The PubMed search identified 11 papers and a search of congress abstracts identified nine records. After removal of duplicates, six records (four papers and two abstracts) met the inclusion/exclusion criteria and are listed in Supplementary Table S1.
The most common (>10%) treatment-related AEs associated with each combination that may be immune-related or associated with axitinib monotherapy included endocrine (hypothyroidism and hyperthyroidism), dermatologic (rash/inflammatory dermatitis and pruritus), gastrointestinal (colitis [diarrhoea], hepatic signals [alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) increased] and nausea), musculoskeletal (inflammatory arthritis [arthralgia]) and general disorders and administration-site conditions (fatigue and infusion-related reaction, Supplementary Table S2).
The currently available guidelines (i.e., ASCO, EAU, ESMO and SITC)19,20,21 provide sufficient guidance, and no further description is required for hypothyroidism, hyperthyroidism, rash/inflammatory dermatitis, pruritus and arthralgia (Supplementary Table S2). Rare AEs of potential interest were also identified, including myasthenia gravis, myocarditis, necrotising fasciitis, pneumonitis, necrotising pancreatitis, cephalalgia and abdominal distension. For these rare AEs, which are mostly seen with ICIs and not TKIs, it is recommended to follow the currently available guidelines.19,20,21
The recommendations herein focus on AEs that require a differential diagnosis of aetiology (i.e., immune-related vs. axitinib-related AEs), and for which there are currently no published guidelines specific to treatments with axitinib–ICI (Supplementary Table S2). Clinical judgement should be exercised while managing AEs, with up- or downranking of AE grade based on patient’s overall clinical status.
Diarrhoea was commonly reported to be related to patients treated with axitinib monotherapy (45%) and in patients treated with combinations of axitinib/avelumab (54%, Supplementary Table S3) and axitinib/pembrolizumab (49%).2,13,14,25,26 Diarrhoea was less common in patients treated with monotherapy avelumab (13%) or pembrolizumab (19%, Supplementary Table S3).7,9 For the combinations axitinib/avelumab and axitinib/pembrolizumab, immune-related diarrhoea was reported in 9 (2%) and 11 (3%) patients, respectively.13,14 Immune-related colitis was reported in 4 (1%) patients treated with axitinib/avelumab. Based on the incidence of immune-related diarrhoea with single-agent ICIs, a higher frequency is anticipated with the combined treatments. The exact aetiology of diarrhoea with these two classes of drugs remains unknown, although aspects of the clinical presentation may be distinct with VEGF-targeted therapy being associated with a slower onset of symptoms, and lack of fever and abdominal pain.
A flow chart describing how to manage diarrhoea induced by treatment with the combination of axitinib and ICIs is presented in Fig. 1. It is first recommended to assess the patient’s condition and general state, and to look for alarming clinical signs (e.g., bloody stools, dehydration, fever, sudden onset of watery diarrhoea with rapidly worsening frequency or clinical deterioration), which should trigger immediate intervention and/or referral to a specialist/hospital. Immune-related colitis, especially at higher grades, requires treatment with corticosteroids. In this paper, we use prednisolone as an example, but other equivalent corticosteroids may be used (e.g., prednisone and methylprednisolone). In patients in whom the clinical index of suspicion for immune-related toxicity is high and in those with clinically alarming signs, empiric treatment with prednisolone (1–2 mg/kg) should be initiated immediately, before disclosure of confirmatory test results. Endoscopic evaluation may be considered at the time of the first occurrence and should be attempted in all patients with refractory diarrhoea. Measuring baseline levels of C-reactive protein, stool lactoferrin and/or calprotectin can help identify patients who are experiencing immune-related colitis.27,28 While the first occurrence of diarrhoea on therapy is often treatment-associated, other causes may be considered. Previous antibiotics or use of immune-suppressive agents increase the risk for infectious causes of diarrhoea, which should be considered during the process of clinical diagnosis. Hence, assessment of Clostridium difficile, ova and parasite, cytomegalovirus or other viral aetiologies should also be considered.
