INTRODUCTION — With the widespread use of immune checkpoint inhibitors (ICI; also known as checkpoint inhibitor immunotherapy) for treatment of various types of cancer, immune-related adverse events (irAEs) are becoming increasingly common.
This topic will review the clinical presentation, diagnosis, and management of ICI colitis. Hepatic, pancreatic, and rare gastrointestinal toxicities of ICI, as well as those involving other organ systems, are presented separately.
●(See "Toxicities associated with checkpoint inhibitor immunotherapy".)
●(See "Mucocutaneous toxicities associated with immune checkpoint inhibitors".)
TERMINOLOGY — ICI diarrhea and colitis are defined by the American Society of Clinical Oncology (ASCO) based on symptoms alone rather than by colonic inflammation that is demonstrated by endoscopic findings (mucosal inflammation with or without ulcers (table 1A-B)) [1]:
●Diarrhea is defined as watery bowel movements with any increase in stool frequency over baseline.
●Colitis involves abdominal pain, mucus or blood in the stools, and fever.
Although the Common Terminology Criteria for Adverse Events (CTCAE, also called "common toxicity criteria") categorizes ICI diarrhea and ICI colitis separately, these conditions (and their management) overlap and they likely represent a spectrum of histopathologic disease [2]. In addition, the terms ICI diarrhea and ICI colitis are often used interchangeably, and the condition has also been referred to as ICI enterocolitis, immune-mediated colitis (IMC), or immune-mediated diarrhea and colitis (IMDC).
For patients who are on ICI therapy alone, clinical variability exists for the grading of these toxicities. At some institutions, each toxicity is graded and reported separately (ie, one grade for diarrhea symptoms and another for colitis symptoms), while other institutions only report the toxicity with the higher grade.
Of note, ICIs can be combined with either chemotherapy or targeted agents to treat certain cancers. The assessment of diarrhea or colitis symptoms in patients receiving such combination therapy may be different from those receiving ICIs alone, since the diarrhea may be related to the additional systemic agent. (See 'Differential diagnosis' below and "Chemotherapy-associated diarrhea, constipation and intestinal perforation: pathogenesis, risk factors, and clinical presentation" and "Management of acute chemotherapy-related diarrhea".)
INCIDENCE AND RISK FACTORS — ICI colitis is the most frequently reported gastrointestinal immune-related adverse event (irAE). The reported overall incidence of ICI colitis ranges between 1 and 25 percent and varies based on the specific agent, dose, and combination of ICIs used [3]. As an example, cytotoxic T-lymphocyte antigen 4 (CTLA-4) inhibitors or combination regimens (eg, nivolumab plus ipilimumab) are associated with increased risk of developing ICI colitis and of progressing to severe colitis compared with single-agent programmed cell death protein 1 (PD-1) or programmed cell death ligand 1 (PD-L1) inhibitors [4-6]. In addition, studies have suggested that higher doses of CTLA-4 therapy are associated with a greater risk of developing colitis [7-9].
Other factors that have been associated with the risk and severity of ICI colitis include the type of malignancy [10], composition of the intestinal microbiome [11], and lack of vitamin D supplementation [12].
PATHOGENESIS — T cells are critical to the efficacy of ICI therapy. The immune checkpoints cytotoxic T-lymphocyte antigen 4 (CTLA-4) and programmed cell death protein 1 (PD-1) both attenuate T cell activation though distinct mechanisms. PD-1 sends a direct inhibitory signal through activation of phosphatases after binding to its ligands programmed cell death ligand 1 (PD-L1) or programmed cell death ligand 2 (PD-L2), and CTLA-4 binds the costimulatory ligands CD80 and CD86, preventing them from interacting with the T cell costimulatory receptor CD28. Additionally, CTLA-4 is constitutively expressed on regulatory T cells (Tregs) and is essential for many of their immune suppressive functions [5]. (See "Principles of cancer immunotherapy".)
Complications from ICI therapy arise from blockade of these important regulatory pathways. Although any organ system can be affected by ICI-related toxicities, these toxicities most often affect the skin, gastrointestinal tract, liver, and the respiratory epithelium [13]. Expanded CD8+ T cells and, to a lesser extent, CD4+ T cells have been found in histopathologic analyses from multiple organs affected by ICI toxicity, and this is consistent with findings in tumors [14-16]. In ICI colitis, CD8+ T cells are likely expanded from the mucosal resident memory cells [17,18]. These expanded cells express markers of proliferation and cytotoxicity and produce the cytokine interferon (IFN)-gamma [17,18]. In addition, ICI colitis is associated with an expansion of IFN-gamma producing Th1 CD4+ T cells, further implicating this cytokine in disease pathogenesis [17,18]. IFN-gamma overexpression is not unique to colitis from ICI as it appears to be critical for antitumor responses as well [19,20]. Tregs are expanded in ICI colitis, though they likely have decreased function, particularly in the setting of CTLA-4 inhibitors [17,18].
The targets of the T cell response in the gastrointestinal tract are not known, but the diversity of T cells identified is consistent with recognition of the microbiome [18]. Alterations in the microbiome have been observed in patients who are susceptible to colitis from ICI therapy, and modulation of the gut microbiome is the basis for studies that use fecal microbiota transplant (FMT) for patients with ICI colitis who do not respond to initial therapy [21,22]. (See 'Nonresponders to initial biologic agent' below.)
In addition to the expansion of resident memory cells, ICI colitis is associated with an inflammatory macrophage population that is responding to and producing tumor necrosis factor (TNF)-alpha and producing the chemokines CXCL9 and CXCL10, both of which likely recruit additional T cells from the blood into the colon, further exacerbating colitis [18]. The importance of these findings is underscored by the response to anti-TNF agents, as well as to inhibitors of T cell trafficking such as vedolizumab.
