INTRODUCTION — Checkpoint inhibitors, immunomodulatory antibodies that are used to enhance the immune system, have substantially improved the prognosis for patients with advanced malignancy.

The primary targets for checkpoint inhibition include:

●Programmed cell death receptor 1 (PD-1) and programmed cell death ligand 1 (PD-L1) – Multiple antibodies against PD-1 and PD-L1 are in development and have shown great promise in multiple malignancies. Nivolumab, pembrolizumab, cemiplimab, and dostarlimab, all of which target PD-1, and atezolizumab, avelumab, and durvalumab, all of which target PD-L1, have been approved in various indications (eg, melanoma, renal cell carcinoma, non-small cell lung cancer, head and neck cancer, urothelial carcinoma, Hodgkin lymphoma, Merkel cell carcinoma, and endometrial cancer, as well as microsatellite instability-high or mismatch repair deficient [dMMR] solid tumors).

●Cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) – Ipilimumab, an anti-CTLA-4 antibody, is approved for use in patients with advanced melanoma, based on a significant improvement in overall survival. Another anti-CTLA-4 antibody (tremelimumab) is under development.

Despite important clinical benefits, checkpoint inhibition is associated with a unique spectrum of side effects termed immune-related adverse events (irAEs) or, occasionally, adverse events of special interest [1,2]. IrAEs include dermatologic, gastrointestinal, hepatic, endocrine, and other less common inflammatory events. IrAEs are believed to arise from general immunologic enhancement, and temporary immunosuppression with glucocorticoids, tumor necrosis factor-alpha antagonists, mycophenolate mofetil, or other agents can be an effective treatment in most cases.

Although rare, fulminant and even fatal toxicities may occur with immune checkpoint inhibitors [3], and therefore, prompt recognition and management is important. The side effects of the checkpoint-blocking antibodies targeting the PD-1 and PD-L1 receptors and CTLA-4 are reviewed here. The management approach to irAEs is presented based on clinical experience, since no prospective trials have been conducted to guide the treatment of irAEs. The majority of data are derived from patients with advanced melanoma who were treated with ipilimumab, nivolumab, and pembrolizumab.

The rationale of immunotherapy and the efficacy of checkpoint inhibitors are presented separately. (See "Principles of cancer immunotherapy" and "Systemic treatment of metastatic melanoma lacking a BRAF mutation" and "Management of advanced non-small cell lung cancer lacking a driver mutation: Immunotherapy".)

GENERAL PRINCIPLES — The American Society of Clinical Oncology (ASCO) organized a multidisciplinary panel that reviewed the literature and proposed both general guidelines and organ-systems-specific recommendations for the management of adverse events associated with checkpoint inhibitor immunotherapy [4]. The general approach is outlined in this section, and more specific recommendations are discussed below for the various toxicities. Our approach is consistent with those recommendations and those of the Society for Immunotherapy of Cancer [5].

Dose modifications and immunosuppressive therapy — In general, treatment of moderate or severe immune-related adverse events (irAEs) requires interruption of the checkpoint inhibitor and the use of glucocorticoid immunosuppression. Patients should be carefully monitored during treatment for initial evidence of grade 1 adverse events. Treatment is based on the severity of the observed toxicity:

●For most patients with grade 2 (moderate) immune-mediated toxicities (excluding endocrinopathies), treatment with the checkpoint inhibitor should be withheld and should not be resumed until symptoms or toxicity is grade 1 or less. Glucocorticoids (prednisone 0.5 mg/kg/day or equivalent) should be started if symptoms do not resolve within a week.

•An exception is for patients who experience grade 2 immune-mediated endocrinopathies. In such patients, depending upon the severity of symptoms, immunotherapy may be withheld until hormone replacement is initiated. Immunotherapy may subsequently be resumed once acute symptoms have resolved and (in the event of adrenal insufficiency or hypophysitis) patients are receiving the equivalent of 10 mg of prednisone per day or less. (See 'Endocrinopathies' below and 'Retreatment after prior toxicity' below.)

●For patients experiencing grade 3 or 4 (severe or life-threatening) immune-mediated toxicities, treatment with the checkpoint inhibitor should be permanently discontinued. High doses of glucocorticoids (prednisone 1 to 2 mg/kg/day or equivalent) should be given. When symptoms subside to grade 1 or less, glucocorticoids can be gradually tapered over at least one month.

In the authors' experience, patients who will benefit from glucocorticoids generally do so within days. If symptoms do not clearly improve, particularly after approximately three days with intravenous glucocorticoids, our approach is to administer infliximab (5 mg/kg) rather than continue with a prolonged course of high-dose intravenous glucocorticoids. If symptoms persist after the first infliximab dose, a second dose of infliximab (5 mg/kg) can be repeated two weeks after the initial dose.

The approach to the management of specific toxicities is discussed in the individual sections below.

Impact of immunosuppressive agents on immunotherapy efficacy — For patients who have required glucocorticoids or other immunosuppressives, data suggest that efficacy of the immune checkpoint inhibitor is not impacted. However, for patients being evaluated for retreatment with immunotherapy after experiencing an irAE, the concurrent use of immunosuppressive therapy is associated with reduced efficacy of immunotherapy. (See 'Retreatment after prior toxicity' below.)

Data are as follows:

●Anti-PD-1 antibodies – IrAEs are significantly less frequent with the anti-programmed cell death receptor 1 (PD-1) antibodies compared with ipilimumab.

In an analysis of 576 patients with advanced melanoma treated in four clinical trials, 24 percent received immunosuppressive therapy for the management of treatment-related adverse events [6]. There was no significant difference in the objective response rate between those who received immunosuppressive treatment and those who did not (29.8 versus 31.8 percent). The median duration of response was not reached in those with immunosuppressive therapy, compared with 22 months in those not requiring immunosuppressive therapy.

●Ipilimumab – The most extensive data with ipilimumab come from a single-institution experience that analyzed the incidence of irAEs and treatment outcomes in 298 melanoma patients treated with ipilimumab (3 mg/kg) outside of a clinical trial setting [7]. IrAEs were seen in 254 patients (85 percent), and 103 patients (35 percent) required glucocorticoids. Anti-tumor necrosis factor-alpha therapy was used in 29 cases (10 percent) who did not respond promptly to glucocorticoids.

The median overall survival was 16.5 months, and the estimated two-year survival rate was 39 percent for the entire cohort. Overall survival was the same in patients who had an irAE compared with those without an irAE, and there was no difference between those requiring glucocorticoids and those not requiring immunosuppressive therapy. The time to treatment failure, defined as the need for alternative therapy or death, was 5.7 months for the entire cohort. As with overall survival, there were no significant differences between those with and without an irAE or between those treated with glucocorticoids and those not receiving glucocorticoids.

Relationship between immunotherapy toxicities and efficacy — Most data suggest that irAEs are associated with either improved efficacy of immunotherapy (such as favorable response rates and prolonged survival [6,8-23]) or similar efficacy, compared with those without irAEs [7,24-27]. Although most clinicians commonly associate irAEs with treatment-related toxicity, the presence of an irAE is also a sign that the immune system is sufficiently activated to hopefully additionally target the patient's cancer. Therefore, management of irAEs should involve the minimum amount of immunosuppression needed to control symptoms and, in certain cases, holding therapy without initiating immunosuppression (eg, select immune-mediated endocrinopathies; asymptomatic hepatitis or pancreatitis). (See 'Dose modifications and immunosuppressive therapy' above and 'Hepatotoxicity' below and 'Exocrine pancreas' below.)

