Background
Immune checkpoint inhibitors (ICIs) are effective in enhancing antitumor response by blocking checkpoint receptors that suppress the immune system, thereby improving clinical and survival outcomes.1–3 However, ICI-induced immune activation can also lead to inflammatory immune-related adverse events (irAEs),4,5 particularly in the gastrointestinal tract. Immune-related enterocolitis (irEC), which occurs in up to 40% of patients depending on the targeted immunoregulatory pathway,6 ranges from mild to severe diarrhea and is a major cause of ICI treatment interruption and discontinuation.1,2,6–8
Current guidelines recommend early initiation of empiric immunosuppression, typically with systemic corticosteroids, for patients on ICIs experiencing grade ≥2 diarrhea.9–11 However, differentiating infectious colitis from irEC is challenging due to overlapping signs and symptoms, such as diarrhea and abdominal pain. Therefore, testing for Clostridioides difficile and other stool pathogens is advised to rule out infectious etiologies, but it is suggested that corticosteroid initiation should not be delayed while awaiting results.10 Despite these recommendations, data are lacking on the prevalence and impact of C difficile infection (CDI) in this setting, including its effects on diarrhea course, severity, and the need for immunosuppressive therapy in patients with suspected irEC. Additionally, the optimal management of patients with CDI remains unclear, with unanswered questions regarding the need for and timing of immunosuppressive therapy in refractory cases.12–14
In this study, we aimed to address this knowledge gap by examining the impact of CDI test results on treatment course and clinical outcomes in ICI-treated patients undergoing evaluation for new-onset diarrhea. We evaluated rates of immunosuppressive therapy for suspected irEC based on the presence or absence of CDI. Additionally, we assessed risk factors associated with both initial and recurrent CDI, differences in clinical severity and antibiotic treatment regimens, and the impact of steroid therapy on outcomes in ICI-treated patients diagnosed with CDI.
Methods
Study Population
This retrospective cohort study included all patients receiving ICI therapy who presented with new-onset diarrhea and underwent CDI PCR-based stool nucleic acid amplification testing (NAAT) at Memorial Sloan Kettering Cancer Center (MSKCC) between July 2015 and July 2021. Patients were excluded if they had a prior diagnosis of inflammatory bowel disease or celiac disease or had previously been treated for irEC. As one of our primary outcomes was steroid initiation following CDI as a proxy for irEC diagnosis, we also excluded patients who had received empiric steroids for any reason <14 days prior to CDI testing. This helped minimize confounding from patients already on steroids due to preexisting irAEs or other medical conditions. This study received approval from the MSKCC Institutional Review Board.
Data Collection
Data were collected on demographics, CDI testing, recent drug exposures within 90 days of CDI testing—including antibiotics, proton pump inhibitors (PPIs), and chemotherapy—severity of diarrhea, immunosuppressive therapy, clinical outcomes, and endoscopic findings. The severity of diarrhea and colitis was graded according to the CTCAE, version 5.0. Severe, fulminant, and recurrent CDI were defined in accordance with consensus guidelines.15,16 The diagnosis of irEC was determined using a combination of clinical signs and symptoms, the impression of treating gastroenterologists and/or oncologists, endoscopic and histologic confirmation, and/or symptom resolution following immunosuppressive therapy. Endoscopic scoring was adapted from the Mayo endoscopic scoring system,17 and colitis histologic subtypes were categorized according to previously described patterns.18 Clinical response was defined as the first observed symptom reduction to CTCAE grade ≤1, with at least a one-grade decrease from initial diagnosis. Sustained response was defined as maintaining CTCAE grade ≤1 symptoms for at least 30 days.
Outcomes
Primary outcomes were the frequency of CDI among ICI-treated patients with diarrhea and the proportion of CDI-positive patients who received immunosuppressive therapy for concurrent irEC, including corticosteroids and second-line biologic agents (eg, infliximab and vedolizumab). Secondary outcomes included risk factors for both initial and recurrent CDI, differences in clinical severity and antibiotic therapy regimens, and the relationship between steroid therapy and clinical outcomes in ICI-treated patients with CDI. All patients were followed for a minimum of 6 months after initial diarrhea onset to assess diarrhea recurrence and the need for subsequent immunosuppressive agents. No need for steroids was therefore defined as no steroids received for diarrhea treatment within 6 months of CDI testing.
Statistical Analysis
Continuous variables were summarized as medians with IQRs, and categorical variables as frequencies with percentages. Group comparisons were performed using Wilcoxon rank sum, Kruskal-Wallis rank sum, Pearson’s chi-square, and Fisher exact tests, with Bonferroni correction applied where appropriate. Univariable and multivariable logistic regression models were used to assess the relationship between CDI positivity, steroid use, and second-line immunosuppressive agent use with various covariates of interest specified a priori based on clinical context. Covariates included age and sex (given prior data showing higher CDI incidence in females and older adults19); type of ICI (given previous associations between combination ICI regimens and increased irAE incidence20); recent antibiotic, PPI, and chemotherapy use (given other studies documenting these drugs as potential independent CDI risk factors21–23); and CTCAE grades of diarrhea and colitis (due to current guidelines recommending escalation to immunosuppressive therapy depending on severity of presentation10). An alpha level of 0.05 was considered statistically significant. All statistical analyses were conducted using R version 4.2.3 (R Foundation for Statistical Computing).
