# Immune-Related Adverse Events in Anti–PD-1 Therapy

## What immune-related adverse events are associated with anti–PD-1 therapy and how are they managed?

Anti-PD-1 therapy is associated with immune-related adverse events (irAEs) affecting 38-69% of patients, involving primarily the endocrine, gastrointestinal, dermatologic, pulmonary, and hepatic systems. Common irAEs include thyroiditis (4-34%), colitis (2.8-22%), rash (1-18%), pneumonitis (3-13%), and hepatotoxicity (0.8-11%). Most irAEs are mild to moderate (grade 1-2), with grade 3-4 events occurring in approximately 7-20% of patients. While most irAEs develop within 12-14 weeks of treatment initiation, delayed onset beyond 12 months occurs in 5.3% of patients, and chronic irAEs persisting beyond 12 weeks after treatment cessation develop in 43.2% of patients, with 85.6% persisting at last follow-up. Endocrinopathies (83%), neurotoxicities (73%), and arthritis (49%) are particularly likely to become chronic, while visceral organ toxicities show lower chronicity rates.

Corticosteroids at 1-2 mg/kg/day represent the cornerstone of irAE management, with approximately 80% of irAEs resolving within a median of 5 weeks with immunosuppressive therapy. Steroid-refractory cases may require additional agents such as infliximab, which can provide symptom relief within 24 hours for gastrointestinal irAEs. Treatment modifications include temporary holds for grade 2 toxicity and permanent discontinuation for grades 3-4, required in 14-15% of patients. Critically, immunosuppressive treatment does not impair anti-tumor efficacy, and tumor progression is less frequent in patients experiencing irAEs. Successful management depends on early recognition and prompt corticosteroid initiation, multidisciplinary collaboration, and systematic monitoring protocols. Patients with preexisting autoimmune disease experience higher rates of disease flares (52%), with severity correlating to baseline immunosuppression intensity.

## Methods

We analyzed 10 sources from an initial pool of 200, using 8 screening criteria. Each paper was reviewed for 6 key aspects that mattered most to the research question.

## Characteristics of Included Studies

The review included 10 sources examining immune-related adverse events associated with anti-PD-1 therapy. Study designs varied from retrospective case series to reviews of clinical trial data, with sample sizes ranging from 27 to 999 patients.

| Study | Full text retrieved? | Study Type | Sample Size | Cancer Type(s) | Follow-up Duration | irAE Grading System |
|-------|----------------------|------------|-------------|-----------------|--------------------|---------------------|
| T. Eigentler et al., 2016 | No | Retrospective analysis of pooled trial data | Not specified | Multiple malignancies | Not specified | Not specified |
| J. Weber et al., 2016 | Yes | Review article | Not specified | Advanced melanoma, NSCLC, renal cell carcinoma | Most grade 3/4 irAEs occur within 12-14 weeks | CTCAE v4.03 |
| G. O’Kane et al., 2017 | Yes | Review of phase III trials | Data from multiple trials | Metastatic NSCLC | Every 12 weeks up to 1 year post-discontinuation | CTCAE v4.0 |
| J. Patrinely et al., 2021 | No | Retrospective multicenter cohort | 387 | Stage III-IV melanoma | Beyond 12 weeks post-discontinuation | Not specified |
| L. Hofmann et al., 2016 | No | Case series | 496 | Metastatic melanoma | Not specified | Not specified |
| D. Iacono et al., 2020 | No | Retrospective single-center | 130 (82% received anti-PD-1) | NSCLC (49%), melanoma (42%), kidney (7%), other (2%) | Jan 2012-Dec 2017 | CTCAE v4.0 |
| C. Owen et al., 2021 | Yes | Retrospective multicenter | 999 | Melanoma | Median 21 months | Severity grades used but version not specified |
| S. Vardhana et al., 2018 | Yes | Review article | Aggregated from multiple trials | Classical Hodgkin lymphoma | Varies by trial | Not specified (grade 3 mentioned) |
| L. Zimmer et al., 2016 | No | Case series | 496 | Metastatic melanoma | Not specified | Not specified |
| S. Hoa et al., 2021 | No | Retrospective multicenter | 27 | Lung cancer and melanoma | Median 11.0 months (IQR 6.0-17.5) | Not specified |

## Anti-PD-1 Agents and Treatment Regimens

The primary anti-PD-1 agents examined were nivolumab and pembrolizumab. Dosing regimens included nivolumab 3 mg/kg IV every 2 weeks and pembrolizumab at 2 mg/kg or 10 mg/kg every 3 weeks. In the Iacono cohort, nivolumab was used in 60% of patients, pembrolizumab in 21%, and atezolizumab in 1%.

