Increased Hemoglobin Associated with VEGF Inhibitors in Advanced Renal Cell Carcinoma

Agents that inhibit vascular endothelial growth factor (VEGF) are now standard of care for the treatment of advanced renal cell carcinoma (RCC). Two VEGF tyrosine kinase inhibitors (TKIs), sunitinib and sorafenib, are now FDA approved for metastatic RCC. Sunitinib increases overall survival compared to interferon-α in the first line setting, and both improve response and progression-free survival (PFS). A monoclonal antibody against VEGF, bevacizumab, is still investigational but has recently been shown to improve PFS in combination with interferon-α. These and other investigational VEGF inhibitors continue to be actively investigated as monotherapies and in combination with other targeted agents for the treatment of advanced RCC among other cancers. Surrogate biomarkers that accurately predict therapeutic VEGF inhibition and anti-tumor response to these inhibitors would be useful. Preclinical work at our institution revealed that extremely stringent VEGF inhibition in mice resulted in increased hemoglobin and erythropoietin (Epo). This Epo production originated in the liver, not the kidney, implicating endogenous VEGF as a previously unsuspected repressor of hepatic Epo synthesis and suggested that hemoglobin and Epo may be non-invasive markers for VEGF inhibition in vivo. We retrospectively reviewed whether hemoglobin levels increased in patients receiving bevacizumab as monotherapy or in combination with epidermal growth factor receptor (EGFR) or dose-reduced (or eventual dose-reduced) VEGF TKIs for advanced RCC. Twelve patients with advanced RCC were eligible for retrospective review. Demographic data, baseline anemia, initial and treatment hemoglobin levels, response, and confounding factors such as recent chemotherapy, blood transfusions, iron supplementation, or growth factor use were assessed. Correlation between the degree of change in hemoglobin from baseline to peak level and PFS were evaluated. The majority of patients were male and treatment naïve (67%). Fifty percent of patients had bone metastases, and 50% had baseline anemia. Five patients were on concurrent treatment with sunitinib, sorafenib, or erlotinib. Hemoglobin rose on bevacizumab in 11 patients. The median change in hemoglobin was 1.6 g/dL (0–2.7). The median percent rise was 12.3%; 10.9% in the monotherapy cohort and 18.6% in the combination group. For all 12 patients, the median time to initial increase in hemoglobin was 35 days (0–168 days) and 82 days (0–476 days) from baseline level to peak level. Median time to peak hemoglobin (first rise to peak) was 68.5 days (0–308 days). Median duration of hemoglobin rise was defined as the first rise in hemoglobin to the first decline in levels after peak and was 113 days (n=10, range: 28–350 days). The best response was disease stabilization in 8 patients. In the metastatic patients, degree of peak increase correlated with longer PFS: a 5–15% increase yielded a 3.1 month median PFS whereas a greater than 15% increase corresponded to an 8.2 month median PFS. In conclusion, the majority of patients (92%) on bevacizumab monotherapy or in combination with dose-reduced VEGF or full-dose EGFR tyrosine kinase inhibitors experienced an increase in hemoglobin from their pre-treatment baseline suggesting a correlative effect. The combination group was expected to have less of an increase in hemoglobin as sunitinib and sorafenib are known to cause anemia via c-Kit inhibition. We hypothesize that the reduced doses of these agents when used in combination with bevacizumab were inadequate to cause significant anemia but were sufficient to induce more complete VEGF inhibition in these patients. This is the first human study to identify increased hemoglobin as a possible consequence of treatment with VEGF inhibitors. The correlation between degree of increase in hemoglobin and longer PFS suggests that hemoglobin may be useful as a surrogate marker of stringent VEGF inhibition and as a predictive factor for clinical response. This relationship warrants validation in larger cohorts.