Immune-related cardiovascular toxicities of PD-1/PD-L1 inhibitors in solid tumors: an updated systematic review and meta-analysis

Purpose The objective of this study was to investigate the risk of cardiovascular toxicities related to PD-1/PD-L1 inhibitors in solid tumors. Methods A literature search was performed following the participants, interventions, comparisons, outcomes, and study design (PICOS) principles, and the study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Data analysis was conducted using Review Manager version 5.4. Results This meta-analysis included 69 randomized controlled trials (RCTs) divided into five groups based on the treatment regimens: PD-1/PD-L1 + chemotherapy versus chemotherapy, PD-1/PD-L1 versus chemotherapy, PD-1/PD-L1 versus placebo, PD-1/PD-L1 + CTLA-4 versus PD-1/PD-L1 and PD-1/PD-L1 + CTLA-4 versus chemotherapy. Compared to chemotherapy treatment alone, PD-1/PD-L1 +chemotherapy significantly increased the risk of hypertension [all-grade (OR = 1.27, 95% CI [1.05, 1.53], p = 0.01); grade 3–5 (OR = 1.36, 95% CI [1.04, 1.79], p = 0.03)], hypotension [all-grade (OR = 2.03, 95% CI [1.19, 3.45], p = 0.009); grade 3–5 (OR = 3.60, 95% CI [1.22, 10.60], p = 0.02)], arrhythmia [all-grade (OR = 1.53, 95% CI [1.02, 2.30], p = 0.04); grade 3–5 (OR = 2.91, 95% CI [1.33, 6.39], p = 0.008)] and myocarditis [all-grade (OR = 2.42, 95% CI [1.06, 5.54], p = 0.04)]. The risk of all-grade hypotension (OR = 2.87, 95% CI [1.26, 6.55], p = 0.01) and all-grade arrhythmia (OR = 2.03, 95% CI [1.13, 3.64], p = 0.02) significantly increased when treated with PD-1/PD-L1 inhibitors compared to the placebo. The risks of cardiovascular toxicities are significantly higher with PD-1+CTLA-4 compared to PD-1 alone (OR = 2.02, 95% CI [1.12, 3.66], p = 0.02). Conclusion PD-1/PD-L1 inhibitor leads to an increased risk of cardiovascular toxicities, especially hypertension, hypotension, arrhythmia, and myocarditis.


Introduction
In recent years, the programmed cell death 1/programmed cell death 1 ligand 1 (PD-1/PD-L1) inhibitor has been used as an immunotherapy and has led to substantial advancements in the prognosis of diverse cancer types (1).It can enhance the immune response by blocking the inhibitory signal of the T cell response and exerting anti-tumor effects (2).However, the enhanced destructive effect of T cells can also damage normal cells and tissues.Clinicians are becoming aware of its adverse effects on almost all organ types (3).Adverse effects often include immune-related pneumonitis, liver damage, endocrine organ abnormalities, and adverse skin reactions (4).Although cardiovascular toxicities, such as myocarditis, arrhythmia, blood pressure abnormalities, and heart failure, are uncommon, their prognoses are poor (5,6).Therefore, additional attention should be paid to cardiovascular toxicity.
PD-1/PD-L1 inhibitors are currently recommended in various therapeutic combinations.Previous reviews and meta-analyses have summarized cardiovascular toxicities associated with different treatment regimens (7,8).The completion of more clinical trials may have affected the original analysis results.The original topic that could not be analyzed because of insufficient data may have to be reoperated and completed.Therefore, given that cardiovascular toxicities are now considered major determinants of prognosis (9), it is necessary to conduct a new meta-analysis for this study.This will further guide the antitumor treatment of patients with solid tumors.

Search strategy and selection criteria
This study was consistent with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (10).Randomized controlled trials (RCTs) on solid tumors with cardiovascular toxicities published between July 2013 and May 2023 were searched based on the principle of PICOS (participants, interventions, comparisons, outcomes, and study design).The following medical subject heading (MeSH) terms were used: nivolumab, pembrolizumab, atezolizumab, tislelizumab, penpulimab, avelumab, durvalumab, camrelizumab, Opdivo, Bavencio, Keytruda, Imfinzi, AK105, MPDL3280A, Tecentriq, MK-3475, and BMS 963558.RCTs mentioned in the relevant reviews and references were also searched to avoid missing data.Five individuals were selected for literature search and data extraction.All conflicts were jointly discussed and resolved by the corresponding author.
The following selection criteria were used: 1) RCTs published between July 2013 and May 2023; 2) participants diagnosed with solid tumors treated with at least one PD-1 or PD-L1 inhibitor; 3) clinical trials reporting all-grade or grade 3-5 adverse effects; 4) research published in English.The exclusion criteria were as follows: 1) no treatment with PD-1/PD-L1; 2) non-RCT studies; 3) RCTs not involving cardiovascular toxicities; 4) single-arm studies without a control group.

