Abstract
Background
Interstitial pneumonitis (IP) is a severe adverse event in patients receiving immunotherapy. Although PD-1/PD-L1 inhibitors and bevacizumab have been widely used in patients with non-small cell lung cancer (NSCLC), the interaction between their combination and IP is less known.
Methods
To investigate the interaction between bevacizumab and PD-1/PD-L1 inhibitors on IP, an observational study between January 2012 and June 2023, US, from the Food and Drug Administration Adverse Event Reporting System (FAERS) database including 55,673 NSCLC patients was performed. The reported proportions of PD-1/PD-L1 inhibitor-associated IP in patients receiving and not receiving bevacizumab treatment were compared.
Results
A total of 23,790 and 4753 patients were treated with PD-1/PD-L1 inhibitors and bevacizumab, among whom 1693 were treated with both PD-1/PD-L1 inhibitors and bevacizumab. The proportions of IP were 7.3% (95% CI 7.1 to 7.5%) in the total population of patients with NSCLC, 11.1% in patients treated with PD-1/PD-L1 inhibitors, and 4.4% in patients treated with bevacizumab. The reported IP proportion was 15.4% for PD-1 inhibitors with bevacizumab, which was greater than the 9.1% for PD-1 inhibitors without bevacizumab, while the opposite trend was observed for PD-L1 inhibitors (4.4% for bevacizumab vs 19.6% for PD-L1 inhibitors without bevacizumab).
Conclusions
Our study showed that bevacizumab was associated with a higher reported proportion of PD-1 inhibitor-associated IP but a lower reported proportion of PD-L1 inhibitor-associated IP. Although this was a post-marketing, observational study with many limitations, bevacizumab might be a proper combination with PD-L1 inhibitors in consideration of IP, especially in patients at high risk of IP. However, this relationship still needs further clinical validation.
Supplementary Information
The online version contains supplementary material available at 10.1007/s00262-025-04158-1.
Keywords: Interstitial pneumonitis, Checkpoint inhibitors, Non-small cell lung cancer, Bevacizumab, Programmed death 1, Programmed cell death 1 ligand 1
Introduction
Immunotherapy, especially PD-1/PD-L1 inhibitors, has been widely used in patients with non-small cell lung cancer (NSCLC). Single immunotherapy is mostly applied to patients with high PD-L1 expression. Immunotherapy combined with chemotherapy and antiangiogenic therapy is the standard of care for NSCLC patients regardless of PD-L1 expression. In the past 20 years, several trials have demonstrated the survival benefit of the addition of the antiangiogenic drug bevacizumab to chemotherapy [1, 2]. In the era of immunotherapy, whether bevacizumab can enhance the efficacy of PD-1/PD-L1 inhibitors is a research hotspot. IMPower150 showed that adding bevacizumab and the PD-L1 inhibitor atezolizumab to chemotherapy improved progression-free survival and overall survival in patients with nonsquamous NSCLC [3]. Another phase 2 study by Provencio et al. [4] showed that atezolizumab plus bevacizumab had a progression-free survival of 13 months in patients with high-tumor-burden nonsquamous NSCLC. ORIENT-31 showed that the bevacizumab biosimilar IBI305 plus the PD-1 inhibitor sintilimab plus chemotherapy could clinically improve progression-free survival in patients with EGFR-mutated NSCLC who progressed after treatment with tyrosine kinase inhibitors.
Interstitial pneumonitis (IP) is a severe adverse event in patients receiving immunotherapy. Once IP occurs, patients suffer treatment delays and life threats. The rates of IP in the KEYNOTE-189 and KEYNOTE-407 groups were 4.4% and 8.3%, respectively, which were significantly greater than those in the control group without immunotherapy [5, 6]. In a real-world study, the incidence of IP was 22% in the first year after treatment with PD-1/PD-L1 inhibitors [7].