Results
Study Cohort
Of 854 ICI-treated patients presenting with new-onset diarrhea and undergoing CDI NAAT, 605 were included in the final analysis, of whom 111 (18.3%) tested positive for CDI (Figure 1). The median age at CDI testing was 65 years (IQR, 56–72). Most patients in both groups were male (52.2%; 316/605), White (81.0%; 490/605), and non-Hispanic (91.2%; 552/605) (Table 1). Patients with CDI had significantly higher rates of prior antibiotic use compared with those without CDI (64.0% vs 46.8%; P=.001). There was also a trend toward higher rates of prior PPI use in CDI-positive patients (47.8% vs 38.%; P=.078). In contrast, chemotherapy exposure did not differ significantly between groups. After adjusting for multiple clinical covariates, recent antibiotic exposure (within 90 days) was associated with a 2-fold increased likelihood of CDI (odds ratio [OR], 2.01; 95% CI, 1.29–3.17; P<.01; Supplementary Table S1, available online in the supplementary materials). The type and combination of ICI therapy were not associated with CDI positivity.
Patient allocation by inclusion criteria and CDI status.
Abbreviations: CDI, Clostridioides difficile infection; IBD, inflammatory bowel disease; ICI, immune checkpoint inhibitor; irEC, immune-related enterocolitis; MSKCC, Memorial Sloan Kettering Cancer Center.
Citation: Journal of the National Comprehensive Cancer Network 23, 5; 10.6004/jnccn.2024.7355
Baseline Characteristics of ICI-Treated Patients Presenting With Diarrhea by CDI Status
| Characteristic | Total Patients n (%) |
CDI-Positive n (%) |
CDI-Negative n (%) |
P Value |
|---|---|---|---|---|
| Total, n | 605 | 111 | 494 | |
| Age, median (IQR), y | 65 (56–72) | 65 (58–73) | 64 (55–72) | .4 |
| Sex | .8 | |||
| Female | 289 (47.8) | 52 (46.8) | 237 (48.0) | |
| Male | 316 (52.2) | 59 (53.1) | 257 (52.0) | |
| Race | .5 | |||
| White | 490 (81.0) | 88 (79.3) | 402 (81.4) | |
| Black | 38 (6.3) | 8 (7.2) | 30 (6.1) | |
| Asian | 43 (7.1) | 6 (5.4) | 37 (7.5) | |
| Othera/Unknown | 34 (5.6) | 9 (8.1) | 25 (5.1) | |
| Ethnicity | .11 | |||
| Non-Hispanic | 552 (91.3) | 97 (87.4) | 455 (92.1) | |
| Hispanic | 37 (6.1) | 8 (7.2) | 29 (5.9) | |
| Unknown | 16 (2.6) | 6 (5.4) | 10 (2.0) | |
| Cancer type | .5 | |||
| Genitourinary | 136 (22.5) | 25 (22.5) | 111 (22.5) | |
| Lung | 64 (10.6) | 15 (13.5) | 49 (9.9) | |
| Gynecologic | 53 (8.8) | 5 (4.5) | 48 (9.7) | |
| Melanoma | 58 (9.6) | 12 (10.8) | 46 (9.3) | |
| Gastrointestinal/Hepatobiliary | 27 (4.5) | 5 (4.5) | 22 (4.5) | |
| Otherb | 267 (44.1) | 49 (44.1) | 218 (44.1) | |
| Antibiotic use <90 d before CDI test | 302 (49.9) | 71 (64.0) | 231 (46.8) | .001 |
| PPI use <90 d before CDI test | 244 (40.3) | 53 (47.8) | 191 (38.7) | .078 |
| Chemotherapy <90 d before CDI test | 318 (52.6) | 58 (52.3) | 260 (52.6) | >.9 |
| Type of ICI | .6 | |||
| PD-(L)1 | 491 (81.2) | 93 (83.8) | 398 (80.6) | |
| CTLA-4 | 14 (2.3) | 3 (2.7) | 11 (2.2) | |
| Combination | 100 (16.5) | 15 (13.5) | 85 (17.2) | |
| Duration of ICI treatment, median (IQR), d | 125 (29–374) | 130 (56–329) | 122 (27–384) | .7 |
| ICI regimen outcome | .2 | |||
| Resumed ICI | 340 (56.2) | 66 (59.5) | 274 (55.5) | |
| Stopped due to irEC | 168 (27.7) | 22 (19.8) | 146 (29.6) | |
| Stopped due to disease progression | 55 (9.1) | 12 (10.8) | 43 (8.7) | |
| Stopped due to other irAEs | 12 (2.0) | 2 (1.8) | 10 (2.0) | |
| Death | 24 (4.0) | 8 (7.2) | 16 (3.2) | |
| Completion of protocol | 6 (1.0) | 1 (0.9) | 5 (1.0) | |
Bold indicates statistically significant P value.
Abbreviations: CDI, Clostridioides difficile infection; ICI, immune checkpoint inhibitor; irAE, immune-related adverse event; irEC, immune-related enterocolitis; PPI, proton pump inhibitor.
Includes Native American, Native Hawaiian or Pacific Islander, and multiple races.
Includes nonmelanoma skin cancer, head and neck/endocrine, hematologic, breast, sarcoma, neuroendocrine, brain/nervous system, and unknown primary.
Clinical Presentations and Outcomes
Of 494 patients who were CDI-negative, 204 (41.3%) were diagnosed with irEC based on clinical and/or histologic criteria, including endoscopic biopsy findings, receipt of immunosuppressive therapy for suspected irEC, or clinical presentation highly suggestive of irEC (Supplementary Table S2). An additional 107 (21.7%) patients had moderate clinical suspicion for irEC but did not ultimately receive steroid treatment due to the transient or low-grade nature of their diarrhea. Non-CDIs were identified in 61 (12.3%) patients, comprising 38 bacterial, 17 viral, and 6 parasitic etiologies. Among the remaining patients, diarrhea was attributed to other drugs in 59 (11.9%) and to other non–drug-related causes in 63 (12.8%), including disease progression, adrenal insufficiency, and diverticulitis.
