ICI-associated diarrhoea and colitis
Also called: immune-mediated diarrhoea and colitis · IMDC · IR-enterocolitis · checkpoint inhibitor colitis · ICI colitis
The commonest serious ICI toxicity and the one most skewed to CTLA-4 blockade. Diarrhoea and colitis are not the same event and do not occur at the same rate — the gap is what decides whether steroids are indicated at all.
Reported frequency
10% — any-grade diarrhoea, anti-PD-1/PD-L1
of patients treated with anti-PD-1/PD-L1 antibodies · Pooled phase I-IV trials, with no clinically relevant difference between compounds
pooled meta-analysis · 35917654
2% — any-grade colitis, anti-PD-1/PD-L1
of patients treated with anti-PD-1/PD-L1 antibodies · Pooled phase I-IV trials. Note this is one fifth the diarrhoea rate in the same population
pooled meta-analysis · 35917654
16% with combination ICIs, 8% with CTLA-4 inhibitors, 1% with PD-(L)1 inhibitors — colitis by ICI class (SITC)
of patients treated with ICIs, stratified by class · SITC GI chapter, citing pooled trial data
guideline · 34172516
approximately 44% with combination therapy, 36% with CTLA-4 inhibitors, 11% with PD-(L)1 inhibitors — diarrhoea by ICI class (SITC)
of patients treated with ICIs, stratified by class · SITC GI chapter, citing pooled trial data
guideline · 34172516
53% grade 3-4 diarrhoea and 32% grade 3-4 colitis — grade 3-4 disease at presentation, endoscopy cohort
of 182 patients endoscoped for immune-mediated diarrhoea and colitis · Single-centre endoscopy cohort. A referred, endoscoped population is enriched for severity and is not the denominator the trial and guideline rates above use
cohort · 30253811
33% — any-grade diarrhoea, ipilimumab
of patients treated with ipilimumab · Pooled phase I-IV trials
pooled meta-analysis · 35917654
7% — any-grade colitis, ipilimumab
of patients treated with ipilimumab · Pooled phase I-IV trials
pooled meta-analysis · 35917654
21%-37% diarrhoea and 4%-8% colitis — any-grade diarrhoea and colitis, ipilimumab plus nivolumab
of patients treated with ipilimumab combined with nivolumab · Pooled phase I-IV trials; the range reflects the regimen used
pooled meta-analysis · 35917654
17%-56% — any-grade diarrhoea, ICI combined with chemotherapy or a TKI
of patients receiving an ICI with chemotherapy or a tyrosine kinase inhibitor · Pooled trials. The corresponding severe colitis rate is 0.5%, and that gap is the point: most of this diarrhoea is not immune colitis
pooled meta-analysis · 35917654
approximately 1% — intestinal perforation among patients with colitis
of patients with ICI-induced colitis · ASCO GI chapter. A complication rate, not a mortality rate
guideline · 34724392
0.5% — severe (grade 3 or higher) colitis, ICI combined with chemotherapy or a TKI
of patients receiving an ICI with chemotherapy or a tyrosine kinase inhibitor · Pooled trials
pooled meta-analysis · 35917654
Severity and class
colitis 11% / 5% / 1% and diarrhoea 10% / 8% / 1% with combination, CTLA-4 and PD-(L)1 therapy — grade ≥3 disease, by ICI class
of patients treated with ICIs, stratified by class · SITC GI chapter, citing pooled trial data — a population rate, not a referral-cohort one
guideline · 34172516
33% diarrhoea and 7% colitis with ipilimumab, versus 10% and 2% with anti-PD-1/PD-L1 — diarrhoea and colitis, CTLA-4 versus PD-1/PD-L1 blockade
of patients in ICI trials pooled by agent class · Systematic review and meta-analysis of phase I-IV trials
pooled meta-analysis · 35917654
Onset
Median 7 weeks from initiation in an endoscopy cohort, and ASCO puts the usual range at 5-10 weeks — but symptoms can occur or recur months after the ICI is stopped. When an ICI is given with cytotoxic chemotherapy, a separate early-onset pattern within 1-2 weeks is described, and attributing that one correctly is what decides the treatment.