The efficacy of bevacizumab plus immunotherapy has been validated by several clinical studies, but the impact of adding bevacizumab to PD-1/PD-L1 inhibitors on IP remains unknown. The TASUKI-52 study showed that adding the PD-1 inhibitor nivolumab to bevacizumab-containing chemotherapy significantly increased the incidence of IP from 1.1 to 7.3% [8]. Therefore, a systematic investigation of the impact of bevacizumab and PD-1/PD-L1 inhibitors on IP is needed. The Food and Drug Administration Adverse Event Reporting System (FAERS) is a publicly available database on adverse events associated with approved drugs. In this study, we compared IP in patients treated with or without bevacizumab and PD-1/PD-L1 inhibitors to investigate this interaction.
Methods
The data from 2012Q4 to 2023Q2 were downloaded and used for further analysis. Patients with NSCLC were selected according to the ‘indi_pt’ column. PD-1/PD-L1 inhibitors were filtered by the following terms: ‘pembrolizumab,’ ‘KEYTRUDA,’ ‘nivolumab,’ ‘OPDIVO,’ ‘LIBTAYO,’ ‘CEMIPLIMAB,’ ‘DOSTARLIMAB,’ ‘TECENTRIQ,’ ‘ATEZOLIZUMAB,’ ‘IMFINZI,’ ‘DURVALUMAB,’ ‘BAVENCIO,’ and ‘AVELUMAB.’ IP defines as ‘idiopathic pulmonary fibrosis,’ ‘interstitial lung disease,’ ‘pneumonitis,’ ‘pneumonitis chemical,’ ‘pulmonary fibrosis,’ and ‘radiation pneumonitis.’ Crude and adjusted odds ratios were calculated by logistic regression. The covariables for calculating adjusted odds ratios were age and sex. Data processing and statistical analyses were performed using R software (version 4.2.2 R Foundation).
Results
A total of 55,673 NSCLC patients were identified from the FAERS database between 2012Q4 and 2023Q2. A total of 23,790 and 4753 patients were treated with PD-1/PD-L1 inhibitors and bevacizumab, respectively (details shown in Table 1). Of the 4753 patients treated with bevacizumab, 1693 were also treated with PD-1/PD-L1 inhibitors. Treatment with bevacizumab was most common in Japan (43.4%), England (11%), the USA (4.7%), and Germany (4.4%).
Table 1.
Characteristics of the 55,673 patients
| Characteristic | With PD-1/PD-L1 inhibitors | Without PD-1/PD-L1 inhibitors | ||
|---|---|---|---|---|
| With bevacizumab (n = 1693) |
Without bevacizumab (n = 22,097) |
With bevacizumab (n = 3060) |
Without bevacizumab (n = 28,823) |
|
| Age, mean (SD), y | 53.12 (24.54) | 63.57 (16.63) | 61.54 (15.03) | 63.94 (15.07) |
| Sex, No. (%) | ||||
| Female | 608 (35.9) | 6580 (29.8) | 725 (23.7) | 9784 (33.9) |
| Male | 831 (49.1) | 13,462 (60.9) | 951 (31.1) | 11,456 (39.7) |
| Not reported | 254 (15) | 2055 (9.3) | 1384 (45.2) | 7583 (26.3) |
| Country, No. (%) | ||||
| Australia | 31 (1.8) | 377 (1.7) | 7 (0.2) | 293 (1) |
| Canada | 7 (0.4) | 560 (2.5) | 7 (0.2) | 534 (1.9) |
| China | 22 (1.3) | 753 (3.4) | 276 (9) | 1600 (5.6) |
| Germany | 74 (4.4) | 1563 (7.1) | 117 (3.8) | 1447 (5) |
| Spain | 43 (2.5) | 384 (1.7) | 23 (0.8) | 356 (1.2) |
| France | 64 (3.8) | 1050 (4.8) | 96 (3.1) | 1206 (4.2) |
| England | 186 (11) | 358 (1.6) | 3 (0.1) | 596 (2.1) |
| India | 0 (0) | 259 (1.2) | 6 (0.2) | 569 (2) |
| Italy | 9 (0.5) | 615 (2.8) | 66 (2.2) | 634 (2.2) |
| Japan | 734 (43.4) | 7581 (34.3) | 582 (19) | 4307 (14.9) |
| Korea | 11 (0.6) | 256 (1.2) | 5 (0.2) | 572 (2) |
| Netherlands | 36 (2.1) | 217 (1) | 53 (1.7) | 412 (1.4) |
| USA | 80 (4.7) | 4267 (19.3) | 687 (22.5) | 7109 (24.7) |