Precisely how each of these factors contributes to epithelial cell damage in ICI colitis is uncertain. However, both TNF-alpha and IFN-gamma may cause cell dysfunction and increased susceptibility to apoptosis. Effector CD8+ T cells may also be causing direct cytotoxicity.
CLINICAL PRESENTATION
Patient presentation
●Timing of symptom onset — For most patients with ICI colitis, symptoms develop within a few weeks to a few months after starting ICI therapy [4,5,23]. Although colitis can present at any time during the course of ICI therapy, it rarely develops extremely early (ie, within the first few days) or extremely late (ie, years) after the initiation of therapy.
●Symptoms — Patients with ICI colitis usually present with diarrhea (ie, loose stools with an increase in stool frequency over baseline (table 1A-B)), although the frequency of diarrhea can vary clinically. In those with more severe disease, diarrhea may be associated with cramping abdominal pain, urgency, and rectal bleeding. Some patients report symptoms such as nausea, vomiting, and early satiety that suggest upper gastrointestinal tract involvement (ie, gastritis and/or enteritis).
Less typical symptoms include fever, abdominal distension, and/or severe abdominal pain. These symptoms should raise suspicion for severe disease and/or a possible complication such as colonic perforation or toxic megacolon. (See "Overview of gastrointestinal tract perforation" and "Toxic megacolon".)
Studies on ICI colitis use an oncology-based grading system called the Common Terminology Criteria for Adverse Events (CTCAE), which relies solely on symptoms and classifies toxicities on a scale from grade 1 (mild) to grade 5 (death (table 1A-B)) [2].
Physical examination — Physical examination may be normal in patients with mild disease, whereas patients with moderate to severe disease may have abdominal tenderness with or without abdominal distension. The presence of guarding, rigidity, or rebound tenderness on examination suggests the presence of a complication such as colonic perforation or toxic megacolon. (See "Evaluation of the adult with abdominal pain".)
Laboratory findings — Routine laboratory tests may be normal, or they may reveal an elevated white blood cell count, an elevated C-reactive protein, or electrolyte abnormalities. Stool inflammatory markers (fecal calprotectin or lactoferrin) may be elevated due to intestinal inflammation [24].
DIAGNOSTIC EVALUATION — The diagnosis of ICI colitis should be suspected in patients who are receiving ICI therapy and who develop diarrhea with or without bleeding or abdominal pain.
Goals of the diagnostic evaluation for patients with suspected ICI colitis are to exclude other causes of colitis and establish the diagnosis. The evaluation should occur promptly (ie, within a few days after symptoms are reported) because colitis may progress quickly without treatment.
The diagnosis of ICI colitis is established clinically by assessing symptoms according to CTCAE and excluding other causes (eg, infection). The diagnosis may be further confirmed with endoscopic evaluation with biopsy [2].
Approach to testing — The approach to diagnostic testing is informed by severity of symptoms and suspicion for disease progression:
●Laboratory studies – For all patients in whom ICI colitis is suspected, laboratory studies including blood tests and stool studies are obtained. (See 'Laboratory studies' below.)
●Endoscopic evaluation – For patients with grade 2 to 4 symptoms and/or an elevated stool inflammatory marker, we typically perform a lower endoscopy (eg, flexible sigmoidoscopy) with biopsies (table 1A-B). While flexible sigmoidoscopy is the initial test of choice, some contributors start with an ileocolonoscopy to also assess for right-sided colitis or ileitis. For patients with a normal flexible sigmoidoscopy or colonoscopy and persistent symptoms, subsequent testing includes upper esophagogastroduodenoscopy (EGD). (See 'Endoscopic evaluation' below.)
●Imaging studies – For patients with a suspected complication (eg, those with fever, severe abdominal pain, abdominal distension, and/or rectal bleeding), we perform cross-sectional imaging (computed tomography [CT] of the abdomen and pelvis) prior to endoscopic evaluation to exclude complications of ICI colitis (eg, bowel perforation or toxic megacolon) as well as potential cancer progression. (See 'Imaging studies' below.)
Laboratory studies
●Blood tests –We obtain complete blood count, complete metabolic panel (CMP), thyroid stimulating hormone (TSH), tissue transglutaminase (TTG) IgA-antibody, total IgA level, and a serum inflammatory marker (eg, C-reactive protein [CRP]). While CRP as a serologic marker of inflammation has largely replaced the erythrocyte sedimentation rate (ESR), some clinicians may only have access to ESR.
For patients with fever, we also obtain blood cultures.
●Stool studies – Stool studies include:
•Testing for enteric pathogens – We typically obtain stool testing for infection (ie, Clostridioides difficile, routine stool cultures [Salmonella, Shigella, Campylobacter, Yersinia], and specific testing for E. coli O157:H7). For patients with risk factors for parasitic infection (eg, recent travel to a region where parasitic infections are endemic), stool microscopy for ova and parasites (three specimens) and a Giardia stool antigen test are also checked. We do not routinely assess for intestinal amebiasis (E. histolytica). Stool analysis for the detection of pathogens is discussed in more detail separately:
-(See "Approach to the diagnosis of cytomegalovirus infection", section on 'Gastrointestinal disease'.)
-(See "Giardiasis: Epidemiology, clinical manifestations, and diagnosis".)
-(See "Approach to the adult with acute diarrhea in resource-rich settings".)
•Stool inflammatory markers – We typically check a stool inflammatory marker (lactoferrin or calprotectin) at initial diagnosis. Elevated stool inflammatory markers can help to confirm a diagnosis of ICI colitis if other causes have also been ruled out. Nevertheless, in patients with elevated markers (regardless of disease severity), lower endoscopy is still necessary to assess for infectious causes of colitis, grade the severity of mucosal inflammation, and clarify the diagnosis if there are conflicting or borderline stool marker results. For patients with grade 1 or 2 symptoms where lower endoscopy is delayed, stool markers can also be used to noninvasively assess for colonic inflammation. (See 'Endoscopic evaluation' below.)