Predictive biomarkers of immunotherapy toxicity — Immunologic biomarkers are being studied as a way to predict the risk of irAEs and as an aid in the early identification of such complications. Examples include interleukin 17 (IL-17) [28], eosinophilia [29,30], and combined toxicity scores based on gene expression profiling of immunologically predictive cytokines [31,32].The optimal predictive biomarker remains to be defined.

Combining or sequencing immunotherapy with other agents — Immunotherapy is being combined with other systemic agents to treat a variety of tumors [33]. Such combinations result in varied toxicity profiles [34]. Clinicians should assess and treat the toxicities associated with each individual agent and remain aware of the potential for development of novel toxicities.

●Immunotherapy plus chemotherapy – Details on the efficacy and toxicity of immunotherapy plus chemotherapy in non-small cell lung cancer are discussed separately. (See "Management of advanced non-small cell lung cancer lacking a driver mutation: Immunotherapy", section on 'Preferred option: Immunotherapy plus chemotherapy'.)

●Immunotherapy plus targeted therapy – Details on the efficacy and toxicity of immunotherapy plus antiangiogenic therapy targeting the vascular endothelial growth factor (VEGF) pathway in advanced renal cell carcinoma are discussed separately. (See "Systemic therapy of advanced clear cell renal carcinoma".)

The combination of immunotherapy with agents targeting the epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK) pathways have been associated with life-threatening irAEs (eg, pneumonitis, hepatitis) in patients with advanced non-small cell lung cancer. This is discussed separately. (See "Systemic therapy for advanced non-small cell lung cancer with an activating mutation in the epidermal growth factor receptor", section on 'Immune-related toxicities with EGFR TKI after immunotherapy' and "Anaplastic lymphoma kinase (ALK) fusion oncogene positive non-small cell lung cancer", section on 'Efficacy of other approaches'.)

In patients with metastatic melanoma, checkpoint inhibitor immunotherapy and BRAF plus MEK inhibitors have been safely combined [35-40] and have regulatory approval in this setting. Further details on these combinations are discussed separately. (See "Systemic treatment of metastatic melanoma with BRAF and other molecular alterations", section on 'Is there a role for combined immunotherapy and targeted therapy?'.)

The combination of ipilimumab with BRAF and MEK inhibitors is not used in clinical practice because it has resulted in severe toxicities such as hepatotoxicity, rash, colitis, and intestinal perforation [35,36,41]. The individual efficacy of these agents in melanoma is discussed separately. (See "Systemic treatment of metastatic melanoma lacking a BRAF mutation" and "Systemic treatment of metastatic melanoma with BRAF and other molecular alterations".)

●Allogeneic stem cell transplant after prior immunotherapy – The role of allogeneic stem cell transplantation after prior immunotherapy is discussed separately. (See "Treatment of extranodal NK/T cell lymphoma, nasal type", section on 'Relapsed or refractory disease' and "Treatment of relapsed or refractory classic Hodgkin lymphoma".)

●Ipilimumab plus GM-CSF – Granulocyte-macrophage colony-stimulating factor (GM-CSF) does not have an established role in either the prevention or treatment of side effects associated with checkpoint inhibition. Although data suggest lower rates of treatment-related adverse events with the addition of GM-CSF to high doses of ipilimumab, these results require validation. (See "Systemic treatment of metastatic melanoma lacking a BRAF mutation", section on 'Other agents'.)

SYSTEMIC ADVERSE EVENTS

Fatigue — Fatigue is among the most common side effects seen, with an estimated overall frequency of 16 to 24 percent for the anti-programmed cell death receptor 1 (PD-1) and anti-programmed cell death ligand 1 (PD-L1) agents [1] and approximately 26 percent in those treated with combination immunotherapy [42] However, the fatigue is generally mild, and severe fatigue is rare as a side effect of these agents. When fatigue is present, it is important to exclude thyroid, pituitary, and other endocrine disorders, such as primary adrenal insufficiency. Fever, chills, and infusion reactions have also been described, but these are also rare.

Infusion-related reactions — Mild infusion-related (grade 1 and 2) side effects have been reported in up to 25 percent of patients treated with anti-PD-1 or anti-PD-L1 agents. For patients treated with avelumab, premedication with acetaminophen and an antihistamine is indicated during the first four cycles, and subsequently as needed [43]. (See "Staging, treatment, and surveillance of Merkel cell carcinoma" and "Staging, treatment, and surveillance of Merkel cell carcinoma", section on 'Avelumab'.)

Immunotherapy is discontinued permanently for patients who experience severe (grade 3) or life-threatening (grade 4) infusion reactions. The reported incidence of grade ≥3 infusion-related reactions due to immunotherapy is rare (less than 2 percent). (See "Infusion-related reactions to therapeutic monoclonal antibodies used for cancer therapy", section on 'Agents targeting the programmed cell death receptor'.)

Cytokine release syndrome — Cytokine release syndrome (CRS) is an acute systemic inflammatory response syndrome characterized by fever with or without multiple organ dysfunction mostly described with chimeric antigen receptor (CAR)-T cell therapy. CRS has also been observed with checkpoint inhibitor immunotherapy, such as nivolumab [44,45]. The management of CRS in patients treated with immune checkpoint inhibitors and other antibody-mediated therapies is discussed separately. (See "Cytokine release syndrome (CRS)", section on 'Antibody-associated CRS'.)

DERMATOLOGIC AND MUCOSAL TOXICITY — Cutaneous immune-related adverse events (irAEs) are the most common toxicity associated with checkpoint inhibitors. While the most common clinical manifestation is an inflammatory skin reaction, other presentations exist such as immunobullous disease, vasculitis, neutrophilic dermatoses, and, rarely, severe cutaneous drug reactions. The management of cutaneous irAEs is based upon severity and impact on the patient's functional status.

Further details on the clinical presentation, diagnosis, and management of cutaneous irAEs are discussed separately. (See "Mucocutaneous toxicities associated with immune checkpoint inhibitors".)

DIARRHEA/COLITIS — Diarrhea is a common clinical complaint in patients undergoing treatment with checkpoint-blocking antibodies. Careful attention to the diagnosis and treatment of the earliest symptoms associated with gastrointestinal toxicity can decrease the risk of more severe toxicity.

Manifestations — Diarrhea/colitis most commonly presents approximately six weeks into treatment, which is later than dermatologic toxicity (figure 1A-B) [46].

The differential diagnosis of patients with diarrhea on treatment with a checkpoint inhibitor includes infections with Clostridioides (formerly Clostridium) difficile, other bacterial infections, or viral pathogens, such as cytomegalovirus (CMV). Diarrhea (increase in stool frequency) is related to but clinically distinct from colitis (abdominal pain, radiographic or endoscopic findings of colonic inflammation).

The incidence of diarrhea is much higher in patients receiving cytotoxic T lymphocyte-associated antigen 4 (CTLA-4)-blocking antibodies compared with inhibition of programmed cell death receptor 1 (PD-1).

●Diarrhea of any grade was reported in approximately 30 percent of patients treated with ipilimumab for melanoma, but severe (grade 3/4) diarrhea occurred in less than 10 percent of cases [47]. The incidence of diarrhea appears to be dose-dependent. In a phase II dose-finding study, the rate of severe diarrhea was higher at the 10 mg/kg dose than with 3 mg/kg (10 versus 1 percent) [48].

In clinical trials of ipilimumab for melanoma, significant colitis was reported in approximately 5 percent of patients [47,48]. Endoscopic findings have revealed mucosal edema with biopsies demonstrating neutrophilic, lymphocytic, or mixed neutrophilic-lymphocytic infiltrates (picture 1) [49,50].

●Diarrhea/colitis appears to be less frequent with PD-1 blockade than with CTLA-4 blockade, with grade 3/4 immune-mediated colitis seen in approximately 1 to 2 percent of cases [51-53].