Data extraction
Five individuals independently obtained the following baseline information from the included studies: year of publication, name of the first author, name of the study, national clinical trial (NCT) number, treatment lines, names of tumors, names of drugs, treatment arms, and the total number of people included in each study.

Publication bias and quality assessments
The Cochrane Collaboration tool was used to evaluate the risk of bias in the RCTs and funnel plots were used to evaluate publication bias (11).Seven sources of bias were evaluated in each RCT: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), and selective reporting (reporting bias).Each domain was independently assigned a 'high', 'low', or 'unclear' risk of bias by all authors, with disag reements adjudicated by the corresponding author.

Heterogeneity assessment and statistical analysis
Review Manager (RevMan) version 5.4.was used to analyze the relevant data using the Mantel-Haenszel method (12).I 2 values were applied to estimate heterogeneity among the included clinical trials, which were classified into three grades: low, moderate, and high (I2 values <25%, 25%-50%, and >50%, respectively) (13).When I 2 was greater than 50%, significant heterogeneity was considered, and the source of heterogeneity was determined by subgroup analysis.Owing to the inherent heterogeneity among the included trials, the random effect (RE) was applied to analyze the odds ratio (OR) and corresponding 95% confidence interval (CI) (14).Funnel plots derived from the fixed effect (FE) model were used to evaluate publication bias.All reported P values were twosided, and P < 0.05 was deemed to be statistically significant.