Presentation
- · Frequent or loose stools, ranging to frank colitis with mucus in the stool, abdominal pain, fever and rectal bleeding
- · Descending colon involved more often than other segments
- · On CT: mesenteric vessel engorgement, bowel wall thickening, fluid-filled colonic distension — findings that overlap inflammatory bowel disease
Differential
- · Infection, which must be excluded before immunosuppression: stool ova and parasites, C. difficile, CMV by PCR, and other infectious studies at grade 2 or above
- · Chemotherapy- or TKI-related diarrhoea, especially the early-onset pattern within 1-2 weeks of starting a combination — the pooled diarrhoea rate on those regimens reaches 56% while severe colitis stays at 0.5%
- · Concomitant NSAID use, reported to increase ICI-induced enterocolitis
- · Inflammatory bowel disease, which shares both the clinical presentation and the radiologic findings
Work-up
Full blood count, comprehensive metabolic panel and faecal lactoferrin · All patients
SITC's grade 1 work-up. Positive faecal lactoferrin and calprotectin >150 μg/g predicted inflammation in a 182-patient cohort — lactoferrin detected histologic inflammation with 90% sensitivity — but SITC states this evidence is not strong enough to use these tests exclusively: a negative result does not excuse endoscopy.
34172516, 30253811
Faecal calprotectin and stool infectious studies including C. difficile and CMV PCR · Grade 2+
The infectious exclusion. Immunosuppressing an infectious colitis is the foreseeable harm, and this is the step that prevents it. SITC names C. difficile and CMV explicitly, alongside stool ova and parasites.
34172516
Flexible sigmoidoscopy or colonoscopy with biopsy · Grade 2+
SITC performs endoscopy at grade ≥3, or at grade 2 persisting ≥5 days. It grades severity objectively rather than by symptom count, and identifies the high-risk features — ulcers deeper than 2 mm, larger than 1 cm, or extensive colonic involvement — that predict needing a biologic. Endoscopy within 7 days of onset was associated with shorter symptom duration (P = 0.026), though the steroid-duration difference did not reach significance (P = 0.053); beyond 30 days it was associated with longer steroids and more recurrence.
34172516, 30253811
Abdominal CT · If a complication is suspected
For a suspected complication — perforation or toxic megacolon — not to make the diagnosis. ESMO recommends against CT for diagnosing enterocolitis on grounds of insufficient sensitivity, and the two positions are compatible only if the question being asked is stated.
34172516
Sources differ here
Work-up
- SITC 2021 (society guideline) — Abdominal CT scan should be obtained in patients with signs and symptoms of colitis complications, such as bowel perforation or toxic megacolon
- ESMO 2022 (society guideline) — A CT scan to diagnose IR-enterocolitis is not recommended because of insufficient sensitivity
Taper
- SITC 2021 (society guideline) — Corticosteroids should be tapered within 4 weeks after improvement of diarrhea or colitis symptoms to grade ≤1 — a ceiling
- ASCO 2021 (society guideline) — Corticosteroids should be tapered over the course of at least 4-6 weeks — a floor, and ASCO's cross-toxicity rule
Reported frequency
- SITC 2021 (society guideline) — Colitis occurs in 16% (11% grade ≥3) of patients treated with combination ICIs; diarrhea in approximately 44% (10% grade ≥3)
- Nielsen 2022 pooled meta-analysis (primary evidence) — 4%-8% colitis and 21%-37% diarrhoea with ipilimumab plus nivolumab, the range reflecting the regimen used
Grade ladder
Grade 1
Continue the ICIOutpatientGrade 1 — increased stool frequency without colitis symptoms.
No corticosteroid at this grade
ESMO treats grade 1 with a low-fibre diet and loperamide and continues the ICI under close supervision; SITC starts corticosteroids only at grade 2 or above. On a combination regimen this rung carries most of the diarrhoea and almost none of the colitis, which is where the meta-analysis warns of overtreatment.
