Heterogeneity in adverse events related to atezolizumab-bevacizumab for hepatocellular carcinoma reported in real-world studies

Abstract

Background & aims: Safety data for patients with hepatocellular carcinoma (HCC) treated with atezolizumab-bevacizumab in the real-world setting remain uncertain. Thus, the aim of this study was to evaluate the incidence of adverse events (AEs) in patients with HCC treated with atezolizumab-bevacizumab in the literature.

Methods: In this systematic review and meta-analysis, we searched PubMed for original studies reporting percentages of AEs in patients with HCC receiving atezolizumab-bevacizumab between 2020 to 2023, using the search terms "Atezolizumab/Bevacizumab", "HCC" and "Adverse events". We summarized the incidence of AEs and performed a meta-analysis in order to evaluate the incidence of AEs reported in the literature.

Results: A total of 30 studies (3,867 patients) were included. The analysis revealed heterogeneity in AE reporting, with arterial hypertension, proteinuria, and fatigue being the most frequently reported AEs whereas incidence of bleeding was reported in 66.7% of the studies and rare immune-related AEs were reported in 26.7% of the studies. The meta-analysis revealed pooled incidence rates of 79% for any grade AEs: 56% for grade 1/2 and 30% for grade ≥3. While the pooled rates of hypertension, anorexia, bleeding, pruritus, rash, and thyroid dysfunction were similar to those reported in the IMbrave150 trial, higher rates were observed in the literature for proteinuria, fatigue, ALT and AST elevations and gastrointestinal perforation. For grade ≥3 AEs, the percentages were consistent with the IMbrave150 trial, except for lower incidences of arterial hypertension and thrombosis in the literature. The exposure-adjusted incidence rates for proteinuria (55.7%), hypertension (45.3%) and fatigue (33.6%) were high. Heterogeneity was observed in the analysis of AEs across articles within the same cohorts of patients.

Conclusion: We observed a significant variability in AE reporting for atezolizumab-bevacizumab treatment in HCC in the literature, underscoring the need for standardized reporting practices.

Impact and implications: Considering the demonstrated safety of atezolizumab-bevacizumab in randomized-controlled trials, this meta-analysis offers valuable insights into reported occurrences of adverse events. Our study highlights significant heterogeneity among studies, underscoring the need to improve adverse event recording. Understanding the incidence and severity of treatment-related adverse events beyond clinical trials is essential for prompt intervention and may help in preventing treatment discontinuation and complications, potentially leading to better outcomes without significantly compromising quality of life due to adverse events.

Keywords: adverse effects; immunotherapy; liver cancer.

Graphical abstract

Fig. 1

Flow chart of the study (adapted from PRISMA 2020 flow diagram). HCC, hepatocellular carcinoma; TARE, trans-arterial radioembolization.

Fig. 2

Raw data on adverse events reported across 30 studies. (A) Heatmap illustrating the presence or absence of descriptions of adverse events, regardless of their grade, across 30 included studies. Each column represents a study analyzed with its annotations. (B) Boxplots displaying the distribution of percentages of each adverse event (raw data) across studies, categorized according to their severity. Boxes represent the IQR of the data, the horizontal line the median value and whiskers the range of the data, extending to 1.5 times the IQR from the quartiles. Individual data points corresponding to each study are plotted and colored differently based on the type of study. On the graph of each adverse event, the percentages reported in the phase Ib and phase III studies have also been included for comparison; however, the data from phase Ib and phase III were not used in the calculation of the median and IQR. ALT, alanine aminotransferase; AST, aspartate aminotransferase.

Fig. 3

Meta-analysis for frequent adverse events occurring during atezolizumab-bevacizumab treatment. Each forest plot represents a specific adverse event category, regardless of the severity grade. (A) Arterial hypertension. (B) Proteinuria. (C) Fatigue. (D) Anorexia. (E) Diarrhea. (F) Rash. Squares indicate estimates; size of squares, study weights; whiskers, 95% CIs; diamonds, mean estimates.

Fig. 4

Meta-analysis for the less frequent adverse events occurring during atezolizumab-bevacizumab treatment. Each forest plot represents a specific adverse event category, regardless of the severity grade. (A) ALT elevation. (B) AST elevation. (C) Fever. (D) Pruritus. Squares indicate estimates; size of squares, study weights; whiskers, 95% CIs; diamonds, mean estimates. ALT, alanine aminotransferase; AST, aspartate aminotransferase.

Fig. 5

Exposure-adjusted incidence rate of adverse events in patients with HCC treated with atezolizumab-bevacizumab. The blue set of columns represents exposure-adjusted rates per 100 PY, while the yellow set represents percentages reported across the seven studies for which the median duration of treatment with atezolizumab-bevacizumab was known. ALT, alanine aminotransferase; AST, aspartate aminotransferase; PY, patient-years.

Fig. 6

Percentage of adverse events reported in studies published by the same research group. Boxplots displayed the distribution of percentages of each any grade adverse event (raw data) reported in different studies published by the same group: RELPEC group (14 studies), European and non-European cohorts (10 studies), Korean cohorts (5 studies), Taiwanese cohorts (2 studies), German cohorts (2 studies) and Japanese cohorts not included in RELPEC group (3 studies). Boxes represent the IQR of the data, the horizontal line the median value and whiskers the range of the data, extending to 1.5 times the IQR from the quartiles. Individual data points corresponding to each study. ALT, alanine aminotransferase; AST, aspartate aminotransferase.