| Other | 396 (23.4) | 3857 (17.5) | 1132 (37) | 9188 (31.9) |
PD-1, programmed death 1; PD-L1, programmed cell death 1 ligand 1
Among the 55,673 patients, 4076 had reported IP, and the proportion was 7.3% (95% CI 7.1 to 7.5%). We then analyzed the impact of age, sex, and drugs on IP (Table 2). A high proportion of patients treated with PD-1/PD-L1 inhibitors had an IP (11.1%), while a low proportion of patients treated with bevacizumab had an IP (4.4%). Patients treated with both PD-1/PD-L1 inhibitors and bevacizumab had significantly lower IP (5.6%) than those treated with PD-1/PD-L1 inhibitors. To determine the interaction effect between bevacizumab and PD-1/PD-L1 inhibitors, we performed multivariable logistic regression incorporating age, sex, bevacizumab, and PD-1/PD-L1 inhibitors. The adjusted odds ratio with bevacizumab was not statistically significant, while it was 2.75 (95% CI 2.56–2.96) with PD-1/PD-L1 inhibitors. The adjusted odds ratio of the interaction effect was 0.57 (95% CI 0.43–0.76), indicating the protective role of bevacizumab against IP (Table 2).
Table 2.
Proportion of Interstitial Pneumonitis in 55,673 Patients with Non-Small Cell Lung Cancer
| Variable | Cases, No (n = 55,673) |
IP (n = 4076) |
Proportion of IP | Crude odds ratio | Adjusteda odds ratio | P Value (Wald test) |
|---|---|---|---|---|---|---|
| Point estimate (95% CI) | Point estimate (95% CI) | Point estimate (95% CI) | ||||
| Age, mean (SD), y | 63.27 (16.39) | 1.01 (1.01–1.01) | 1.00 (1.00–1.00) | < 0.001 | ||
| Sex | ||||||
| Female | 19,369 | 1048 | 5.4 (5.1–5.7) | 1 (Reference) | 1 (Reference) | |
| Male | 25,028 | 2341 | 9.4 (9.0–9.7) | 1.80 (1.67–1.95) | 1.52 (1.41–1.64) | < 0.001 |
| Not reported | 11,276 | 687 | 6.1 (5.7–6.5) | 1.13 (1.03–1.25) | 1.46 (1.31–1.61) | < 0.001 |
| Bevacizumab | 4753 | 211 | 4.4 (3.9–5.1) | 0.57 (0.49–0.65) | 0.84 (0.69–1.02) | 0.08 |
| PD-1/PD-L1 inhibitors | 23,790 | 2646 | 11.1 (10.7–11.5) | 2.66 (2.49–2.85) | 2.75 (2.56–2.96) | < 0.001 |
| Bevacizumab plus PD-1/PD-L1 inhibitors | 1693 | 95 | 5.6 (4.6–6.8) | 0.57 (0.43–0.76) | < 0.001 |
IP, interstitial pneumonitis; PD-1, programmed death 1; PD-L1, programmed cell death 1 ligand 1. aMultivariable logistic regression incorporating age, sex, bevacizumab, and PD-1/PD-L1 inhibitors
To further investigate the differences in the interactions between PD-1 and PD-L1 inhibitors, we grouped patients who received PD-1 or PD-L1 inhibitors (Table S1). The reported IP proportion was 15.4% for PD-1 inhibitors with bevacizumab, which was greater than the 9.1% for PD-1 inhibitors without bevacizumab, while the opposite trend was observed for PD-L1 inhibitors (4.4% for bevacizumab vs 19.6% for PD-L1 inhibitors without bevacizumab). The crude odds ratios were 1.83 (95% CI 1.14–2.81) and 0.19 (95% CI 0.14–0.24) for PD-1 inhibitors and PD-L1 inhibitors combined with bevacizumab, respectively
Durvalumab has been approved for use in many countries and regions, and maintenance of durvalumab without bevacizumab is the standard of care after chemoradiotherapy. Considering that radiotherapy might impact the proportion of IPs in patients receiving durvalumab, we stratified patients according to treatment with durvalumab or atezolizumab (Table S2). The reported proportions of IPs were 5.9% for durvalumab with bevacizumab and 4.7% for atezolizumab with bevacizumab, both of which were lower than those for patients without bevacizumab (33.0% for durvalumab and 10.7% for atezolizumab). The crude odds ratios were 0.13 (95% CI 0.01–0.62) and 0.48 (95% CI 0.36–0.62) for durvalumab and atezolizumab, respectively.