-Elevated stool markers – Calprotectin is borderline elevated if levels are between 50 and 120 mcg/g and elevated if greater than 120 mcg/g [25]. Lactoferrin is elevated when positive for inflammation, although some laboratories provide quantitative results. If the stool marker is elevated (in particular, lactoferrin due to its high sensitivity), colonic inflammation is likely to be found on lower endoscopy. In one study including 182 patients with suspected ICI colitis, the sensitivity for elevated fecal lactoferrin to predict endoscopic or histologic inflammation was 70 and 90 percent, respectively [24].
-Normal stool markers – For patients whose stool markers are normal (calprotectin levels <50 mcg/g or negative lactoferrin), we typically defer lower endoscopy (especially if symptoms have completely resolved). However, the absence of elevated stool inflammatory markers does not exclude the possibility of diarrhea due to ICI colitis or ICI enteritis. [26].
Stool inflammatory markers can also be used for subsequent monitoring of treatment response. (See 'Monitoring patients with symptomatic remission' below.)
Imaging studies — In patients who are suspected to have a complication (eg, fever, severe abdominal pain, abdominal distension, and/or rectal bleeding), we perform cross-sectional imaging (CT scan of the abdomen and pelvis). In such patients, abdominal CT can exclude complications of ICI colitis as well as disease progression [27]. (See 'Complications' below.)
Patients on ICI therapy may be incidentally diagnosed with colitis on routine CT imaging obtained during cancer restaging. Such patients should also be appropriately evaluated for potential etiologies of the colitis, including ICI colitis. (See 'Differential diagnosis' below and 'Diagnostic evaluation' above.)
When imaging features of ICI colitis are present, they typically include colonic inflammation demonstrated by wall thickening and mucosal hyperenhancement in a diffuse or segmental distribution [27]. However, for most patients, CT does not demonstrate specific features of ICI colitis, and studies have shown poor correlation between imaging and endoscopic findings [26,28].
Endoscopic evaluation — For patients with grade 2 to 4 symptoms (table 1A-B), we typically perform a lower endoscopy (eg, flexible sigmoidoscopy) to evaluate for mucosal disease and obtain biopsies to exclude infection (eg, cytomegalovirus). (See "Approach to the diagnosis of cytomegalovirus infection", section on 'Gastrointestinal disease'.)
Because most patients have left colon involvement, a flexible sigmoidoscopy is the initial test of choice rather than ileocolonoscopy because the diagnosis can be established while minimizing the risk of colonic dilation and perforation [29,30]. Alternatively, some contributors start with an ileocolonoscopy, as the incidence of isolated right-sided colitis or ileitis is approximately 10 percent. Patients with persistent diarrhea and a normal flexible sigmoidoscopy or colonoscopy should be evaluated with EGD.
Endoscopic features of ICI colitis are usually nonspecific and range from a normal endoscopic appearance to mucosal inflammation with or without ulceration (picture 1).
Although there is no consensus for the approach to grading the endoscopic severity of ICI colitis, the Mayo Clinic ulcerative colitis (UC) endoscopic scoring system is frequently used in clinical practice (calculator 1) [31]. In addition, the following endoscopic features are associated with increased risk for disease that does not respond to glucocorticoids [32], overall worse outcomes, and the need for second-line therapy to achieve clinical improvement [24,26]:
●Large ulcers >1 cm
●Deep ulcers >2 mm
●Multiple ulcers (3 or more)
●Extensive colitis (ie, mucosal involvement proximal to the left colon)
During lower endoscopy, biopsies are routinely obtained from the left colon and rectum even if endoscopically normal in appearance to assess for histologic evidence of inflammation. Histologic findings in most patients with acute ICI colitis include neutrophil and/or eosinophil infiltration, epithelial apoptosis, cryptitis, and crypt abscesses. Some patients have histologic features of chronic colitis such as crypt architectural distortion, basal lympho-plasmocytosis, granulomas, and Paneth cell metaplasia [24,26].
Some patients with ICI colitis have minimal mucosal inflammation without ulceration but have more predominant histologic inflammation [33-35]. In addition, some patients lack both endoscopic and histologic inflammation on initial endoscopic evaluation [24]. Such patients may have either enteritis (inflammation of the small intestine) which can be confirmed with EGD and duodenal biopsies or with ileocolonoscopy and ileal biopsies; an alternative cause of diarrhea; or an uncommon condition known as "noncolitis ICI-related diarrhea" that is characterized by ICI-related diarrhea without colitis [26].
There is also a subset of patients with ICI colitis who have histologic features of microscopic colitis (lymphocytic infiltration) with or without visible mucosal inflammation [34,35], and such patients are discussed below. (See 'ICI-mediated microscopic colitis' below.)
DIFFERENTIAL DIAGNOSIS — The differential diagnosis for acute ICI colitis is broad and includes etiologies that share a similar clinical presentation and overlapping symptoms (diarrhea, abdominal pain).
●Infectious colitis or enterocolitis – Patients with infectious colitis or enterocolitis (small bowel inflammation) typically report the acute onset of diarrhea, and stool studies are performed to evaluate for enteric pathogens (eg, Clostridioides difficile, Salmonella). (See "Approach to the adult with acute diarrhea in resource-rich settings".)