Patients who had significant diarrhea/colitis during CTLA-4 blockade have subsequently been treated with PD-1 blockade using nivolumab without developing diarrhea/colitis [54,55]. (See 'Retreatment after prior toxicity' below.)

Management — Patients being treated with a checkpoint inhibitor should be counseled on the importance of maintaining oral hydration if diarrhea develops. If symptoms persist for more than three days or increase and no infectious causes are identified, prompt assessment and use of oral or intravenous glucocorticoids are required.

●Mild (grade 1) symptoms (fewer than four stools per day over baseline) can be managed symptomatically. Some clinicians feel that the American Dietary Association colitis diet and anti-motility agents (loperamide or oral diphenoxylate atropine sulfate) can be helpful for mild symptoms. Budesonide may be helpful in the early treatment of mild noninfectious diarrhea symptoms that persist but do not escalate after two to three days of dietary modification and anti-motility agents.

●Colonoscopy may be helpful if grade 2 symptoms (increase of four to six stools per day over baseline) or greater occur or in situations where the diagnosis is unclear. Treatment should be initiated if colitis is observed.

●For patients with severe or life-threatening enterocolitis (grade 3/4, increase of seven or more stools per day over baseline or other complications), treatment with checkpoint inhibitors should be permanently discontinued. High doses of glucocorticoids should be given.

●For patients with colitis that is refractory to intravenous glucocorticoids, we obtain repeat infectious work-up, including assessment of CMV colitis. CMV infection or reactivation has been reported in patients experiencing glucocorticoid-refractory immune-mediated colitis with checkpoint inhibitors such as nivolumab [56]. (See "Approach to the diagnosis of cytomegalovirus infection" and "Approach to the diagnosis of cytomegalovirus infection", section on 'Gastrointestinal disease'.)

●If patients do not improve with intravenous glucocorticoids after approximately two to three days on intravenous glucocorticoids, early, prompt treatment with a glucocorticoid-sparing immunosuppressive agent (eg, infliximab) is warranted to rapidly resolve colitis symptoms and reduce the potential infection risk associated with longer durations of glucocorticoids alone [57].

Infliximab is administered at an initial dose of 5 mg/kg. If symptoms persist after the first infliximab dose, a second dose of infliximab is repeated two weeks after the initial dose [58-61]; the dose and schedule is based on experience treating inflammatory bowel disease [62].

It is rare to administer more than two doses of infliximab for immune-mediated colitis, and patients who require more therapy should consider alternative agents. In such cases refractory to infliximab, vedolizumab or mycophenolate [63] may be helpful [63-65]. In preliminary results from one observational study comparing vedolizumab with infliximab in 150 patients with immunotherapy-induced colitis refractory to glucocorticoids, vedolizumab was associated with lower recurrence rates of colitis and less steroid use, although clinical remission rates were similar between the two treatment arms [66].

●Observational data also suggest efficacy of other therapies for patients with refractory immune-mediated colitis, such as tofacitinib, a janus kinase inhibitor [67,68], ustekinumab, an anti-interleukin 12/23 antibody [69], and fecal microbiota transplantation [70]. However, more data are required before incorporating such therapies into the routine care of this patient population. Further details on the use of these therapies in patients with ulcerative colitis are discussed separately. (See "Management of moderate to severe ulcerative colitis in adults", section on 'Tofacitinib' and "Management of moderate to severe ulcerative colitis in adults", section on 'Ustekinumab' and "Management of moderate to severe ulcerative colitis in adults", section on 'Investigational therapies'.)

●In very rare cases, colitis can result in bowel perforation, potentially requiring colostomy.

Prophylactic treatment with the matrix-release glucocorticoid, budesonide is not recommended for the prevention of diarrhea/colitis. This approach was studied as a way to reduce the incidence and/or severity of diarrhea in a double-blind, placebo-controlled phase II study [71]. The rates of diarrhea were similar in both study arms.

Management recommendations are summarized in the American Society of Clinical Oncology (ASCO) guidelines table (table 1) [4].

HEPATOTOXICITY

Manifestations — Elevations in serum levels of the hepatic enzymes, aspartate aminotransferase (AST) and alanine aminotransferase (ALT) can be seen with both cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) and programmed cell death receptor 1 (PD-1) blockade. Most episodes are asymptomatic laboratory abnormalities, but occasionally patients have an associated fever. Rarely, elevations in total bilirubin are seen as well, usually in association with a prolonged period of AST and ALT increase. Among patients that develop liver-related toxicities, the most common time of onset is 8 to 12 weeks after initiation of treatment, although early or delayed events may also be seen (figure 1A-B) [72].

The reported rates of AST and ALT elevations with CTLA-4 blockade have varied in different trials and study populations but typically are less than 10 percent [48,73,74]. In the pivotal phase III study of ipilimumab as monotherapy at 3 mg/kg for patients with advanced melanoma, the rate of elevated AST/ALT was only approximately 1 to 2 percent with no reported grade 3/4 events [47]. On the other hand, hepatotoxicity is more common with the nivolumab plus ipilimumab combination, with approximately 20 percent of patients experiencing grade 3 AST and ALT elevations with the ipilimumab 3 mg/kg and nivolumab 1 mg/kg combination and <5 percent grade 3 with the ipilimumab 1 mg/kg, nivolumab 3 mg/kg combination [75].

In observational studies of PD-1-blocking antibodies, the rates of inflammatory hepatitis are less than 5 percent, and grade 3/4 toxicity is rarer [53,76,77]. The frequency of liver-related laboratory abnormalities, especially grade 3/4 events, may be more common in combination with other agents such as dacarbazine, vemurafenib, or vascular endothelial growth factor receptor (VEGFR) tyrosine kinase inhibitors [78]. (See 'Combining or sequencing immunotherapy with other agents' above.)

Radiographic findings are not typically seen in patients with abnormal liver function tests. However, computed tomography (CT) scan may show mild hepatomegaly, periportal edema, or periportal lymphadenopathy [79]. Biopsies of some patients with hepatotoxicity have demonstrated severe panlobular hepatitis with prominent perivenular infiltrate with endothelialitis. A primary biliary pattern with mild portal mononuclear infiltrate around proliferated bile ductules has also been reported [79,80].

Management — Management recommendations for hepatotoxicity are summarized in the American Society of Clinical Oncology (ASCO) guidelines table (table 1) [4].

Universal hepatitis B screening should be performed prior to initiation of immunotherapy (as for all anticancer therapy [81]), with management according to the algorithm (algorithm 1). Hepatic function (transaminases and bilirubin) should be monitored prior to each dose [35]. When a patient has an elevated AST and/or ALT, we assess for viral or drug-induced causes of hepatitis. Current prescription, over-the-counter, complementary and alternative medications, and herbal supplements should be carefully reviewed for association with hepatotoxicity. (See "Complementary and alternative therapies for cancer".)

Additionally, for a cholestatic pattern or grade 2 or higher transaminitis, cross-sectional imaging of the liver with CT or magnetic resonance imaging (MRI) should be obtained to exclude new or progressive liver metastases [77].

If no other etiology is obvious, then prompt treatment with glucocorticoids following an established algorithm is the next step. Hepatitis may persist for quite some time and may require prolonged or repeated glucocorticoid tapers (a minimum of three weeks treatment is suggested) and/or additional immunosuppression.

●Grade 2 hepatic toxicity – AST or ALT >2.5 times the upper limit of normal (ULN) but ≤5 times the ULN, or total bilirubin >1.5 times the ULN but ≤3 times the ULN. Treatment with the checkpoint inhibitor should be withheld.

