Do peripheral neuropathies differ among immune checkpoint inhibitors? Reports from the European post-marketing surveillance database in the past 10 years

Although the immunotherapy advent has revolutionized cancer treatment, it, unfortunately, does not spare cancer patients from possible immune-related adverse events (irAEs), which can also involve the peripheral nervous system. Immune checkpoint inhibitors (ICIs), blocking cytotoxic T-lymphocyteassociated protein 4 (CTLA-4), programmed cell death protein 1 (PD-1), or programmed cell death ligand 1 (PD-L1), can induce an immune imbalance and cause different peripheral neuropathies (PNs). Considering the wide range of PNs and their high impact on the safety and quality of life for cancer patients and the availability of large post-marketing surveillance databases, we chose to analyze the characteristics of ICI-related PNs reported as suspected drug reactions from 2010 to 2020 in the European real-world context. We analyzed data collected in the European pharmacovigilance database, Eudravigilance, and conducted a systematic and disproportionality analysis. In our study, we found 735 reports describing 766 PNs occurred in patients treated with ICIs. These PNs included Guillain-Barré syndrome, Miller-Fisher syndrome, neuritis, and chronic inflammatory demyelinating polyradiculoneuropathy. These ADRs were often serious, resulting in patient disability or hospitalization. Moreover, our disproportionality analysis revealed an increased reporting frequency of PNs with tezolizumab compared to other ICIs. Guillain-Barré syndrome is a notable potential PN related to ICIs, as it is associated with a significant impact on patient safety and has had unfavorable outcomes, including a fatal one. Continued monitoring of the safety profile of ICIs in real-life settings is necessary, especially considering the increased frequency of PNs associated with atezolizumab compared with other ICIs.


Introduction
Peripheral neuropathies (PN) include several pathologic conditions that affect the peripheral nervous system and result from damage to peripheral nerves. Nerve injury can cause a wide range of disorders, mainly characterized by various degrees of alterations in sensitivity, pain, muscle strength and endurance, osteotendinous reflexes, and/or fine motor skills (1).
The symptoms and signs of PNs depend on the type of damaged nerves. Even if the hands and the feet are most frequently involved in PNs, the gastrointestinal, cardiovascular, and urogenital systems can also be affected (2). In particular, bowel, bladder, or digestive problems, in addition to drops in blood pressure causing dizziness, are other possible signs and symptoms of PN related to damage to autonomic nerves. So, these conditions can have a significant impact on the quality of life of patients. PNs causes can be different, like inherited and hereditary aspects, traumatic injuries, infections, metabolic problems, nutritional deficiencies, or toxic exposure, but also drugs or specific diseases can induce them (2)(3)(4)(5)(6)(7)(8). Sensorimotor polyneuropathies are often experienced by cancer patients (9), and these can be related to their oncologic condition, due to tumor invasion or compression exercised on nerves, or can be associated with a paraneoplastic effect (10). Furthermore, PNs can often also be iatrogenic, being associated with treatments (11). PNs are well-known as adverse events of classic chemotherapy or radiotherapy, which can damage healthy nerve tissue (12). Uncommonly, neuropathies can also be immune-mediated. The different possible causes make a correct differential diagnosis difficult (13).