Combination therapy with anti-CTLA-4 inhibition (ipilimumab) was examined in several studies. Owen et al. reported 20 patients who experienced delayed irAEs after initial combination with anti-CTLA-4. In the Hoa cohort of patients with preexisting autoimmune disease, 2 patients received sequential anti-CTLA-4 therapy.

## Incidence and Spectrum of Immune-Related Adverse Events

### Overall Incidence

The incidence of any irAE varied across studies. Patrinely et al. reported that 69.0% of patients experienced acute irAEs during treatment, with 19.5% developing grades 3-5 events. Iacono et al. found that 38% of patients developed irAEs. In patients with preexisting autoimmune disease, Hoa et al. observed PAD exacerbations in 52% of cases.

### Organ System Involvement and Specific irAEs

| Organ System | Specific irAEs | Incidence | Severity | Time to Onset | Source |
|--------------|----------------|-----------|----------|---------------|-------|
| Endocrine | Thyroiditis, hypothyroidism | 4-7%; most common (34%); 9-16% hypothyroidism | Typically grade 1-2 | 9-12 weeks | O’Kane et al., Weber et al., Iacono et al., Vardhana et al. |
| Endocrine | Hypophysitis | Not specified | Grade 1-2 | 9-12 weeks | Weber et al. |
| Endocrine | Diabetes mellitus | Rare | Not specified | Not specified | Hofmann et al. |
| Gastrointestinal | Colitis, diarrhea | 14%; 2.8-2.9% high-grade; 2-3%; 22% delayed | Grade 3-4 common | ~7 weeks; 3.5-5.3 months | Weber et al., Iacono et al., Vardhana et al., Owen et al. |
| Gastrointestinal | Enteritis, gastritis, esophagitis | Not specified | Not specified | ~7 weeks | Weber et al. |
| Dermatologic | Rash, cutaneous toxicity | 18%; 1-9%; 18% delayed | Variable | ~2 weeks | Weber et al., Iacono et al., Vardhana et al., Owen et al. |
| Dermatologic | Lichen planus | Rare | Not specified | Not specified | Hofmann et al. |
| Pulmonary | Pneumonitis | 3%; 13% delayed | High-grade | 15.1-31.1 weeks median; 3.3-3.5 months | Weber et al., O’Kane et al., Vardhana et al., Owen et al. |
| Hepatic | Hepatotoxicity, hepatitis | 2.3% nivolumab, 0.8% pembrolizumab; up to 11%; 0.7-1.8% autoimmune hepatitis | High-grade | 9-12 weeks; 1-3 months | Weber et al., O’Kane et al., Vardhana et al. |
| Neurologic | Encephalitis, myasthenia-like syndrome | Not specified | Variable | Not specified | O’Kane et al. |

### Chronic and Delayed irAEs

Patrinely et al. provided detailed characterization of chronic irAEs, defined as those persisting at least 12 weeks after therapy cessation. They found that 43.2% of patients developed chronic irAEs, with most being mild (96.4% grade 1-2) and persisting until last follow-up in 85.6% of cases. Certain organ systems were particularly prone to chronicity: endocrinopathies in 83.0% of cases, neurotoxicities in 73.3%, ocular events in 62.5%, xerostomia in 52.9%, and arthritis in 48.9%. In contrast, visceral organ irAEs showed lower chronicity rates, with colitis becoming chronic in only 13.6% of cases, of which 66.7% eventually resolved.