●Grade 3 or greater hepatic toxicity – AST or ALT >5 times the ULN, or total bilirubin >3 times the ULN. Treatment should be permanently discontinued.

●In rare cases, elevations in AST and ALT are refractory to glucocorticoid therapy, and mycophenolate mofetil (500 mg every 12 hours) may be an additional medication to administer concurrently with glucocorticoids. The use of antithymocyte globulin therapy has also been described [82].

●Infliximab should not be given to patients with elevated AST/ALT since infliximab itself carries a risk of hepatotoxicity.

PNEUMONITIS — Pneumonitis is an uncommon but potentially severe or fatal complication of treatment with checkpoint inhibitor immunotherapy [83-88]. In some circumstances, pneumonitis can also present as a delayed immune-related adverse event (irAE) (ie, occurring more than one year after initiating therapy [89]). Drug-induced pneumonitis is a diagnosis of exclusion, and alternative diagnoses, including infection and malignancy, need to be excluded.

Manifestations — The clinical spectrum of pneumonitis is illustrated by a series of 43 patients who developed pneumonitis after being treated with an anti-programmed cell death receptor 1 (PD-1) or anti-programmed cell death ligand 1 (PD-L1) monoclonal antibody, either alone or in combination with an anti-cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) agent [85]. Melanoma and non-small cell lung cancer were the most common malignancies treated in this series.

Key observations included the following:

●The overall incidence of pneumonitis was 5 percent (43 patients out of a total of 915 treated with checkpoint blockade), including 3 percent of those treated with anti-PD-1 or anti-PD-L1 monotherapy and 10 percent of those treated with a combination that included an anti-CTLA-4 antibody. The incidence of pneumonitis was similar in patients treated for melanoma and non-small cell lung cancer.

●The duration of treatment prior to the development of pneumonitis was variable, with a median of 2.8 months (range 9 days to 19 months), and was earlier for those treated with combination rather than monotherapy (median 2.7 versus 4.6 months). The most common presenting symptoms were dyspnea and cough (53 and 35 percent, respectively), while one-third of patients were asymptomatic.

●There were no characteristic radiographic or pathologic features that were associated with pneumonitis due to checkpoint inhibitor immunotherapy (figure 2). Chest computed tomography (CT) was the preferred imaging modality, and routine chest radiography did not detect a new radiologic abnormality in almost one-fourth of cases.

●Of the 43 patients diagnosed with pneumonitis, 31 (72 percent) had either grade 1 or grade 2 involvement (table 2). Overall, 37 of 43 (86 percent) improved by simply withholding additional checkpoint inhibition or treating with immunosuppression, while five (0.5 percent of the total population) worsened and died, despite therapy. Of these five cases, infectious complications or progression of tumor appeared to be the proximal cause of death in four instances.

Pulmonary toxicity may also manifest as a radiation recall pneumonitis limited to previously irradiated areas of the lung [90]. This phenomenon can occur when treatment is initiated years after radiation therapy.

Management — The management of pneumonitis is summarized in the American Society of Clinical Oncology (ASCO) guidelines table (table 3) [4].

The outcomes from the 43 patients diagnosed with pneumonitis provide some insight into the optimal management of this complication [85]. Overall, 15 of 17 asymptomatic patients (grade 1) were successfully managed by withholding the checkpoint inhibitor, while 2 of the 17 and all 14 with grade 2 pneumonitis were successfully treated with glucocorticoids. All 12 patients with grade 3 or higher pneumonitis were initially treated with glucocorticoids. Five patients in this group required additional immunosuppression (infliximab with or without cyclophosphamide), but all five ultimately died. The median starting dose of prednisone was 50 mg, and the median duration of treatment required was 68 days (range 20 to 154 days).

There are no prospective clinical trials that have defined the optimal treatment approach. Based on the observations in this series, our empiric approach to treatment includes the following:

●For asymptomatic, grade 1 pneumonitis (table 2), we generally withhold drug for two to four weeks with close follow-up. If symptoms arise or there is radiographic progression, glucocorticoids are appropriate.

●Patients with grade 2 or higher pneumonitis should have their drug withheld and be treated using glucocorticoids with close follow-up. Additional immunosuppression may be used in patients with worsening of pneumonitis, although the benefit of this approach is uncertain.

Special considerations during the COVID-19 pandemic — The COVID-19 pandemic has increased the complexity of cancer care. Important issues in areas where viral transmission rates are high include balancing the risk from treatment delay versus harm from COVID-19, minimizing the use of immunosuppressive cancer treatments whenever possible, mitigating the negative impacts of social distancing during care delivery, and appropriately and fairly allocating limited health care resources. Patients undergoing treatment with checkpoint inhibitor immunotherapy are of special concern, given that treatment-related pneumonitis may mimic COVID-19, and confusion as to the correct diagnosis may delay the initiation of glucocorticoids. These issues and other recommendations for cancer care during active phases of the COVID-19 pandemic are discussed separately. (See "COVID-19: Considerations in patients with cancer".)

ENDOCRINOPATHIES — Inflammation of the pituitary, thyroid, or adrenal glands as a result of checkpoint blockade often presents with nonspecific symptoms such as nausea, headache, fatigue, and vision changes. The incidence of endocrinopathies has been difficult to precisely state due to variable methods of assessment, diagnosis, and monitoring in different clinical trials. The most common endocrinopathies are hypothyroidism, hyperthyroidism, and hypophysitis.

In a systematic review and meta-analysis that included 7551 patients in 38 randomized trials, the overall incidence of clinically significant endocrinopathies is approximately 10 percent of patients treated with checkpoint inhibitors [91]. The frequency of specific endocrinopathies and the relationship to different agents are discussed in this section.

The management of endocrine adverse events involving the thyroid, adrenal, pituitary, or endocrine pancreas is summarized in the American Society of Clinical Oncology (ASCO) guidelines table (table 4) [4].

Autoimmune thyroid disease — Thyroid function should be monitored prior to each dose of a checkpoint inhibitor. Autoimmune thyroid disease can be manifested as primary hypothyroidism secondary to a destructive thyroiditis or as hyperthyroidism associated with Graves' disease.

Hypothyroidism — Most commonly, thyroid disorders present with nonspecific symptoms such as fatigue. Since these symptoms can be vague, distinguishing primary thyroid disorders from secondary hypothyroidism (typically a result of hypophysitis) is critical to a thorough differential diagnosis. Typically, a high thyroid-stimulating hormone (TSH) with low free thyroxine (T4) indicates primary hypothyroidism, and a low TSH with low free T4 indicates hypophysitis. Occasionally, thyroiditis with transient hyperthyroidism (low TSH and high free T4) may be followed by more longstanding hypothyroidism (high TSH and low free T4).

For patients treated with ipilimumab, nivolumab or pembrolizumab, atezolizumab, and the combination of ipilimumab plus nivolumab, the incidence rates for hypothyroidism were 3.8, 7.0, 3.9, and 13.2 percent, respectively [91]. There was no significant difference in incidence with nivolumab and pembrolizumab (6.5 and 7.9 percent, respectively).

Management of primary hypothyroidism typically involves replacement with thyroid hormone (levothyroxine) and endocrinology consultation. For cases of acute thyroiditis, a short period of high-dose glucocorticoids (1 mg/kg of prednisone or equivalent) may be helpful, but there is little strong evidence to suggest that this prevents longer term thyroid dysfunction.

Hyperthyroidism — Persistent primary hyperthyroidism is significantly less frequent than hypothyroidism and should be treated similarly to primary hyperthyroidism. (See "Graves' hyperthyroidism in nonpregnant adults: Overview of treatment".)

