Mucosal healing as an index of colitis activity: Back to histological healing for future indices

Discussions of mucosal healing as an index of response to therapy for ulcerative, Crohn's, and indeterminate colitis seem endless. Recent excellent reviews on the topic include those by Rutgeerts et al,1 Rubin,2 Kane et al,3 Schitzler et al,4 and Pineton de Chambrun et al.5 In these reviews, 80%–90% of the references are from the most recent decade and they include minimal criteria for histological healing. The gamut of criteria for “mucosal healing” includes: Healed mucosa to the eye of the endoscopist (“I know it when I see it”). Much improved appearance on endoscopy but still a duskiness or scattered erosion with most the mucosa appearing normal (not reproducible). Normal appearance in some areas previously involved, but not in others (not reproducible). Perfectly normal mucosa with residual one or many pseudopolypi. A fecal or serological marker specific to healing or healed mucosa. In none of the clinical trials which emphasize “mucosal healing” as the index of response to a specific drug treatment are any of these validated as primary or secondary endpoints. While mainly healed mucosa to the eye of the endoscopist has been the most satisfying objective confirmation to support the clinical response, the influence of a number of variables remains uncertain. What proportion of the success is due to the trial drug? What period of time on therapy is necessary for the success to endure? Why does the required dose of each drug responsible for the remission vary so greatly, and does some biochemical or immunological change in the natural course of the disease alter the early response from success to loss of success and even failure? We have seen the unexpected success of sulfasalazine for ulcerative colitis (UC) and documented differences from other sulfonamides. We have seen dramatic clinical responses to corticosteroids, which led to complacency in pursuit of research until there was confirmation that these drugs had no maintenance value. We have confirmed the efficacy of immunosuppressives but not clarified why their success is delayed in time. The gap has now been filled by the availability of anti-tumor necrosis factor (TNF) biologicals. But now we are seeing loss of response to the anti-TNF and struggling with the choice of the best routes with which to correct it. More trials will therefore be necessary to help most patients all of the time. This pertains to the responses to drug combinations as well as individual drugs. The eye of the beholding endoscopist is clearly not a perfect tool for determining mucosal healing. What experienced endoscopist has not concluded and reported that the mucosa is now “normal,” only to then receive a report from the pathologist that not merely does inflammation persist but it is “severe,” even when the biopsies are taken from multiple areas of normal-appearing colon. This criterion of healing might be satisfactory for comparison in time for response to therapy in an individual case, but not for mucosal healing as an entity and certainly not to be used as an index of response to therapy in trials. I personally believe that histological healing is a minimal criterion for mucosal healing and preferably this information should be derived from multiple biopsy sites of previous inflammation. If one reviews the current literature on this subject, rarely is a reference found to the work of pathologists from decades ago when they were faced with this same problem. Let us recall that Truelove and Richards6 in 1956 and Dick and Grayson7 and Matts8 in 1961 told us that a large number of patients with UC in clinical remission had abnormal sigmoidoscopic and histological findings, both of which could be severe, but at those times there was no clarification as to whether persistent, histological mucosal abnormality correlated with duration of freedom from symptoms. Dick et al9 reviewed 200 cases of UC after a minimum period of 5 years from onset to determine what proportion of mucosal abnormalities persist in the absence of symptoms. Their criteria for abnormal findings included: 1) an abnormal glandular pattern; 2) an excess of chronic inflammatory cells (small round cells and plasma cells); and 3) more than an occasional polymorphonuclear leukocyte. Mucosal biopsy in 79 patients free of symptoms revealed abnormal histological appearance in 71 (90%). There was no relationship between the period for which the patient had been symptom-free and histological changes. Morson10 in his 1972 New England Journal of Medicine review of rectal biopsies in inflammatory bowel disease (IBD) again emphasized that UC in clinical and sigmoidoscopic remission could be abnormal histologically. He included signs of mucosal atrophy such as loss of parallelism and separation of tubules or branching of the crypt. He recognized