Patients with CDI were more likely than those without to present with grade 3–4 diarrhea (34.2% vs 22.9%; P=.01) and grade 3–4 colitis (25.2% vs 17.2%; P=.05; Table 2). They were also more likely to report higher rates of abdominal pain (68.5% vs 33.4%; P<.001), fever (50.5% vs 19.8%; P<.001), and bloody stools (23.4% vs 4.9%; P<.001) than their CDI-negative counterparts. Despite these more severe clinical presentations, patients who were CDI-positive were less likely to receive steroids (24.3% vs 36.6%) or second-line immunosuppressive agents (8.1% vs 16.4%) for treatment of concomitant irEC after CDI diagnosis. Treatment regimens stratified by CDI testing status are shown in Figure 2. Furthermore, although patients with CDI demonstrated more severe initial presentations, median time to diarrhea clinical response (8 vs 7 days; P=.6), sustained clinical response (8 vs 8 days; P=.9), and rates of recurrent diarrhea (9.9% vs 8.9%; P=.7) did not significantly differ.
Clinical Presentation and Outcomes of ICI-Treated Patients Presenting With Diarrhea by CDI Status
| Characteristic | Total Patients n (%) |
CDI-Positive n (%) |
CDI-Negative n (%) |
P Value |
|---|---|---|---|---|
| Total, n | 605 | 111 | 494 | |
| Time from ICI therapy initiation to diarrhea onset, median (IQR), d | 85 (28–196) | 92 (27–189) | 80 (29–196) | .7 |
| Presenting associated symptoms | ||||
| Abdominal pain | 241 (39.8) | 76 (68.5) | 165 (33.4) | <.001 |
| Fever | 154 (25.5) | 56 (50.5) | 98 (19.8) | <.001 |
| Bloody stools | 50 (8.3) | 26 (23.4) | 24 (4.9) | <.001 |
| Highest grade of diarrhea | .01 | |||
| I/II | 454 (75.0) | 73 (65.8) | 381 (77.1) | |
| III/IV | 151 (25.0) | 38 (34.2) | 113 (22.9) | |
| Highest grade of colitis | .05 | |||
| I/II | 492 (81.3) | 83 (74.8) | 409 (82.8) | |
| III/IV | 113 (18.7) | 28 (25.2) | 85 (17.2) | |
| Duration of diarrhea categories | .6 | |||
| Acute (<14 d) | 418 (69.1) | 81 (73.0) | 337 (68.2) | |
| Persistent (14–30 d) | 95 (15.7) | 16 (14.4) | 79 (16.0) | |
| Chronic (>30 d) | 92 (15.2) | 14 (12.6) | 78 (15.8) | |
| Underwent lower endoscopy | 133 (22.0) | 14 (12.6) | 119 (24.1) | .01 |
| Time to endoscopic assessment, median (IQR), d | 12 (3–31) | 16 (3–50) | 10 (3–31) | .7 |
| Mayo endoscopic score, N | 133 | 14 | 119 | <.01 |
| Score 0 | 24 (18.0) | 4 (28.6) | 20 (16.8) | |
| Score 1 | 65 (48.9) | 2 (14.3) | 63 (52.9) | |
| Score 2 | 36 (27.1) | 4 (28.6) | 32 (26.9) | |
| Score 3 | 8 (6.0) | 4 (28.6) | 4 (3.4) | |
| Colitis histologic subtype, N | 133 | 14 | 119 | .08 |
| Active colitis | 45 (33.8) | 5 (35.7) | 35 (29.4) | |
| Chronic active colitis | 40 (30.1) | 2 (14.3) | 43 (36.1) | |
| Microscopic colitis | 12 (9.0) | 0 (0.0) | 12 (10.1) | |
| GVHD-like (apoptosis) | 7 (5.3) | 2 (14.3) | 5 (4.2) | |
| Normal mucosa | 29 (21.8) | 5 (35.7) | 24 (20.2) | |
| Hospitalization for diarrhea | 225 (37.2) | 46 (41.4) | 179 (36.2) | .3 |
| Steroid treatment for concomitant irEC | 208 (34.4) | 27 (24.3) | 181 (36.6) | .01 |
| Duration of steroid treatment, median, (IQR), d | 41 (24–72) | 42 (24–95) | 41 (23–70) | .8 |
| Time from CDI testing to steroid initiation, median (IQR), d | 1 (0–6) | 2 (0–12) | 1 (0–6) | .7 |
| Second-line immunosuppressivesa | 90 (14.9) | 9 (8.1) | 81 (16.4) | .03 |
| Time to clinical response,b median (IQR), d | 7 (4–14) | 8 (4–13) | 7 (4–15) | .6 |
| Time to sustained response,c median (IQR), d | 8 (4–16) | 8 (4–20) | 8 (4–16) | .9 |
| Recurrence of ICI-related diarrhea | 55 (9.1) | 11 (9.9) | 44 (8.9) | .7 |
| Time from diarrhea onset to recurrence, median (IQR), d | 1 (0–6) | 2 (0–12) | 1 (0–6) | .7 |
Bold indicates statistically significant P value.
Abbreviations: CDI, Clostridioides difficile infection; ICI, immune checkpoint inhibitor; GVHD, graft-versus-host disease; irEC, immune-related enterocolitis.
Second-line immunosuppressive agents included infliximab and vedolizumab.
Clinical response defined as the first observed decrease to grade ≤1 symptoms with at least one grade decrease from initial diagnosis.
Sustained response defined as grade ≤1 symptoms for ≥30 consecutive days.
Treatment strategy for ICI-treated patients presenting with diarrhea by CDI status.
Abbreviations: CDI, Clostridioides difficile infection; ICI, immune checkpoint inhibitor.
