Evaluation of Cardiac Adverse Events with Nivolumab Using a Japanese Real-World Database

Nivolumab has been used for the treatment of various types of cancers and has achieved improvements in overall survival. However, nivolumab can cause a variety of adverse events (AEs). Among these, cardiac-specific AEs have received little attention in clinical trials, despite their life-threatening potential. The present study aimed to determine the risk of nivolumab-induced cardiac AEs, time to onset, incidence rates, and post hoc outcomes using the Japanese Adverse Drug Event Report database. We analyzed data for the period between April 2004 and March 2021. Data on cardiac AEs were extracted and relative risk of AEs was estimated using the reporting odds ratio (ROR). We analyzed 1,772,494 reports and identified 18,721 reports of AEs caused by nivolumab. Of these, 409 reports involved cardiac AEs. Signals were detected for four cardiac AEs: myocarditis; pericardial effusion; pericarditis; and immune-mediated myocarditis. Among these, myocarditis was the most frequently reported (35.0%) and included fatal cases. A histogram of times to onset showed nivolumab-associated AEs occurring 41–127 days after starting administration, with outlier cases of myocarditis or pericardial effusion occurring after more than one year, both with catastrophic consequences. This study focused on cardiac AEs caused by nivolumab as post-marketing AEs. Myocarditis and pericardial effusion have been associated with some fatal cases after administration of nivolumab. Patients should be monitored for signs of onset for these AEs, not only at the start of administration, but also over an extended period after nivolumab administration.


Introduction
Nivolumab is a programmed cell death protein-1 (PD-1)-targeted monoclonal antibody, representing a breakthrough immunotherapy for some cancers. This agent has been used for the treatment of various types of cancers and has achieved improvements in overall survival [1][2][3][4][5][6]. In Japan, nivolumab has been approved for the treatment of advanced solid tumors, including malignant melanoma, non-small cell lung cancer, renal cell carcinoma, Hodgkin lymphoma, head and neck cancer, gastric cancer, malignant pleural mesothelioma, colorectal cancer, and esophageal cancer. Nivolumab has been used as a monotherapy or in combination with chemotherapies and has been reported to prolong both overall and progression-free survival in clinical trials [2][3][4][5][6].
However, nivolumab can cause a variety of adverse events (AEs) [7][8][9][10]. Among these, cardiac immunerelated AEs (irAE) are rare, but also associated with high mortality rates [11][12][13][14]. Myocarditis reportedly showed a low incidence among AEs of immune checkpoint inhibitors (ICIs) in clinical trials but is increasingly being cited in the literature as their use increases [13,14]. On the other hand, many cases are considered to go undiagnosed, so the actual number is expected to be much higher. In the future, cancer immunotherapy is expected to develop further, and the occurrence of undiagnosed cardiac complications is thus a concern. In contrast, cardiac-specific AEs attributable to nivolumab have received little attention in clinical trials, despite their life-threatening potential. Furthermore, even though nivolumab has been widely used in patients since its launch, detailed information on cardiacspecific AEs from post-marketing monitoring has not been reported. Inadequate management of AEs may force discontinuation of nivolumab treatment until the events can be controlled, which may in turn incur disadvantages to the patient, such as decreased therapeutic efficacy. In other words, survival outcomes for patients are also affected.
There are already studies using real-world databases to evaluate AEs after nivolumab administration [11,15]. Ji et al. clarified disproportionality signals for less common AEs in other organ systems, including cardiac systems, etc., that were also detected, and concluded that the results of their studies agreed with clinical observations, suggesting the usefulness of pharmacovigilance in "real-world" safety monitoring [15]. Mascolo et al, in their analysis of spontaneous reports in Eudravigilance, also reported that ICI-induced cardiac AEs were serious and had unfavorable outcomes, and nivolumab was the only ICI with a small increased reporting frequency of individual case safety reports with cardiac AEs compared to all other ICIs [11]. However, information on the cardiovascular AEs with nivolumab is still lacking, especially in Japanese subjects. We therefore conducted this study to determine the risk of nivolumab-induced cardiac AEs, time to onset, incidence rates, and post hoc outcomes using the Japanese Adverse Drug Event Report (JADER) database of the Pharmaceuticals and Medical Devices Agency (PMDA). Identification of safety signals for anticancer drugs using spontaneous adverse event reporting databases may provide a valid way to hypothesize about possible drug-drug relationships with unknown or potential AEs.