In immunocompromised patients, cytomegalovirus (CMV) infection can mimic ICI colitis or enterocolitis. This can also be an important cause of recurrent/refractory colitis in patients undergoing ICI colitis treatment. (See "Epidemiology, clinical manifestations, and treatment of cytomegalovirus infection in immunocompetent adults", section on 'Gastrointestinal manifestations'.)
●ICI-related enteritis – ICI-related enteritis (involving the small bowel) can also present with symptoms that mimic ICI colitis. ICI enteritis can be evaluated endoscopically with esophagogastroduodenoscopy (EGD) and/or ileocolonoscopy.
●Hyperthyroidism – Patients with hyperthyroidism and/or an endocrinologic immune-related adverse event (irAE) may present with gastrointestinal symptoms that are similar to ICI colitis (loose stools). We routinely screen for thyroid dysfunction by checking thyroid stimulating hormone (TSH) as part of the initial evaluation for suspected ICI colitis. (See "Diagnosis of hyperthyroidism" and "Toxicities associated with checkpoint inhibitor immunotherapy", section on 'Endocrinopathies'.)
●Ischemic colitis – The clinical presentation of ischemic colitis may overlap with ICI colitis and includes lower abdominal pain and bloody diarrhea or hematochezia. The pattern of mucosal involvement on CT scan of the abdomen, endoscopic appearance of the mucosa on lower endoscopy (edematous, friable mucosa; erythema; and interspersed pale areas), and histologic evidence of ischemic colitis can establish the diagnosis of ischemic colitis. (See "Colonic ischemia", section on 'Diagnosis'.)
●Celiac disease – Celiac disease is a small bowel disorder characterized by mucosal inflammation and villous atrophy which occur in response to dietary gluten. Patients with celiac disease may present with gastrointestinal symptoms that are similar to ICI colitis (eg, diarrhea, abdominal pain). Serologic evaluation (ie, tissue transglutaminase-IgA antibody) is typically performed to detect celiac disease in adults. The clinical manifestations and diagnosis of celiac disease are presented separately. (See "Epidemiology, pathogenesis, and clinical manifestations of celiac disease in adults" and "Diagnosis of celiac disease in adults".)
Celiac disease may manifest either as a de novo diagnosis unrelated to ICI therapy or as a rare manifestation of immunotherapy-related toxicity. (See "Hepatic, pancreatic, and rare gastrointestinal complications of immune checkpoint inhibitor therapy", section on 'Rare GI toxicities'.)
●Malignancy involving the GI tract – Some gastrointestinal (GI) malignancies present with diarrhea and/or rectal bleeding. Examples include intestinal involvement from the primary cancer or metastatic disease (eg melanoma metastases to the GI tract) [36], or a separate primary cancer that could affect the colon (eg, colorectal cancer, neuroendocrine tumors, gynecologic malignancies). (See "Imaging studies in melanoma", section on 'Gastrointestinal tract'.)
●Pancreatic insufficiency – Patients with pancreatic insufficiency may report chronic diarrhea with pale, greasy, voluminous, foul-smelling stools and unintentional weight loss despite adequate food intake. Abdominal pain is unusual. We screen for maldigestion due to pancreatic insufficiency with fecal elastase. (See "Approach to the adult patient with suspected malabsorption".)
●Irritable bowel syndrome – Patients with undiagnosed irritable bowel syndrome (IBS) may have the condition triggered by the stress of cancer treatment. Such patients may be incidentally diagnosed with this condition during the assessment for ICI colitis. (See "Clinical manifestations and diagnosis of irritable bowel syndrome in adults".)
●Diarrhea due to chemotherapy or antiangiogenic agents administered with ICI – Patients who are treated with a combination of ICI and chemotherapy may present with diarrhea that is specifically related to chemotherapy alone. Chemotherapy-related diarrhea often presents with a waxing and waning pattern of increased bowel movements and/or loose stools that correlates with treatment cycles. This pattern can help to distinguish it from ICI diarrhea or ICI colitis, which typically present with persistent diarrhea that occurs independently of administering or holding subsequent ICI cycles. For patients who undergo endoscopic evaluation and biopsy, chemotherapy-related diarrhea demonstrates a pattern of toxic injury with minimal inflammation on histology. (See "Chemotherapy-associated diarrhea, constipation and intestinal perforation: pathogenesis, risk factors, and clinical presentation" and "Management of acute chemotherapy-related diarrhea".)
Immunotherapy can also be administered in combination with targeted therapy (eg, antiangiogenic therapy in patients with advanced or renal cell carcinoma). Symptoms of diarrhea or colitis in patients on this combination therapy is more commonly due to the antiangiogenic agent rather than the immunotherapy. (See "Chemotherapy-associated diarrhea, constipation and intestinal perforation: pathogenesis, risk factors, and clinical presentation", section on 'Risk with molecularly targeted agents'.)
For patients on combination therapy with symptoms of diarrhea, our initial approach is to hold the antiangiogenic agent and provide supportive care. In patients whose symptoms do not improve after holding antiangiogenic therapy, diarrhea due to ICI colitis should be suspected. For such patients who are being evaluated for possible immunosuppressive therapy, we obtain endoscopic evaluation with biopsies. Diarrhea due to targeted agents typically reveals a lack of toxic mucosal injury on endoscopic evaluation; however, biopsies are still necessary because the absence of mucosal injury does not exclude the possibility of ICI colitis.
MANAGEMENT
Goals — The goals of treatment for patients with ICI colitis are resolution of symptoms, avoiding complications, and either continuing or reintroducing ICI therapy when possible. (See 'Resuming immune checkpoint inhibitors' below.)
Our approach to therapy for ICI colitis is informed by the Common Terminology Criteria for Adverse Events (CTCAE) grade severity, which is consistent with guidelines from the American Society of Clinical Oncology (ASCO) [37]. For patients with different grading for diarrhea and colitis, we use the higher grade to guide treatment decisions.