Among immune-mediated PNs, those induced by immunotherapies have recently emerged. Although the advent of oncological immunotherapy has revolutionized cancer treatment, leading to long-lasting tumor responses, it, unfortunately, does not spare cancer patients from possible adverse events, like PNs. Indeed, this new type of immune-driven neuropathy is added to the toxic neuropathies associated with exposure to traditional chemotherapies (13)(14)(15). Immune checkpoint inhibitors (ICIs), which block cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), programmed cell death protein 1 (PD-1), or programmed cell death ligand 1 (PD-L1), can also induce immune imbalance and cause different immune-related adverse events (16), also in the peripheral nervous system (Figure 1) (17)(18)(19)(20)(21)(22). From our previous study, PNs emerged as the second nervous subclass of ICIrelated neurological complications more frequently described in the European pharmacovigilance database (23). Therefore, given the wide range of PNs and their impact on patients' quality of life and the availability of large pharmacovigilance databases (24), we decided to analyze in more detail the characteristics of ICI related PNs emerging from 10 years of ICI-use in a real-world context. clinical data, even the rarest ones. The analysis of ADR cases collected in these databases allows for the extraction of important information useful for clinicians' prompt differential diagnosis and management. Analysis of these databases allows for constant drug safety surveillance in order to identify any new important safety signal coming from the real-world context. Therefore, for our analysis, we retrieved from the EV database all safety cases reporting PN complications that occurred in patients treated with at least one ICI and were collected in EV from 01/01/2010 until 07/ 02/2020. The ICIs included in our analysis were the ones authorized by the European Medicines Agency (EMA) until January 2020: ipilimumab, nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, and avelumab, in mono-or combination therapies.
The EV is the European pharmacovigilance database, managed by the EMA. It collects all individual case safety reports (ICSRs) describing cases of suspected ADR related to drugs or vaccines. In fact, there are different levels of access to the data via the EMA website (www.adrreports.eu), the most comprehensive of which requires specific authorization from the agency. According to recent pharmacovigilance legislation, ICSR present in the EV database can be reported both by a healthcare professional and a non-healthcare professional (e.g., a citizen or other professional figures). The reported adverse events are categorized according to the Medical Dictionary for Regulatory Activities (MedDRA). According to a hierarchical structure, this dictionary is characterized by five different levels, from the lowest one (the lowest level term, LLT), up to the highest one (the System Organ Class, SOC). Thus, it is possible to select specific cases in EV by searching for SOCs. In light of this, we selected all cases reporting at least one ICI as a suspected drug and an ADR belonging to the SOC "Neurological disorders", focusing our analysis on those cases that described at least one neurological complication belonging to the MedDRA "Peripheral neuropathies" High-Level Group Terms (HLGT).

Descriptive analysis
ADR reports were analyzed for patient characteristics, describing age group, gender, and therapies, and differentiating suspected ICI (ipilimumab, nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, or combination ICI treatments) from other suspected or concomitant drugs (Level II ATC). Moreover, reports were categorized by source, including reporter type (Healthcare Professional, HCP, or Non-Healthcare Professional, N-HCP), and the country for regulatory purposes, distinguished into the European Economic Area and the Non-European Economic Area. Our aim was to compare which type of PN was more frequently reported as a suspected ADR for each ICI treatment. PNs were described in terms of action taken to manage them, outcome, and severity, with severity criteria specified in accordance with the International Council on Harmonization E2D guidelines. In particular, an ADR was considered serious when it resulted in death, hospitalization or its prolongation, severe or permanent disability, or congenital anomalies/birth defects, or if it was a life-threatening or clinically relevant event. The outcome of neurological complications was classified as "recovered/resolved", "recovering/resolving", "recovered/resolved with sequelae", "not recovered/not resolved", "fatal", and "unknown". Moreover, PNs were categorized according to the MedDRA High-Level Terms (HLT), and the different neurological diagnoses (p-term) included in each reference HLT.

Disproportionality analysis
In order to compare the frequency of PNs reporting between each ICI or ICI class, we performed] disproportionality analysis applying the Reporting Odds Ratio (ROR) of 95% CI and considered a statistically significant signal if the lower limit of the 95% CI of a ROR exceeded 1.0. For our analysis, we considered the first and most frequent PNs belonging to each different HLT.
The reporter distribution remained the same for each ICI, for which more than 80% of ICSRs were reported by HCPs. Ipilimumab-related ICSRs showed a slightly higher percentage of N-HCPs (14.7%) as the reported source compared to the other ICIs (Table 1).
