Mesalamine and cholestyramine for immune checkpoint inhibitor-mediated diarrhea and colitis

Immune checkpoint inhibitors (ICI) are effective against various malignancies. However, adverse events including diarrhea and colitis can lead to significant morbidity and mortality. Recommendations for the management of ICI mediated diarrhea and colitis include steroids and biologics. Given their associated risks, this study evaluated the role of the non-immunosuppressive agents, mesalamine and or cholestyramine. This is a retrospective, descriptive, single-center study of adults who developed ICI diarrhea and colitis between 2010 and 2020 at MD Anderson Cancer Center. Clinical data and outcomes were compared between those treated with the non-immunosuppressive therapies mesalamine and/or cholestyramine alone versus those who received additional immunosuppression with steroids and biologics. Our sample comprised 66 patients wherein, the mean age was 63 years, 71% were males, and 97% had stage III/IV cancers. Fourteen patients were treated successfully with non-immunosuppressive therapy. They had grade 1–3 diarrhea and 1–2 colitis with no difference in the rate of histologic colitis compared to those who received immunosuppressive therapy. They had less CTLA-4 inhibitor-based therapy (36% vs. 73%, p = 0.034), delayed onset of symptoms (159 vs. 64 days, p = 0.011), lower fecal calprotectin levels (56 vs. 234, p = 0.012) and were more likely to resume ICI therapy (64% vs. 25%, p = 0.006). Mesalamine and/or cholestyramine may be effective for mild ICI diarrhea and colitis among patients with delayed symptom onset with lower colonic inflammatory burden. Prospective studies randomizing patients with mild colitis between mesalamine/cholestyramine and immunosuppressive treatment are warranted to assess their efficacy and safety.


Introduction
Immune checkpoint inhibitors (ICI) are highly efficacious cancer therapies that target regulators of the immune system. However, by augmenting the immune response, they can impact multiple organ systems and cause immune-related adverse effects (irAEs). Gastrointestinal toxicities are common, reported in up to 27% of cases and include ICI-mediated diarrhea and colitis, which can be associated with treatment cessation, morbidity and mortality (Postow et al. 2018;Schneider et al. 2021;Wang et al. 2018a).
Current guidelines from oncology societies and GI organization for the diagnosis and management of ICI diarrhea and colitis are based on symptoms described by the clinical severity grade based on the Common Terminology Criteria for Adverse Events (CTCAE) (Schneider et al. 2021;Bellaguarda and Hanauer 2020;Dougan et al. 2021;National Cancer Institute 2017;Som et al. 2019;Haanen et al. 2017;Puzanov et al. 2017). However, studies show that symptom severity of diarrhea and colitis does not correspond with objective evidence of inflammation (Som et al. 2019;Wang et al. 2018b). Endoscopic evaluation is, therefore, often essential for evaluation and management of ICI diarrhea and colitis. Histologic findings of ICI colitis include acute, chronic, or microscopic colitis with infiltration of neutrophils, lymphocytes, plasma cells and eosinophils (Bellaguarda and Hanauer 2020). The pathogenicity of ICI colitis is speculated to be multifactorial, with various contributing factors including autoimmune, genetics, and microbiome to name a few. Some of the disease distribution and histologic findings of ICI colitis overlap with those of inflammatory bowel diseases (IBD) and microscopic colitis (Dougan et al. 2021;Som et al. 2019;Coutzac et al. 2017;Kubo et al. 2017;Choi et al. 2019). While emerging studies are demonstrating the mechanism of enteric injury by ICI therapy which will assist in effective target therapy, (Sasson et al. 2021;Luoma 2020) current treatment options overlap with treatment regimens for IBD and other irAEs (Schneider et al. 2021;Haanen et al. 2017;Puzanov et al. 2017;Nishida et al. 2019).