Discussion
This meta-analysis included recently completed RCTs and provided updated information on the cardiotoxicity of PD-1/PD-L1 inhibitors.With a larger sample size and more detailed subgroups, this study provided several novel findings, indicating that the combination of PD-1/PD-L1 inhibitors with chemotherapy carries a considerably higher risk of myocarditis and hypotension than conventional chemotherapy alone.An increasing number of people are now paying attention to the cardiovascular toxicities of PD-1/PD-L1, and this study provides strong supporting evidence for these concerns.Additionally, it assists doctors in making preliminary assessments of the potential causes of these side effects when they detect cardiovascular issues in patients.This, in turn, allows for a more significant TABLE 2 The risk of all-grade myocardial infarction, heart failure, pericardial diseases, embolism, thrombosis and vasculis: subgroup analyses were carried out based on PD-1/PD-L1.improvement in patient prognosis without compromising their antitumor treatment.Additionally, this study supports previous metaanalyses (7,8) and preclinical evidence (9) (92, 93), highlighting the substantial increase in cardiovascular toxicities associated with PD-1/ PD-L1 inhibitors.Flow cytometry and metabolomic assays revealed that PD-1/PD-L1 treatment in mice resulted in an increase in the overall lymphocyte count and changes in lipid metabolism within the cardiac tissue.These findings provide evidence that PD-1/PD-L1 disrupts immune homeostasis and energy production in the heart (9).Furthermore, single-cell sequencing revealed that endothelial cells constituted the majority of cells in the cardiac interstitium.Notably, these endothelial cells, along with cardiomyocytes and vascular endothelial cells, exhibit high levels of PD-L1 expression on their surfaces (92,93).The use of PD-1/PD-L1 inhibitors can enable T cells to nonselectively target normal cells in the heart.Consequently, these factors increase the risk of cardiovascular toxicity.This study demonstrated a notable increase in the risk of hypertension with the use of PD-1/PD-L1 inhibitors in combination with chemotherapy (22, 24, 25, 29, 31, 32, 35-37, 40, 42-47, 51).This trend was specifically observed in the subgroups of PD-1 inhibitors, first-line treatment, and urothelial carcinoma (UC), which has not been reported in previous meta-analyses.This phenomenon may be attributed to the immune-enhancing effects of PD-1/PD-L1 inhibitors.Owing to the high expression of PD-L1 on vascular endothelial cells (94), medications that enhance non-specific attack by T cells can also cause damage to vascular endothelial cells.This weakens the ability of cells to regulate blood pressure, leading to blood pressure fluctuations (95).However, the exact mechanism requires further investigation.In addition, while PD-1/PD-L1 did not exhibit statistically significant outcomes compared with chemotherapy or placebo, it can be inferred that PD-1/PD-L1 carries a reduced risk of inducing hypertension compared with the placebo group.This novel fact should be applied in clinical settings; when hypertension occurs after using PD-1/PD-L1, initial focus should be on identifying factors unrelated to this medication, such as potential drug interactions, unhealthy lifestyle choices, underlying health conditions, age, or gender.
Despite the lack of significant differences in the risk of heart failure among the treatment regimens in this study (20, 22, 25, 31, 32, 34, 37, 45-47, 49, 62, 63, 65, 67, 78, 84), the potential detrimental effects of PD-1/PD-L1 on cardiac function should not be overlooked.Michel et al. (9) observed that six of seven patients with stage IV progressive melanoma treated with PD-1 had decreased left ventricular ejection fraction (LVEF) and exhibited no significant signs of myocarditis four weeks after the first treatment.In addition, this study also concluded that PD-1/PD-L1 alone (68,71,75,76,78,83,84) or in combination with chemotherapy (25,29,31,36,40,42) leads to an appreciably higher risk of hypotension, which was first reported in a metaanalysis, and could not be ruled out as a manifestation of reduced ejection following a decrease in cardiac function due to PD-1/PD-L1.This trend was particularly evident in the PD-1 + chemotherapy, PD-L1 alone, first-line treatment, or breast cancer subgroups.In addition to diminished cardiac pumping, hypotension cannot exclude the less common drug-induced hypersensitivity syndrome (DIHS), which results from excessive activation of T-cell function by immune checkpoint inhibitors (ICIs) (96).Vasodilation and increased permeability of the vessel wall lead to plasma extravasation, which reduces the intravascular blood volume and vasogenic hypotension.However, the exact mechanisms remain to be further elucidated.
In a comparison of PD-1/PD-L1 + chemotherapy versus chemotherapy (21-24, 29, 30, 32, 36, 40-42, 46, 47) and PD-1/ PD-L1 versus placebo (68,71,72,75,76,78,83,84), the use of PD-1/PD-L1-related therapy was associated with a considerably increased risk of arrhythmias.Particularly in the NSCLC subgroup, the combination of PD-1/PD-L1 inhibitors with chemotherapy led to a notably higher occurrence of all-grade or grade 3-5 arrhythmia (21,36).This is broadly consistent with the results of previous meta-analyses or reviews by Herrmann and Liu et al. (7,97).In addition, although there was no statistically significant difference in the risk of arrhythmia between PD-1/PD-L1 inhibitors and chemotherapy, the two PD-1 inhibitors, nivolumab and pembrolizumab, exhibited a lower risk of arrhythmia than docetaxel.Thus, more important with docetaxel is the prevention of several serious complications, such as myocardial ischemia due to abnormal heart rhythms.Additionally, positive results may be obtained concerning the apparent subjective discomfort experienced by the patients.Currently, physicians can easily ascertain abnormal heart rhythms an d c o l l e c t t h e s e d a t a u s i n g H o l t e r ( 2 4 h d y n a m i c electrocardiogram) or other devices.However, additional fundamental research is required to investigate the mechanisms by which PD-1/PD-L1 affects the cardiac conduction system.
Clinical evidence has indicated that immunotherapy can cause myocarditis, which should be taken seriously.The severity of immune-associated myocarditis varies from mild cases without apparent inflammation to severe cases that may be associated with heart failure, cardiogenic shock, and a high mortality rate in the case of rapidly progressing fulminant myocarditis (98, 99).Hu et al. concluded that immunotherapy drastically increased the risk of myocardial disease compared with conventional antitumor therapy (100).This is the first study to provide evidence that the combination of PD-1/PD-L1 inhibitors and chemotherapy is associated with an elevated risk of myocarditis (17, 21-25, 28, 30, 31, 33, 37, 38, 50, 69, 91).However, no positive results were obtained in the subgroup analysis, which should be conducted in additional RCTs.The exact mechanism of immune-associated myocarditis remains unclear, but some preclinical studies have made some conjectures, such as inflammation due to T-cell activation (101).Given the poor prognosis of this disease, more clinical data and basic research are required.
The combination of PD-1/PD-L1 and CTLA-4 blockade substantially enhances the immune responses and survival rates in certain cancers (102).However, it also increases the risk of adverse effects.This study found that the risk of cardiovascular toxicity following PD-1 combined with CTLA-4 treatment was noticeably higher than following PD-1 treatment alone, and these results were consistent with prior findings.Preclinical trials have revealed that when PD-1 on the surface of myocardial cells binds to PD-L1 on the surface of T lymphocytes, it prevents T lymphocytes from attacking the myocardium.CTLA-4, on the other hand, prevents lymphocyte proliferation and spread.Therefore, the simultaneous inhibition of both pathways inevitably leads to indiscriminate T lymphocyte attacks on myocardial tissue, resulting in an increased risk of cardiovascular toxicity with the combined use of ICIs (103).Further research is required to decrease the occurrence of adverse event while maintaining the efficacy of the combination.
Cardiovascular toxicities associated with ICIs can be indicated by several biomarkers, including inflammatory markers such as Creactive protein, erythrocyte sedimentation rate, and white blood cell count, as well as cardiac injury markers like troponin I, creatine kinase-MB, and brain natriuretic peptide.The development of ICI adverse effects is attributed to excessive enhancement of immune function, leading to inadvertent harm to normal cells.In response, we initially administered symptomatic treatments involving a variety of immunosuppressive agents, including corticosteroids, cytotoxic drugs, calcineurin inhibitors, and biologics.Secondly, the severity of the adverse effects needs to be assessed to determine whether temporary or permanent discontinuation of the medication is warranted.In addition, screening specific patients before initiating treatment can help prevent adverse effects.For instance, it is not recommended for individuals with autoimmune diseases, organ transplant recipients, patients with active hepatitis, or elderly patients to use ICIs.Furthermore, patients with pre-existing cardiovascular disorders should be monitored (104).
This meta-analysis further refined the cardiovascular toxicity of PD-1/PD-L1 through a comprehensive analysis of 69 RCTs.Moreover, there was no heterogeneity or insignificant heterogeneity among the RCTs included in this meta-analysis; thus, the results were reliable.However, this study had some limitations.Only 11% of the original studies searched reported the above cardiovascular toxicity events.In an initial comparison of morbidity data, PD-1/PD-L1 treatment resulted in a higher number of cardiovascular adverse events than conventional treatment.However, the final meta-analysis did not yield positive results.First, it can be inferred that PD-1/PD-L1 therapy is safe.However, it should also be noted that cardiovascular adverse events may not have received sufficient attention from doctors and patients, resulting in patients not seeking medical treatment promptly or first consulting physicians not collecting data on time.Therefore, due to the lack of sufficient sample size, this study was unable to collect baseline information for subgroup analyses of additional possible risk factors or to shed light on the specifics of chemotherapy.Furthermore, this meta-analysis exclusively included RCTs; most of these only reported a greater than certain percentage of cardiovascular toxicities, which may lead to the underreporting of some rare diseases with low incidence.

