Atezolizumab-induced myasthenia gravis and myositis: a fatal case report and review of the literature

Abstract

Immune checkpoint inhibitors can cause immune-related adverse events, with neurologic manifestations such as myasthenia gravis (MG) and myositis predominantly reported with PD-1 inhibitors, while cases linked to PD-L1 inhibitors remain rare. An 84-year-old man with hepatocellular carcinoma developed severe MG and myositis after atezolizumab treatment. Despite intensive therapy, including corticosteroids, plasma exchange, and intravenous immunoglobulin, his condition worsened, and he died. MG in cancer patients poses treatment challenges due to tumor progression and functional decline. Given the limited reports on atezolizumab-induced MG, further case accumulation is essential to improve understanding of its clinical course and optimize management strategies.

Introduction

Immune checkpoint inhibitors (ICIs) have transformed cancer treatment by targeting immune cells instead of cancer cells, marking a paradigm shift in oncology. Following their introduction, ICIs have played a crucial role in various cancer therapies. These treatments involve antibodies that counteract specific immune checkpoint molecules, including cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4), programmed cell death protein (PD-1), and its ligand PD-L1. By blocking these checkpoints, they prevent tumor cells from evading immune surveillance and enhance antitumor immune responses. However, these therapies are known to induce immune-related adverse events (irAEs). In particular, PD-1 inhibitors frequently lead to neurological irAEs, with myasthenia gravis (MG) and myositis being commonly reported [9, 19]. In contrast, cases of MG and myositis associated with PD-L1 inhibitors are rare. This report presents the first fatal case of a patient who developed MG and myositis after PD-L1 antibody treatment, despite undergoing multidisciplinary therapeutic interventions.

Case report

An 84-year-old Japanese man presented with muscle weakness. He had a medical history of hypertension and hepatitis C virus infection. He had previously undergone a subsegmental resection of hepatic segment 8 for hepatocellular carcinoma. Six months later, the tumor recurred in segment 6, leading to a decision for systemic chemotherapy. He received three cycles of atezolizumab and bevacizumab while maintaining an Eastern Cooperative Oncology Group performance status of 0, indicating independent activities of daily living (ADL). Before the first cycle, testing for anti-acetylcholine receptor antibodies was negative. After the third cycle, muscle weakness appeared on day 7, followed by a progressive decline in food intake and worsening proximal muscle weakness over the next week, leading to a neurology consultation. On examination, his body temperature was 36.7 °C, blood pressure was 147/95 mmHg, heart rate was 96 bpm, and SpO₂ was 93% on room air. The use of accessory respiratory muscles was observed during breathing. Forced vital capacity was 1.62 L, which was 57.2% of the predicted value, and the forced expiratory volume in one second percentage was 75.9%, indicating respiratory impairment. Neurological examination revealed no muscle atrophy, involuntary movements, pain, or swelling. There were no pressure ulcers that could explain the elevated creatine kinase (CK) levels. Ptosis and diplopia were absent, and no obvious cranial nerve abnormalities were observed. The manual muscle test showed proximal-dominant muscle weakness. Blood tests revealed markedly elevated creatine kinase levels, but autoantibodies associated with myositis were negative. Anti-acetylcholine receptor (AchR) and anti-muscle-specific kinase (MuSK) antibodies, associated with myasthenia gravis, were also negative (Table 1). Nerve conduction studies showed no abnormalities, but the edrophonium test was positive. The repetitive stimulation test was negative for both the orbicularis oculi and abductor pollicis brevis muscles. Transthoracic echocardiography showed no cardiac dysfunction. The patient presented with a progressive, purely motor deficit predominantly affecting proximal muscles. Neurologic conditions, such as systemic deterioration due to tumor progression, drug-induced effects, paraneoplastic syndromes, Lambert–Eaton myasthenic syndrome, and antibody-positive forms of myositis, were ruled out based on antibody testing and other diagnostic findings. The edrophonium test produced a clear clinical improvement, leading to a diagnosis of seronegative MG. Furthermore, elevated serum creatine kinase levels suggested overlapping features of inflammatory myopathy. Cases of myositis and MG have been reported as irAEs associated with PD-1 inhibitors. Considering that atezolizumab is a PD-L1 inhibitor, the involvement of a comparable irAEs mechanism was inferred. From day 2, the patient received three cycles of methylprednisolone pulse therapy (1000 mg/day for 3 days) with 4-day intervals. Following this, prednisolone was initiated at 1 mg/kg and gradually tapered, with tacrolimus (3 mg/day) administered concurrently. Plasma exchange (PLEX) was initiated on the same day as methylprednisolone pulse therapy and performed three times, followed by intravenous immunoglobulin therapy (IVIG) (4 g/kg over 5 days). Despite treatment, motor dysfunction progressed, and ptosis and diplopia developed. Respiratory function did not improve, and mechanical ventilation was initiated on day 4. CK levels improved and normalized after treatment, but despite multidisciplinary interventions, his condition did not improve. His general condition deteriorated due to pneumonia, and he died on day 24. Autopsy was not performed, because consent was not obtained.