For patients treated with ipilimumab, nivolumab or pembrolizumab, atezolizumab, and the combination of ipilimumab plus nivolumab, the incidence rates for hyperthyroidism were 1.7, 3.2, 0.6, and 8 percent, respectively [91].

Hypophysitis

Clinical presentation and diagnosis — Typically, hypophysitis is manifested by clinical symptoms of fatigue and headache. The diagnosis is established by low levels of the hormones produced by the pituitary (adrenocorticotropic hormone [ACTH], TSH, follicle-stimulating hormone [FSH], luteinizing hormone [LH], growth hormone [GH], prolactin). (See "Clinical manifestations of hypopituitarism".)

Laboratory findings differentiate hypophysitis from primary adrenal insufficiency (manifested by low cortisol or inappropriate cortisol stimulation test and high ACTH) and primary hypothyroidism (manifested by low free T4 and high TSH). The diagnosis of hypophysitis is also supported radiographically by enhancement and swelling of the pituitary gland (image 1) [92,93].

For patients treated with ipilimumab, nivolumab or pembrolizumab, atezolizumab, and the combination of ipilimumab plus nivolumab, the incidence rates for hypophysitis were 3.2, 0.4, <0.1, and 6.4 percent, respectively [91].

Management — When hypophysitis is suspected, a course of high-dose glucocorticoids (1 mg/kg of prednisone daily) given during the acute phase may result in reversal of the inflammatory process in some cases and prevent the need for longer term hormone replacement.

In most patients, however, long-term supplementation of the affected hormones is necessary due to secondary hypothyroidism (treated with levothyroxine) or secondary hypoadrenalism (treated with replacement doses of hydrocortisone, typically 20 mg each morning and 10 mg each evening). In some cases, patients can be successfully weaned from replacement steroids over time [94]. (See "Treatment of hypopituitarism".)

Adrenal insufficiency — The most critical endocrinopathy is adrenal insufficiency, which can cause dehydration, hypotension, and electrolyte imbalances (hyperkalemia, hyponatremia) and constitutes an emergency. Adrenal insufficiency is rare and has been reported in 0.7 percent of patients treated in randomized clinical trials [91].

When an adrenal crisis is suspected, intravenous glucocorticoids and immediate hospitalization is warranted. Consultation with an endocrinologist, aggressive hydration, and evaluation for sepsis are also critical. (See "Clinical manifestations of adrenal insufficiency in adults" and "Treatment of adrenal insufficiency in adults".)

Type 1 diabetes mellitus — Treatment with checkpoint inhibitors has been associated with acute onset of type 1 diabetes mellitus in approximately 0.2 to 0.9 percent of cases [91,95]. As examples, in several case series, patients typically presented with severe hyperglycemia or diabetic ketoacidosis; all required insulin therapy at diagnosis and remained insulin-dependent for diabetic control [95,96].

It is important to monitor glucose with each dose of immunotherapy. Patients who develop immunotherapy-induced type 1 diabetes mellitus are typically treated with insulin therapy. In contrast to other immune-related adverse events, treatment with glucocorticoids or infliximab is not effective in these patients, due to the almost complete destruction of the pancreatic beta cells by immunotherapy [95]. Immunotherapy-related diabetes may be suggested when the C-peptide level is low in the setting of hyperglycemia. (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Treatment" and "Management of blood glucose in adults with type 1 diabetes mellitus".)

OPPORTUNISTIC INFECTIONS

Risk with immunosuppressive therapy — For most patients with no underlying pulmonary conditions receiving glucocorticoids for less than six weeks to treat an uncomplicated irAE, we do not suggest Pneumocystis pneumonia (PCP) prophylaxis. Although PCP prophylaxis is often administered to patients on prolonged courses of glucocorticoids for other indications, it may not be necessary for those receiving them for an uncomplicated irAE. The incidence of opportunistic infections in patients with an irAE is low in observational studies, ranging between 2 and 7 percent [97,98]. Additionally, glucocorticoids in this patient population theoretically do not induce greater immunosuppression, but rather reduce immune activation from the checkpoint inhibitor immunotherapy. This approach differs from the typical use of PCP prophylaxis in other HIV-uninfected patients. (See "Treatment and prevention of Pneumocystis pneumonia in patients without HIV", section on 'Prophylaxis'.)

However, prolonged immune suppression may put certain patients at risk for unusual or opportunistic infections. We offer PCP prophylaxis to patients receiving glucocorticoids for an irAE in the setting of combined chemotherapy plus immunotherapy; those with underlying pulmonary conditions receiving glucocorticoids for an uncomplicated irAE; or those with a complicated irAE (eg, those requiring longer than six weeks of glucocorticoids, or those not responsive to glucocorticoids requiring additional immunosuppressive therapy). The role of prophylactic antifungal or antiviral therapies in these patients requires further study.

●In one series of 740 patients treated for advanced melanoma [97], approximately 80 percent of patients had been treated with ipilimumab, either alone or in combination with nivolumab. Serious infections were reported in 54 cases (7.3 percent). Specific infectious etiologies were bacterial, viral, fungal, and parasitic in 46, 6, 5, and 1 cases, respectively.

These infections were predominantly seen in association with glucocorticoids or infliximab. The incidence of serious infections was lower in those treated with nivolumab and pembrolizumab alone compared with ipilimumab or ipilimumab plus nivolumab, but this may have reflected the lower incidence of serious side effects necessitating immunosuppression.

Further details on the use of Pneumocystis pneumonia prophylaxis in HIV-uninfected patients are discussed separately. (See "Treatment and prevention of Pneumocystis pneumonia in patients without HIV".)

LESS COMMON IMMUNE-RELATED ADVERSE EVENTS — Treatment with checkpoint inhibitors has also been associated with less common side effects in other organs. In some instances, these side effects have been severe or fatal.

Kidney — Acute kidney injury (AKI) is a rare but potentially serious complication of checkpoint inhibitor immunotherapy [99]. For patients with severe immune checkpoint inhibitor-associated acute kidney injury (ICPi-associated AKI), discontinuation of checkpoint inhibitor immunotherapy and treatment with glucocorticoids are indicated.

The estimated incidence of ICPi-associated AKI is approximately 1.5 to 5 percent [100-102]. The most common reported underlying pathology is acute tubulointerstitial nephritis (image 2), but immune complex glomerulonephritis and thrombotic microangiopathy have also been observed [103-106].

In one retrospective observational study of 138 patients, ICPi-associated AKI was diagnosed approximately 14 weeks after initiation of immunotherapy [100,103]. One meta-analysis also suggested an association between PD-1 inhibitor treatment and hypocalcemia (RR 10.9, 95% CI 1.4-84.2) [99]. (See "Clinical manifestations and diagnosis of acute interstitial nephritis", section on 'Drugs'.)

The management of kidney toxicities is summarized in the American Society of Clinical Oncology (ASCO) guidelines table (table 5) [4].

Exocrine pancreas — Monitoring serum amylase and lipase in asymptomatic patients is not recommended unless pancreatitis is suspected clinically. Glucocorticoid treatment is not indicated in patients with modest asymptomatic elevations in serum amylase and lipase, as long as there are no other signs or symptoms of pancreatic inflammation.

Elevated levels of serum amylase and lipase have been reported in many patients in trials of cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) and PD-1 blockade (approximately 10 to 15 percent grade 3/4), but these laboratory findings do not fulfill the criteria for acute pancreatitis and most of these patients are asymptomatic. The clinical significance of elevated amylase/lipase in the absence of associated symptoms remains unknown [107,108]. (See "Clinical manifestations and diagnosis of acute pancreatitis".)