Conclusion
The combination of PD-1/PD-L1 with chemotherapy increases the risk of hypertension, hypotension, arrhythmia, and myocarditis.The incidence of hypotension or arrhythmia associated with PD-1/ PD-L1 inhibitors was substantially higher than that associated with placebo.When hypertension is observed in patients receiving PD-1/ PD-L1 inhibitors, factors other than ICIs should be considered as potential contributors in the first instance.
subgroup analyses were conducted according to types of tumors.Funnel plot depicting the risk of hypertension in PD-1/PD-L1 versus chemotherapy.(B) The risk of hypertension of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.Funnel plot depicting the risk of hypertension in PD-1/PD-L1 versus placebo.(C) The risk of hypertension of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.

SUPPLEMENTARY FIGURE 3
Funnel plots depicting the risk of hypotension in PD-1/PD-L1 + chemotherapy versus chemotherapy.(A1) The risk of hypotension of allgrade: subgroup analyses were conducted according to PD-1/PD-L1.(A2) The risk of hypotension of all-grade: subgroup analyses were conducted according to types of tumors.(A3) The risk of hypotension of grade 3-5: subgroup analysis was conducted according to PD-1/PD-L1.Funnel plot depicting the risk of hypotension in PD-1/PD-L1 versus placebo.(B) The risk of hypotension of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.Funnel plot depicting the risk of hypotension in PD-1/PD-L1 versus chemotherapy.(C) The risk of hypotension of grade 3-5: subgroup analysis was conducted according to PD-1.