Table 1.

Blood and cerebrospinal fluid results

Blood count	Patient’s value	Reference value	Unit	Biochemistry/immunology	Patient’s value	Reference value	Unit
White blood cell count	8.36	3.3–8.6	× 103/μL	Total protein	5.9	6.6–8.1	g/dL
Red blood cell count	4.89	4.35–5.55	× 106/μL	Albumin	2.9	4.1–5.1	g/dL
Hemoglobin	16.1	13.7–16.8	g/dL	Creatine phosphokinase	4494	59–248	U/L
Platelet count	168	158–348	× 103/μL	CK-MB	202	< 5	ng/mL
Neutrophils	83	–	%	Blood urea nitrogen	37.7	8–20	mg/dL
Lymphocytes	7.9	–	%	Creatinine	0.72	0.65–1.07	mg/dL
Monocytes	6.2	–	%	Sodium	142	138–145	mEq/L
Eosinophils	0.6	–	%	Potassium	5	3.6–4.8	mEq/L
Basophils	0.8	–	%	Chloride	107	101–108	mEq/L
Erythrocyte sedimentation rate	17	2–10	mm/hr	C-reactive protein	1.11	< 0.14	mg/dL
				Anti-AChR antibody	< 0.2	< 0.2	nmol/L
				Anti-MuSK antibody	< 0.01	< 0.02	nmol/L
Cerebrospinal fluid analysis				Antinuclear antibody	< 40	< 40	titer
Cell count	1	< 5	cells/μL	Anti-DNA antibody	< 2.0	< 6.0	IU/mL
Mononuclear cells %	100	–	%	Anti-Scl-70 antibody	Negative	–	–
Segmented neutrophils %	0	–	%	Anti-RNA polymerase III antibody	< 5 (negative)	< 28	U/mL
Protein	28.3	10–40	mg/dL	Anti-Jo-1 antibody	Negative	–	–
Glucose	63	40–75	mg/dL	Anti-ARS antibody	< 5.0	< 25.0	U/mL
IgG index	0.43	< 0.73	–	PR3-ANCA	< 1.0	< 3.5	U/mL
				MPO-ANCA	< 1.0	< 3.5	U/mL

Abbreviations: AChR, acetylcholine receptor; ANCA, anti-neutrophil cytoplasmic antibody; CK-MB, creatine kinase-MB; CSF, cerebrospinal fluid; IgG, immunoglobulin G; MPO, myeloperoxidase; MuSK, muscle-specific kinase; PR3, proteinase 3

Discussion

PD-1 is expressed on T cells, B cells, and NK cells [4]. PD-L1, the ligand for PD-1, suppresses T-cell activation upon binding. Tumor cells evade immune surveillance by expressing PD-L1; however, blocking PD-L1 enhances antitumor immune responses, which can lead to tumor elimination [2]. However, the fact that PD-1 knockout mice spontaneously develop autoimmune diseases [21] raises concerns that blocking PD-1 signaling in cancer therapy could contribute to autoimmune manifestations. In fact, disrupting immune system balance through these therapies can lead to irAEs, which may affect various organs and tissues [13]. To date, there have been many reports of MG and myositis associated with PD-1 inhibitors, but reports of PD-L1 are rare. Among patients treated with PD-L1 inhibitors, the frequency of severe irAEs was significantly lower compared to those receiving PD-1 inhibitors [10, 18]. This difference is presumed to reflect several underlying mechanisms, highlighting the complexity of immune checkpoint blockade. While both antibody classes inhibit the PD-1/PD-L1 interaction, their additional molecular targets differ: PD-L1 inhibitors inhibit the interaction between PD-L1 and B7-1 (CD80), whereas PD-1 inhibitors interfere with PD-1 binding to PD-L2 [11, 17]. The distinct binding partners of these molecules may help explain the variation in irAE, and could partly account for the differing incidence of myasthenia gravis and myositis observed between PD-1 and PD-L1 inhibitors.

Most cases of MG and myositis reported after ICI treatment have been linked to PD-1 inhibitors [9, 19]. Although MG and myositis are distinct diseases, their coexistence is uncommon. However, when they arise as irAEs, they often present with overlapping features [9, 19]. In cases of MG induced by the PD-1 inhibitor nivolumab, the positivity rate for anti-AChR antibodies is low [7], and the condition typically develops after an average of 3.6 treatment cycles [9]. Although atezolizumab targets PD-L1 instead of PD-1, the patient exhibited similar characteristics, including MG symptoms, myositis with CK elevation, negative anti-AChR antibodies, and onset after three treatment cycles. These findings suggest that this case falls within the disease spectrum of MG and myositis reported with nivolumab treatment.