- · Low-fibre diet and loperamide
- · Continue the ICI under close medical supervision
- · Review NSAIDs and any concurrent chemotherapy or TKI as the more likely cause
Escalate when: Progression to grade 2, or any colitis symptom — blood, mucus, fever or abdominal pain.
Grade 2
Hold the ICIOutpatientGrade 2 — increased stool frequency meeting CTCAE grade 2, or diarrhoea accompanied by colitis symptoms.
Corticosteroid 1 mg/kg/day (oral)
Taper within 4 weeks, starting once: Diarrhoea or colitis symptoms improve to grade 1 or lower.
No step size or interval is carried: the source guideline states a total window and a response marker, and does not specify a cadence.
Second line
- Infliximab 5 mg/kg, three doses at 0, 2 and 6 weeks — No response to corticosteroids within 3-5 days, recurrence after the corticosteroid taper, or a severe ulcerative presentation on colonoscopy. ESMO calls it the drug of choice for non-responders with acute, severe colitis. Screen for HBV, HIV and TB first — but do not delay the dose for pending results.
- Vedolizumab 300 mg, three doses at 0, 2 and 6 weeks — Symptoms persisting after the second infliximab dose — SITC holds the third and switches. An indirect meta-analytic comparison found infliximab and vedolizumab equally effective, but SITC states no head-to-head studies exist; ESMO notes a slightly delayed response with vedolizumab. Same HBV, HIV and TB screening applies.
- · Exclude infection before immunosuppressing — stool studies including CMV PCR
- · Endoscopy if symptoms persist 5 days or more, and earlier endoscopy was associated with a shorter steroid course
- · The hold at this rung is ASCO's cross-toxicity rule — ICPi therapy may be suspended for most grade 2 toxicities — not a colitis-chapter statement; SITC's GI bullets address withholding only at grade 3 or above
Sources differ here
Corticosteroid dose
- SITC 2021 (society guideline) — The initial dose of prednisone should be 1 mg/kg/day (or equivalent) for grade 2 diarrhea or colitis
- ASCO 2021 (society guideline) — Once diarrhea symptoms are grade 2 or higher, or accompanied by apparent colitis symptoms, corticosteroid at 1-2 mg/kg is still the first-line treatment option
- ESMO 2022 (society guideline) — Grade 2 colitis should be treated with oral CSs
Work-up
- SITC 2021 (society guideline) — Flexible sigmoidoscopy and/or colonoscopy with biopsy should be performed for patients with diarrhea or colitis symptoms of grade ≥3 or with persistent (≥5 days) diarrhea or colitis symptoms of grade 2
- ESMO 2022 (society guideline) — Flexible sigmoidoscopy or colonoscopy and biopsies in patients treated with ICIs experiencing grade >1 diarrhoea should be carried out
Escalate when: No response to corticosteroids within 3-5 days, recurrence after the taper, or a severe ulcerative presentation on colonoscopy.
Grade 3
Hold the ICIAdmitGrade 3 — severe increase in stool frequency or severe colitis symptoms.
Corticosteroid 1–2 mg/kg/day (oral)
Taper within 4 weeks, starting once: Diarrhoea or colitis symptoms improve to grade 1 or lower.
No step size or interval is carried: the source guideline states a total window and a response marker, and does not specify a cadence.
Second line
- Infliximab 5 mg/kg, three doses at 0, 2 and 6 weeks — No response to corticosteroids within 3-5 days, recurrence after the corticosteroid taper, or a severe ulcerative presentation on colonoscopy. ESMO calls it the drug of choice for non-responders with acute, severe colitis. Screen for HBV, HIV and TB first — but do not delay the dose for pending results.
- Vedolizumab 300 mg, three doses at 0, 2 and 6 weeks — Symptoms persisting after the second infliximab dose — SITC holds the third and switches. An indirect meta-analytic comparison found infliximab and vedolizumab equally effective, but SITC states no head-to-head studies exist; ESMO notes a slightly delayed response with vedolizumab. Same HBV, HIV and TB screening applies.