Source of Data
We used data from JADER public releases [16][17][18][19]. This database, which includes AE cases, is available for free download from the PMDA website (https:// www. pmda. go. jp/ nglish/ index. html, https:// www. info. pmda. go. jp/ fukus ayoudb/ CsvDo wnload. jsp). We looked at AE reports for the period between April 2004 and March 2021. The JADER data structure comprises four different datasets: patient demographics (DEMO); drug information (DRUG); AEs (REAC); and medical history (HIST). Adverse events were coded using the Medical Dictionary for Regulatory Activities/Japanese version 24.1 (http:// www. pmrj. jp/ jmo/ php/ indexj. php), and Preferred Terms (PTs) were used to indicate terms in the REAC table. All PTs selected in this study are AEs already listed in the technical data sheets of the nivolumab (in "the Drug package insert" and "the Interview Form") in Japan or based on previous studies [12,14,[20][21][22].
We first deleted duplicate cases from the DRUG and REAC columns referenced by previous studies [23]. The identifying number for each AE case was then utilized to integrate case data from the database. The terms "suspected medicine," "concomitant medicine," and "interaction" were used to categorize drug contributions to AEs. Only cases classed as "suspected pharmaceuticals" were extracted.
We reviewed the JADER, which contains reports of spontaneous AEs reported to the PMDA, to determine if a link existed between nivolumab and the occurrence of cardiac AEs.

Statistical Analyses
For the disproportionality analysis of cardiac AEs, the reported odds ratio (ROR) was calculated for data on cardiac AEs with ≥ 10 reported cases. In the spontaneous reporting database, ROR is widely utilized as a measure of the relative risk of AEs.
We created 2 × 2 tables based on two classifications using the analytic data table: the presence or lack of "cardiac AEs"; and the presence or absence of suspected nivolumab use. The ROR was determined by dividing the reported rate of nivolumab-related AEs by the rate of the same AEs reported for all other medicines in the database. The signal of AEs was considered positive if the lower limit of the 95% confidence interval (95% CI) of the ROR was > 1 [24].
The time to onset of AEs was calculated and the number of cases was counted for reports in which the date of onset of AEs, date of start of administration, and date of end of administration were described in year/month/day or year/ month format [23]. Onset time was calculated as "(onset date of AE) − (administration start date) + 0.5" in principle [25]. In the event of a period of non-administration for more than one year, the date of first administration for the most recent continuous administration period was used. The time to onset of AEs for analysis was limited to 2 years (730 days). The Weibull distribution is represented by a scale parameter α and a shape parameter β. The scale parameter α represents the scale of the distribution function, as the quantile in which 63.2% of AEs occur [26]. A larger α value indicates a wider distribution, while a smaller value indicates a narrower distribution. The shape parameter β represents the change in hazard over time in the absence of a reference population. Depending on the value of shape parameter β, an upper limit of the 95% CI for β < 1 indicates that the hazard increases initially, then decreases (early failure type), a 95% CI including 1 indicates that the hazard remains constant throughout the exposure period (random failure type), and a lower limit of the 95% CI > 1 indicates that the hazard increases over time (wear-out failure type). Clinical outcomes were divided in three groups: "recovered/remission", "not recovered/with sequelae" and "death". Additionally, the time-to-onset profiles for AEs among three groups were compared using the Kaplan-Meier method with the log-rank test. The difference was considered significant at P < 0.05. All statistical analyses were performed using JMP Pro ® version 16.2 (SAS Institute, Cary, NC, USA).

Cardiac AEs Associated with Nivolumab
We used ID numbers to link three tables: DRUG (3,875,874 reports); REAC (1,096,193 reports); and DEMO (693,295 patients). Duplicate data were deleted from the DRUG and REAC databases [17]. Adverse events can have one of three causes: "suspected drugs", "concomitant drugs", and "interaction" are all terms used to describe medications that are taken together. All data in the category of "suspected drugs" were retrieved and used as the basis for the study "table of data" (1,772,494 reports).
We reviewed this data table and found 18,721 reports of nivolumab-related AEs, including 409 cardiac AEs (Fig. 1). Table 1 lists the patient characteristics. Male patients made up about 76.8% of the total. Cardiac AEs were common for individuals aged in their 70s, according to the age distribution of the study sample (37.4%).
Among the types of cardiac AEs caused by nivolumab, reported numbers of cases of myocarditis, cardiac failure, pericardial effusion, atrial fibrillation, blood pressure decreased, acute myocardial infarction, pericarditis,

Time to Onset of Cardiac AEs with Nivolumab
Histograms of the times to onset for the four detected cardiac AE signals showed that AEs occurred in the median range from 41 to 127 days after nivolumab administration (Fig. 2). The median time to onset was 41 days (interquartile range [IQR] 23-73 days) for myocarditis, 64 days (IQR 24-148 days) for pericardial effusion, 127 days (IQR 86-364 days) for pericarditis, and 121 days (IQR 36-206 days) for immune-mediated myocarditis. The Table 2 Numbers of reports and RORs of the cardiac adverse events (AEs) related to nivolumab "Cases" indicates the number of reported cases of cardiac AEs. Italicized P-values represent statistically significant results. We used more than 10 reports for each type of cardiac AEs. All analyzed data were obtained from the Japanese Adverse Drug Event Report database. The hypothesis tests were two-sided, and statistical significance was set at P < 0.05. P-values were calculated using Fisher's exact test ROR reporting odds ratio, 95% CI 95% confidence interval  (Table 4). Myocarditis and pericardial effusion developed earliest in the "death" group among three groups ("recovered/remission", "not recovered/with sequelae" and "death" group), and the log-rank test demonstrated a significant difference in their time transitions (p < 0.001, p = 0.027, respectively) ( Table 4).