Citation: Journal of the National Comprehensive Cancer Network 23, 5; 10.6004/jnccn.2024.7355
In a multivariable logistic regression model (Supplementary Table S3), patients who were CDI-positive had significantly lower odds of subsequently receiving steroids for suspected irEC compared with those who were CDI-negative (OR, 0.45; 95% CI, 0.27–0.75). Similarly, steroid initiation was less likely in those who had chemotherapy within 90 days prior to CDI testing (OR, 0.50; 95% CI, 0.34–0.72) and those who experienced a clinical response to antibiotics alone (OR, 0.003; 95% CI, 0.001–0.008). In contrast, CTLA-4–based immunotherapy and higher grades of diarrhea were positively associated with steroid use. CDI positivity, grade 3–4 colitis, and prior steroid use for irEC were also independent predictors of biologic use for irEC (Supplementary Table S4).
Antibiotic Treatment Regimens for CDI
All 111 CDI-positive patients were treated with antibiotics: 63 (56.8%) received oral vancomycin, 17 (15.3%) oral metronidazole, and 31 (27.9%) a combination of oral vancomycin and intravenous metronidazole (Supplementary Table S5). Most patients treated with metronidazole monotherapy (14/17) were diagnosed with CDI before May 2018, reflecting shifts in CDI treatment guidelines.16 More intensive dual therapy with vancomycin and metronidazole was used more frequently in severe CDI cases (48.4% vs 30.2% for vancomycin monotherapy and 17.7% for metronidazole monotherapy; P=.02). The proportion of patients requiring steroids or biologic agents for treatment of irEC did not differ by antibiotic regimen, and there were no significant differences in time to clinical response or sustained clinical response across antibiotic regimens.
Overall CDI recurrence was 19.8% (22/111) and did not differ between antibiotic regimens. The median time to recurrence was 46 days (IQR, 31–87). Of the 22 recurrent cases, 12 (54.5%) were diagnosed early within 2 to 8 weeks of the original CDI, while the remaining 10 were diagnosed >8 weeks after the initial infection. Based on prior data, these late recurrences were classified as true recurrences rather than distinct reinfections.24 CDI recurrence was diagnosed based on PCR positivity in combination with clinical history judged by the treating physician to be most compatible with CDI recurrence and warranting a repeat course of antibiotics. No univariable associations were found between CDI recurrence and index CDI severity, type of antibiotic used for index CDI, or use of steroids for concurrent index CDI-irEC. However, in the time period between the index and recurrent CDI episodes, antibiotic use for any indication (OR, 6.21; 95% CI, 1.76–21.97), PPI use (OR, 3.57; 95% CI, 1.02–12.51), and neutropenia (OR, 5.34; 95% CI, 1.11–25.76) were all significantly associated with CDI recurrence (Supplementary Table S6).
Concurrent CDI and irEC
Of the 111 patients who were CDI-positive, 84 (75.7%) achieved symptom resolution with antibiotics alone. The remaining 27 (24.3%) were treated with antibiotics in addition to corticosteroids and/or biologics for clinically suspected (n=17; displaying signs and symptoms highly compatible with a leading diagnosis of irEC at the discretion of the treating gastroenterologist and/or oncologist) or histologically confirmed (n=10) concurrent irEC (Table 3). Median cycle threshold (CT) values, which inversely correlate with the log amount of C difficile bacterial load in the PCR sample and have been associated with the presence of toxin and increased CDI severity, were similar between groups (26.6 for those with concurrent CDI and irEC [CDI-irEC] and 27.7 for CDI alone), with all values below the previously defined cutoff of 28.0. Patients with CDI-irEC were more likely to receive CTLA-4–based immunotherapy compared with those with CDI alone (27.0% vs 9.5%; P<.01). Additionally, those with CDI-irEC were more likely to present with grade 3–4 diarrhea (66.7% vs 23.8%; P<.01), grade 3–4 colitis (70.4% vs 10.7%; P<.01), abdominal pain (88.9% vs 61.9%; P=.01), and bloody stools (48.2% vs 15.5%; P<.0111). Patients with CDI-irEC also had higher rates of chronic diarrhea lasting 30 days (40.8% vs 3.6%; P<.01), higher rates of hospitalization for diarrhea (70.4% vs 32.1%; P<.01), longer times to clinical response (13 vs 6 days; P<.01), and sustained clinical response (25 vs 6 days; P<.01). CDI recurrence rates and treatments were similar between groups. Predictors of steroid use among CDI-positive patients included combination CTLA-4 + PD-1–based immunotherapy (OR, 5.11; 95% CI, 1.63–16.5; P<.01) and grade 3–4 diarrhea (OR, 6.40; 95% CI, 2.55–17.1; P<.01) and colitis (OR, 19.8; 95% CI, 7.05–61; P<.01; Table 4). In contrast, antibiotic response was significantly associated with a lower likelihood of steroid initiation (OR, 0.08; 95% CI, 0.04–0.19).