Mild disease (grade 1) — We treat patients with grade 1 symptoms (ie, increase of ≤3 watery stools daily over baseline and no abdominal pain (table 1A-B)) with supportive care while continuing ICI therapy because grade 1 symptoms are generally mild and self-limiting. We do not initiate systemic glucocorticoids in these patients.
Supportive care includes:
●Diet – We advise patients to maintain adequate oral hydration and adhere to a bland, lactose-free diet with no/reduced caffeine.
●Symptomatic therapy – For patients with stool studies that are negative for infection, antidiarrheal agents can be used as needed as a temporary measure in patients with diarrhea only and no colitis-related symptoms. We typically use loperamide 2 to 4 mg every four hours as needed, not to exceed 16 mg in 24 hours.
However, some contributors avoid the use of antidiarrheal agents for all patients due to the risk of toxic megacolon. In this situation, alternatives to loperamide include cholestyramine 4 grams twice daily.
●ICI therapy – ICI therapy is typically continued. However, holding ICI therapy and resuming it after symptoms resolve may be a reasonable option that is individualized and informed by oncologist preference and indication for ICI therapy.
Patients with grade 1 symptoms who do not respond to supportive care are managed similarly to those with grade ≥2 symptoms. (See 'Moderate to severe disease (grade 2 to 4)' below.)
Moderate to severe disease (grade 2 to 4)
Initial management — For patients with moderate to severe disease (grade 2 to 4 symptoms (table 1A-B)), it is important to initiate therapy promptly, as symptoms and disease severity can escalate rapidly without treatment. Initial measures include (algorithm 1):
●Discontinue ICI therapy – We hold ICI therapy temporarily for grades 2 and 3 symptoms and typically discontinue it permanently for grade 4 symptoms [37-39]. ICI antibodies are still likely to exert substantial biologic activity for weeks to months after stopping treatment, unless immunosuppressive therapy is administered.
●Obtain gastroenterology consultation
●Initiate systemic glucocorticoid therapy – We confirm that gastrointestinal infection has been excluded based on stool studies prior to initiating systemic glucocorticoid therapy. (See 'Laboratory studies' above.)
The route of administration for glucocorticoid therapy is informed by the disease severity and need for inpatient hospitalization:
•For patients with grade 2 or 3 diarrhea but without systemic symptoms (eg, fever, hypotension) and who are tolerating oral intake (ie, fluids and bland diet), we suggest initiating oral glucocorticoid therapy. We begin oral prednisone at 1 mg/kg/day (with no specific maximum dose) or an equivalent agent. When symptoms improve to grade 1 diarrhea (<4 loose stools daily over baseline), we gradually taper prednisone by 10 mg every five to seven days with the goal of discontinuing prednisone over four to six weeks (table 2). Glucocorticoids should not be used long-term due to significant side effects. (See "Major side effects of systemic glucocorticoids".)
Patients who do not have clinical improvement with oral prednisone in three days are escalated to intravenous glucocorticoid therapy or biologic therapy. Hospitalization of these patients is often necessary for prompt treatment.
•Patients with grade 4 symptoms and patients with grade 2 or 3 symptoms who have systemic symptoms (eg, fever, hypotension) and/or are not tolerating oral intake require inpatient hospitalization. For such patients, we give methylprednisolone 1 to 2 mg/kg daily intravenously in two divided doses every 12 hours with a maximum total daily dose of 120 mg.
Most patients who respond to intravenous glucocorticoids will have symptomatic improvement (ie, fewer stools, less abdominal pain) with a goal of achieving grade 1 symptoms within three days after starting therapy. For patients who have no clinical improvement within three days of starting systemic glucocorticoids, we initiate second-line therapy with either infliximab or vedolizumab. (See 'Subsequent management (biologic agents)' below.)
For patients who respond (ie, improvement to grade 1 symptoms), we discontinue intravenous glucocorticoids and begin an equivalent dose using an oral formulation such as prednisone prior to hospital discharge.
Hospitalized patients are reassessed daily to determine if they are eligible to be discharged. Patients must meet all criteria listed below before they can be discharged from the hospital:
•Normalization of vital signs (ie, resolution of fever, hypotension)
•<4 stools per day over baseline (grade 1 symptoms)
•Resolution of severe abdominal pain
•Tolerance of oral diet
The use of systemic glucocorticoids is based on retrospective series [10,40-42], clinical experience, and expert opinion [37]. Randomized trials evaluating glucocorticoid therapy for patients with ICI colitis are lacking.
Subsequent management (biologic agents) — For patients who do not have symptomatic response within three days of receiving intravenous glucocorticoid therapy or those with high-risk endoscopic features (eg, colonic ulcers with Mayo Endoscopic score 3 or greater (calculator 1)), we suggest the addition of a biologic agent (either infliximab or vedolizumab) to glucocorticoid therapy rather than glucocorticoids alone or other agents. In hospitalized patients with glucocorticoid refractory ICI colitis, data suggest that efficacy is comparable between these two agents [43]. Therefore, the selection of infliximab versus vedolizumab is based on clinician preference, patient comorbidities, the type of cancer, and the need for rapid treatment response.
We obtain pretreatment screening prior to initiating biologic agents, and this is discussed separately. (See "Overview of dosing and monitoring of biologic agents and small molecules for treating ulcerative colitis in adults", section on 'Pretreatment screening'.)
●Infliximab – For most patients, we initiate infliximab at an initial dose of 5 mg/kg intravenously, at zero, two, and six weeks. Patients with grade 4 colitis and severe hypoalbuminemia (serum albumin <2.5 g/dL) may be treated with an initial dose of 10 mg/kg. While most UpToDate contributors administer three induction doses of infliximab, others do not administer all three doses unless there are persistent symptoms.
