Impact of Antihypertensive Therapy on Recurrence Risk of Ovarian Cancer for Bevacizumab-associated Hypertension

Antiangiogenic therapies, such as bevacizumab, have significantly improved the treatment landscape of various cancers, including gynecological types. Bevacizumab is a humanized monoclonal IgG1 antibody targeting all forms of human vascular endothelial growth factor-A (VEGF), essential for angiogenesis in healthy and cancerous tissues. This drug inhibits angiogenesis primarily by blocking VEGF from activating its receptors, VEGFR, on endothelial cells, thus limiting tumor growth by cutting off their blood supply and aiding the effectiveness of cytotoxic drugs by reducing tumor interstitial pressure and abnormal blood vessels. For advanced ovarian cancer, bevacizumab was found to improve the median progression-free survival of 4 months. However, targeting tumor vessels will eventually modify the patients' vasculature, leading to hypertension in 2 to 19% of patients as per a meta-analysis. VEGF-VEGFR pathway inhibition can cause hypertension by inhibiting the production of nitric oxide in the arterial wall or by capillary rarefaction and increased afterload. Several retrospective studies support a correlation between the occurrence of a bevacizumab-associated hypertension (BIH) and outcomes, highlighting the link between tumor and patient vasculature. However, bevacizumab-associated hypertension can lead, in some rare cases, to major adverse cardiovascular events (MACE). Multiple treatments have been proposed to control this adverse reaction. Calcium channel blockers (CCBs) such as amlodipine have been proposed and offers a way to control bevacizumab-associated hypertension. Angiotensin-converting enzyme inhibitors (ACEi), which targets the renin-angiotensin-aldosterone system (RAAS) to control hypertension, could also constitute an interesting option and could be more efficient. However, there is both caution against the use of CCBs and ACEis, since nifedipine has been shown to induce VEGF secretion in vitro, and ACEi (enalapril and perindopril) have been associated in a case report with CA125 increase. Current guidelines vary on the preferred class of medication for initial management of antiangiogenic-associated hypertension. The European Society of Cardiology (ESC) guidelines favor ACEis as the first-line treatment, whereas guidelines for gynecological cancers suggest CCBs for managing BIH in patients without other comorbidities. This discrepancy highlights the need for direct comparisons of oncologic and cardiologic outcomes in this context. Randomized controlled trials (RCTs), despite being the gold standard, face practical and financial constraints that may limit their use. In the absence of RCT, observational data can serve as a crucial alternative for estimating real-world causal effects. However, observational studies come with challenges, such as confounding factors due to the lack of randomization and the risk of immortal-time bias. A recent advancement in addressing these challenges is the use of emulated trials, which strive to replicate the conditions of a target RCT. This approach, employing propensity-score adjustment methods, has shown promise in extracting the highest level of evidence from observational data. The SNDS, covering 98.8% of the French population, offers extensive and detailed demographic, health condition, and healthcare utilization data, enabling a thorough and representative analysis of healthcare outcomes in oncology. In this study, to address measured confounding at baseline, an exact copy of each patient's record will be created and allocated to each arm of the study, ensuring identical baseline characteristics across the study groups. Clones will be artificially censored when their observed trajectories deviated from the treatment strategy of the arm to which they were assigned. This informative artificial censoring introduces selection bias over time, which will be addressed using inverse probability of censoring (IPC) weights. Standardized mean differences (SMDs) will be derived to assess adjustment quality after IPC weighting. The per-protocol average treatment effect (ATE) will be estimated by the one- and three-year differences in PFS probability and three-year restricted mean survival time (RMST) differences in the IPC-weighted population. This will be visually assessed using weighted Kaplan-Meier survival curves. Similar steps will be performed for overall survival (OS) at one, three, and five years for survival probability differences, and at five years for RMST. This study focuses on comparing the effects of antihypertensive drugs - ACEi vs CCBs- on the risk of relapse in patients undergoing adjuvant maintenance treatment with bevacizumab for ovarian cancer after debulking surgery. It employs an emulation of a clinical trial using data from the French National Health Data System (SNDS). Secondary objectives will include examining the causal impact of these antihypertensive classes on overall survival, the incidence of MACE and the time to treatment failure.