Looking at all ICSRs, the male gender was more frequently represented (N=458; 62.3%). Similarly, when our dataset was also stratified for each suspected drug ICI, the male gender was reported in more than 50% of ICSRs, particularly for those related to ipilimumab (68.8%) or nivolumab (66.3%). Only in four ICSRs related to the association of ipilimumab/pembrolizumab, the female gender was represented in 75.0% of ICSRs (Table 1).
In terms of the primary country of origin for regulatory purposes, the Non-European Economic Area was the most representative one for all ICSRs (N=463; 63.0%), especially for those related to atezolizumab (37 out of 51; 72.5%) and pembrolizumab (138 out of 193; 71.5%) ( Table 1).
We categorized all reported PNs, such as ICI-related irADRs, according to the MedDRA HLTs in Table 2. "Peripheral neuropathies NEC" (N=496; 64.8%) were the most common HLT.
In the majority of ICSRs, the reported PNs resulted in serious ADRs (95.2%), most of which were medically significant (N=335; 43.7%), or caused or prolonged hospitalization (N=228; 29.8%). Moreover, PNs resulted in death in 81 cases (10.6%) or disability in 43 cases (5.6%) ( Table 3). In particular, the higher percentages of disabling PNs were associated with pembrolizumab (N=15; 7.4%) and ipilimumab (N=8; 7.0%). In the same table, we reported the distribution of PNs by outcome. The outcome was unknown in 45.5% of cases, especially for PNs related to ipilimumab therapy (55.3%). Overall, 22.2% of ICI-related PNs had unfavorable outcomes, defined as "not resolved" or "resolved with sequelae" at the time of reporting (20.9% and 1.3%, respectively). Positive outcomes were reported in 28.5% of adverse events. In particular, 9.7% of PNs (N=74) were completely resolved, while 18.8% (N=144) improved. In contrast, a fatal outcome was observed in 3.8% (N=29) and Guillain-Barrésyndrome was the most common PN (N=12; 41.4%) among those with fatal outcomes.
In 131 ICSRs, suspected drugs other than ICIs were reported ( Table 1); in the majority of cases, only one (N=51) or two (N=49) suspected drugs other than ICIs were reported. Looking at each ICI, suspect drugs were reported in 40 of 264 nivolumab-related ICSRs  Table 1). The most frequently reported other suspect drugs were other antineoplastic agents (N=206; 73.3%), in particular taxanes (N=56) and platinum compounds (N=51). The distribution of other suspected drugs for Level II ATC is reported in Table 4.
In the majority of ICSRs (N=521; 70.9%), there were no other concomitant medications, which were only reported in 214 ICSRs. More than five concomitant drugs were reported in 93 ICSRs, mainly related to nivolumab (N=37), followed by pembrolizumab (N=25). Generally, the most frequently reported concomitant drug classes were other antineoplastic agents (N=126; 10.9%), followed by analgesics (N=99; 8.5%) and drugs for acid-related disorders (N=83; 7.2%). The distribution of concomitant drugs for Level II ATC is reported in Figure 2.
Finally, when considering the reported actions taken to manage the adverse events, drug withdrawal was reported in 53.6% of cases (N=394), while the suspected ICI dose remained unchanged in 7.2% of cases (N=53). Only one report described a peripheral neuropathy managed with the ICI dosage reduction. This case was referred to an adult male patient (age group 18-64 years) affected by Hodgkin's disease in treatment with nivolumab (3 mg/kg/iv).