Outcomes After Occurrence of AEs
After commencement of the four AEs, percentages of outcomes (recovery, remission, not recovered, with sequelae, death, and unclear) are presented in Fig. 3. Myocarditis and pericardial effusion both included cases with catastrophic consequences.

Discussion
This study focused on cardiac AEs caused by nivolumab, and those AEs for which signals were detected were myocarditis, pericardial effusion, pericarditis, and immune-mediated myocarditis. Among these, myocarditis was the most frequently reported (35.0%, 143/409: including immunemediated myocarditis, 37.7%, 154/409) and included fatal cases. In our study, the median time to event was 41 days (IQR 23-73 days), suggesting early occurrence. Myocarditis caused by nivolumab has been reported in several clinical trials [13,14,27] and the results of the present study are consistent with the clinical findings. In previous studies, most cases of and fatalities from myocarditis also occurred shortly after initiating ICI therapy [27]. In their systematic review, Atallah-Yunes et al. clarified that most cases and fatalities of myocarditis occurred shortly after initiation of ICI therapy [13]. Ganatra et al. also reported ICI-associated myocarditis typically occurred early, and the number of cases was likely underestimated [28]. Clinicians should thus be alert for the development of myocarditis immediately after nivolumab administration. In addition, continuous monitoring is recommended throughout the treatment period, as myocarditis develops at a constant rate, with some cases occurring approximately 17 months after administration. Japanese Adverse Drug Event Report is useful for the time-to-onset analysis [29][30][31]. Therefore, this study provides useful information for clinicians.
Of the four AEs for which signals were detected, fatal outcomes were also observed for pericardial effusion. The Weibull distribution showed that incidents of pericardial effusion developed at a constant rate. A previous study reported that ICI-induced pericardial effusion carries a substantial mortality rate and prompt diagnosis, and management are critical [32]. Canale et al. reported pericardial effusion as a relatively frequent AE of immunotherapy in non-small cell lung cancer [33]. Saade et al. also reported that cardiotoxic AEs, initially thought to be rare, are more often encountered paralleling the expanding use of ICIs. Among the cardiotoxic AEs, pericardial effusion deserves attention as it may present with unusual symptomatology [34]. Vittorio et al. proposed that nivolumab might have been responsible for a sudden worsening of symptoms in the setting of prolonged remission [35]. In this study, as for pericarditis, although no fatal cases were extracted, there were reports of cases with unrecovered or sequelae. Salem et al. revealed in their observational, retrospective, pharmacovigilance study that pericardial diseases (pericarditis, pericardial effusion, and tamponade) are associated with ICIs [36]. They also observed over-representation of pericardial diseases, especially in patients receiving anti-PD-1 or anti-PDL-1 therapy for lung cancer with one plausible hypothesis that radiotherapy and immunotherapies might synergize in patients with lung cancer [36]. Clinicians must therefore remain vigilant for these rare AEs regardless of the duration of drug exposure or time since cessation of therapy. Similarly, continuous monitoring throughout the entire administration period is recommended. Although anti-PD-1 treatment is usually regarded as less toxic than chemotherapy [37,38], a wide spectrum of life-threatening immune-related AEs may still occur, and clinical vigilance is required.
The present results must be considered in light of some limitations. First, the JADER database is based on selfreported data, which may contribute to reporting biases such as over-and underreporting. That is, serious AEs may have been reported preferentially. Second, the lack of comprehensive medical records and medication histories limits the scope of analysis, as dosages, durations, clinical laboratory data, severity of AEs, severity of disease and more information on concomitant medications of nivolumab use were unavailable. As for concomitant medications, we could not rule out the possibility that AEs may have been caused by concomitant use of anticancer drugs or other ICIs Fig. 3 Percentages of the four adverse events (AEs) associated with nivolumab by outcome (particularly in combination with ipilimumab). Third, most cases were men (76.8%) aged in their 60s (35.0%) and 70s (37.4%). Accordingly, male sex and higher age cannot be ruled out as causal factors for cardiac AEs. Fourth, potential confounding, selection, and information biases cannot be fully excluded from this study. However, the results of this study were based on extracted data in which nivolumab was judged to be the suspect drug by the reporter (physician or pharmacist) who knew the details of the clinical course in each specific case. Thus, our report provides useful information for monitoring cardiac AEs caused by nivolumab.

Conclusion
We focused on cardiac AEs caused by nivolumab as postmarketing AEs. Myocarditis and pericardial effusion have been associated with some fatal cases after administration of nivolumab. Patients should be monitored for signs of the onset of these AEs not only at the start of administration, but also for an extended period thereafter.