Background
Phase III, double-blind, placebo-controlled RCT. 2710 patients with resected stage II (26%) or stage III (74%) colon cancer. Tested whether adding bevacizumab to adjuvant mFOLFOX6 would improve DFS, based on its efficacy in metastatic CRC. Bevacizumab was given for 1 year total (beyond the 6 months of chemotherapy).
Interventions and follow up
Arm A: mFOLFOX6 ×12 cycles (6 mo) + bevacizumab 5 mg/kg IV q2wk ×26 cycles (1 yr total)
Arm B: mFOLFOX6 ×12 cycles + placebo q2wk ×26 cycles
Primary endpoint: Disease-free survival (DFS)
Median follow-up: 35.6 mo
Arm B: mFOLFOX6 ×12 cycles + placebo q2wk ×26 cycles
Primary endpoint: Disease-free survival (DFS)
Median follow-up: 35.6 mo
Results
DFS: 3-yr DFS 77.4% vs 75.5%; HR 0.89, 95% CI 0.76–1.04, P=.15 — not significant
OS: HR 0.95 — not significant
DFS curve pattern: Early separation favoring bevacizumab during treatment, then convergence after discontinuation at 1 year
OS: HR 0.95 — not significant
DFS curve pattern: Early separation favoring bevacizumab during treatment, then convergence after discontinuation at 1 year
Adverse events
Vascular/renal: Grade ≥3 hypertension 10.9% (bevacizumab) vs 2.5% (placebo); proteinuria 0.9% vs 0.1%; arterial thromboembolism 1.5% vs 0.8%
Surgical/overall: Wound complications 1.3% vs 0.2%; overall grade ≥3 rate 52.3% vs 49.8%
Surgical/overall: Wound complications 1.3% vs 0.2%; overall grade ≥3 rate 52.3% vs 49.8%
Conclusions
Adding bevacizumab to adjuvant mFOLFOX6 did not significantly improve DFS in resected stage II/III colon cancer. The DFS curves separated during bevacizumab administration but converged after its discontinuation at 1 year — suggesting a cytostatic rather than curative effect — without translating into lasting DFS or OS benefit.
Key Limitations
The convergence of DFS curves after bev discontinuation is a fundamental limitation and raises questions about the biological rationale for anti-VEGF therapy in the adjuvant (micrometastatic) disease setting. The DFS effect during treatment did not persist — a hallmark of anti-angiogenic therapy in this context. Stage II patients (26%) were included, limiting generalizability to stage III. These data, alongside AVANT, collectively closed the door on bevacizumab in adjuvant CRC.
Clinical Context
NSABP C-08 was one of several definitive negative trials of bevacizumab in the adjuvant colon cancer setting. AVANT (de Gramont, Lancet Oncol 2012) showed similar results, and together these trials established that anti-VEGF therapy has no role post-resection in colon cancer. Bevacizumab remains standard only in metastatic CRC. These negative results are often cited in discussions of anti-angiogenic therapy biology and the adjuvant-metastatic discordance in solid tumors.
References