The management of ICI diarrhea and colitis is based on retrospective studies with recommendations for conservative therapy for grade 1 diarrhea and colitis, steroids for persistent symptoms and at clinical discretion for grade 2, and steroids with a prolonged steroid taper of 4-8 weeks with possible addition of biologics for grade 3 and 4 (Schneider et al. 2021;Haanen et al. 2017;Puzanov et al. 2017). However, safety concerns about immunosuppressant therapy with steroids and infliximab have been increasingly recognized and need to be considered (Postow et al. 2018;Schneider et al. 2021;Favara et al. 2020;Del Castillo et al. 2016). High dose and long-term use of steroids are known to increase the risk for multiple metabolic derangements such as hyperglycemia, osteoporotic fractures, and reports of a two-fivefold increased risk for infection (Fardet et al. 2016) Studies have demonstrated similar concerns in patients with irAEs; Favara et al. reported a rate of infection of 15%, and hyperglycemia in 10% in patients treated with steroids for ICI colitis (Favara et al. 2020). Furthermore, some studies have raised the concern that immunosuppression may reduce the antitumor efficacy of ICI with decreased rates of progressionfree survival and overall survival (Postow et al. 2018;Del Castillo et al. 2016;Mouri et al. 2021). Given the potential risks associated with immunosuppressive therapy, it is of clinical significance to investigate the role of non-immunosuppressive treatment modalities in the management of ICI diarrhea and colitis. Currently, the evidence on the use of non-potent immunosuppressive agents is limited to small retrospective studies (Som et al. 2019;Haanen et al. 2017;Kubo et al. 2017;Choi et al. 2019;Abu-Seih et al. 2018a;Kikuchi et al. 2019;Yamauchi et al. 2018;Hughes et al. 2019).
Mesalamine is a 5-aminosalicylate agent (5-ASA) with anti-inflammatory properties, thought to play a role in maintaining mucosal integrity in the intestine with high concentrations in colonic cells. Given its localized function, it is not considered to carry systemic immunosuppressive effects (Rousseaux et al. 2005). The effectiveness of 5-ASA depends on its locally absorbed mucosal concentration for which various formulations have been developed to be released in the gut (Rousseaux et al. 2005;Steinhart et al. 2007). Currently, 5-ASA agents are considered first-line therapy for induction and remission in patients with mild to moderate Ulcerative Colitis (UC) given their favorable efficacy and safety profile, (Rubin et al. 2019;Ko et al. 2019) with doses of 2-3 g per day or higher being more effective and adding rectal mesalamine for more extensive and left sided colonic disease (Ko et al. 2019). The AGA also recommends mesalamine as a second-line therapy for microscopic colitis including ICI induced mild microscopic colitis as well as the use of cholestyramine for microscopic colitis (Dougan et al. 2021, Nguyen et al. 2016. Excessive bile acid in the gastrointestinal tract is a common and confounding etiology of diarrhea lead to mucosal injury. Cholestyramine, a bile acid sequestrant, is effective in the treatment of bile acid diarrhea (Arasaradnam et al. 1 3 2018). Microscopic colitis is one disease state that has been associated with bile acid malabsorption with some studies showing cholestyramine monotherapy or with mesalamine to be effective. (Ung et al. 2000;Calabrese et al. 2007).
Given the rapid expansion of indications for ICI, there is a critical need to better understand the varying manifestations of gastrointestinal toxicities and explore safer treatment modalities to minimize complications from its treatments. To address this concern, we performed a retrospective study of patients with ICI-mediated diarrhea and colitis at a tertiary cancer center to describe the characteristics of patients who successfully responded to non-immunosuppressive treatment with mesalamine and/or cholestyramine as compared to those who were treated with immunosuppressive therapy.

Patient selection and data collection
We retrospectively reviewed the institutional database to identify consecutive patients who received both ICI and mesalamine and/or cholestyramine from January 1, 2010, through June 30, 2020 at The University of Texas MD Anderson Cancer Center. Inclusion criteria included (1) age 18 years or older, (2) cancer diagnosis (3) cancer treatment with ICI (4) ICI-mediated diarrhea and colitis diagnosis, and (5) treatment with mesalamine and/or cholestyramine for ICI-mediated diarrhea and colitis. Patients who had a concurrent active gastrointestinal infection were excluded. ICImediated diarrhea and colitis were defined based on current guidelines from the American Society of Clinical Oncology (Schneider et al. 2021). Pathologic ICI colitis was determined based on histology findings on endoscopic evaluation.

