologic inflammation in the right colon without any endoscopic evidence of inflammation, and in patients with endoscopic evidence of left-sided colitis, 34% had histologic inflammation from right-sided biopsies. Conversely, in colonic segments with an MES of 0, only 6% had histologic evidence of inflammation. Bryant, et al. also demonstrated disagreement between endoscopic and histologic remission.⁵ In this cohort of 91 patients, 25% of those in endoscopic remission had evidence of histologic activity on biopsy. Only 29% of patients were in remission by all 3 clinical, endoscopic, and histologic parameters.

Other studies examined the association between histologic activity and important outcomes such as relapse and colorectal neoplasia. Riley and colleagues followed 82 patients in endoscopic or near-complete endoscopic remission over 12 months.⁶ Patients with evidence of acute inflammatory infiltrate on biopsy had a 52% chance of relapse; similar rates were seen in patients with other features of inflammation such as crypt abscesses, mucin depletion, and breaches in surface epithelium. Similarly, Bitton et al. conducted a multicenter prospective study where patients underwent serial colonoscopy or sigmoidoscopy with biopsies.⁷ Despite being in clinical and endoscopic remission at baseline, patients with presence of basal plasmacytosis and crypt atrophy on biopsy had shorter times to relapse compared with patients without these features. Studies have also examined the association between active inflammation and colorectal neoplasia in UC. One study matched 86 patients with controls and found a significant correlation between both endoscopic and histologic inflammation and colorectal neoplasia; after adjusting for confounding variables, histologic inflammation was the only variable associated with dysplasia.⁸ A more recent study demonstrated a similar correlation between histologic inflammation and risk of colorectal neoplasia that was also independent of endoscopic activity.⁹

Currently, we know that histologic activity is associated with relapse and dysplasia, a finding that may be independent of endoscopic activity. We also know that newer agents under investigation such as ustekinumab and ozanimod for UC improve histology. However, mucosal healing, defined as both endoscopic and histologic remission, is uncommon in these trials. In the phase 3 trial of ustekinumab for patients with moderately to severely active UC, only 35.6% of patients achieved histologic improvement and 19% of patients achieved mucosal healing at week 8.¹⁰ Similarly, the TOUCHSTONE study found low rates of histologic remission (Geboes score < 2) at week 8 (22.4%), with only a modest increase to 31.3% at week 32 in patients treated with high-dose ozanimod.¹¹

We have no controlled trials demonstrating that treating to histologic remission improves outcomes. We can, however, tell patients that achieving histologic remission is associated with an excellent prognosis. If histologic remission is to become an end point in clinical practice, standardization of pathology reports will be needed. Future treat-to-target studies are needed to determine if practitioners should make changes in treatment based on histologic activity alone. Furthermore, information is needed on the proportion of patients achieving histologic remission with optimized treatment. Until those studies are complete, we agree with the STRIDE recommendations that histologic remission as a target of treatment is aspirational. We do not recommend changes in treatment based on histologic activity only, although closer monitoring with traditional biomarkers including fecal calprotectin may be warranted in these situations to identify recurrent endoscopic activity and/or clinical symptoms sooner.

Supported by: None.

Conflicts of Interest: R.K. Cross has participated in advisory boards and consulting for Abbvie, LabCorp, Janssen, and Pfizer. R.K. Cross has a research grant with Abbvie.