Discussion
We found the interactions between bevacizumab and PD-1/PD-L1 inhibitors. The impact of bevacizumab on the IP differed depending on the immune checkpoint inhibitor. In brief, bevacizumab increases the IP proportion when combined with PD-1 inhibitors but reduces the IP proportion when combined with PD-L1 inhibitors.
The associations of immune checkpoint inhibitors with IP have not been well elucidated. The destruction of immune homeostasis in pulmonary tissue is thought to be one of the major causes [9]. Aberrantly activated T cells induce autoimmunity and nonspecific inflammation in normal lung tissues. PD-1 inhibitors activate T cells by blocking their cell surface PD-1 [10]. When combined with an anti-VEGF monoclonal antibody, bevacizumab normalizes vessels, leading to increased T cell infiltration. These infiltrated T cells might aggravate IP.
Reports on the relationship between bevacizumab and IP are limited. Lind et al. reported a high rate of IP in patients receiving radiotherapy concurrent with bevacizumab [11]. The authors proposed that bevacizumab might increase radiation-induced cell killing and downregulate DNA repair genes. Another study showed that bevacizumab mitigated IP via PD-1/PD-L1 inhibitors by reducing alveolar exudate leakage [12]. Additionally, Pang et al. conducted a single-arm trial evaluating the efficacy of bevacizumab in patients with severe COVID-19 [13], and the results showed that bevacizumab significantly improved oxygenation and reduced the pulmonary leakage area on chest computed tomography. Bolandi et al. showed that bevacizumab could downregulate VEGF expression to diminish bronchial inflammation [14].
The mechanism by which bevacizumab reduces PD-L1 but increases PD-1 inhibitor-associated pneumonitis is unclear. To further determine the interaction between bevacizumab and PD-1/PD-L1 inhibitors, we performed similar analyses in hepatocellular carcinoma patients from the FAERS database. Both PD-1 and PD-L1 inhibitors were associated with increased IP when combined with bevacizumab (supplement data S1). The difference between hepatocellular carcinoma and non-small cell lung cancer (NSCLC) indicates the complex mechanism of IP. The presence of multiple antigens between NSCLC and normal lung tissues might also explain this difference. Another potential explanation is that different combination therapies (pre- and post-immunotherapy) might impact IP in patients receiving PD-1 and PD-1 inhibitors with or without bevacizumab.
Unlike other anti-VEGFR antibodies (e.g., ramucirumab), bevacizumab uniquely mitigates PD-1 inhibitor-associated risks through a tripartite mechanism: vessel normalization, inflammatory cytokine suppression, and Treg modulation. While alternative anti-VEGFR agents may engage similar pathways, their effects exhibit target-dependent distinctions. Future correlative studies should investigate whether these mechanistic differences translate to clinically significant variations in IP incidence when combined with PD-1 or PD-L1 inhibitors.