Immune-related pancreatitis has been reported in a patient treated with CTLA-4 blockade [109]. In this patient, who already had a prior history of gallstone pancreatitis, abdominal symptoms were accompanied by laboratory evidence of elevated amylase and lipase.

Neurologic — A wide range of neurologic syndromes have been associated with checkpoint blockade involving ipilimumab and PD-1 pathway inhibitors [110-112]. Case series suggest that neurotoxicity occurs in approximately 1 to 14 percent of patients, with the highest rates associated with the use of combined immunotherapy using nivolumab plus ipilimumab [113-115]. As an example, in a review of approximately 9000 patients, the incidence of neurologic immune-related adverse events (irAEs) was 4 percent with CTLA-4 blockade, 6 percent with PD-1 inhibitors, and 12 percent with combined CTLA-4 and PD-1 inhibition [5].

Headache and peripheral sensory neuropathy are the most commonly encountered symptoms [5]. In regards to more severe toxicity, cases of Guillain-Barré syndrome are particularly notable [116], and one case resulted in a treatment-related death in a postsurgical adjuvant study of ipilimumab [117]. (See "Adjuvant and neoadjuvant therapy for cutaneous melanoma", section on 'Agents not commonly used'.)

Other reported severe neurologic complications include myasthenia gravis [1,118], posterior reversible encephalopathy syndrome (PRES) [119], aseptic meningitis [120], enteric neuropathy [121], transverse myelitis [122], pancerebellitis [123], autoimmune encephalitis [124], and cranial and peripheral neuropathies. (See "Paraneoplastic and autoimmune encephalitis" and "Paraneoplastic syndromes affecting spinal cord, peripheral nerve, and muscle", section on 'Checkpoint inhibitor-associated neuropathies'.)

Neurologic irAEs typically develop within three months of starting immune checkpoint inhibitor therapy, although the time of onset may vary [125]. Serious neurologic irAEs should be treated with glucocorticoids, even when the non-immunotherapy-induced version of the same neurologic disorder is not typically managed with glucocorticoids (such as Guillain-Barré syndrome). Consultation with neurology is indicated to consider additional treatment, such as plasmapheresis and intravenous immunoglobulin. Early recognition and intervention are key to reducing severity and duration of toxicity. As an example, in one observational study of 35 patients treated for neurologic irAEs, symptoms resolved in 26 patients (75 percent), with a median time to resolution of approximately one month [114].

The management of patients with immunotherapy-related neurologic toxicity is summarized in the ASCO guidelines table (table 6) [4].

Neurologic toxicities associated with other molecularly targeted and biologic agents are discussed separately. (See "Neurologic complications of cancer treatment with molecularly targeted and biologic agents".)

Cardiovascular toxicity — Cardiotoxicity may develop in the absence of a history of significant cardiac risk factors and may be associated with a more general myositis as well as other irAEs. Venous thromboembolism may also be seen, although its relationship to checkpoint inhibitor immunotherapy is less clear.

The time to onset was variable, but fatal myocarditis has been reported after a single treatment with the combination of nivolumab plus ipilimumab [126]. In pharmacovigilance studies, the incidence of myocarditis was higher in patients treated with the combination of nivolumab plus ipilimumab compared with nivolumab alone (0.27 versus 0.06 percent).

High-dose steroids have been used to treat cardiac complications, but symptoms may progress in some cases despite aggressive therapy. Immediate transfer to a coronary care unit or, if available, cardiac transplant unit should be considered for patients with elevated troponin or conduction abnormalities. The early institution of cardiac transplant rejection doses of steroids (methylprednisolone 1 g every day) and the addition of either mycophenolate, infliximab, antithymocyte globulin, or abatacept [127] should be considered in patients without an immediate response to high-dose steroids.

The management of cardiovascular toxicities is summarized in the ASCO guidelines table (table 7) [4].

Hematologic — Red cell aplasia, neutropenia, thrombocytopenia, acquired hemophilia A, and cryoglobulinemia have been described in patients treated with checkpoint inhibitors [128-134]. As with other irAEs, the standard approach is initial glucocorticoid treatment with addition of other immune-suppressing agents if symptoms are steroid refractory.

Autoimmune hemolytic anemia (AIHA) has been associated with the use of multiple immunotherapy agents (eg, atezolizumab, nivolumab, pembrolizumab, and ipilimumab) [135,136]. The treatment approach to immunotherapy-induced AIHA is similar to that of AIHA due to other drugs, and it includes immediate discontinuation of immunotherapy and initiation of glucocorticoid-based therapy. Further details regarding the diagnosis and management of this condition are discussed separately. (See "Drug-induced hemolytic anemia".)

Hemophagocytic lymphohistiocytosis (HLH) has been reported in patients receiving immunotherapy with nivolumab, ipilimumab, and/or pembrolizumab [137,138]. This is a rare but potentially fatal syndrome of excessive immune activation resulting in multi-organ failure, including cytopenias and bleeding. Further data are needed to confirm the association of either the PD-1 pathway or CTLA-4 inhibition with the development of HLH. Details regarding the treatment of HLH are discussed separately. (See "Treatment and prognosis of hemophagocytic lymphohistiocytosis".)

The management of more common hematologic toxicities is summarized in the ASCO guidelines table (table 8) [4].

Eye — CTLA-4 blockade with ipilimumab has been associated with eye inflammation, which can be manifested by episcleritis, conjunctivitis, uveitis, or orbital inflammation. The incidence is less than 1 percent, and symptoms can include photophobia, pain, dryness of the eyes, and blurred vision.

Intraocular inflammation (uveitis) following treatment with pembrolizumab or nivolumab is a rare but clinically important event described in approximately 1 percent of treated patients. Although the available data are limited, the risk of eye disorders may be aggravated when drugs of both checkpoint inhibitor classes are combined. (See "Ocular side effects of systemically administered chemotherapy", section on 'Anti-PD-1 and PD-L1 agents'.)

An ophthalmology consultation is recommended, and treatment with topical glucocorticoids (eg, 1 percent prednisolone acetate suspension) may be helpful. Oral glucocorticoids can be used for severe (grade 3/4) or refractory cases. The management of eye toxicity is summarized in the ASCO guidelines table and discussed in detail separately (table 9). (See "Ocular side effects of systemically administered chemotherapy", section on 'Ipilimumab'.)

Rheumatologic and musculoskeletal — A wide range of rheumatologic toxicities has been observed with checkpoint inhibition immunotherapy. These include myositis, inflammatory arthritis, salivary gland dysfunction (sicca syndrome), and vasculitis, among others [139-143]. The incidence of these side effects has not been clearly determined. (See "Rheumatologic complications of checkpoint inhibitor immunotherapy".)

Myositis can also be seen with checkpoint inhibition immunotherapy and is occasionally severe/fatal. Given the possibility of severe cases, Health Canada issued a safety alert urging recognition and management of myositis [144]. Patients who have significant myositis should be evaluated for myocarditis since these syndromes occasionally manifest together.

The management of musculoskeletal toxicities is summarized in the ASCO guidelines table (table 10) [4].

RETREATMENT AFTER PRIOR TOXICITY — Retreatment with immunotherapy may be safely offered to some patients with significant immune-related adverse events (irAEs) during initial treatment that consists of either a cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) inhibitor and/or a programmed cell death receptor 1 (PD-1)/programmed cell death ligand 1 (PD-L1) checkpoint inhibitor [54,55,145-150].