Baseline demographic and oncological characteristics
Pertinent demographic data (age, gender, race/ethnicity, and body mass index) were extracted from electronic health records and endoscopy databases. Information related to medical comorbidities, oncologic history, and cancer treatments was also collected. Comorbidities included coronary artery disease, chronic obstructive pulmonary disease, hypertension, diabetes mellitus, and hypercholesterolemia. The collected oncologic data included cancer type, stage, and treatments (including the types of ICIs, total cycles, and duration and resumption of ICI therapy).

Clinical presentation, endoscopic and histologic evaluation of immune checkpoint inhibitor-mediated diarrhea and colitis
ICI-mediated diarrhea and colitis characteristics were presented in terms of duration and severity of diarrhea and colitis based on the Common Terminology Criteria for Adverse Events (CTCAE) v5. (National Cancer Institute 2017) Information about non-gastrointestinal ICI-related irAEs was also included. When available, endoscopic findings observed on initial evaluation were noted. Features of endoscopy findings were categorized as normal, ulcers, and non-ulcerative inflammation (e.g., erythema, friability, erosions, inflammatory exudate, loss of vascular pattern, and edema). The histologic findings were described as normal, acute active colitis (e.g., cryptitis, crypt abscess, apoptosis, eosinophilic infiltration, and intraepithelial neutrophil infiltration), chronic active colitis (e.g., crypt architectural distortion, basal lymphoplasmacytosis, and Paneth cell metaplasia), or microscopic colitis (e.g., intraepithelial lymphocytic infiltration and subepithelial collagen bands). The endoscopic and histologic categorizations were based on our prior publication of endoscopic evaluation of ICI-mediated histologic colitis and current description of histologic abnormalities (Bellaguarda and Hanauer2020;Wang et al. 2018a).

Treatment and outcomes of immune checkpoint inhibitor-mediated diarrhea and colitis
All patients who received mesalamine/cholestyramine as treatment for ICI-mediated diarrhea and colitis in the study window were included based on the inclusion criteria, and were divided into two groups: patients treated with immunosuppressive therapy and patients successfully treated with the non-immunosuppressive therapies mesalamine and/or cholestyramine in order to evaluate the characteristics of patients who respond to mesalamine and/or cholestyramine alone. Patients treated with immunosuppressive therapy included those treated with systemic steroids with or without additional immunosuppressive therapy with infliximab and vedolizumab after initial or combined treatment with mesalamine or cholestyramine. Clinical outcomes of ICImediated diarrhea and colitis, including the need for hospitalization, duration of hospital stay, symptom resolution, recurrence of diarrhea and colitis, resumption of ICI therapy and death were examined.

Statistical analysis
We performed a descriptive statistical analysis of our data. Categorical variables were summarized with frequencies and percentages. Continuous variables were summarized with medians and interquartile ranges (IQRs). The Fisher exact test or Chi-square test was used to assess associations between categorical variables. Continuous variables were compared using the Wilcoxon rank-sum test. All the statistical methods were included in the footnote of each result table. Statistical analysis was conducted using SPSS (version 24.0; IBM Corporation, Armonk, NY). All statistical evaluations were two sided, and p values of less than 0.05 were considered statistically significant.

Ethical considerations
The study was conducted with approval from the Institutional Review Board at MD Anderson Cancer Center. Informed consent from patients was waived for this study.

Patient characteristics
A total of 66 eligible patients were included for data analysis (Fig. 1). The median age was 63 years at the time of ICImediated diarrhea and colitis diagnosis. Seventy-one percent of the patients were male, and more than 90% were Caucasian. Sixty-three patients (96%) had solid tumors as primary malignancy with 97% of these having advanced cancer stage (III or IV) at the onset of diarrhea and colitis. Of the solid malignancies, genitourinary cancer was the most common (58%) followed by melanoma (17%), with others including lung, head and neck, endocrine, and gastrointestinal cancers. These cases were present proportionately in both the non-immunosuppressive group and the immunosuppressive group (p = 1.000). Three patients had hematologic malignancies, and they were all treated with immunosuppressive therapy. All-cause mortality was 42%, which was primarily related to the underlying malignancy. The median follow-up time was 241 days for the non-immunosuppressive group vs. 360 days for the immunosuppressive group (p = 0.435). None of the baseline characteristics were significantly associated with immunosuppressive or non-immunosuppressive treatment (Table 1).

