Abstract
Background/Aim: Durvalumab is a human monoclonal antibody targeting programmed cell death ligand 1. It is classified as an immune checkpoint inhibitor and has shown high efficacy as maintenance therapy after chemoradiation for stage III non-small-cell lung cancer and as the primary treatment for small-cell carcinoma. Interstitial lung disease is the most common adverse event leading to durvalumab discontinuation. Hence, this study was aimed at assessing the incidence and timing of durvalumab-induced lung toxicity by using the Japanese Adverse Drug Event Report (JADER) database. Patients and Methods: Adverse Adverse events (AEs) of durvalumab reported from August 2018 to March 2021 were extracted. Data on lung AEs were analysed to estimate relative risk using reporting odds ratios (RORs) and 95% confidence interval (CIs). Furthermore, the times of onset of signs of lung toxicity were also estimated. Results: Overall, 2,162 AEs attributable to durvalumab were obtained. Of these, 1,239 were lung toxicities, the most common among which were pneumonia, interstitial lung disease, and radiation-associated pneumonitis. The corresponding RORs (95% CIs) for these signs were 271.50 (244.79-301.11), 5.96 (5.29-6.72), and 713.21 (595.04-854.85), respectively. The median (interquartile range) times of onset were 32.5 (28.5-35.5), 31.5 (28.5-41.5), and 28.5 (28.5-30.5) days, respectively. Conclusion: Among the AEs of durvalumab, pneumonia, interstitial lung disease, and radiation-induced pneumonitis were associated with high RORs, suggesting a strong causal relationship with durvalumab. Interstitial lung disease and radiation-induced pneumonitis most often occurred approximately 30 days after treatment initiation, suggesting that monitoring for adverse events during this period is important.
Durvalumab is a human monoclonal antibody targeting programmed cell death ligand 1 (PD-L1). It enhances antitumour immune responses and inhibits tumour growth by inhibiting binding of PD-L1 to its receptor, PD-1, and is considered an immune checkpoint inhibitor (ICI). Durvalumab is used as maintenance therapy after chemoradiation in patients with unresectable locally advanced non-small-cell lung cancer (NSCLC). The PACIFIC trial compared durvalumab to placebo as intensified therapy in patients with stage III NSCLC treated with platinum-based chemoradiation. The median progression-free survival periods associated with durvalumab and placebo were 16.8 and 5.6 months, respectively (hazard ratio=0.52, 95% confidence interval (CI)=0.42-0.65; p<0.001) (1). Another indication for durvalumab is extensive-stage small-cell lung cancer (ES-SCLC). The CASPIAN trial evaluated durvalumab in combination with etoposide and platinum-based anticancer agents in patients with ES-SCLC who had not received prior chemotherapy (2). The median duration of progression-free survival associated with durvalumab and placebo in these patients was 13.0 and 10.3 months, respectively (hazard ratio=0.73, 95% CI=0.59-0.91; p<0.001).
Although durvalumab contributes to improving the curative effect of conventional chemoradiotherapy and the response rate for ES-SCLC, adverse events (AEs) including immune-related adverse events (irAEs) should also be noted. In the PACIFIC study, 68.0% of durvalumab-treated patients experienced some type of adverse reaction, with the most common ones being rash (15.4%), hypothyroidism (10.5%), diarrhoea (9.7%), and lung toxicity (9.7%) (1). In particular, lung toxicity was noted in 8.4% of patients as a serious AE and was the most frequent one that led to treatment discontinuation (rate: 9.4%) in the PACIFIC trial. The use of durvalumab led to a 33.9% (+9.1%) increase in lung toxicity compared to that of placebo (24.8%) (1). Interactions among radiation therapy, cytotoxic anticancer agents including platinum, and immune checkpoint inhibitors (ICIs) that result in lung toxicity remain unknown; however, the management of lung toxicity due to durvalumab, especially after chemoradiotherapy, is a major problem. The median time to onset of interstitial lung disease in the PACIFIC and CASPIAN trials was 55 and 141 days, respectively. However, the range was extremely wide (1-406 days and 47365 days, respectively), and the time of onset of symptoms remained unknown. Furthermore, it is unclear whether the characteristics of durvalumab-induced lung toxicity differ from those of lung toxicity caused by other ICIs. A comparison of irAEs due to ICIs in advanced NSCLC by a network meta-analysis showed that, compared to pembrolizumab, durvalumab was associated with the second highest incidence of lung toxicity (3). Furthermore, a comparison of irAEs associated with ICIs by using the Food and Drug Administration Adverse Event Reporting System indicated that lung toxicity due to durvalumab was associated with the highest mortality rate (4). The PACIFIC trial suggests that the incidence of interstitial lung disease may be higher in its Japanese subset than in the foreign population (toxicity of all grades: 74% vs. 23%; toxicity of grade ≥3: 7% vs. 4%). However, these data were based on a small number of Japanese patients. Only 72 (15%) out of the 475 patients in the PACIFIC trial received durvalumab, and only 34 (6%) out of the 537 patients in the CASPIAN trial were Japanese. A more objective and detailed profile of durvalumab-induced lung toxicity in Japanese patients remains to be determined. Thus, alertness to lung toxicity is important for the safe use of durvalumab, and this requires validation of a detailed profile of durvalumab-induced lung toxicity in a larger patient population.