Discussion
ICI-related neuropathies represent one of the most frequently reported neurological complications in the European database (23). Given how in our study the PNs described in the majority of retrieved ICSRs were categorized as serious, these adverse events represent an important safety concern for the therapeutic use of ICIs. Severe motor or sensory neuropathies are among the complications of ICIs that may require permanent discontinuation of ICI therapy. Moreover, even if ICI withdrawal may not be necessary, the occurrence of PNs can have a critical impact on the patient's quality of life and lead to negative outcomes. Immunotherapy-related neuropathies represent a small subset of autoimmune PNs due to an aberrant immune response against components of the peripheral nervous system. This abnormal immune response induced by immunotherapeutic agents can lead to various adverse manifestations, characterized by possible demyelination processes or axonal damages (25, 26). In our dataset, both types of injuries emerged, although demyelinating PNs were more frequently reported than axonal injuries associated with ICIs. Among the demyelinating disorders, we found several cases of Guillain-Barreś yndrome, demyelinating polyneuropathy, and chronic inflammatory demyelinating polyradiculoneuropathy (CIDP). We also found some cases of ICI-related Miller-Fisher syndrome, a rare variant of Guillain-Barrésyndrome characterized by a triad of symptoms: ophthalmoplegia, ataxia, and areflexia (26). Fortunately, these neurologically adverse events are rare. According to a systematic review conducted by Yan Li et al., from 1990 up to 2021, only 33 ICI-related case reports of Guillain-Barrésyndrome or its subtypes, such as Miller-Fisher syndrome (MFS), acute inflammatory demyelinating polyneuropathy (AIDP), and acute motor axonal neuropathy (AMAN), were described in the literature (27). In our analysis, a few cases of AMAN and motor-sensory axonal neuropathy emerged, which were mainly related to pembrolizumab. Although rare, these adverse events require special attention from clinicians given their potentially serious consequences. In addition to the characteristic damage type, PNs can also be categorized based on the time of onset and duration. According to our results, the acute forms were more frequently reported as ICI-related ADRs than the chronic ones. Among the latter, our dataset described some cases of chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), an immune-mediated progressive neuropathy characterized by electrodiagnostic evidence of peripheral nerve demyelination, and only one case of multifocal motor neuropathy (MMN). According to our results, these chronic forms were mainly related to nivolumab treatment, either as monotherapy or in association with ipilimumab. According to Okada et al., polyradiculoneuropathy induced by ICIs presents specific clinical features, such as severe motor weakness affecting both legs symmetrically and resulting in gait disturbance, and an effective response to steroid treatment for its management. In addition, the authors suggested its early detection by electrodiagnostic evaluation of demyelination and cerebrospinal fluid, typically characterized by elevated lymphocytes (28). Other case reports have been described in the literature describing acute sensorimotor neuropathy and polyneuropathy occurring in patients treated with ipilimumab (25, 29). The possible relationship between the occurrence of neuropathy events and the administration of ICIs is supported by a close temporal association between symptomatic onset and a positive response to corticosteroids or immunomodulatory therapy (30). Among the acute PNs emerging from our analysis, GBS is noteworthy. In particular, it was reported in 20% of ICSRs. Moreover, Guillain-Barrésyndrome was associated with a significant impact on patient safety, as it was the most   "-" means that 0 cases were reported.
prevalent PN with adverse outcomes, including fatalities. As reported in the literature, there is an increased incidence of fatal Guillain-Barreś yndrome in patients treated with immune checkpoint inhibitors (31). Overall, fatal cases in our dataset were more frequently related to nivolumab and ipilimumab. Moreover, the majority of disabling PNs were also related to anti-PD-1 agents, nivolumab, and pembrolizumab. Overall, the majority of all ICSRs were related to these three ICIs, which are the older approved ICIs and therefore the most widely used and longest used in clinical practice. In the same way, the higher incidence of tumor pathologies like lung cancer, melanoma, and renal cell cancer in elderly male patients should be considered when looking at the gender and age distribution of the reported PNs. In fact, these tumors are more common in adult or elderly men. However, the data in the literature regarding the greater susceptibility of men to ICI-related irADRs, including PNs, are contradictory and still controversial, requiring further investigation (32).