Clinical characteristics and endoscopy
In our cohort, 14 cases were successfully treated with 3-4 g oral mesalamine daily and or cholestyramine alone and 52 cases were treated with immunosuppressive treatment following or in addition to mesalamine and/or cholestyramine (Table 2). In the immunosuppressive group, all received steroids and 38 cases received additional biologic therapy. While some cases received vedolizumab instead of infliximab, these cases were included in the immunosuppressive group due to concurrent systemic steroids. Two patients that received budesonide were also included in the immunosuppressive group due to prior initial exposure to prednisone. In the immunosuppressive group, 11 cases had escalation of therapy from mesalamine to steroids with 8 of these also receiving biologics. In 8 of the 11 cases, steroids were started after 3-5 days of persistent or worsening symptoms on mesalamine and in three cases, immunosuppressive therapy was not started until after 10 days of mesalamine treatment. The remaining 41 cases received mesalamine and/or cholestyramine concurrently with immunosuppressive therapy. The median time from ICI therapy initiation to the time of diarrhea or colitis diagnosis was longer among patients in the non-immunosuppressive group compared to those in the immunosuppressive group (159 days vs. 64 days; p = 0.011). The total median duration of diarrhea and colitis symptoms was approximately 20 days and was similar between the two groups (p = 0.141). Patients managed with non-immunosuppressive therapy were more likely to have been treated with PD-1/L1 monotherapy versus CTLA-4 monotherapy or CTLA-4 combined therapy (64% vs. 1% vs. 4%; p = 0.034). Interruptions of ICI treatment due to ICI-mediated diarrhea and colitis were more common in the immunosuppressive group (90% vs. 64%, p = 0.016).
There was no significant difference in the peak grade of diarrhea and colitis between the two groups based on CTCAE grading; with 50% vs. 29% grade 1-2 diarrhea for those on non-immunosuppressive vs. immunosuppressive and 100% vs. 83% grade 1-2 colitis for those on nonimmunosuppressive vs. immunosuppressive therapy. In the non-immunosuppressive group, there were no cases with grade 4 diarrhea and no cases with grade 3-4 colitis. When comparing the grade of ICI-mediated diarrhea and colitis between the 14 cases successfully treated with non-immunosuppressive therapy to the 11 cases who were escalated to steroids from mesalamine/cholestyramine; those started on steroids were noted to have worsening of symptoms with a higher rate of grade 3 diarrhea (91% vs. 50%, p = 0.042) and grade 3 colitis (64% vs. 0%, p = 0.026) ( Table 3). There was no difference in endoscopic or histologic findings of colitis, fecal calprotectin levels or outcomes between those successfully treated with non-immunosuppressive therapy versus those with escalation to steroids.
In total, thirty-eight patients (58%) also suffered from non-gastrointestinal irAEs involving other organs including hepatobiliary, endocrine, pulmonary, dermatologic, and hematologic systems. The incidence of other irAEs was similar in both groups (p = 0.209). Of the 52 patients treated with steroids for immunosuppression, four cases had concurrent irAEs at the time of ICI diarrhea and colitis diagnosis; however, steroids and biologics were initiated primarily for the diarrhea and colitis. In addition, there was one case that had already been on hydrocortisone for adrenal insufficiency but was started on high dose prednisone for the treatment of ICI diarrhea and colitis.