Adverse event (AE) data obtained in clinical trials before drug approval is information obtained from a relatively small patient population with clearly defined backgrounds. In actual clinical practice, however, the use of an approved drug in many patients with a variety of backgrounds reveals previously unknown profiles of AEs. Therefore, pharmacovigilance, which is aimed at monitoring drug safety, is important for the optimal use of all drugs (5-7). Pharmacovigilance is performed using a spontaneous reporting system that reflects actual clinical practice (8). In Japan, the Pharmaceuticals and Medical Devices Agency (PMDA) has established the ‘Japanese Adverse Drug Event Report (JADER)’ database as a spontaneous reporting system. The AE profiles revealed by analysis of this database can be used to provide immediate information to healthcare professionals and patients, thereby reducing the risk of ADRs. This study was aimed at retrospectively analysing the JADER database to determine the type and time profile of intercurrent toxicities in durvalumab-treated patients.
Patients and Methods
Data source. Data from the public releases of the JADER database, which contains spontaneous AE reports submitted to the PMDA, were used (9-12) to investigate the association between durvalumab and lung toxicity. The JADER database is available for free download from the PMDA website (http://www.pmda.go.jp). The data structure of the JADER database consists of four datasets: Patient demographic information (DEMO), drug information (DRUG), AEs (REAC), and medical history (HIST). The AEs in the JADER database are coded according to the terminology preferred by the Medical Dictionary for Regulatory Activities/Japanese version 24.1 (www.pmrj.jp/jmo/php/indexj.php).
AEs recorded between August 2018 and March 2021 were analysed in this study. Duplicate cases were removed from the DRUG and REAC tables, as described by our previous report using the same methodology (13). Then the identification number of each AE case was used to merge corresponding case data from the DRUG, REAC, and DEMO tables. Medications that contributed to the AEs were classified as ‘suspected medicine’, ‘concomitant medicine’, and ‘interaction’. Only cases that were classified as ‘suspected medicines’ were extracted.
Statistical analyses. Data on lung toxicity with more than five reported cases were extracted, and the relative risk of AEs was estimated using the reporting odds ratio (ROR). The ROR is frequently used in spontaneous reporting databases as an indicator of the relative risk of AEs. An analysis data table was used, and 2×2 tables were constructed based on two classifications: The presence or absence of lung toxicity, and the presence or absence of suspected durvalumab use. The ROR was calculated by dividing the reported rate of AEs attributable to durvalumab by the reported rate of the same AEs attributable to all other drugs reported in the database. The AE signals were considered positive when the lower limits of the 95% CIs of the RORs were greater than 1 (14).
The time to AE onset was calculated and the number of cases was counted for reports in which the dates of AE onset, start of treatment administration, and end of treatment administration were described in the year/month/day or year/month format (13). The time to onset was calculated as ‘(onset date of AE)–(treatment administration start date)+0.5 in principle (15). When there was a period of non-administration of treatment for more than 1 year, the first treatment administration date from the most recent continuous administration period was used. The time to AE onset 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; it is the quantile in which 63.2% of AEs occur (16). A large value of α indicates a wide distribution, while a small value indicates a narrow distribution. The shape parameter β represents the change in hazard over time in the absence of a reference population. Depending on the value of the shape parameter β, the hazard is interpreted in the standard way as follows. That is, β <1 indicates that the hazard is initially high and then decreases, β1 indicates that the hazard is constant throughout the exposure period, and β >1 indicates that the hazard increases over time (16).