Concerning the biological plausibility of ICI-related PNs, results from experimental and clinical studies confirm the possible immunemediated pathogenesis of these disorders. In particular, inhibition of Treg cell activity may be one potential mechanism for breaking immune tolerance. It has been shown that T-reg cells expressing PD-1 and CTLA-4 receptors on their surfaces are involved in self-tolerance processes. AntiPD-1 and antiCTLA-4 inhibitors block immune checkpoints that break physiological immune tolerance, with a consequent hyperproliferation and hyperactivation of immune effector cells, which can lead to the development of neurological adverse events, such as peripheral neuropathies (33). Humoral and/or cellular immune mechanisms against Schwann cell/myelin antigens may underlie or participate in the pathogenesis of ICI-related PNs (34,35). This pathogenetic mechanism should be linked to the phenomenon of molecular mimicry in terms of cross-reactivity between the tumor antigens and similar epitopes on healthy peripheral nerve cells. In addition, epitope spreading (36, 37) is one of the suggested mechanisms. PNs can be stimulated to increase the production of inflammatory cytokines, such as TNF-alpha and IL-17, by activated effector T-cells and their consequent effects on the structures of the peripheral nervous system (29). As suggested by Xi Chen et al., another possible hypothesis may be related to a pre-existing neuropathy, which may be induced by previous or concurrent chemotherapy or due to comorbidities such as diabetes. A pre-existing condition could increase the patient's susceptibility to immune-mediated ICIs complications, which could worsen after ICIs treatments (30). In our dataset, 107 ICSRs reported at least one anticancer agent as other suspect drugs, which is considered a predisposing factor or alternative cause of PNs. Si Zhihua et al. described a higher risk of developing peripheral neuropathy when PD-1/PD-L1 inhibitors were used in combination with chemotherapy (38). Similarly, Yuan Tian et al. also showed an increased incidence trend of neurological toxicities, especially grade 3-5 peripheral neuropathy, with anti-PD-1 and anti-PD-L1 plus chemotherapy (39).
Finally, our disproportionality analysis surprisingly revealed an increased frequency of reports of peripheral neuropathy associated with atezolizumab compared to other ICIs, both in monotherapy and in combination therapy with ipilimumab/nivolumab. Although some case reports of neuropathy following atezolizumab treatments were reported in the literature (40)(41)(42), these differences in neurological adverse events reporting with anti-PD-L1 agents, in particular atezolizumab, compared to other ICIs have not been previously documented in the literature. Thus, the cause of this higher frequency of reporting remains largely unknown, requiring further investigation.

Study strengths and limitations
Our study has inherent limitations due to the post-marketing surveillance system, as it was based on data from the European pharmacovigilance reporting system. For example, considering that safety reports are mainly sent in a spontaneous way by patients and physicians, they may be affected by the so-called underreporting phenomenon (43). In addition, the data reported in spontaneous ICSRs may often be incomplete (44). Moreover, information about the patients' previous predisposing conditions is not available. These can only be inferred from the reported concomitant and suspected medications. As we did not have full-level access to EV, we were unable to analyze the time to onset or time to resolution, or the median age of the patients, which was only reported as an age group. Despite these limitations, pharmacovigilance databases are useful tools for monitoring the safety of medicines and represent a significant source of information (44). Spontaneous reporting systems allow for a better characterization of the safety profile of drugs and overcome the inherent limitations of clinical trials (45,46). In recent years, the efforts of regulatory agencies and the scientific community have greatly improved the quantity, quality, and timeliness of ADRs reporting. These advances have increased the value of spontaneous reporting systems for post-marketing pharmacovigilance purposes. It is worth emphasizing that drug and vaccine safety signals require further ad hoc confirmatory Drugs reported as concomitants in the Individual Case Safety Reports (ICSRs) collected in Eudravigilance, categorized by