The choice to retreat is dependent on multiple factors, including the severity and nature of the initial irAE, its degree of responsiveness to systemic immunosuppression (eg, glucocorticoids or infliximab), and the availability of alternative treatment options. Data are limited on the specific patient populations who should not be offered retreatment, and clinical judgment is necessary. For example, patients who survive a frequently fatal irAE (eg, myocarditis) are not routinely offered retreatment. Similarly, retreatment is discouraged in those receiving steroid doses equivalent to prednisone 10 mg daily or higher for treatment of an initial irAE, as concurrent use of prednisone is associated with reduced efficacy of immunotherapy [151]. (See "Management of advanced non-small cell lung cancer lacking a driver mutation: Immunotherapy", section on 'Impact of steroids on efficacy of immunotherapy'.)

Data are also limited for the benefits of immunotherapy retreatment in those experiencing an initial irAE, particularly in those who have had a complete or sustained response to the initial regimen and do not require further intervention. Clinicians who offer this approach should also provide a careful risk-benefit discussion to patients who are candidates for retreatment. For these patients, the optimal choice of retreatment agent varies in clinical practice, although patients who experience severe toxicity from initial CTLA-4 blockade are typically offered retreatment with single-agent PD-1 or PD-L1 monotherapy rather than repeat CTLA-4 blockade. Dose reductions of immunotherapy are not recommended with retreatment, as this approach has not been assessed in clinical trials.

When rechallenged with checkpoint inhibitor therapy, the rate of recurrent irAEs ranged between 18 and 88 percent [54,145-148,152], with the highest rates among those receiving CTLA-4 blockade following discontinuation of PD-1/PD-L1 blockade for toxicity. However, a majority of the irAEs were mild and managed successfully with immunosuppressive therapy, such as glucocorticoids or immunomodulatory agents (such as infliximab or vedolizumab). The following sequences of retreatment strategies after prior toxicity have been investigated:

●Retreating with PD-1/PD-L1 inhibition after prior PD-1/PD-L1 inhibitor toxicity – In one study of 38 patients with lung cancer rechallenged with a PD-1 inhibitor, 52 percent developed recurrent irAEs, with one-half of these experiencing their previous irAE. However, over one-half of the recurrent irAEs were grade ≤2, and a majority resolved or improved with immunosuppressive therapy [147]. Two other studies of recurrent irAEs with PD-1/PD-L1 inhibitor retreatment reported incidence rates of 37 and 55 percent [54,148].

●Retreating with PD-1/PD-L1 inhibition after prior CTLA-4 blockade toxicity

•In a single-arm, open-label, phase II trial (CheckMate 172), approximately 1000 patients with advanced melanoma who progressed on prior ipilimumab subsequently received nivolumab at 3 mg/kg every two weeks for up to two years [153]. Among the subset of 84 patients with an initial grade ≥3 ipilimumab-related irAE, none experienced recurrent grade ≥3 diarrhea or colitis.

•Among 67 patients with advanced melanoma and a major irAE from ipilimumab who were subsequently retreated with pembrolizumab or nivolumab, 34 percent (23 patients) developed a new irAE, whereas only 3 percent (two patients) had a recurrence of the ipilimumab-related irAE [145].

●Retreating with PD-1/PD-L1 inhibition after prior dual CTLA-4/PD-1 inhibitor toxicity

•In one study, 80 patients with advanced melanoma and initial irAEs from combined CTLA-4 and PD-1 inhibitors were subsequently retreated with either single-agent nivolumab or pembrolizumab [146]. Among these patients, 18 percent experienced a recurrent irAE, including one patient who died of Steven-Johnson syndrome. Other grade ≥3 recurrent irAEs included hepatitis (7 percent); colitis, dermatitis, and elevated lipase (3 percent each); and hypophysitis (1 percent).

•In another study of approximately 170 patients with melanoma, lung cancer, or genitourinary cancer treated with immunotherapy, 88 patients developed immune-mediated colitis on initial CTLA-4 blockade; for approximately one-half of these patients, the initial regimen included both a CTLA-4 and a PD-1 inhibitor [54]. Among the 64 patients eventually retreated with a single-agent PD-1/PD-L1 inhibitor, the rate of recurrent colitis was 27 percent. However, in the entire study population, severe toxicity was rare (grade ≥3 diarrhea in 4 percent), and most cases were successfully managed with immunosuppressive therapy (84 percent with glucocorticoids and 13 percent with infliximab or vedolizumab).

●Retreating with CTLA-4 blockade after prior PD-1/PD-L1 inhibitor toxicity – In the study discussed above, 79 patients who initially developed immune-mediated colitis after initial therapy with a PD-1/PD-L1 inhibitor and were subsequently treated with CTLA-4 blockade experienced high rates of recurrent colitis (88 percent) [54].

Further studies are needed to clarify the most effective and safest approach to immunotherapy retreatment.

PATIENTS WITH VULNERABILITIES TO IMMUNOTHERAPY TOXICITIES

Preexisting autoimmune disease — In patients with preexisting autoimmune disease, there are limited data on the safety and efficacy of checkpoint inhibitor immunotherapy [27,154]. Such patients were excluded from clinical trials evaluating immunotherapy due to concerns about exacerbating underlying autoimmune disease or immune-related adverse events (irAEs) [155]. Observational studies suggest that most patients with autoimmune disease can safely receive checkpoint inhibitor immunotherapy [27,145,156-160]. However, compared to those without autoimmune disease, these patients may be at higher risk for specific irAEs (eg, immune-mediated colitis in those with inflammatory bowel disease [IBD]) or discontinuing immunotherapy due to irAEs [27]; they may also be at higher risk for developing an exacerbation of their underlying autoimmune disorders [145,155-157,159].

Clinicians should offer such patients a cautious risk-benefit discussion prior to initiating immunotherapy and evaluate various factors including the effectiveness of immunotherapy for the underlying malignancy; performance status and comorbidities; and type and severity of the autoimmune condition. Patients should also be evaluated for contraindications to immunotherapy (eg, poorly controlled autoimmune disease, active immunosuppression using the equivalent of prednisone 10 mg daily or higher) that require alternative therapies for their cancer [161]. Immunotherapy should be used with extreme caution or avoided in patients with potentially life-threatening autoimmune conditions [162]. We also recommend discussion of treatment options with both the patient and the clinician responsible for treating their autoimmune condition before determining a specific plan of action. Further details on the general approach to managing irAEs in patients with specific autoimmune disorders are discussed separately. (See "Rheumatologic complications of checkpoint inhibitor immunotherapy", section on 'General principles of evaluation and management' and "Systemic treatment of metastatic melanoma with BRAF and other molecular alterations", section on 'Defining immunotherapy eligibility'.)

In a prospective observational cohort study of 4367 patients with advanced melanoma, 415 patients (10 percent) with autoimmune disease were identified [27]. This cohort included 227 patients with rheumatologic disease (rheumatoid arthritis, systemic lupus erythematosus, and others); 143 patients with endocrine disease (hypo- or hyperthyroidism, Graves' disease); 155 patients with IBD; 8 patients with "other" autoimmune conditions, and 20 with multiple autoimmune conditions. Of note, patients with autoimmune disease and active immunosuppression were more likely to receive alternative initial therapy (eg, BRAF plus MEK inhibitors) rather than immunotherapy.

Among the 55 percent who received immunotherapy (228 patients), the incidence of grade ≥3 irAEs was similar to that seen in a cohort of patients with advanced melanoma and no autoimmune disease, regardless of the choice of immunotherapy (30 percent each for a cytotoxic T lymphocyte-associated antigen 4 [CTLA-4] inhibitor; 17 versus 13 percent for a programmed cell death receptor 1 [PD-1] inhibitor; and 44 versus 48 percent for combination immunotherapy). Among those treated with PD-1 inhibitors, patients with autoimmune disease were more likely to discontinue immunotherapy due to irAEs compared with those without autoimmune disease (17 versus 9 percent). Patients with IBD treated with PD-1 inhibitors were more likely to develop immunotherapy-induced colitis (6 of 31 patients [19 percent]) versus those with other (non-IBD) autoimmune diseases (3 percent) and those without autoimmune disease (2 percent), which is consistent with prior studies [155]. However, data on acute flares of patients' preexisting autoimmune condition were not reported.