Clinical Presentation and Outcomes of CDI-Positive Patients Treated With Antibiotics and Steroids Versus Antibiotics Alone
| Characteristic | Total CDI-Positive Patients n (%) |
Antibiotics + Steroids n (%) |
Antibiotics Alone n (%) |
P Value |
|---|---|---|---|---|
| Total | 111 | 27 | 84 | |
| Time from ICI therapy initiation to diarrhea onset, median (IQR), d | 92.3 (168.3) | 122.7 (158.2) | 88.2 (151.6) | .62 |
| Presenting associated symptoms | ||||
| Abdominal pain | 76 (68.5) | 24 (88.9) | 52 (61.9) | .01 |
| Fever | 56 (50.5) | 15 (55.6) | 41 (48.8) | .5 |
| Bloody stools | 26 (23.4) | 13 (48.2) | 13 (15.5) | <.01 |
| Highest grade of diarrhea | <.01 | |||
| I/II | 73 (65.8) | 9 (33.3) | 64 (76.2) | |
| III/IV | 38 (34.2) | 18 (66.7) | 20 (23.8) | |
| Highest grade of colitis | <.01 | |||
| I/II | 83 (74.8) | 8 (29.6) | 75 (89.3) | |
| III/IV | 28 (25.2) | 19 (70.4) | 9 (10.7) | |
| Duration of diarrhea categories | <.01 | |||
| Acute (<14 d) | 81 (73.0) | 10 (37.0) | 71 (84.5) | |
| Persistent (14–30 d) | 16 (14.4) | 6 (22.2) | 10 (11.9) | |
| Chronic (>30 d) | 14 (12.6) | 11 (40.8) | 3 (3.6) | |
| Underwent lower endoscopy | 14 (12.6) | 10 (24.3) | 4 (4.8) | <.01 |
| Time to endoscopic assessment, median (IQR), d | 17 (3–56) | 7 (1–140) | 35 (9–56) | .67 |
| Mayo endoscopic score | 14 | 10 | 4 | .14 |
| Score 0 | 4 (28.6) | 1 (10.0) | 3 (75.0) | |
| Score 1 | 2 (14.3) | 2 (20.0) | 0 (0) | |
| Score 2 | 4 (28.6) | 3 (30.0) | 1 (25.0) | |
| Score 3 | 4 (28.6) | 4 (40.0) | 0 (0) | |
| Colitis histologic subtype | 14 | 10 | 4 | .2 |
| Active colitis | 2 (14.3) | 1 (10.0) | 1 (25.0) | |
| Chronic active colitis | 5 (35.7) | 5 (50.0) | 0 (0) | |
| GVHD-like (apoptosis) | 2 (14.3) | 1 (10.0) | 1 (25.0) | |
| Normal mucosa | 5 (35.7) | 3 (30.0) | 2 (50.0) | |
| Hospitalization for diarrhea | 46 (41.4) | 19 (70.4) | 27 (32.1) | <.01 |
| Second-line immunosuppressives | 9 (8.1) | 9 (33.3) | 0 (0) | <.01 |
| Clostridioides difficile strain | 93 | 23 | 70 | .09 |
| Hypervirulenta | 8 (8.6) | 0 (0) | 8 (11.4) | |
| Other | 85 (91.4) | 23 (100) | 62 (88.6) | |
| Threshold cycle value, median (IQR) | 27.4 (24.4–33.2) | 26.2 (23.9–33.5) | 27.7 (24.5–33.2) | .73 |
| Severe CDI | 34 (30.6) | 9 (33.3) | 25 (29.8) | .73 |
| Fulminant CDI | 11 (9.9) | 2 (7.4) | 9 (10.7) | .62 |
| Initial CDI antibiotic treatment regimen | .84 | |||
| Vancomycin monotherapy | 61 (55.0) | 14 (51.9) | 47 (56.0) | |
| Metronidazole monotherapy | 17 (15.3) | 4 (14.8) | 13 (15.4) | |
| Vancomycin + metronidazole | 31 (27.9) | 8 (29.6) | 23 (27.4) | |
| Vancomycin + fidaxomicin | 2 (1.8) | 1 (3.7) | 1 (1.2) | |
| Time to clinical response, median (IQR), d | 8 (4–13) | 13 (8–26) | 6 (3–10) | <.01 |
| Time to sustained clinical response, median (IQR), d | 8 (4–20) | 25 (10–63) | 6 (3–11) | <.01 |
| Recurrence of CDI | 22 (19.8) | 5 (18.5) | 17 (20.2) | .85 |
| Recurrent CDI treatment regimen | .45 | |||
| Vancomycin monotherapy | 14 (63.6) | 2 (40.0) | 12 (70.6) | |
| Metronidazole monotherapy | 1 (4.6) | 0 (0) | 1 (5.9) | |
| Vancomycin + metronidazole | 4 (18.2) | 2 (40.0) | 2 (11.8) | |
| No antibiotics | 2 (9.1) | 1 (20.0) | 1 (5.9) | |
| Fecal microbiota transplant | 1 (4.6) | 0 (0) | 1 (5.9) | |
Abbreviations: CDI, Clostridioides difficile infection; GVHD, graft-versus-host disease; ICI, immune checkpoint inhibitor; irEC, immune-related enterocolitis.
Defined as Clostridioides difficile ST-1 and ST-11.
Univariable Associations With Steroid and Second-Line Immunosuppressive Use for Suspected Concomitant irEC in CDI-Positive Patients (N=111)
| Characteristic | Steroids (n=27) | Second-Line Immunosuppressives (n=9) | ||
|---|---|---|---|---|
| OR (95% CI) | P Value | OR (95% CI) | P Value | |
| Age at CDI test | 0.99 (0.95–1.02) | .4 | 0.97 (0.92–1.02) | .2 |
| Sex | ||||
| Male | Ref | Ref | ||
| Female | 1.07 (0.45–2.56) | .9 | 0.90 (0.21–3.59) | .9 |
| Chemotherapy <90 d before CDI test | 0.66 (0.27–1.58) | .4 | 0.23 (0.03–1.03) | .079 |
| Type of ICI | ||||
| PD-(L)1 | Ref | Ref | ||
| CTLA-4 | 8.94 (0.81–199) | .081 | 8.80 (0.37–109) | .1 |
| Combination | 5.11 (1.63–16.5) | .005 | 4.40 (0.82–20.4) | .062 |
| Time from ICI initiation to diarrhea onset | 1.00 (1.00–1.00) | .5 | 1.00 (1.00–1.00) | .9 |
| Highest grade of diarrhea | ||||
| I/II | Ref | Ref | ||
| III/IV | 6.40 (2.55–17.1) | <.001 | 8.02 (1.82–55.9) | .012 |
| Highest grade of colitis | ||||
| I/II | Ref | Ref | ||
| III/IV | 19.8 (7.05–61.8) | <.001 | 7.27 (1.77–36.7) | .008 |
| Severe CDI | 1.18 (0.45–2.94) | .7 | 0.63 (0.09–2.76) | .6 |
| Fulminant CDI | 0.67 (0.10–2.81) | .6 | 1.15 (0.06–7.28) | .9 |
| Response to antibiotics | 0.09 (0.04–0.19) | <.001 | 0.10 (0.06–0.18) | <.001 |
Bold indicates statistically significant P value.