The differential efficacy observed between bevacizumab combined with PD-1 versus PD-L1 inhibitors stems from fundamental biological heterogeneity in their immunomodulatory mechanisms and the complex target interactions within the tumor microenvironment (TME). The superior clinical outcomes with bevacizumab plus PD-L1 blockade may be attributed to their complementary actions: bevacizumab reverses VEGF-mediated dendritic cell (DC) dysfunction and Treg infiltration, while PD-L1 inhibitors directly enhance co-stimulatory signaling to synergistically activate T cells. This mechanistic synergy is particularly pronounced in VEGF-high tumors (e.g., EGFR-mutant NSCLC), providing a molecular rationale for precision immunotherapy combinations. Future studies should explore novel biomarkers to further optimize personalized treatment strategies.
The occurrence of IP is a multifactor and complicated process. The baseline pulmonary function, previous treatment, and sequence of PD-1/PD-L1 inhibitors that might influence IP were not reported in the FARES database. Additionally, the database was organized by self-reports from pharmaceutical companies, clinicians, and patients, leading to reporting bias. Severe adverse effects, especially those of high concern, were reported to be preferred. The reported ratio in this study was not the occurrence rate but rather the relative ratio in all reported cases. Our study showed that bevacizumab decreased PD-L1 inhibitor-associated IP, but the causality was not reliable. This relationship still needs further clinical validation.
It is well documented that the incidence of immune-mediated pneumonitis (IP) following PD-1 or PD-L1 inhibitor therapy varies across racial and ethnic groups. However, detailed race/ethnicity data from our patient cohort were not systematically collected or analyzed. We acknowledge this as a study limitation, as our primary focus was evaluating bevacizumab’s impact on IP risk using available data from PD-1/PD-L1 inhibitor cohorts. Future large-scale prospective studies specifically designed to collect comprehensive data on race, ethnicity, genetic ancestry, and detailed environmental/clinical covariates would be valuable to elucidate the relative contributions of these factors to IP development during PD-1/PD-L1 inhibitor therapy.
In conclusion, bevacizumab combined with different checkpoint inhibitors has different IP risks. Bevacizumab increases PD-1 inhibitor-related IP but reduces PD-L1 inhibitor-related IP.
Supplementary Information
Below is the link to the electronic supplementary material.
Author contributions
Lingling Wang, Yasi Xu, and Lucheng Zhu were involved in writing of the original draft and editing; Yanyan Zhao and Bing Xia helped in methodology and data acquisition; Lucheng Zhu and Yasi Xu contributed to data analyses and interpretation; Lucheng Zhu was involved in study concept and design, project administration; manuscript revision was done by Lingling Wang. All the authors have read and approved the final manuscript.
Funding
This study was supported by grants from the Construction Fund of Key Medical Disciplines of Hangzhou (2025HZGF06) and National Natural Science Foundation of China (82373889). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Data availability
The original contributions presented in the study are included in the article material. Further inquiries can be directed to the corresponding authors.
Declarations
Conflict of interest
The authors declare that they have no conflicts of interest.
Ethical approval
FAERS is a publicly anonymized database, and institutional review board approval was not needed.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Linling Wang and Yasi Xu have contributed equally to this work.