that the time taken for an attack of colitis to resolve in response to therapy varied but made no distinction between responses to the kinds of drug treatment available at the time. The next step was reported by Korelitz and Sommers (197411 and 197612) utilizing the technique of cell counting in addition to histological features and comparing the response to three forms of drug therapy: salicylazosulfapyridine, prednisone, and 6-mercaptopurine (6-MP). The cell counting included 500 mucosal epithelial cells (mucous, goblet cells, nonmucous cells, mitotic cells, pyknotic cells, karyorrhectic cells) and 500 lamina propria connective tissue cells (fibroblasts, macrophages, mast cells, lymphocytes, plasma cells, eosinophils, neutrophils). UC sigmoidoscopy positive on medication was compared with sigmoidoscopy negative on medication. Controls included UC sigmoidoscopy positive on no therapy and sigmoidoscopy negative biopsies of patients without IBD. The biopsies of 55 patients with negative sigmoidoscopies after therapy (compared with the 115 treated cases that still had positive sigmoidoscopy) showed that mucous goblet cells were increased and fragmented nuclei (evidence of epithelial degeneration) were significantly decreased (both P < 0.001), and the lamina propria was altered in the sigmoidoscopy negative cases after therapy by a significant decrease in eosinophils (P < 0.001). For sulfasalazine specifically, there was a decrease of all leukocytes and fewer plasma cells. For the prednisone treated group the results were the same, but the most striking reductions of leukocytes were the neutrophils. For the 6-MP group the results were again similar but neutrophils were significantly fewer than after either sulfasalazine or prednisone, and in addition there was an increase in mast cells (still not yet understood). Combinations of two drugs did not result in a statistically greater increase in mucous goblet cells or decrease in fragmented epithelial cells or total leukocytes than any of the three drugs given individually. Sulfasalazine had its major role in decreasing chronic inflammation as suggested by the reduction in plasma cells. Acute inflammation was most repressed by 6-MP and prednisone but greater after 6-MP alone than when combined with other drugs as supported by the marked decrease in neutrophils. Macrophages that are indicators of resolving inflammation were also most significantly increased after drug combinations that included 6-MP, perhaps also stimulated by the increase of epithelial nuclear fragmentation associated with this drug. In 1991, Riley et al13 called attention to the increased risk of relapse of UC when biopsies revealed an acute inflammatory infiltrate than when such an infiltrate was absent or when the infiltrate was limited to chronic inflammatory cells. D'Haens et al14 in 1999 reported that the disappearance of neutrophils after infliximab treatment of Crohn's disease was “remarkable.” In 2000 Geboes et al15 found that the presence of neutrophils in the epithelium correlates well with crypt destruction and erosions and emphasized that the combination of endoscopy and biopsies provides a better indication of activity than endoscopy alone, especially in noninflamed mucosa. It has also been shown that there is preferential suppression of neutrophils in complete blood differential counts (CBCs) when leukopenia is induced by 6-MP during the treatment of IBD,16 and that all drugs (6-MP, prednisone, and sulfasalazine) lead to a reduction in neutrophils when measured in rectal mucosa biopsy specimens with successful response to therapy.17 Bitton et al18 presented evidence that histologic healing may improve clinical outcomes in UC. They called attention specifically to the presence of plasmacytosis on biopsy increasing the likelihood of relapse. This corresponds to the finding of decreased plasma cells in response to therapy found in the mucosal cell count study mentioned earlier11,12; both support the role of histological healing. When the technique of mucosal cell counts was introduced it was recognized that it would not be appealing to the general pathologist, who performs a service to the hospital and does not have sufficient time for a focus such as this. Nevertheless, the need for a histological process has become better focused as we search for an accurate indicator of mucosal healing and recognizing that the endoscopic appearance is insufficient for this purpose. As suggested 36 years ago, cell counting could be modified, shortened, and validated so that it would be much less time-consuming. Just as total and differential counts of blood leukocytes are useful in evaluating inflammatory conditions, similar methods applied to the analyses of inflamed tissue are also of value. Then it might serve as the index of mucosal healing in response to drug therapy that is so badly needed for clinical trials.