A total of 50 (96%) of 52 cases in the immunosuppressive group and 12 (86%) of the 14 in the non-immunosuppressive group underwent initial endoscopy evaluation for ICI diarrhea and or colitis. The non-immunosuppressive group was more likely to have a normal colon on colonoscopy (54% vs. 31%) with no ulceration (0% vs. 18%) and a similar rate of non-ulcer inflammation (46% vs. 55%); however, these findings did not reach statistical significance. Furthermore, there was no correlation between endoscopic findings of ulceration and grade of diarrhea (p = 0.862) or colitis (p = 0.524). Of the total 62 cases that underwent colonoscopy evaluation, 56 cases (85%) were found to have histologic colitis on pathology evaluation; 45 cases (90% of 50) in the immunosuppressive group and 11 cases (92% of 12) in the non-immunosuppressive group. Although the immunosuppressive group was more likely to have acute colitis (70% vs. 42%) and a lower rate microscopic colitis (6% vs. 33%), the overall rate of acute, chronic and microscopic colitis was not statistically different between both groups (p = 0.053). There was also no correlation between the finding of acute colitis on histology and grade of diarrhea (p = 0.862) or colitis (p = 0.243). When comparing histologic colitis between those successfully treated with non-immunosuppressive therapy to the cases that were escalated to steroids, there was no difference in the rate of histologic colitis or subtypes of colitis. 38 cases (76% of 50) in the immunosuppressive group were on steroids prior to or started on the same day of colonoscopy, before histology results were available. Absence of histologic abnormality   was noted in a total six cases, 1 in the non-immunosuppressive group and 5 in the immunosuppressive group. Of the 5 cases treated with immunosuppression but normal histology, all had already been initiated on steroid therapy for suspected ICI-related diarrhea without objective colitis between 3 and 27 days prior to colonoscopy (4 of the 5 were on steroids ≤ 7 days of colonoscopy) and continued steroid taper course for 5-30 days. There was no difference between the rate of obtaining endoscopic examination between the nonimmunosuppressive and immunosuppressive group (86% vs. 96%, p = 0.195). Calprotectin was measured in seven (50%) non-immunosuppressive cases and 12 (23%) immunosuppressive cases at the time diarrhea and colitis diagnosis and found to be significantly higher in those treated with immunosuppressive therapy (234, IQR: 117-382 vs. 56, IQR: 23-99, p = 0.012).

Immune checkpoint inhibitor-mediated diarrhea and colitis outcomes
A total of 56 patients (85%) required hospitalization for diarrhea or colitis for a median duration of 7 days ( Table 2). Most of these patients (88%) achieved clinical improvement after initiation of treatment either with mesalamine and/or cholestyramine or with immunosuppressive therapy. There was no significant difference in regards to these outcomes between the two groups. Of the 66 cases, those on non-immunosuppressive therapy were more likely to resume ICI therapy (64% vs. 25%, p = 0.006). A total of 31 (47%) patients had ICI diarrhea and colitis recurrence, and there was no difference in the rate of recurrence between the two groups. However, those on non-immunosuppressive therapy were more likely to have recurrence after ICI resumption; whereas, those on immunosuppressive therapy were more likely to have recurrence even without ICI resumption (p = 0.023). When comparing the 11 cases that had delayed steroid initiation after initial trial of non-immunosuppressive therapy compared to the total 52 cases who received steroids, there was no difference in the duration of steroid use (38 days vs. 45 days), biologic therapy use (82% vs. 73%), need for hospitalization (100% vs. 88%) or recurrence of ICI diarrhea or colitis (36% vs. 50%).
Of the 52 cases treated with immunosuppression, there were a total of 11 cases (21%) with infection while on steroid treatment or within 30 days of steroid and/or infliximab treatment completion. In comparison, there were no cases of infection within 30 days of completion of mesalamine and/or cholestyramine in the non-immunosuppressive group; however, this was not a statistically significant difference (p = 0.102). Types of infection included urinary tract infection, pneumonia, perirectal abscess, candida esophagitis, oral herpes simplex out-break, and subsequent development of Enteropathogenic Escherichia coli colitis and Clostridioides difficile colitis after a negative initial infectious workup. There were no reports of adverse events within 30 days of non-immunosuppressive therapy completion. However, of the 11 cases initially started on mesalamine and then escalated to steroids, 1 case initially presented with grade 1 diarrhea and colitis but went on to have progression of symptoms within 30 days requiring hospitalization and steroid initiation.