In the absence of alarming signs, axitinib dosing should be withheld for 24–48 h, and the patient should be monitored closely for clinical outcomes. The short half-life of axitinib in the plasma (2.5– 6.1 h) allows for a quick decrease in plasma concentrations after withholding the drug and, as a result, quick recovery from axitinib-related AEs.29 If the diarrhoea is axitinib-induced, the symptoms should be partially or completely better within 48 h and, if not, it could potentially be immune-related. If there is improvement (i.e., a decrease in diarrhoea) after withholding axitinib and monitoring for 48 h, consider optimisation of supportive measures, intermittent dosing (e.g., breaks of 24–48 h from axitinib as needed for diarrhoea) and/or axitinib dose reduction. The duration of periodic breaks should be driven by severity, and by relief and reoccurrence upon breaks. In principle, intermittent dosing allows maintenance of a higher axitinib dose compared with dose reduction. Breaks should consider time to occurrence/resolution of diarrhoea, and should be individualised to the patient’s needs, rather than a fixed, discontinuous schedule for all patients.
If there is no improvement after withholding axitinib and monitoring for 48 h, assume immune-related diarrhoea and follow existing guidelines.20,21,24 For grade 1–2 axitinib-related diarrhoea, consider continuing therapy with dietary adjustment and/or antidiarrhoeal medication supplementation. Axitinib treatment can be withheld or reduced due to persistent, low-grade diarrhoea. Patients whose diarrhoea cannot be managed despite dietary adjustment and antidiarrhoeal medication, and who present with signs and symptoms of dehydration, should be considered for extended diagnostics; referral to a specialist/hospitalisation may be considered, regardless of the grade of the event. In cases of bloody stools, high-volume watery stools or sudden onset of severe diarrhoea, immediate action is required. Corticosteroid treatment should be initiated, along with hospitalisation and referral to specialist care if appropriate.
Increased transaminases are considered diagnostic for hepatic toxicity, and may have multiple causes, including immune toxicity. Progression of liver metastases is of principal interest in the differential diagnosis, but the incidence of primary treatment failure with axitinib plus ICI remains a rare event (10.5–11.5%).13,14 De novo occurrence of elevated transaminases during treatment in the absence of new concomitant medication or other potential hepatotoxins is highly suspicious for treatment-related toxicity and should be treated as such, without delay. On the contrary, in patients with previous exposure to corticosteroids, tests for infectious causes should be included in the diagnostic algorithm. Activation of viral pathogens should be considered in patients at risk (i.e., previous exposure to immunosuppressant medications), and anti-hepatitis A virus immunoglobulin (Ig) M, hepatitis B surface antigen (HBsAg), anti-HBsAg, anti-hepatitis B core antigen IgM and anti-hepatitis C virus (HCV) Ig should be assayed. In the case of positive HCV results, HCV RNA should be assayed.
The incidence of treatment-related ALT and AST increases, respectively, was reported in 9% and 6% of patients treated with axitinib monotherapy, in 13% and 11% of patients treated with axitinib/avelumab and in 24% and 23% of patients treated with combinations of axitinib/pembrolizumab (Supplementary Table S3).2,13,14 The FDA label of the combinations axitinib/avelumab and axitinib/pembrolizumab includes special warnings for hepatic toxicity and hepatitis.30,31
A flow chart describing how to manage axitinib plus ICI treatment-induced hepatitis is presented in Fig. 2. It is first recommended that the patient’s condition and general state is assessed. For grade 1–2 hepatitis, withhold axitinib and monitor closely for symptoms for 48–72 h before making a final judgement on whether it is axitinib or immune-related. Liver function tests should then be rechecked. If there is improvement of hepatitis after withholding axitinib, consider further treatment interruptions and axitinib dose reduction as described previously for diarrhoea.
If there is no clinical improvement after withholding axitinib, both drugs should be stopped, and corticosteroid treatment considered. Worsening hepatitis after withholding axitinib while continuing avelumab or pembrolizumab indicates that the ICI agents are related to the AE; in these cases, corticosteroid treatment should be initiated. For grade 2 immune-related transaminitis, low-dose prednisolone (0.5 mg/kg) is recommended, with twice-weekly monitoring of liver function tests. If no improvement is seen within a 2-week window, addition of mycophenolate mofetil (MMF) 0.5–1 g twice daily is recommended (Fig. 3). Because there is an observed delay in the response to MMF treatment (~5–7 days), it is important to consider timely initiation of MMF and maintaining prednisolone use once MMF is started. In the case of inadequate response within 2 weeks, dose escalation of prednisolone to 1–2 mg/kg once daily is mandated. Immunosuppression should be adapted to the clinical needs in the case of clinical deterioration or severe worsening of abnormalities.
In cases with grade 3–4 hepatitis, interruption of both agents and the initiation of high-dose prednisolone 1–2 mg/kg is recommended. For these patients and for any patients with concomitant bilirubin elevation, a full investigation of hepatic function, including imaging and referral to a hepatologist, is indicated.