Objective response rates and overall survival were similar for those with and without autoimmune disease, although this could be biased by the imbalance of baseline treatment characteristics between the two groups. The impact of other immunosuppressive agents (eg, infliximab or vedolizumab) on clinical outcomes is not known. In addition, approximately one-third of the study population was comprised of patients with thyroid disease (for whom there is less concern regarding irAEs). However, this study did not include patients with more rare, potentially fatal autoimmune conditions/irAEs (eg, myositis, myasthenia gravis, and Guillain-Barré syndrome) [3]. (See 'Relationship between immunotherapy toxicities and efficacy' above.)

Older adult patients — Checkpoint inhibitor immunotherapy appears to have similar efficacy and toxicity in those ≥65 years and those <65 years of age, and chronologic age alone should not preclude the use of these agents [163-166].

A meta-analysis of nine randomized trials that compared nivolumab, pembrolizumab, or atezolizumab with chemotherapy or targeted therapy in solid tumors (non-small cell lung cancer, melanoma, renal cell carcinoma, head and neck cancer) analyzed efficacy in 5458 patients [163]. There was a consistent overall survival advantage for immunotherapy across all trials (hazard ratio [HR] 0.69, 95% CI 0.63-0.74), and there was no difference between those <65 years (HR 0.68, 95% CI 0.61-0.75) and those ≥65 years (HR 0.64, 95% CI 0.54-0.76). An analysis of progression-free survival based on age was available in four trials. There was no significant difference between the two age cohorts (HR 0.73, 95% CI 0.61-0.88, and HR 0.74, 95% CI 0.60-0.92, respectively).

Limitations of this meta-analysis include the potential lack of applicability to those with an impaired performance status or significant comorbidity, the limited number of patients 75 years or older, and the lack of relative toxicity data in the older adult population versus younger patients.

SUMMARY AND RECOMMENDATIONS

●Impact of checkpoint inhibitor immunotherapy – Checkpoint inhibitors that target cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) and programmed cell death receptor 1 (PD-1) are having a dramatic impact on the care of patients with advanced melanoma, lung cancer, urothelial cancers, and kidney cancers, among others. (See "Systemic treatment of metastatic melanoma lacking a BRAF mutation" and "Treatment of metastatic urothelial cancer of the bladder and urinary tract" and "Management of advanced non-small cell lung cancer lacking a driver mutation: Immunotherapy" and "Systemic therapy of advanced clear cell renal carcinoma".)

●Common immune-related adverse events (irAEs) – Treatment is associated with irAEs that typically are transient but occasionally can be severe or fatal. The most common and important irAEs are dermatologic, diarrhea/colitis, hepatotoxicity, pneumonitis, and endocrinopathies, although other sites can also be affected. (See 'Dermatologic and mucosal toxicity' above and 'Diarrhea/colitis' above and 'Hepatotoxicity' above and 'Pneumonitis' above and 'Endocrinopathies' above and 'Less common immune-related adverse events' above.)

•Rapid identification of irAEs and prompt initiation of local or systemic immunosuppression can optimize outcomes.

•Side effects are more common with the anti-CTLA-4 antibody ipilimumab than with the anti-PD-1 agents (nivolumab, pembrolizumab) or anti-programmed cell death ligand 1 (PD-L1) agents (atezolizumab, durvalumab, avelumab). The combination of nivolumab plus ipilimumab is associated with more toxicity than either agent alone. (See 'General principles' above.)

●Treatment approach – In general, treatment of moderate or severe irAEs requires interruption of the checkpoint inhibitor and the use of glucocorticoid immunosuppression. Treatment is based on the severity of the observed toxicity (see 'General principles' above):

•Grade 2 irAEs – For patients with most grade 2 (moderate) immune-mediated toxicities, treatment with the checkpoint inhibitor should be withheld and should not be resumed until symptoms or toxicity is grade 1 or less. Glucocorticoids (prednisone 0.5 mg/kg/day or equivalent) should be started if symptoms do not resolve within a week.

-Grade 2 endocrinopathies – For those with grade 2 immune-mediated endocrinopathies, immunotherapy may be withheld until hormone replacement is initiated and subsequently resumed once acute symptoms have resolved and (in the event of adrenal insufficiency or hypophysitis) patients are receiving the equivalent of 10 mg of prednisone per day or less.

•Grade 3 or 4 irAEs – For patients experiencing grade 3 or 4 (severe or life-threatening) immune-mediated toxicities, treatment with the checkpoint inhibitor should be permanently discontinued. High doses of glucocorticoids (prednisone 1 to 2 mg/kg/day or equivalent) should be given. When symptoms subside to grade 1 or less, glucocorticoids can be gradually tapered over at least one month.

•Refractory toxicity – If glucocorticoids are not effective in treating immunotherapy-related diarrhea/colitis after approximately three days, infliximab (5 mg/kg) may be considered. Infliximab should not be given to patients with immune-mediated hepatitis. (See 'Hepatotoxicity' above.)

●Communication – Frequent and consistent communication between patients, caregivers, and the clinical team is vital to successful irAE management.

●Immunotherapy plus other agents – Immunotherapy in combination with other systemic agents may result in varied toxicity profiles. Clinicians should assess and treat the toxicities associated with each individual agent and remain aware of the potential for development of novel toxicities. (See 'Combining or sequencing immunotherapy with other agents' above.)

●Indications for PCP prophylaxis – For patients with no underlying pulmonary conditions receiving glucocorticoids for less than six weeks to treat an uncomplicated irAE, we suggest observation rather than pharmacologic Pneumocystis pneumonia (PCP) prophylaxis (Grade 2C).

We offer PCP prophylaxis to patients receiving glucocorticoids for an irAE in the setting of combined chemotherapy plus immunotherapy; those with underlying pulmonary conditions receiving glucocorticoids for an uncomplicated irAE; or those with a complicated irAE (eg, those requiring longer than six weeks of glucocorticoids or additional immunosuppressive therapy). (See 'Opportunistic infections' above.)

●Preexisting autoimmune disease – In patients with preexisting autoimmune disease receiving immunotherapy, limited observational data suggest a similar incidence of overall irAEs compared with those without autoimmune disease, but a potentially higher risk for developing specific irAEs (eg, immune-mediated colitis in patients with inflammatory bowel disease) and/or discontinuing therapy. Although data suggest that most patients with autoimmune disease can safely receive immunotherapy, clinicians should offer a cautious risk-benefit discussion prior to initiating therapy, evaluate for contraindications (eg, poorly controlled autoimmune disease, active immunosuppression) that require alternative therapies, and involve the clinician treating the patient's autoimmune condition. (See 'Preexisting autoimmune disease' above.)

●Retreatment after prior toxicity – A majority of patients with significant irAEs during initial treatment with either a CTLA-4 inhibitor and/or a PD-1/PD-L1 checkpoint inhibitor may be safely retreated with immunotherapy after a risk-benefit discussion. The choice to retreat is dependent on multiple factors, including the severity and nature of the initial irAE, its degree of responsiveness to systemic immunosuppression, the clinical response to the initial immunotherapy regimen, and the availability of alternative treatment options. Data are limited on the effectiveness of this retreatment approach. (See 'Retreatment after prior toxicity' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Jedd Wolchok, MD, PhD, who contributed to an earlier version of this topic review.