When the clinical characteristics were compared between patients who resumed ICI therapy and those who did not (Supplemental Table 1), patients who had more severe colitis (grade 3) were less likely to resume ICI (p = 0.024), and patients who achieved clinical improvement on non-immunosuppressive therapy were more likely to resume ICI therapy (p = 0.006). Although the overall rate of ICI-mediated diarrhea and colitis recurrence was not significantly different between the immunosuppressive and non-immunosuppressive groups (50% vs. 36%, p = 0.342), recurrence among the non-immunosuppressive group was associated with ICI resumption in four of the five cases. The median duration from ICI resumption to ICI-mediated diarrhea and colitis recurrence was 45 days (IQR 39-80). Two patients were started on immunosuppressive therapy to prevent diarrhea and colitis recurrence upon ICI resumption.

Discussion
Immune checkpoint inhibitor-mediated diarrhea and colitis, a commonly encountered toxicity from ICI therapy, ranges in the degree of severity and can result in cancer treatment interruption, morbidity, and mortality (Schneider et al. 2021;Wang et al. 2018b). Current recommendations for the management of ICI-mediated diarrhea and colitis are based on CTCAE grading of symptoms and frequently involve the use of immunosuppressive therapy (Schneider et al. 2021;Haanen et al. 2017;Puzanov et al. 2017;Del Castillo et al. 2016). In this study, we described 14 patients with ICI diarrhea and colitis who responded to the non-immunosuppressive therapies mesalamine and/or cholestyramine. Compared to those treated with immunosuppressive therapy, a higher percentage of patients treated with mesalamine/cholestyramine received PD-1/L1 monotherapy, had delayed onset of symptoms and resumption of ICI therapy. There was no difference in CTCAE grade of diarrhea/colitis, histologic colitis, rates of hospitalization or recurrence of symptoms.
Although there was no significant difference in the CTCAE grade of diarrhea or colitis between these two treatment groups, the 14 cases treated with non-immunosuppressive therapy had fewer grade 3 diarrhea and no cases of grade 3 and 4 colitis. There were also no cases with ulceration, suggesting a lower inflammatory burden. The histologic findings varied among microscopic, chronic and acute colitis. Although 18% of the cases treated with immunosuppressive therapy had ulceration on endoscopy and higher rates of acute colitis, there was no significant difference in the endoscopic or histologic findings between the two groups. One would hypothesize that patients who may not require immunosuppressive therapy are more likely to have lower grade symptoms and lower grade inflammatory burden on colonoscopy evaluation such as absence of ulcer. However, this is not reflected from our findings likely due to the following reasons: (1) mesalamine and/or cholestyramine is not routinely used for immune-mediated diarrhea and colitis, and these therapies were used at the discretion of physicians with initiation of steroids for grade 2 at the discretion of the physician; (2) there is poor correlation between clinical symptoms and endoscopic and/or histologic findings; therefore, individual parameter may not provide accurate measurement of severity of colitis; (Som et al. 2019;Wang et al. 2018a) (3) this is a small sample size cohort, which added limitation to categorize patients into groups based on a set criteria for each treatment. Notably, 9% of our cohort had no histologic colitis which is in line with other studies reporting that up to 15% of cases with ICI-related gastrointestinal toxicity (diarrhea only) do not have objective colitis on histology (Bellaguarda and Hanauer 2020). Nevertheless, we present a cohort of patients with histologic findings of colitis who had symptom resolution without immunosuppressive therapy, which suggests that there may be a subset of patients with ICI diarrhea and colitis who may not require immunosuppression for management.
Our findings re-demonstrated the lack of correlation between clinical symptoms of ICI diarrhea and colitis and endoscopy/histologic findings, which is in line with prior studies (Bellaguarda and Hanauer 2020;Som et al. 2019;Wang et al. 2018b). Endoscopic and histologic inflammation is mostly present in patients with clinical symptoms. And while cases with absence of colitis on histology but clinical symptoms alone are uncommonly encountered, may still be a presentation of ICI-related GI toxicity if other alternative etiologies have been excluded. The requirement of aggressive immunosuppressive treatment on these patients appeared to be less frequent (Wang et al. 2018a). Given the range of ICI diarrhea and colitis clinical manifestations as well as the varying histologic findings, early endoscopic evaluation is needed to improve the measurement accuracy on the severity and subtype of histologic colitis which may be important for choosing an appropriate and effective non-IMS vs. IMS therapy. In addition, non-invasive markers to monitor disease activity ought to be considered. Recent studies have shown calprotectin levels to be predictive of the inflammatory burden seen on endoscopic and histologic evaluation of ICI colitis at the time of diagnosis and after treatment (Zou et al. 2021). In our study, fecal calprotectin level at the time of ICI diarrhea and colitis diagnosis was significantly higher in those treated with immunosuppressive therapy (234 vs. 56, p = 0.012); however, this association is limited by our sample size, varying histologic findings and study design. Larger prospective studies are needed to evaluate the utility of fecal calprotectin in ICI diarrhea and colitis and its association with histologic colitis.