therapeutic reference class (Level II ATC). A02, Antacids; A03, Combinations of psycholeptics and antispasmodics; A04, Antiemetics and antinausea; A05, Biliary and hepatic therapy; A06, Drugs for constipation; A07, Antidiarrheals, anti-inflammatory/anti-infective agents; A10, Drugs used in diabetes; A11, Vitamins; A12, Mineral supplements; B01, Antithrombotic agents; B02, Antihemorrhagics; B03, Antianemic preparations; B05, Blood substitutes and perfusion solutions; C01, Cardiac therapy; C02, Antihypertensives; C03, Diuretics; C05, Vasoprotectives; C07, Beta blocking agents; C08, Calcium channel blockers; C09, Agents acting on the renin-angiotensin; C10, Lipid modifying agents; D07, Corticosteroids. dermatological preparations; D10, Antiacne preparations; D11, Other dermatological preparations; G01, Gynecological anti-infectives and antiseptics; G02, Other gynecologicals; G03, Sex hormones and modulators of the genital system; G04, Urologicals; H02, Corticosteroids for systemic use; H03, Thyroid therapy; H04, Pancreatic hormones; J01, Antibacterials for systemic use; J02, Antimycotics for systemic use; J05, Antivirals for systemic use; J06, Immune sera and immunoglobulins; L01, Antineoplastic agents; L02, Endocrine therapy; L03, Immunostimulants; L04, Immunosuppressants; M01, Anti-inflammatory and antirheumatic products; M02, Topical products for joint and muscular pain; M03, Muscle relaxants; M04, Antigout preparations; M05, Drugs for treatment of bone diseases; N01, Anesthetics; N02, Analgesics; N03, Antiepileptics; N04, Anti-Parkinson drugs; N05, Psycholeptics; N06, Psychoanaleptics; R01, Nasal preparations; R02, Throat preparations; R03, Drugs for obstructive airway diseases; R05, Cough and cold preparations; R06, Antihistamines for systemic use; S01, Ophthalmologicals; V01, Allergens; V03, All other therapeutic products; V04, Diagnostic agents. Six drugs (0.5%) reported as concomitant were officinal drugs.
investigations based on different study designs in order to validate them and evaluate the hypothetically necessary regulatory actions required (47, 48).

Conclusion
PNs are one of the most frequent ICIs-induced neurological ADRs. We analyzed data from the European pharmacovigilance database on the association between ICIs and PNs. In our study, we found a total of 766 PNs in 735 patients treated with immunotherapy. These included GBS, Miller-Fisher syndrome, carpal tunnel syndrome, neuritis, and demyelinating polyneuropathy. These ADRs were often serious, resulting in disability or hospitalization for the patient. Given the negative outcomes associated with the reported PNs, preventing their occurrence and obtaining earlier treatment would certainly improve the quality of life of patients with a better chance of a complete recovery and also reduce the associated healthcare costs. Moreover, our disproportionality analysis revealed an increased frequency of reported PNs after treatment with atezolizumab compared to other ICIs. This result requires further investigation to better characterize this potential risk. In this context, continuous monitoring of the safety profile of ICIs in real-life settings and A B C FIGURE 3 Reporting odds ratio (ROR) for disproportionality analysis of the most frequently reported peripheral neuropathies as suspected ADRs associated with ICIs. ROR was performed for the following p-term "neuropathy peripheral" (A), "Guillain-Barre syndrome" (B) and "demyelinating polyneuropathy" (C).
conducting pharmacovigilance studies serve as essential instruments to identify specific safety signals. Tools to evaluate and monitor the safety of these new therapeutic approaches in clinical use are crucial to improving public health and reducing costs to healthcare systems. Pharmacovigilance remains a key component of effective public health programs aimed at ensuring the safer use of medical interventions for patients. In this context, the description and analysis of irADRs collected in pharmacovigilance datasets can increase the knowledge of possible ICI-related ADRs, making them more easily identifiable and better managed by clinicians and oncologists.
Finally, advances in immunotherapy have greatly and positively changed the way cancer is managed and controlled. However, despite all their benefits, scientific evidence demonstrates that irADRs related to ICIs are a common cause of disability in oncology patients and require continuous monitoring.

Data availability statement
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: www.adrreports.eu.