Owen et al. examined delayed irAEs occurring more than 12 months after treatment initiation, with an estimated incidence of 5.3%. The median onset was 16 months (range 12-53). Delayed irAEs were often high-grade, with 39% being grade 3 or higher. The most common delayed irAEs were colitis (22%), rash (18%), and pneumonitis (13%). Among patients developing delayed irAEs, 74% were still on anti-PD-1 at onset, 12% were less than 3 months from last dose, and 14% were more than 3 months from last dose. Early irAEs had occurred in 58% of patients who later developed delayed irAEs, affecting a different organ system in 86% of these cases.

## Management of Immune-Related Adverse Events

### Treatment Modifications

Anti-PD-1 therapy management varied by irAE severity. O’Kane et al. recommended holding treatment for grade 2 symptoms and permanent discontinuation for grades 3-4. Weber et al. noted that symptoms must improve to baseline or grade 1 toxicity before resuming treatment. Vardhana et al. advised delaying or stopping therapy depending on IMAE severity. Across studies, 14-15% of patients required permanent ICI discontinuation due to irAEs.

### Immunosuppressive Therapy

Corticosteroids were the primary immunosuppressive treatment across all studies. Weber et al. recommended oral prednisone at 1-2 mg/kg/day for moderate symptoms, with IV methylprednisolone for severe cases requiring hospitalization. O’Kane et al. specified 0.5-1 mg/kg/day corticosteroids for grade 2 symptoms, 1-2 mg/kg/day for grade 3, and IV corticosteroids for grade 4. Eigentler et al. noted that methylprednisolone or equivalent glucocorticoids managed most AEOSI effectively.

### Healthcare Utilization

The management of irAEs required substantial healthcare resources. Iacono et al. documented 373 unscheduled accesses to healthcare facilities, with 24% (89 accesses) due to irAEs; 78 were unplanned oncology consultations and 11 were emergency department visits. IrAEs led to hospitalization in 14 patients for a cumulative 118 days, with colitis requiring the longest hospitalization (range 4-31 days). The management required 67 specialist consultancies and additional diagnostic examinations.

### Supportive Care and Monitoring

Beyond immunosuppression, supportive care measures included hydration and loperamide for diarrhea, gastroprotection with proton pump inhibitors or histamine blockers. O’Kane et al. recommended surveillance every 12 weeks up to 1 year after immunotherapy discontinuation.

### Treatment Outcomes

#### Resolution and Recovery

Resolution rates varied by irAE type and management approach. Weber et al. reported that approximately 80% of irAEs resolved within a median of 5 weeks with immune-modulating medication, except for endocrinopathies which often persisted.

## Mortality

Fatal irAEs were uncommon but documented. Patrinely et al. reported one patient each with fatal myocarditis and neurotoxicity. Owen et al. documented two irAE-related deaths: one from encephalitis occurring during anti-PD-1 therapy and one from multiple organ irAEs with onset 11 months after cessation. O’Kane et al. noted that deaths from pneumonitis had been reported.

## Impact on Cancer Treatment

The relationship between irAE management and cancer outcomes appeared favorable. O’Kane et al. reported that steroid treatment did not affect PD-1 inhibitor efficacy. Owen et al. found minimal impact on cancer outcomes from delayed irAEs or immunosuppressive therapy. Hoa et al. observed that tumor progression was not associated with immunosuppressive drug exposure before or after ICI initiation and was numerically less frequent in patients with irAEs.

## Synthesis

The evidence reveals several important patterns regarding immune-related adverse events with anti-PD-1 therapy. While overall incidence rates of any irAE ranged from 38-69% across studies, the variation appears largely explained by differences in irAE definitions, follow-up duration, and population characteristics. Studies with longer follow-up naturally captured more events, particularly delayed and chronic irAEs.

The spectrum of irAEs showed consistency across cancer types, with endocrine, gastrointestinal, dermatologic, and pulmonary systems most frequently affected. However, certain populations demonstrated distinct patterns. The Hodgkin lymphoma population studied by Vardhana et al. showed particular vulnerability to pneumonitis (3%), likely related to prior bleomycin and radiation exposure. Patients with preexisting autoimmune disease experienced high rates of disease flares (52%), with severity correlating to pre-treatment immunosuppression intensity, suggesting that baseline immune dysregulation predicts irAE severity.