Infliximab is preferred for patients who require a rapid treatment response and those with concurrent inflammatory arthritis. Patients can demonstrate a partial clinical response (ie, fewer stools, tolerating oral intake) to infliximab as early as 72 hours and generally within less than one week.
Pretreatment screening, dosing, monitoring, and adverse effects associated with anti-tumor necrosis factor (TNF) agents are discussed separately:
•(See "Tumor necrosis factor-alpha inhibitors: An overview of adverse effects".)
Contraindications to the use of anti-TNF therapies (briefly summarized) include the following (see "Treatment of axial spondyloarthritis (ankylosing spondylitis and nonradiographic axial spondyloarthritis) in adults", section on 'Use of TNF inhibitors'):
•Active uncontrolled infection
•Latent (untreated) tuberculosis
•Demyelinating disease (eg, multiple sclerosis, optic neuritis)
•Heart failure
•Although anti-TNF therapies may be associated with a risk of cancer, we do not avoid infliximab in patients with hematologic malignancies if it is indicated for the treatment of ICI colitis and if administered over a limited period of time. (See "Tumor necrosis factor-alpha inhibitors: Risk of malignancy".)
●Vedolizumab – We typically use standard induction dosing of vedolizumab (300 mg, intravenously, at zero, two, and six weeks). While most UpToDate contributors administer three doses of vedolizumab, others do not administer all three doses unless there are persistent symptoms. Patients typically demonstrate a clinical response (ie, fewer stools, tolerating oral intake) within one to two weeks. Vedolizumab is preferred in patients with hematologic malignancies or for those with an ongoing tumor response to ICI therapy where the impact of systemic immunosuppression needs to be minimized. It can also be used in those who are able to tolerate a slightly slower treatment response, since the clinical response for vedolizumab is less rapid than that seen for infliximab. However, it is important to note that patients with ICI colitis respond faster than those with other forms of inflammatory bowel disease (IBD).
Vedolizumab is a gut selective anti-integrin agent that is not generally associated with systemic immunosuppression that is seen with anti-TNF agents. Pretreatment screening and adverse effects associated with vedolizumab are discussed separately. (See "Overview of dosing and monitoring of biologic agents and small molecules for treating ulcerative colitis in adults", section on 'Anti-integrin antibodies'.)
Most patients who respond to the induction dosing regimen of a biologic agent do not require long-term biologic therapy. In addition, while most guidelines suggest a glucocorticoid taper over four to six weeks, one contributor shortens the glucocorticoid taper to three to four weeks to decrease the risk of adverse effects related to concurrent glucocorticoid use and biologic therapy (eg, infection) [26,44].
Observational studies suggest that infliximab and vedolizumab were similarly beneficial for patients with ICI colitis [45-47]. In a study including 184 patients with ICI colitis who received glucocorticoid therapy followed by a biologic agent, rates of clinical remission were not significantly different for patients treated with vedolizumab compared with infliximab (89 versus 88 percent) [45]. Although vedolizumab was associated with shorter duration of glucocorticoid therapy compared with infliximab (35 versus 50 days), the time to achieve clinical response was longer with vedolizumab (17.5 versus 13 days).
Nonresponders to initial biologic agent — For patients with ICI colitis who do not respond to glucocorticoids plus the initial biologic agent (ie infliximab or vedolizumab), we typically switch to a biologic agent from a different class (ie, from infliximab to vedolizumab or vice versa) [47].
For patients who do not respond to first- or second-line drug therapy, the efficacy and safety of other agents has not been well established [31]. However, limited data from case reports suggest that subsequent options may include ustekinumab [48], tofacitinib [49,50], or fecal microbiota transplantation [22,51].
Resuming immune checkpoint inhibitors — For patients with grade 2 or 3 ICI colitis who achieved symptomatic remission, the decision to reintroduce ICI therapy and selection of ICI agent is individualized and requires a multidisciplinary approach with input from specialists in oncology and gastroenterology.
For patients with grade 4 colitis, we recommend permanently discontinuing ICI therapy in most cases.
The approach to retreatment for patients with history of ICI toxicity is discussed separately. (See "Toxicities associated with checkpoint inhibitor immunotherapy", section on 'Retreatment after prior toxicity'.)
For patients who resume ICI therapy, the overall risk of recurrent ICI diarrhea and colitis is approximately 34 percent, including grade ≥3 diarrhea (5 percent), and grade ≥2 colitis (20 percent) [52]. Risk factors associated with recurrence include the use of an anti-cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) inhibitor as the resumed agent and history of severe colitis that required immunosuppressive therapy to achieve remission.
In contrast, limited data suggest that using a biologic agent (eg, infliximab and vedolizumab) while resuming ICI therapy was associated with a lower risk of recurrent ICI colitis [53,54]. This approach remains investigational, and we await further studies prior to using it routinely in clinical practice.
If ICI colitis occurs with ICI rechallenge, the management is similar to treatment of the index episode.
Monitoring patients with symptomatic remission — For most patients who achieved symptomatic improvement (ie, grade 1 symptoms or less) with pharmacologic therapy and for whom ICI therapy will be resumed, we typically repeat endoscopic evaluation to assess for mucosal healing:
Selecting a procedure — The choice of procedure is based on the prior sites of inflammation. For example, patients who had left-sided colitis on initial lower endoscopy can be monitored with flexible sigmoidoscopy. Patients with isolated right-sided colitis or ileitis are monitored with ileocolonoscopy. Patients with isolated small bowel disease are monitored with esophagogastroduodenoscopy (EGD) and small bowel biopsies or with ileocolonoscopy with ileal biopsies, depending on the initial disease distribution.