In this study, the patients who responded to mesalamine and/or cholestyramine reflect similar reports of smaller case reports. This favorable response has been observed in patients with grade 1-3 diarrhea, grade 1-2 colitis and different histologic subtypes of ICI colitis including acute, chronic, and microscopic colitis as reported in our study (Kubo et al. 2017;Choi et al. 2019;Abu-Seih et al. 2018b;Kikuchi et al. 2019;Yamauchi et al. 2018). The AGA has also endorsed mesalamine as a second-line agent and trial of cholestyramine for diarrhea in conventional microscopic colitis, (Nguyen et al. 2016) with success reported from the use of mesalamine and cholestyramine in the management of ICI induced microscopic colitis in case studies (Ung et al. 2000;Calabrese et al. 2007).
The exact pathophysiology of ICI colitis is not fully understood, but there is emerging data demonstrating a higher proportion of activated CD8 + tissue resident T cells in ICI colitis versus a higher proportion of B lineage cells in IBD as well as different T cell cytokine expressions (Sasson et al. 2021;Luoma et al. 2020). The exact mechanism of action of mesalamine is also not fully understood, but it is thought to inhibit localized cytokine production, inflammatory mediators and target peroxisome proliferator-activated receptor gamma which plays a role in maintaining mucosal integrity in the intestine (Rousseaux et al. 2005). Given that mesalamine is an effective first-line therapy for mild to moderate UC, (Rubin et al. 2019;Nguyen et al. 2016) future studies are warranted to evaluate if mesalamine has a role in ICI-related mucosal injury through its role on local mucosal immune action. In addition, other therapies need further evaluation for their efficacy including the role of pancreatic enzymes for ICI-related diarrhea due to pancreatic exocrine insufficiency and budesonide given its high first pass metabolism with lower systemic absorption (Hughes et al. 2019).
Immunosuppressive therapy with systemic steroids and biologics such as infliximab are known to have significant side effects including the risk of infection. In our study, of the 14 cases treated with mesalamine and/or cholestyramine, there were no cases of infection during or within 30 days of therapy completion. However, this finding is limited by the small cohort, which does not provide a full representation of side effects associated with these drugs. On the contrary, of the 52 patients treated with steroid with or without infliximab, 21% had confirmed infection during therapy or within 30 days of completing steroid or biologic therapy. In 1 3 comparison, an infection rate of 13-15% has been reported previously in patients treated with immunosuppressant after ICI therapy, while only 2% in those who did not receive additional immunosuppressive therapy (Favara et al. 2020;Del Castillo et al. 2016) Furthermore, Wang et al. reported a rate of infection in patients treated with steroids for greater than 30 days for ICI diarrhea and colitis to be approximately 40% vs. 26% for those on steroids for less than 30 days ). These findings highlight the importance of minimizing immunosuppressive therapies when possible with a particular consideration to the length of steroids use, especially among those patients with no objective histologic colitis who generally had mild disease course (Wang et al. 2018b) Furthermore, mesalamine and cholestyramine are generally well tolerated; patients may rarely experience headache, fever, rash, nausea, or vomiting, flatulence, constipation, with rare reports of hepatoxicity, pancreatitis, and interstitial nephritis related to mesalamine therapy (Rubin et al. 2019;Xie et al. 2019;Barkun et al. 2013). In the present study, there were no significant adverse events related to mesalamine and cholestyramine; however, the small cohort of cases limits these findings. Nonetheless, given the favorable safety profile of mesalamine and cholestyramine, future studies may also consider evaluating the role of non-immunosuppressive therapy prophylactically to prevent severe diarrhea and colitis in at risk patients.