- · Hospitalise; ESMO specifies intravenous corticosteroids at this grade
- · Endoscopy with biopsy — high-risk features predict needing a biologic
- · SITC permits temporarily withholding rather than discontinuing the ICI at grade 3 or above
Sources differ here
Corticosteroid dose
- SITC 2021 (society guideline) — The initial dose of prednisone should be 1 mg/kg/day (or equivalent) for grade 2 diarrhea or colitis and 1–2 mg/kg/day for grades 3 and 4. Grade 4 diarrhea or colitis should initially be treated with IV corticosteroids
- ESMO 2022 (society guideline) — Grade 3-4 colitis should be treated by hospitalisation, with i.v. CSs [IV, A]
Escalate when: No response within 3-5 days moves to infliximab. Abdominal CT if perforation or toxic megacolon is suspected.
Grade 4
Hold — consider permanent discontinuationAdmitGrade 4 — life-threatening consequences, including perforation or toxic megacolon.
Corticosteroid 1–2 mg/kg/day (intravenous)
Taper within 4 weeks, starting once: Diarrhoea or colitis symptoms improve to grade 1 or lower.
No step size or interval is carried: the source guideline states a total window and a response marker, and does not specify a cadence.
Second line
- Infliximab 5 mg/kg, three doses at 0, 2 and 6 weeks — No response to corticosteroids within 3-5 days, recurrence after the corticosteroid taper, or a severe ulcerative presentation on colonoscopy. ESMO calls it the drug of choice for non-responders with acute, severe colitis. Screen for HBV, HIV and TB first — but do not delay the dose for pending results.
- Vedolizumab 300 mg, three doses at 0, 2 and 6 weeks — Symptoms persisting after the second infliximab dose — SITC holds the third and switches. An indirect meta-analytic comparison found infliximab and vedolizumab equally effective, but SITC states no head-to-head studies exist; ESMO notes a slightly delayed response with vedolizumab. Same HBV, HIV and TB screening applies.
- · Intravenous corticosteroids from the outset
- · Surgical review for perforation or toxic megacolon
- · ASCO recommends permanent discontinuation for grade 4 toxicities in general — its cross-toxicity rule, not a colitis-specific statement
Sources differ here
ICI action
- SITC 2021 (society guideline) — ICIs may be temporarily withheld (instead of discontinued) in patients experiencing grade ≥3 diarrhea or colitis symptoms. These patients may be re-challenged with ICIs if their symptoms are stable (grade ≤1 or baseline) with <10 mg/day of prednisone (or equivalent)
- ASCO 2021 (society guideline) — In general, permanent discontinuation of ICPis is recommended with grade 4 toxicities — ASCO's cross-toxicity rule, not a colitis-specific statement
Escalate when: Already the top rung; failure to respond to a biologic is a repeat-endoscopy and surgical question.
Rechallenge
individualized
A third of patients who resumed an ICI after immune-mediated colitis had it recur, and recurrence was not a mild event: 82% needed immunosuppression and every one of them stopped the ICI permanently. The class matters — recurrence was less frequent after resuming anti-PD-1/PD-L1 than anti-CTLA-4 — and so does the timing, with a median 53 days from resumption to recurrence, which is the window worth watching.
34% overall; 44% after anti-CTLA-4 and 32% after anti-PD-1/PD-L1 — recurrent immune-mediated diarrhoea and colitis after resuming an ICI
of 167 patients who resumed an ICI after immune-mediated colitis · Retrospective multicentre cohort, 32 resuming anti-CTLA-4 and 135 anti-PD-1/PD-L1
cohort · 31163011
- · Symptoms stable at grade 1 or baseline on less than 10 mg/day of prednisone or equivalent
- · Repeat endoscopy before resuming
- · A multidisciplinary decision — ESMO frames GI rechallenge as case-by-case rather than protocolised
Deliberately not carried
- ladder[].steroid.taper step cadence — SITC gives the ceiling and the trigger — tapered within 4 weeks after symptoms improve to grade 1 or lower — and no step size or interval. No anchor guideline states a cadence for any organ. The ceiling itself is stored as a null lower bound rather than a 4-week window, because ASCO states at least 4-6 weeks for the same decision and the two are opposite instructions sharing a digit; see the taper divergence rows.