SUPPLEMENTARY FIGURE 4
Funnel plots depicting the risk of arrhythmia in PD-1/PD-L1 + chemotherapy versus chemotherapy.(A1) The risk of arrhythmia of all-grade: subgroup analyses were conducted according to PD-1/PD-L1.(A2) The risk of arrhythmia of all-grade: subgroup analyses were conducted according to types of tumors.(A3) The risk of arrhythmia of grade 3-5: subgroup analyses were conducted according to PD-1/PD-L1.(A4) The risk of arrhythmia of grade 3-5: subgroup analyses were conducted according to types of tumors.Funnel plot depicting the risk of arrhythmia in PD-1/PD-L1 versus chemotherapy.(B) The risk of arrhythmia of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.

SUPPLEMENTARY FIGURE 5
Funnel plot depicting the risk of arrhythmia in PD-1/PD-L1 versus placebo.(A1) The risk of arrhythmia of all-grade: subgroup analyses were conducted according to PD-1/PD-L1.(A2) The risk of arrhythmia of all-grade: subgroup analyses were conducted according to treatment lines.

SUPPLEMENTARY FIGURE 6
Funnel plot depicting the risk of myocarditis in PD-1/PD-L1 versus chemotherapy.(A) The risk of myocarditis of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.Funnel plot depicting the risk of myocarditis in PD-1/ PD-L1 versus placebo.(B) The risk of myocarditis of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.Funnel plot depicting the risk of myocarditis in PD-1/PD-L1 + chemotherapy versus chemotherapy.(C) The risk of myocarditis of all-grade: subgroup analysis was conducted according to PD-1/ PD-L1.Funnel plot depicting the risk of cardiovascular toxicities in PD-1/PD-L1 + CTLA-4 versus PD-1/PD-L1.(D) The risk of cardiovascular toxicities of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.Funnel plot depicting the risk of cardiovascular toxicities in PD-1/PD-L1 + CTLA-4 versus chemotherapy.(E) The risk of cardiovascular toxicities of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.

2
FIGURE 2Forest plots depicting the risk of hypertension in PD-1/PD-L1 + chemotherapy versus chemotherapy.(A1) The risk of hypertension of all-grade: subgroup analyses were conducted according to PD-1/PD-L1.(A2) The risk of hypertension of grade 3-5: subgroup analyses were performed based on PD-1/PD-L1.(A3) The risk of hypertension of grade 3-5: subgroup analyses were performed based on types of tumors.Forest plot depicting the risk of hypertension in PD-1/PD-L1 versus chemotherapy.(B) The risk of hypertension of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.Forest plot depicting the risk of hypertension in PD-1/PD-L1 versus placebo.(C) The risk of hypertension of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.

3 Forest
FIGURE 3 Forest plots depicting the risk of hypotension in PD-1/PD-L1 + chemotherapy versus chemotherapy.(A1) The risk of hypotension of all-grade: subgroup analyses were conducted according to PD-1/PD-L1.(A2) The risk of hypotension of all-grade: subgroup analyses were conducted according to types of tumors.(A3) The risk of hypotension of grade 3-5: subgroup analysis was conducted according to PD-1/PD-L1.Forest plot depicting the risk of hypotension in PD-1/PD-L1 versus placebo.(B) The risk of hypotension of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.Forest plot depicting the risk of hypotension in PD-1/PD-L1 versus chemotherapy.(C) The risk of hypotension of grade 3-5: subgroup analysis was conducted according to PD-1.

4
FIGURE 4 Forest plots depicting the risk of arrhythmia in PD-1/PD-L1 + chemotherapy versus chemotherapy.(A1) The risk of arrhythmia of all-grade: subgroup analyses were conducted according to PD-1/PD-L1.(A2) The risk of arrhythmia of all-grade: subgroup analyses were conducted according to types of tumors.(A3) The risk of arrhythmia of grade 3-5: subgroup analyses were conducted according to PD-1/PD-L1.(A4) The risk of arrhythmia of grade 3-5: subgroup analyses were conducted according to types of tumors.Forest plot depicting the risk of arrhythmia in PD-1/PD-L1 versus chemotherapy.(B) The risk of arrhythmia of all-grade: subgroup analysis was conducted according to PD-1/PD-L1.

5 6
FIGURE 5Forest plots depicting the risk of arrhythmia in PD-1/PD-L1 versus placebo.(A1) The risk of arrhythmia of all-grade: subgroup analyses were conducted according to PD-1/PD-L1.(A2) The risk of arrhythmia of all-grade: subgroup analyses were conducted according to treatment lines.

TABLE 1
The baseline characteristics of the RCTs included in this meta-analysis (Total of 69 clinical trials).

TABLE 3
The risk of all-grade myocardial infarction, heart failure, pericardial diseases, embolism, thrombosis and vasculis: subgroup analyses were carried out based on treatment lines.