References
- 1.Sandler A, Gray R, Perry MC et al (2006) Paclitaxel-carboplatin alone or with bevacizumab for non-small-cell lung cancer. N Engl J Med 355(24):2542–2550. 10.1056/NEJMoa061884 [DOI] [PubMed] [Google Scholar]
- 2.Zhou C, Wu YL, Chen G et al (2015) Beyond: a randomized, double-blind, placebo-controlled, multicenter, phase III study of first-line carboplatin/paclitaxel plus bevacizumab or placebo in Chinese patients with advanced or recurrent nonsquamous non-small-cell lung cancer. J Clin Oncol 33(19):2197–2204. 10.1200/jco.2014.59.4424 [DOI] [PubMed] [Google Scholar]
- 3.Socinski MA, Jotte RM, Cappuzzo F et al (2018) Atezolizumab for first-line treatment of metastatic nonsquamous NSCLC. N Engl J Med 378(24):2288–2301. 10.1056/NEJMoa1716948 [DOI] [PubMed] [Google Scholar]
- 4.Provencio M, Ortega AL, Coves-Sarto J et al (2023) Atezolizumab plus bevacizumab as first-line treatment for patients with metastatic nonsquamous non-small cell lung cancer with high tumor mutation burden: a nonrandomized controlled trial. JAMA Oncol 9(3):344–353. 10.1001/jamaoncol.2022.5959 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gandhi L, Rodríguez-Abreu D, Gadgeel S et al (2018) Pembrolizumab plus chemotherapy in metastatic non-small-cell lung cancer. N Engl J Med 378(22):2078–2092. 10.1056/NEJMoa1801005 [DOI] [PubMed] [Google Scholar]
- 6.Novello S, Kowalski DM, Luft A et al (2023) Pembrolizumab plus chemotherapy in squamous non-small-cell lung cancer: 5-year update of the phase III KEYNOTE-407 study. J Clin Oncol 41(11):1999–2006. 10.1200/jco.22.01990 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tiu BC, Zubiri L, Iheke J et al (2022) Real-world incidence and impact of pneumonitis in patients with lung cancer treated with immune checkpoint inhibitors: a multi-institutional cohort study. J Immunother Cancer. 10.1136/jitc-2022-004670 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sugawara S, Lee JS, Kang JH et al (2021) Nivolumab with carboplatin, paclitaxel, and bevacizumab for first-line treatment of advanced nonsquamous non-small-cell lung cancer. Ann Oncol 32(9):1137–1147. 10.1016/j.annonc.2021.06.004 [DOI] [PubMed] [Google Scholar]
- 9.Zhang Z, Zhou J, Verma V et al (2021) Crossed pathways for radiation-induced and immunotherapy-related lung injury. Front Immunol 12:774807. 10.3389/fimmu.2021.774807 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bagchi S, Yuan R, Engleman EG (2021) Immune checkpoint inhibitors for the treatment of cancer: clinical impact and mechanisms of response and resistance. Annu Rev Pathol 16:223–249. 10.1146/annurev-pathol-042020-042741 [DOI] [PubMed] [Google Scholar]
- 11.Lind JS, Senan S, Smit EF (2012) Pulmonary toxicity after bevacizumab and concurrent thoracic radiotherapy observed in a phase I study for inoperable stage III non-small-cell lung cancer. J Clin Oncol 30(8):e104–e108. 10.1200/jco.2011.38.4552 [DOI] [PubMed] [Google Scholar]
- 12.Iwai T, Sugimoto M, Patel H, Yorozu K, Kurasawa M, Kondoh O (2021) Anti-VEGF antibody protects against alveolar exudate leakage caused by vascular hyperpermeability, resulting in mitigation of pneumonitis induced by immunotherapy. Mol Cancer Ther 20(12):2519–2526. 10.1158/1535-7163.Mct-21-0031 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Pang J, Xu F, Aondio G et al (2021) Efficacy and tolerability of bevacizumab in patients with severe Covid-19. Nat Commun 12(1):814. 10.1038/s41467-021-21085-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bolandi SM, Abdolmaleki Z, Assarehzadegan MA (2021) Anti-angiogenic properties of bevacizumab improve respiratory system inflammation in ovalbumin-induced rat model of asthma. Inflammation 44(6):2463–2475. 10.1007/s10753-021-01516-w [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The original contributions presented in the study are included in the article material. Further inquiries can be directed to the corresponding authors.