Abbreviations: CDI, Clostridioides difficile infection; ICI, immune checkpoint inhibitor; irEC, immune-related enterocolitis; OR, odds ratio.
CDI-irEC Versus irEC Alone
Among patients treated with immunosuppressives for irEC (n=208), those with CDI-irEC (n=27; 13%) had more severe clinical presentations. They had higher rates of grade 3–4 diarrhea (66.7% vs 33.1%; P<.01) and colitis (70.4% vs 18.8%; P<.01), as well as more severe endoscopic colitis findings, with Mayo scores of 2 to 3 reported in 70.0% compared with 40.6% in CDI-negative patients treated for irEC (P<.01) (Supplementary Table S7). Although time to steroid response was slower in the CDI group, there were no significant differences in the duration of steroid treatment or the need for second-line immunosuppressives.
Discussion
In the largest study to date of C difficile in ICI-treated patients presenting with diarrhea, CDI was common, occurring in 18% of this population. By analyzing patients at the time of initial diarrhea presentation prompting CDI testing, we aimed to address the clinical conundrum of the role of CDI testing during irEC evaluation and how test results influence clinical care pathways. We observed that CDI-positive patients presented with more severe diarrhea and colitis symptoms than CDI-negative patients, yet the majority (76%) recovered with antibiotics alone, without requiring immunosuppressive treatment. This finding is critical, as unnecessary immunosuppression not only carries well-documented side effects but also may inhibit optimal ICI antitumor responses.25–27 These patients’ presentation with severe gastrointestinal symptoms would have frequently prompted immediate initiation of immunosuppressive medications for suspected irEC, in accordance with society guidelines. This underscores the critical need for routine CDI testing in this patient population to prevent unnecessary immunosuppressive treatment. Among CDI-positive patients, recent antibiotic use was the only identified risk factor.
Understanding the clinical course of CDI in ICI-treated patients is imperative to direct appropriate therapy, given the challenge of distinguishing C difficile colitis from irEC due to their overlapping clinical presentations but significantly different treatment approaches. Indicators such as fever and bloody stools were key in differentiating CDI-related diarrhea from ICI-induced diarrhea. Therefore, all patients on ICIs with new-onset diarrhea—particularly those presenting with these high-risk features and recent antibiotic exposure—should first undergo CDI testing to properly guide further workup and avoid unnecessary endoscopic evaluation and immunosuppressive therapy when CDI is confirmed.
Furthermore, while most patients in our CDI cohort demonstrated diarrhea resolution with antibiotics alone, the need for additional immunosuppressive therapy was common (24%), with frequent occurrence of high-grade colitis and bloody diarrhea among those who were also treated for irEC (CDI-irEC). This underscores the importance of close follow-up of patients and early introduction of immunosuppressive treatment if there is an inadequate response to CDI-directed therapy. In such cases, prompt lower endoscopic evaluation may also prove valuable by allowing for histologic confirmation of irEC and guiding the need for immunosuppression.
We hypothesize that patients treated with both antibiotics and immunosuppressives may represent cases of concomitant CDI-irEC or C difficile colonization. Similar to findings in patients with inflammatory bowel disease, CDI has been posited to be not only a complication of irEC but also a potential trigger in patients undergoing immunotherapy.28 Such an association may be due to CDI predisposing patients toward colonic autoimmune destruction and gut microbiome dysregulation via enhanced recruitment of T cells and inflammatory agents.29 This relationship may explain our data showing that patients with CDI-irEC had more severe presentations than those with irEC or CDI alone.
Similar to our data, previous retrospective studies reported more severe clinical presentations, longer durations of symptoms, and greater rates of steroid and biologic treatment failures in patients with concurrent CDI-irEC compared with those with CDI alone.28,30 As such, the presence of CDI with superimposed irEC may lead to a more severe and prolonged disease course associated with an increased need for immunosuppression.
Our study is unique from these prior studies in several important ways. Notable strengths include a large sample size and a homogeneous cohort focused on ICI-treated patients presenting for the initial workup of diarrhea, excluding those who had received pretest empiric immunosuppressive therapy, which could confound the interpretation of clinical outcomes. Our primary objective was to assess how CDI testing results influence care pathways in the diagnostic algorithm for new-onset diarrhea in ICI-treated patients. As such, we chose to exclude patients already receiving treatment for irEC, because this represents a different clinical scenario and question—those with ongoing symptoms despite colitis therapies such as corticosteroids. In this context, we observed significantly higher rates of antibiotic responsiveness and a lower need for immunosuppressive therapy compared with prior studies. This finding underscores the importance of testing for C difficile in ICI-treated patients with diarrhea, ideally before the initiation of immunosuppressive therapy, to ensure appropriate treatment is administered and to avoid unnecessary immunosuppression. Furthermore, our study contributes novel insights into the impact of different CDI antibiotic regimens on outcomes and utilizes multivariable logistic regression models to assess the association between CDI positivity and immunosuppressive agent use and various a priori–defined clinically relevant covariates.