In cases of MG associated with PD-1 inhibitor treatment, corticosteroids are the most commonly used therapy, followed by IVIG and PLEX, reflecting the standard approach for managing severe MG and other neuromuscular disorders [9]. However, no large-scale prospective studies have been conducted to determine the optimal treatment strategy, even for nivolumab, a PD-1 inhibitor with relatively more reported cases. The half-life of atezolizumab is 27 days [6], and that of nivolumab is 25 days [3]. In this case, given the need to reduce the drug remaining in the body, treatment was initiated with corticosteroids and plasma exchange (PLEX), followed by IVIG. To date, five cases of atezolizumab-associated myositis [1, 12, 14, 15] and only three cases of MG-associated myositis have been reported [16, 20, 22] (Table 2). Of the two cases with detailed treatment descriptions, one required prolonged mechanical ventilation, while the other suffered cardiac arrest on day 3, both indicating a severe disease course. Atezolizumab-induced MG may be more likely to present with severe manifestations. Differences in binding affinity between anti-PD-1 and anti-PD-L1 antibodies may influence the depth and duration of checkpoint inhibition, potentially explaining the observed clinical disparities [23]. However, the actual frequency of mild cases remains unclear, as such cases may be underrecognized or underreported. Further accumulation of well-characterized cases is needed to clarify the full clinical spectrum. When MG develops as an irAE, prolonged hospitalization is often required, significantly impairing patients' ADL. Moreover, in cases where underlying malignancies cause cardiopulmonary dysfunction, long-term ventilator weaning can be particularly challenging—an issue not commonly encountered in primary MG or myositis. In some cases, aggressive interventions, such as intubation and intensive management, were not pursued [19]. On the other hand, some reports suggest that the occurrence of irAEs is associated with improved survival rates [5, 8]. Given the limited number of reported cases, predicting the clinical course remains difficult, complicating decisions on whether to pursue mechanical ventilation and intensive therapy. In the present case, the patient had no underlying respiratory malignancy and was fully independent in ADL before onset, prompting a multidisciplinary treatment approach. However, the patient's condition did not improve and eventually resulted in death. It remains uncertain whether MG as an irAE induced by atezolizumab is treatable and what clinical course it follows, and accumulating case reports can provide valuable insights for future cases.

Table 2.

Clinical characteristics of myasthenia gravis induced by atezolizumab

	Our case	Reference 14	Reference 15	Reference 16
Age	84	87	56	83
Sex	Male	Male	Male	Male
Medical history	Hypertension	Hypertension, dyslipidemia, and chronic kidney disease	Not reported	Type 2 diabetes
Cancer type	Hepatocellular carcinoma	Urothelial carcinoma	Small cell lung cancer	Hepatocellular carcinoma
Type of adverse events	Myasthenia gravis, myositis	Myasthenia gravis, myositis	Myasthenia gravis	Myasthenia gravis, myositis, myocarditis, autoimmune autonomic ganglionopathy
Dose number at onset	3	2	3	1
Days after last dose	7 days	Not reported	Not reported	7 weeks
Concomitant chemotherapy	Bevacizumab	None	Carboplatin and etoposide	Bevacizumab
MG-related antibodies	Anti-AChR antibody (−), MuSK antibody (−)	Anti-AChR antibody (−), MuSK antibody (−)	Anti-AChR antibody (+), MuSK antibody (−)	Anti-AChR antibody (+), MuSK antibody (−), Kv1.4 antibody (+)
Other antibodies	Anti-ARS antibody (−), ANA (−), anti-RNA polymerase III antibody (−), Jo-1 (−), Scl-70 (−)	Anti-striated muscle antibody (+), VGCC (−)	Anti-striated muscle antibody (−), titin (−)	Anti-striated muscle antibody (+), Kv1.4 (+), titin (+), Jo-1 (−), PL-7 (−), PL-12 (−), EJ (−), KS (−), TIF-1γ (−)
Tensilon test	Positive	Not reported	Not reported	Negative
Repetitive stimulation test	No abnormalities	Not reported	Decreased	No abnormalities
Acute phase treatment	Steroids, Tacrolimus, PLEX, IVIG	IVIG, pyridostigmine	Steroids, IVIG, pyridostigmine	Steroids, IVIG
Course of treatment	Died on the 24th day under ventilator management	Died on the 3rd day, no resuscitation per family's wishes	Improved, details unknown	Discharged on the 193rd day after ventilator support

Abbreviations: AChR, acetylcholine receptor; ANA, antinuclear antibody; ARS, aminoacyl-tRNA synthetase; IVIG, intravenous immunoglobulin; PLEX, plasma exchange; MuSK, muscle-specific kinase; VGCC, voltage-gated calcium channel; TIF-1γ, transcription intermediary factor 1-gamma

Although MG and myositis as irAEs of PD-1 inhibitors are frequently reported, they can also develop with PD-L1 inhibitors. Unlike primary MG or myositis, patients with malignancies may have comorbid systemic deterioration associated with cancer progression, making ventilator weaning and overall management of their condition challenging. Most reported cases of atezolizumab-induced myasthenia gravis have been severe; however, the true clinical spectrum may be broader, with milder cases potentially going unrecognized. Further case accumulation can help clarify its frequency, severity, and treatment course.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The data supporting the findings of this report are available from the corresponding author upon reasonable request and with approval from the co-authors and the Ethics Committee of Ina Central Hospital.

Declarations

Conflict of interest

None.