The risk associated with immunosuppressive therapy needs to be counterbalanced by the risk of worsening pathologic colitis progression that may be unresponsive to non-immunosuppressive therapy and delay cancer therapy. In our study, 11 of the 52 immunosuppressive cases had a delayed start of steroid treatment after persistent or worsening symptoms on non-immunosuppressive therapy. Eight of the 11 escalated to steroids within 3-5 days and three were over 10 days, which may have been related to insufficient communication to physicians. One case was associated with worsening of symptoms from grade 1 diarrhea and colitis on mesalamine ultimately requiring hospitalization within 30 days. Otherwise, there were no significant major adverse effects or impaired outcome from this delayed treatment transition compared to all cases that received immunosuppressive therapies. The optimal timing and duration of nonimmunosuppressive trial before escalation needs further investigation.
There is a paucity of evidence determining when to resume ICI therapy and its safety in regard to the recurrence of potentially more severe irAEs. ICI diarrhea and colitis recurrence was reported in 34% of cases in a multi-center study, with risk factors including treatment resumption with a CTLA-4 agent, need for immunosuppressive therapy for the index episode of ICI diarrhea and colitis, and longer duration of symptoms (Abu-Sbeih et al. 2019) In comparison, of the 14 patients treated with non-immunosuppressive therapy in our cohort, the overall rate of recurrence was 36%, which was similar to those treated with immunosuppression.
The retrospective nature and small sample size of this study limits the power of our findings and conclusion. We were unable to measure cancer-related outcomes between those treated with non-immunosuppressive and those treated with immunosuppressive treatment given the study's limited power. The limited cases treated with non-immunosuppressive therapy also limit our ability to effectively ascertain the side effect profile, complications and overall survival of this group. Furthermore, our sample size limits our ability to draw conclusions about which subtypes of ICI-mediated histologic colitis may be more responsive to non-immunosuppressive therapy. We suspect that our data are also in part limited by the current practice patterns that are not inclusive of the use of mesalamine and cholestyramine for the treatment of ICI diarrhea and colitis. Given the aforementioned reason, the associations presented in this study are subject to bias related to patient selection for mesalamine and cholestyramine treatment who may have had milder symptoms at the time of evaluation by treating physician. In addition, due to our study design we were unable to control for all the confounding variables adequately (e.g., cancer type, performance status, use of other non-ICI chemotherapy, and colitis recurrence status). The potential difference of these confounding variables between the two study groups, although not as evident currently due to small sample size, could contribute to the dramatic impact on the colitis outcome and alter our conclusion. Future studies are warranted to provide high quality data by inclusion of more comprehensive variables for the outcome measures. Moreover, given the possibility that there is a component of diarrhea symptom with otherwise normal gastrointestinal function such as bile acid diarrhea and pancreatic exocrine insufficiency, we recognize the importance and critical need of future studies further differentiating the etiology of diarrhea and their characteristics in patients exposed to ICI therapy to determine the most appropriate management plan.
Despite these limitations, this study suggests that nonimmunosuppressive therapy may have a therapeutic role in the management of mild ICI diarrhea and colitis. Lastly, in the present study we only comment on the therapeutic use of oral mesalamine formulations with lack of data on the topical options. Future prospective studies could randomize patients diagnosed with CTCAE grade 1-2 diarrhea/ colitis to treatment with a non-IMS regimen such as with mesalamine with different routes of administration (oral and enema) versus steroids following the fecal calprotectin assessment with early endoscopic/histologic evaluation when indicated at time of diagnosis. Clinical outcomes, endoscopic improvement and sides effects profile can also be measured and compared between the two groups.

Conclusion
Non-immunosuppressive therapy with mesalamine and/or cholestyramine could be effective in the management of a select subgroup of patients with grade 1-2 immune checkpoint inhibitor-mediated diarrhea and colitis. In particular, patients who have delayed onset of symptoms after PD-1/ L1 monotherapy and lower colonic inflammatory burden. Prospective studies are warranted to further determine the patient population that can benefit from the treatment of mesalamine and/or cholestyramine, the optimal therapeutic dose, treatment duration and safety.