Recurrence of CDI was common, occurring in 20% of ICI-treated patients. Although the severity of the index CDI, the type of antibiotic used for the index CDI, and concurrent steroid use for CDI-irEC did not predict CDI recurrence, we observed significant associations between CDI recurrence and antibiotic use for any indication, PPI use, and neutropenia during the time period between the index and recurrent CDI episodes. Patients with recurrent CDI after initial resolution experienced response to CDI treatment without requiring immunosuppression. This suggests that CDI testing should be performed upon symptoms recurrence, especially in individuals with subsequent exposure to antibiotics and gastric acid suppression. Additionally, the availability of toxin testing would be important to exclude transient diarrhea in C difficile–colonized individuals, for whom recurrent antibiotics may be superfluous. Lastly, emerging evidence also suggests the promising efficacy of fecal microbiota transplantation for treating recurrent CDI, with or without concurrent irEC, in patients undergoing ICI therapy. This treatment approach warrants further investigation.15,31
In summary, based on our data, we recommend that patients on immunotherapy who present with new-onset diarrhea undergo prompt CDI testing, especially those with recent antibiotic exposure and high-risk features such as grade 3–4 diarrhea, abdominal pain, fever, and bloody stools. If the CDI test is positive, appropriate antibiotics should be initiated as the first-line treatment, with close follow-up. For cases of persistent or worsening colitis, endoscopic evaluation is warranted, and escalation to empiric corticosteroid therapy should be strongly considered, especially for patients treated with CTLA-4 inhibitors.
Our study was limited by its retrospective design. The availability of only CDI NAAT testing and not C difficile toxin testing during the study period limited our ability to definitively distinguish between colonization and active infection. However, it is important to note that from July 2015 to July 2021, the sole use of NAAT was the most prevalent practice in US acute care hospitals, particularly in oncology facilities, due to the risk of missed CDI diagnoses with toxin-based testing only.32 It is also well established that clinically significant CDI can manifest as NAAT-positive/toxin-negative on diagnostic testing.33 Moreover, despite the lack of overt toxin testing, we leveraged the quantitative capacity of PCRs to measure CT values, which inversely correlate with the log amount of target in the sample. Lower CT values correlate with the presence of free toxin and increased CDI severity.34 Our cohort’s median CT values were all below the cutoff of 28.0, which has been previously demonstrated to have a sensitivity of 91% for predicting the presence of free toxin and shown to predict 77% cytotoxin-positive CDI cases, as well as 91% and 100% of severe and complicated CDI episodes, respectively.35 Thus, the lower CT values of the CDI-positive patients in our study serve as a surrogate for toxin positivity and suggest that these patients did in fact have toxin-positive and clinically relevant CDI rather than colonization alone. High response rates to antibiotics alone further suggest that most cases were true active infections and not simply colonization acting as a red herring for irEC or self-limited diarrhea. Nonetheless, future studies are needed to explore the role of toxin positivity in determining the need for and response rates to antibiotic and immunosuppressive therapies. Furthermore, although CDI testing is a routine, guideline-supported practice at our institution for patients on ICIs with new-onset diarrhea, we only included patients who underwent CDI testing. As a result, we may have missed a subgroup of patients whose providers did not order this testing or whose symptoms resolved prior to submission of stool testing. Finally, rates of fidaxomicin treatment for primary CDI in this cohort were low due to our study encompassing a period prior to the updated 2021 Infectious Diseases Society of America guidelines recommending fidaxomicin as a first-line therapy.36
Conclusions
In the largest analysis of CDI testing in patients with diarrhea on ICIs to date, CDI was found to be common, particularly linked to prior antibiotic therapy, and often presented with symptoms such as abdominal pain, fever, and bloody stools. Notably, 76% of patients were successfully treated with antibiotics alone. Nonetheless, an important minority of patients, especially those on CTLA-4–based immunotherapy and presenting with bloody stools, required the addition of immunosuppressive therapy for irEC. Given the overlapping clinical presentations of CDI and irEC, it is critical to test for C difficile in ICI-treated patients who present with new-onset diarrhea. Regular and timely workups will allow for appropriate management strategies and help reduce unnecessary immunosuppression. Future studies are needed to further elucidate the role of newer 2-step CDI assays incorporating toxin testing in patients on ICIs with diarrhea and to optimize the timing and interplay of conservative management, antibiotics, and immunosuppression.
References
- 1.↑
Postow MA, Callahan MK, Wolchok JD. Immune checkpoint blockade in cancer therapy. J Clin Oncol 2015;33:1974–1982.
- 2.↑
Dougan M. Checkpoint blockade toxicity and immune homeostasis in the gastrointestinal tract. Front Immunol 2017;8:1547.
- 3.↑
Seidel JA, Otsuka A, Kabashima K. Anti-PD-1 and anti-CTLA-4 therapies in cancer: mechanisms of action, efficacy, and limitations. Front Oncol 2018;8:86.
- 4.↑
Dougan M, Luoma AM, Dougan SK, et al. Understanding and treating the inflammatory adverse events of cancer immunotherapy. Cell 2021;184:1575–1588.
- 5.↑
Postow MA, Sidlow R, Hellmann MD. Immune-related adverse events associated with immune checkpoint blockade. N Engl J Med 2018;378:158–168.
- 6.↑
Collins M, Soularue E, Marthey L, et al. Management of patients with immune checkpoint inhibitor-induced enterocolitis: a systematic review. Clin Gastroenterol Hepatol 2020;18:1393–1403.e1.
- 7.↑
Pauken KE, Dougan M, Rose NR, et al. Adverse events following cancer immunotherapy: obstacles and opportunities. Trends Immunol 2019;40:511–523.
- 8.↑
Haanen J, Carbonnel F, Robert C, et al. Management of toxicities from immunotherapy: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2017;28:iv119–142.
- 9.↑
Dougan M, Wang Y, Rubio-Tapia A, et al. AGA clinical practice update on diagnosis and management of immune checkpoint inhibitor colitis and hepatitis: expert review. Gastroenterology 2021;160:1384–1393.
- 10.↑
Thompson JA, Schneider BJ, Brahmer J, et al. NCCN Guidelines Insights: Management of Immunotherapy-Related Toxicities, Version 1.2020. J Natl Compr Canc Netw 2020;18:230–241.
- 11.↑
Puzanov I, Diab A, Abdallah K, et al. Managing toxicities associated with immune checkpoint inhibitors: consensus recommendations from the Society for Immunotherapy of Cancer (SITC) Toxicity Management Working Group. J Immunother Cancer 2017;5:95.