Results
Incidence of lung toxicity associated with durvalumab. We combined three tables, DRUG (3,875,874 reports), REAC (1,096,193 reports), and DEMO (693,295 patients) by using the ID number and removed duplicate data from the DRUG and REAC tables (5). AE causes were categorised as follows: ‘suspected drugs’, ‘concomitant drugs’, and ‘interaction’. All data for the category ‘suspected drugs’ were extracted and used as the ‘data table’ (1,772,494 reports).
We analysed this data table and obtained 2,162 reports of AEs caused by durvalumab. Of these, 1,239 lung toxicities were reported to be associated with durvalumab (Figure 1). All patients had unresectable locally advanced NSCLC or advanced SCLC. Patient characteristics are shown in Table I. Approximately 72.6% of the patients were male. According to the age distribution of the study population, lung toxicity was more common in patients in their 70s (41.5%), followed by those in their 60s (27.4%).
The process for constructing the data analysis table. AEs: Adverse events; DRUG: drug information; REAC: adverse events information; DEMO: patient demographic information.
Characteristics of the patients exhibiting lung toxicity due to durvalumab (1,239).
Pulmonary toxicity due to durvalumab was reported as pneumonitis, interstitial lung disease, radiation-induced pneumonitis, lung injury, respiratory failure, pneumonia, bacterial pneumonia, pneumothorax, pleural effusion, metastasis to the lung, and Pneumocystis jirovecii-associated pneumonia. Pneumocystis jirovecii-associated pneumonia, interstitial pneumonia, and radiation-induced pneumonia accounted for 26.64%, 14.66% and 10.27% of the total number of reported cases. Significant signals with 95% CI of ROR >1 were found for all AEs except pneumonia, pleural effusion and Pneumocystis jiroveci-associated pneumonia. Of these, the RORs for radiation-induced pneumonitis and pneumonitis were particularly large at 713.21, (95% CI=595.04-854.85, p<0.001) and 271.50 (95% CI=244.79-301.11, p<0.001) respectively (Table II).
Numbers of reports and reporting odds ratios (ROR) for lung toxicity due to durvalumab. More than five reports were used for each type of pulmonary toxicity. All analysed data were obtained from the Japanese Adverse Drug Event Report database.
Time to onset of lung toxicity due to durvalumab. A histogram of the time to onset of the nine detected lung toxicity signals showed that they occurred from a median of 28.5 to 60.5 days after durvalumab administration (Figure 2).
A histogram of the time of onset of lung toxicities including pneumonitis, interstitial lung disease, radiation pneumonitis, lung disorder, respiratory failure, bacterial pneumonia, and pneumothorax in patients treated with durvalumab.
The median times to onset (interquartile range=25-75%) of pneumonitis, interstitial lung disease, radiation pneumonitis, lung disorder, respiratory failure, bacterial pneumonia, and pneumothorax caused by durvalumab were 32.5 (28.5-35.5), 31.5 (28.5-41.5), 28.5 (28.5-30.5), 42.5 (24.5-63.5), 54.0 (26.5-87.5), 65.0 (22.5-215.5), and 60.5 (29.5-212.5) days, respectively, after durvalumab administration. The Weibull distribution the histogram of the time to onset showed that the values of the shape parameter β for pneumonitis, interstitial lung disease and lung disorder was <1, whereas those for radiation-induced pneumonitis, respiratory failure, bacterial pneumonia, and pneumothorax were >1 (Table III).
Medians times to onset of adverse events and Weibull parameters of lung toxicity. The detected lung toxicity signals were analysed to determine the time to onset.
Outcome after the occurrence of AEs. Outcomes (recovery, remission, not recovered, with sequelae, death, unclear) after the onset of eight AEs are shown in Figure 3. Of the nine items for which signals were detected, fatal outcomes were recorded for seven AEs.
The percentage of outcomes of four adverse events associated with durvalumab therapy.