Procedure timing — The timing of follow-up endoscopic evaluation is based on the treatment used. For those treated with glucocorticoids alone, we repeat endoscopic evaluation approximately two to four weeks after patients have discontinued glucocorticoid therapy. For those treated with biologic therapy, we repeat endoscopic evaluation after three doses of biologic therapy, or approximately eight weeks.
Postprocedure follow-up — The approach to postprocedure management is determined by endoscopic findings and the therapy used to achieve symptomatic improvement:
●Endoscopic remission – Most patients with complete endoscopic remission (ie, no ulcers, no active inflammation) who were previously on biologic agents do not require maintenance therapy.
●Active mucosal disease – For patients with active mucosal disease but who have sustained symptomatic remission after a course of glucocorticoids alone, we monitor and retreat only if symptoms recur.
For those with active mucosal disease who required biologic therapy to achieve symptomatic remission, we typically continue the biologic agent as maintenance therapy every four to eight weeks (with shorter intervals [ie, every four weeks] for those who develop recurrent symptoms or are planning to resume ICI therapy), and we repeat endoscopic evaluation after one to two maintenance doses.
Alternatives to endoscopic evaluation — For patients who cannot tolerate or do not want to undergo surveillance lower endoscopy, a stool inflammatory marker such as fecal calprotectin is a noninvasive alternative [24,25], although clinical practice varies for this approach. Stool inflammatory markers are obtained at initial diagnosis and every two to three months during immunosuppressive therapy for ICI colitis. If the calprotectin value is ≤116 microg/g, we typically taper glucocorticoids and discontinue biologic therapy if it had been initiated. Limited data suggest that stool inflammatory markers correlate with endoscopic disease activity. In one study, the specificity of a calprotectin value ≤116 microg/g to predict endoscopic remission was 94 percent [25].
The goal of endoscopic monitoring is to target endoscopic remission, based on mucosal healing rather than on symptoms alone. Observational studies suggest that achieving endoscopic and histologic remission with pharmacologic therapy (ie, glucocorticoids and ≥3 doses of a biologic agent) was associated with lower risk of disease recurrence [55].
For the rare patient who continues biologic therapy as maintenance, additional monitoring is performed to screen for complications related to immunosuppression. Laboratory monitoring for patients on infliximab or vedolizumab includes complete blood count, complete metabolic panel, and C-reactive protein (CRP) with each infusion, and checking hepatitis B virus (HBV) serologies and interferon-gamma release assay annually. Colonoscopy is performed after one year of therapy and then periodically, based on symptoms.
COMPLICATIONS — If left untreated or not recognized promptly, patients with ICI colitis can develop severe disease and serious complications [6]:
●Toxic megacolon — Toxic megacolon is a potentially lethal complication of ICI colitis, that is characterized by nonobstructive colonic dilatation plus systemic toxicity. The diagnosis and management of toxic megacolon is discussed separately. (See "Toxic megacolon".)
●Colonic perforation — Patients with ICI colitis who develop colonic perforation require colectomy, although this complication is rare.
SPECIAL POPULATIONS
ICI-mediated microscopic colitis — Some patients with ICI colitis have histologic features that resemble lymphocytic or collagenous colitis, and this condition is referred to as ICI-mediated microscopic colitis [33-35]. Patients with ICI-mediated microscopic colitis can present with a wide range and severity of clinical symptoms. Endoscopic examination may demonstrate visible mucosal inflammation which is not seen in patients with non-ICI microscopic colitis [33-35]. (See "Microscopic (lymphocytic and collagenous) colitis: Clinical manifestations, diagnosis, and management", section on 'Diagnostic approach'.)
Initial pharmacologic management typically includes an oral glucocorticoid (ie, oral budesonide or prednisone). For patients who are discontinuing ICI therapy, we treat with budesonide 9 mg daily for six weeks, followed by a three-week taper. For patients who remain on ICI therapy, we continue budesonide 9 mg daily indefinitely (table 3). (See "Microscopic (lymphocytic and collagenous) colitis: Clinical manifestations, diagnosis, and management", section on 'Initial approach'.)
Pre-existing chronic colitis
●Patients with inflammatory bowel disease – A history of inflammatory bowel disease (IBD) is not an absolute contraindication to treatment with ICIs. Patients with IBD still gain therapeutic benefit from ICI therapy, as data demonstrate cancer-specific survival in this population is similar to that reported in clinical trials of ICIs in the general population [56,57]. However, such patients are at increased risk for GI irAEs or flare of IBD. Therefore, collaborative input from both oncology and gastroenterology is necessary. Further details on the use of ICIs and the risk of irAEs in patients with autoimmune disease are discussed separately. (See "Toxicities associated with checkpoint inhibitor immunotherapy", section on 'Pre-existing autoimmune disease'.)
Patients with pre-existing autoimmune diseases have previously been excluded from clinical trials evaluating ICI therapy. Therefore, management recommendations and colitis outcomes among patients with IBD are mostly based on observational studies [56,58]. As an example, one multicenter retrospective study evaluated the use of ICI therapy in 102 patients with IBD (approximately one-half with a diagnosis of ulcerative colitis and the other half with Crohn disease) [57]. Approximately 43 percent of patients were not on IBD-directed therapy at the time of ICI initiation. Most patients had either melanoma or lung cancer and received programmed cell death protein 1 (PD-1)/ programmed cell death ligand 1 (PD-L1) inhibitors. In this study, the rate of GI irAEs was 41 percent, with 21 percent having severe symptoms. This rate is considerably higher than that noted in other control cohorts without IBD (approximately 11 percent) [57] or in clinical trials of PD-1/PD-L1 inhibitors.
Patients with cancer and IBD who are interested in the use of ICIs may also enroll in clinical trials, where available. A multicenter clinical trial is actively evaluating the use of ICIs in cancer patients with IBD (NCT03816345).