- 12.↑
Ma W, Gong Z, Abu-Sbeih H, et al. Outcomes of immune checkpoint inhibitor-related diarrhea or colitis in cancer patients with superimposed gastrointestinal infections. Am J Clin Oncol 2021;44:402–408.
- 13.↑
Piralla A, Lunghi G, Ardissino G, et al. FilmArray GI panel performance for the diagnosis of acute gastroenteritis or hemorragic diarrhea. BMC Microbiol 2017;17:111.
- 14.↑
Magahis PT, Satish D, Esther Babady N, et al. Prevalence of enteric infections in patients on immune checkpoint inhibitors and impact on management and outcomes. Oncologist 2024;29:36–46.
- 15.↑
Kelly CR, Fischer M, Allegretti JR, et al. ACG clinical guidelines: prevention, diagnosis, and treatment of Clostridioides difficile infections. Am J Gastroenterol 2021;116:1124–1147.
- 16.↑
McDonald LC, Gerding DN, Johnson S, et al. Clinical practice guidelines for clostridium difficile infection in adults and children: 2017 update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA). Clin Infect Dis 2018;66:e1–48.
- 17.↑
Cheung VTF, Gupta T, Olsson-Brown A, et al. Immune checkpoint inhibitor-related colitis assessment and prognosis: can IBD scoring point the way? Br J Cancer 2020;123:207–215.
- 18.↑
Pai RK, Pai RK, Brown I, et al. The significance of histological activity measurements in immune checkpoint inhibitor colitis. Aliment Pharmacol Ther 2021;53:150–159.
- 19.↑
Lessa FC, Mu Y, Bamberg WM, et al. Burden of Clostridium difficile infection in the United States. N Engl J Med 2015;372:825–834.
- 20.↑
Larkin J, Chiarion-Sileni V, Gonzalez R, et al. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma. N Engl J Med 2015;373:23–34.
- 21.↑
Kwok CS, Arthur AK, Anibueze CI, et al. Risk of Clostridium difficile infection with acid suppressing drugs and antibiotics: meta-analysis. Am J Gastroenterol 2012;107:1011–1019.
- 22.↑
Fuereder T, Koni D, Gleiss A, et al. Risk factors for Clostridium difficile infection in hemato-oncological patients: a case control study in 144 patients. Sci Rep 2016;6:31498.
- 23.↑
Raza S, Baig MA, Russell H, et al. Clostridium difficile infection following chemotherapy. Recent Pat Antiinfect Drug Discov 2010;5:1–9.
- 24.↑
Kamboj M, Khosa P, Kaltsas A, et al. Relapse versus reinfection: surveillance of Clostridium difficile infection. Clin Infect Dis 2011;53:1003–1006.
- 25.↑
Faje AT, Lawrence D, Flaherty K, et al. High-dose glucocorticoids for the treatment of ipilimumab-induced hypophysitis is associated with reduced survival in patients with melanoma. Cancer 2018;124:3706–3714.
- 26.↑
Verheijden RJ, May AM, Blank CU, et al. Association of anti-TNF with decreased survival in steroid refractory ipilimumab and anti-PD1-treated patients in the Dutch Melanoma Treatment Registry. Clin Cancer Res 2020;26:2268–2274.
- 27.↑
Bai X, Hu J, Betof Warner A, et al. Early use of high-dose glucocorticoid for the management of irAE is associated with poorer survival in patients with advanced melanoma treated with anti-PD-1 monotherapy. Clin Cancer Res 2021;27:5993–6000.
- 28.↑
Vuillamy C, Arnault JP, Fumery M, et al. Clostridium difficile infection and immune checkpoint inhibitor-induced colitis in melanoma: 18 cases and a review of the literature. Melanoma Res 2023;33:192–198.
- 29.↑
Gupta A, Wash C, Wu Y, et al. Diagnostic modality of Clostridioides difficile infection predicts treatment response and outcomes in inflammatory bowel disease. Dig Dis Sci 2021;66:547–553.
- 30.↑
Vasavada S, Panneerselvam K, Amin R, et al. Clostridioides difficile infection in cancer patients receiving immune checkpoint inhibitors. Ann Gastroenterol 2022;35:393–399.
- 31.↑
Elkrief A, Waters NR, Smith N, et al. Immune-related colitis is associated with fecal microbial dysbiosis and can be mitigated by fecal microbiota transplantation. Cancer Immunol Res 2024;12:308–321.
- 32.↑
Centers for Disease Control and Prevention. Short summary: testing for C. difficile and Standardized Infection Ratios, National Healthcare Safety Network, 2019. Accessed February 24, 2025. Available at: https://www.cdc.gov/nhsn/pdfs/ps-analysis-resources/Cdiff-testing-sir-508.pdf
- 33.↑
Guh AY, Fridkin S, Goodenough D, et al. Potential underreporting of treated patients using a Clostridioides difficile testing algorithm that screens with a nucleic acid amplification test. Infect Control Hosp Epidemiol 2024;45:590–598.
- 34.↑
Kaltsas A, Simon M, Unruh LH, et al. Clinical and laboratory characteristics of Clostridium difficile infection in patients with discordant diagnostic test results. J Clin Microbiol 2012;50:1303–1307.
- 35.↑
Kamboj M, Brite J, McMillen T, et al. Potential of real-time PCR threshold cycle (C(T)) to predict presence of free toxin and clinically relevant C. difficile infection (CDI) in patients with cancer. J Infect 2018;76:369–375.
- 36.↑
Johnson S, Lavergne V, Skinner AM, et al. Clinical practice guideline by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA): 2021 focused update guidelines on management of Clostridioides difficile infection in adults. Clin Infect Dis 2021;73:e1029–1044.