Discussion
In this study, we evaluated assessment of the incidence and timing of durvalumab lung toxicity by using the JADER database. Our results include 317 cases of interstitial lung disease and 222 cases of radiation-induced pneumonitis, which are larger numbers than those reported in two previous clinical trials (1, 2). Regarding the types of pulmonary toxicity, besides interstitial lung disease and radiation-induced pneumonitis, diverse signals were detected for lung-related AEs associated with durvalumab, including lung inflammation, lung damage, respiratory failure, bacterial pneumonia, and pneumothorax. In addition, in this study, we provided details on the timing of the common onset of lung toxicity. The median times of onset of interstitial lung disease and radiation pneumonitis were 31.5 and 28.5 days, respectively. These times were shorter than that for durvalumab-related lung toxicities in the PACIFIC trial, which was 55 days, and similar to that for death due to worsening pulmonary toxicity in five cases in the same trial (3-43 days). These findings suggest that durvalumab-related lung toxicity may require early attention. A drug with a shape parameter β of <1 estimated from the Weibull distribution indicates that side-effects are more likely to occur in the early stages of treatment. On the other hand, a value of β >1 indicates that side-effects occur as the duration of treatment administration increases. In the present study, the β value for interstitial lung disease due to durvalumab was 0.89 (95% CI=0.81-0.98), indicating that symptoms occurred early in the course of treatment. This was reflected in a median time to onset of interstitial lung disease of 31.5 days. On the other hand, the β value for radiation-induced pneumonitis was 1.17 (95% CI=1.06-1.29), indicating that the AE occurred over the duration of treatment. These findings provide important information for the safe use of durvalumab. In addition to interstitial lung disease and radiation pneumonitis, the common time to onset of lung-related AEs such as lung inflammation, lung injury, respiratory failure, bacterial pneumonia, and pneumothorax is valuable information for monitoring the side-effects of durvalumab. In particular, the risk factors for lung toxicities include pre-existing lung lesions (especially interstitial pneumonia), history of lung irradiation, respiratory infections, smoking history, respiratory compromise, receipt of oxygen, and elderly age (17); therefore, early monitoring of patients with these factors may be important.
The information on treatment outcomes obtained here is also valuable. The recovery and remission of interstitial lung disease and radiation pneumonitis were noted in more than half of the cases. The most common interstitial lung disease associated with ICIs is organising pneumonia with lymphocytic infiltration (18). In general, organising pneumonia is more responsive to steroid therapy and has a good prognosis, which may be reflected in the results of the present study. Cases of a sustained therapeutic response after interstitial pneumonitis due to ICIs have also been reported (19). ICIs are considered highly effective in patients with irAEs (20, 21). However, caution should be exercised when re-administering these drugs after the development of interstitial lung disease because as many as 30% of patients experience symptom relapse (22). Lung toxicity of grade 3 or more requiring limitation of activities of daily living, oxygenation, and emergency treatment should result in durvalumab discontinuation even if symptoms improve to the baseline status with the use of steroids or immunosuppressive agents.
There are some limitations of this study. Firstly, because the JADER database is based on self-reports, unlike in clinical trials or observational studies, it was not possible to track all patients who received treatment in this analysis. Furthermore, only positive results were likely over-reported. Secondly, the calculated data for the date of onset do not reflect all reported data because data that did not include the treatment initiation date were excluded. Thirdly, the risk factors affecting lung toxicity mentioned earlier were not evaluated in the JADER database. Therefore, it is unclear whether the patient cohort evaluated in this study differs from the general patient population. Despite these limitations, the results of this study provide important information on the incidence of interstitial lung disease associated with durvalumab use in Japan.
Conclusion
Data from a comprehensive survey using pharmacovigilance methods are considered more reflective of actual clinical practice compared with those for a limited number of patients for a limited period, as in the case of clinical trials. The data presented in this study are important for the accumulation and analysis of AEs of durvalumab since the drug’s introduction in Japan in 2018. This study reveals that lung toxicity due to durvalumab is most likely to occur around 30 days after treatment administration, suggesting that monitoring for adverse events during this period is important. Our findings should make physicians and pharmacists more aware of the details of pulmonary AEs during durvalumab treatment. These detailed data will also serve as important information to alert patients using durvalumab.
Acknowledgements
The Authors are grateful to Professor Yoshihiro Uesawa at the Department of Medical Molecular Informatics, Meiji Pharmaceutical University, and to Tadashi Hirooka (TAIHO PHARMA Corporation) for his lecture on Hirooka methods using the JADER database.
Footnotes
Authors’ Contributions
Junya Sato and Tadashi Shimizu: Data curation, writing - original draft, review, and editing. Kana Nakano: Data curation, writing - review, and editing. Mayako Uchida: Conceptualization, data curation, supervision, writing - review, and editing.
Conflicts of Interest
All Authors declare no conflicts of interest.
- Received May 15, 2022.
- Revision received May 29, 2022.
- Accepted June 1, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