Fatigue is a symptom with multidimensional causes, which may include endocrine abnormalities as a reversible cause.32 The overlap of fatigue between both treatment modalities and cancer itself implies a diagnostic dilemma. In the absence of endocrine dysfunction and non-axitinib-related fatigue, other reversible causes and cancer-related fatigue should be considered.
Treatment-related fatigue was commonly reported in patients treated with axitinib monotherapy (27%) as well as in those treated with combination axitinib/avelumab (36%) or axitinib/pembrolizumab (30%, Supplementary Table S3).2,13,14 Fatigue was not reported as immune-related in the combination phase III trials.2,13,14 In the phase I trial of axitinib/avelumab, treatment-related fatigue was reported in 46% of patients, and no cases of fatigue were reported as immune-related.26 In the phase I trial of axitinib/pembrolizumab, treatment-related fatigue was reported in 73% of patients and immune-related fatigue in 12%.25 Treatment-related fatigue was also common in patients treated with avelumab (18%) or pembrolizumab monotherapies (25%, Supplementary Table S3).7,9
Fatigue can be an early sign of multiple AEs and causalities (e.g., hypothyroidism). Treatment-related hypothyroidism was reported in patients treated with axitinib monotherapy (21%) as well as in those treated with combination axitinib/avelumab (24%) or axitinib/pembrolizumab (32%).2,13,14 Proper endocrine surveillance throughout the course of treatment identifies reversible causes of fatigue. Sudden onset, rapid deterioration within a few days, gradual increase in the setting of stable disease, de novo and/or severe cephalgia or de novo visual impairment should trigger investigations of the hormonal axis (thyroid-stimulating hormone, free triiodothyronine, free thyroxine, total testosterone, oestrogen and cortisol). Abnormalities should trigger hormone replacement, and referral to endocrinology should be considered. In any other case, axitinib should be withheld for 48–72 h to assess for improvement before investigating further (Fig. 4). If improvement of fatigue occurs upon stopping axitinib, intermittent treatment or dose reductions for axitinib should be considered.
Cardiovascular AEs occur frequently during treatment with axitinib, which may include hypertension (49%) or chest pain (5%). Immune-related cardiac events are rare (<1%), but have been reported to be fatal.13,14 The spectrum of immune-related cardiac events is broad and includes arrhythmias and myocarditis. Myocarditis was reported in 0.06% of patients treated with nivolumab and 0.27% of patients treated with the combination nivolumab/ipilimumab.33 Clinical signs of dyspnoea, heart failure or arrhythmias should prompt immediate attention and diagnostic workup, and early referral to a specialist should be considered. Electrocardiogram, troponin I or T, N-terminal pro-brain natriuretic peptide and creatinine kinase are recommended as diagnostics to assess cardiac symptoms. Additional workup may be warranted.
Axitinib-associated hypertension is frequent and may be associated with the cardiovascular events listed previously. Therefore, continuous monitoring and early medical treatment with antihypertensive agents (e.g., calcium channel blockers or angiotensin-converting enzyme inhibitors) is recommended for axitinib-associated hypertension. The objective is to achieve a blood pressure window of 120/80–140/95 mmHg. Treatment often requires antihypertensive combinations, but usually results in sufficient response. Dose reduction of axitinib is rarely needed in order to control hypertension. Notably, holding axitinib for toxicity as noted previously can result in hypotension in patients who had previous intensification of their antihypertensive regimen. Thus, careful monitoring of blood pressure during axitinib holds, and consideration of holding antihypertensive agent(s) concomitantly should be considered.
Cardiac symptoms in the absence of hypertension may be immune-related and should prompt further investigation and cardiologic consultation, as symptoms may be masked. The exact rate of immune-related cardiac toxicities with the combination of axitinib and ICI is currently not known, and proper monitoring is recommended.
Where available, cardiac magnetic resonance imaging may reveal evidence of inflammatory changes, but the role of imaging modalities remains vague, and biopsy should be considered in some patients. Due to the high mortality associated with immune-related myocarditis, aggressive management needs to be initiated with high-dose prednisolone (1–2 mg/kg). Cardiac monitoring in an intensive-care unit should be considered, given the high risk for high-grade conduction delays and ventricular arrhythmias. Additional immunosuppression may be required in severe cases or based on the myocardial biopsy findings, although it is not advisable to delay treatment and wait for the full investigative findings. The decision to treat should be made clinically.