- ladder[].gradeDefinition — the CTCAE stool-frequency bands — Each rung names its CTCAE grade without the discriminators that define it — stools per day over baseline, incontinence, whether hospitalisation is indicated — which makes the grade 2 definition circular on its face. CTCAE v5.0 carries them, but it has no PubMed record and no anchor chapter reprints them, so entering them would be an uncited number on the field that decides which rung a patient lands on.
- mortality — None of the cited sources reports a case-fatality rate for ICI colitis. The perforation rate carried in the frequency list is a complication rate, not a mortality rate, and is not repurposed as one.
- onset.rangeWeeksLow / onset.rangeWeeksHigh — The cohort reports a median with no dispersion, and ASCO's 5-10 weeks is a modal window rather than the spread of that median. Both are stated in the note in the terms their sources used.
- guidelinePmids — Empty although all three anchors are rows in guidelines.ts: their stored keyRecommendation strings are renal-scoped, so linking them would surface kidney advice on a colitis card. The GI chapters are carried in citations with their own quotes.
Sources
- Nielsen DL, Juhl CB, Chen IM, Kellermann L, Nielsen OH (2022) Immune checkpoint Inhibitor-Induced diarrhea and Colitis: Incidence and Management. A systematic review and Meta-analysis.Cited for: Class-stratified incidence of diarrhoea versus colitis, the divergence between the two when an ICI is combined with chemotherapy or a TKI, and the equivalence of infliximab and vedolizumab.Supporting text: the PubMed abstract (checkable at the link above)
- Abu-Sbeih H, Ali FS, Luo W, Qiao W, Raju GS, Wang Y (2018) Importance of endoscopic and histological evaluation in the management of immune checkpoint inhibitor-induced colitis.Cited for: Median time to onset, the grade distribution at presentation, what counts as a high-risk endoscopic feature, and the association between early endoscopy and a shorter steroid course.Supporting text: the PubMed abstract (checkable at the link above)
- Abu-Sbeih H, et al. (2019) Resumption of Immune Checkpoint Inhibitor Therapy After Immune-Mediated Colitis.Cited for: Recurrence rate after resuming an ICI, its class split, the time from resumption to recurrence, and what recurrence cost the patients it happened to.Supporting text: the PubMed abstract (checkable at the link above)
- Haanen J, et al. (2022) Management of toxicities from immunotherapy: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-upCited for: The grade-indexed ICI action and treatment for IR-enterocolitis, the endoscopy threshold, the recommendation against CT for diagnosis, and the multidisciplinary framing of resumption.Supporting text: the full text
- Brahmer JR, et al. (2021) Society for Immunotherapy of Cancer (SITC) clinical practice guideline on immune checkpoint inhibitor-related adverse eventsCited for: The class-stratified diarrhoea and colitis rates with their grade ≥3 shares; the faecal-marker test performance AND the explicit warning against using those markers in place of endoscopy; the grade-indexed work-up including infectious exclusion; the endoscopy threshold; corticosteroid dose and the ≤4-week taper CEILING by grade; the infliximab and vedolizumab schedules with their pre-treatment screening; the absence of any head-to-head comparison between those two agents; and the rechallenge prerequisite. GI SCOPE.Supporting text: the full text
- Schneider BJ, et al. (2021) Management of Immune-Related Adverse Events in Patients Treated With Immune Checkpoint Inhibitor Therapy: ASCO Guideline UpdateCited for: Presenting symptoms, the onset window, the NSAID association, the radiologic and histologic picture, and the grade-2-or-higher corticosteroid dose. GI SCOPE.Supporting text: the full text + the PubMed abstract (checkable at the link above)