●Patients with microscopic colitis – In addition to IBD, other autoimmune GI diseases have also been recognized. In one retrospective cohort study, among four patients with pre-existing microscopic colitis who were treated with ICI therapy, three patients developed a colitis flare that required holding cancer treatment and the use of immunosuppressants [56].
IS THERE A ROLE FOR PROPHYLAXIS? — We do not use pharmacologic prophylaxis (eg, glucocorticoids or biologic agents) to prevent ICI-related colitis, because data to support this approach is limited [59,60]. Further prospective studies are necessary.
As an example, in a trial of 115 patients with advanced melanoma treated with ipilimumab, the use of budesonide prophylaxis did not result in lower rates of grade ≥2 diarrhea compared with placebo (33 versus 35 percent) [59]. These results are likely generalizable to patients with other types of cancer as well as with use of other ICIs routinely in clinical practice.
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Management of toxicities due to checkpoint inhibitor immunotherapy".)
SUMMARY AND RECOMMENDATIONS
●Clinical presentation – Immune checkpoint inhibitor (ICI) colitis is the most frequently reported gastrointestinal (GI) immune-related adverse event. Patients with ICI colitis usually present with diarrhea of variable clinical frequency. Symptoms usually develop within a few weeks to a few months after starting therapy. (See 'Clinical presentation' above.)
•In patients with more severe disease, diarrhea may be associated cramping abdominal pain, urgency, and rectal bleeding.
•Less typical symptoms such as fever, abdominal distension, and/or severe abdominal pain should raise suspicion for severe disease and/or possible colonic perforation or toxic megacolon. (See 'Complications' above.)
●Diagnosis – The diagnosis of ICI colitis should be suspected in patients who are receiving ICI therapy and who develop diarrhea with or without bleeding or abdominal pain. The diagnosis is established clinically by assessing symptoms according to Common Terminology Criteria for Adverse Events (CTCAE (table 1A-B)) and excluding other causes (eg, infection). The diagnosis may be further confirmed with endoscopic evaluation with biopsies. (See 'Diagnostic evaluation' above.)
•Laboratory testing – For all patients in whom ICI colitis is suspected, laboratory studies including blood tests and stool studies are obtained. (See 'Laboratory studies' above.)
•Lower endoscopy with biopsies – For patients with grade 2 to 4 symptoms and/or an elevated stool inflammatory marker, we typically perform a lower endoscopy (eg, flexible sigmoidoscopy) with biopsies. While flexible sigmoidoscopy is the initial test of choice, some contributors start with an ileocolonoscopy to also assess for right-sided colitis or ileitis. Patients with persistent symptoms but with a normal flexible sigmoidoscopy or colonoscopy can subsequently be evaluated with upper esophagogastroduodenoscopy (EGD). (See 'Endoscopic evaluation' above.)
•Imaging studies – For patients with a suspected complication (eg, those with fever, severe abdominal pain, abdominal distension, and/or bleeding), we perform cross sectional imaging (CT of the abdomen and pelvis) prior to endoscopic evaluation to exclude complications of ICI colitis (eg, bowel perforation or toxic megacolon) as well as potential cancer progression. (See 'Imaging studies' above.)
●Management – The initial management of ICI-mediated colitis is based on severity of disease using the Common Terminology Criteria for Adverse Events (CTCAE (table 1A-B)). For patients treated with ICI therapy alone and with different grading for diarrhea and colitis, we use the higher grade to guide treatment decisions. (See 'Management' above.)
•Mild disease (grade 1) – Patients with mild (grade 1) symptoms are managed with supportive care and may typically continue ICI therapy. We do not initiate systemic glucocorticoids. (See 'Mild disease (grade 1)' above.)
•Moderate to severe disease (grade 2 to 4) – Patients with moderate to severe (grade 2 to 4) symptoms are managed as follows (see 'Moderate to severe disease (grade 2 to 4)' above):
-We hold ICI therapy temporarily for grades 2 to 3 symptoms and typically discontinue it permanently for grade 4 symptoms.
-We confirm that GI infection has been excluded based on stool studies prior to initiating systemic glucocorticoids.
-For patients with grade 2 to 3 diarrhea without systemic symptoms (eg, fever, hypotension) and who are tolerating oral intake, we suggest initiating glucocorticoid therapy (Grade 2C). We begin oral prednisone at a dose of 1 mg/kg by mouth daily) or an equivalent agent.
-Patients who do not have clinical improvement on oral prednisone after three days; those with grade 4 symptoms; and those with grade 2 to 3 symptoms who have systemic symptoms and/or are not tolerating oral intake are treated with intravenous glucocorticoid therapy and typically require hospitalization. We treat with methylprednisolone 1 to 2 mg/kg daily (in two divided doses with a maximum total daily dose of 120 mg (algorithm 1)).
•Subsequent management – For patients who do not have symptomatic improvement within three days of initiating glucocorticoids or for those with high-risk endoscopic features (eg, colonic ulcers with Mayo Endoscopic score 3 or greater (calculator 1)), we suggest the addition of a biologic agent (either infliximab or vedolizumab) to glucocorticoid therapy rather than glucocorticoids alone or other agents (Grade 2C). The choice between infliximab and vedolizumab is based on clinician preference, patient comorbidities, and the type of cancer, as efficacy is comparable between the two agents. (See 'Subsequent management (biologic agents)' above.)
●Monitoring for remission – For most patients who achieved symptomatic improvement (ie, grade 1 symptoms or less) with pharmacologic therapy and for whom ICI therapy will be resumed, we repeat endoscopic evaluation to assess for mucosal healing. (See 'Monitoring patients with symptomatic remission' above.)
