Effectiveness of Bevacizumab With First-Line Combination Chemotherapy for Medicare Patients With Stage IV Colorectal Cancer

Before 1998, intravenous fluoropyrimidine therapy was the only efficacious option for metastatic colorectal cancer, extending median survival from 6 months without therapy to 1 year.¹ Over the last 13 years, additional drugs have entered the landscape, including two other cytotoxic drugs (irinotecan and oxaliplatin) and targeted monoclonal antibodies (bevacizumab, cetuximab, and panitumumab). First-line randomized controlled trials demonstrated that adding either irinotecan or oxaliplatin to fluoropyrimidines improves median survival by 2 to 4 months.^2–4 In the United States, a weekly bolus regimen of irinotecan, fluorouracil (FU), and leucovorin (IFL) was initially embraced as the standard regimen for chemotherapy-naive patients. Subsequent trials demonstrated that infusional fluoropyrimidine regimens with either oxaliplatin (infusional FU, leucovorin, and oxaliplatin [FOLFOX]) or irinotecan (FU, leucovorin, and irinotecan [FOLFIRI]) are more efficacious^5,6 and less toxic⁵ than IFL, leading to a shift from IFL to FOLFOX or FOLFIRI in the mid-2000s.

Bevacizumab, an antibody against the vascular endothelial growth factor, was initially approved by the US Food and Drug Administration (FDA) in 2004 with FU-based chemotherapy as first-line treatment of metastatic colorectal cancer.⁷ The pivotal trial demonstrated that bevacizumab added to IFL improved median survival from 15.6 to 20.3 months (P < .001).⁸ In a subsequent trial, bevacizumab added to fluoropyrimidine and oxaliplatin improved median overall survival more modestly (21.3 months with bevacizumab v 19.9 months without bevacizumab; P = .08).⁹

The trials that led the FDA to approve bevacizumab in first-line metastatic colorectal cancer treatment addressed the question of clinical efficacy in patients who met stringent eligibility criteria and were typically younger and healthier than the typical patient with metastatic colorectal cancer. Effectiveness studies examine the impact of treatment in the context of usual care settings, often in populations more diverse by age, race, and health status. To understand the balance of benefits and harms in such a nonclinical trial setting, we used the Surveillance, Epidemiology, and End Results (SEER) -Medicare linked database to compare the effectiveness of cytotoxic chemotherapy treatment with and without bevacizumab in newly diagnosed stage IV colorectal cancer.

Randomized clinical trials have demonstrated that bevacizumab improves overall and/or progression-free survival when added to first-line cytotoxic chemotherapy in metastatic colorectal cancer.^8,9,17 Using the SEER-Medicare linked database, we evaluated the effectiveness of bevacizumab with combination chemotherapy in the Medicare population. The addition of bevacizumab to fluoropyrimidine and either oxaliplatin or irinotecan led to a statistically significant improvement in overall survival, although the benefit seemed limited to patients receiving irinotecan-based chemotherapy regimens and not evident in patients treated with oxaliplatin-based chemotherapy. The addition of bevacizumab to cytotoxic chemotherapy increased the rates of strokes and GI perforations but not cardiac events.

Bevacizumab was initially tested in colorectal cancer in a small, randomized phase II trial with FU and leucovorin.¹⁸ The results led to a phase III trial of IFL with or without bevacizumab that demonstrated a significant improvement in median overall survival from 15.6 to 20.3 months (HR, 0.66; P < .001).⁸ By the time of FDA approval of bevacizumab, IFL was no longer the preferred first-line regimen for colorectal cancer. There was an increasing shift from IFL to FOLFOX based on results from North Central Cancer Treatment Group trial 9741, showing a survival advantage with FOLFOX compared with IFL (HR, 0.74; P = .0014).⁶ IFL was also supplanted in practice by FOLFIRI. Likely because of these shifts in cytotoxic therapy backbone, the FDA approved bevacizumab in combination with intravenous FU-based chemotherapy, not restricting usage with only IFL but enabling usage of bevacizumab with FU alone, IFL, FOLFIRI, or FOLFOX. Perhaps because bevacizumab is a biologic agent, and not a cytotoxic agent, both oncologists and the FDA were more willing to combine it with an infusional FU and irinotecan regimen or oxaliplatin-based therapy, initially in the absence of efficacy data. For many oncologists, FOLFOX and bevacizumab became the regimen of choice in the United States. Two nonrandomized, observational studies of patients receiving bevacizumab with combination chemotherapy (including IFL, FOLFOX, FOLFIRI, or capecitabine and oxaliplatin) confirmed progression-free survival of 9.6 to 10.8 months,^19,20 with one cohort reporting comparable progression-free survival (although reduced overall survival) in elderly patients.¹⁹ However, these cohorts did not collect data from patients not receiving bevacizumab to serve as a comparison. Furthermore, most subsequent phase III trials of first-line therapy incorporate bevacizumab in both arms of the study.^7,21,22

The goal of effectiveness research is to understand the outcomes of treatments as they are adopted by real-world practitioners who face everyday treatment decisions in the face of multiple studies and patients whose attributes do not exactly correspond to the clinical trial eligibility criteria. However, data from a phase III trial of oxaliplatin-based therapy with or without bevacizumab⁹ and our effectiveness results suggest that the magnitude of benefit combining bevacizumab with oxaliplatin is more modest than originally observed for IFL and bevacizumab. This difference by chemotherapy backbone may suggest a differential benefit when bevacizumab is added to irinotecan compared with oxaliplatin, although such a biologic interaction has not been reported.

We anticipated that the rates of major toxicities such as stroke, perforation, and cardiovascular events would have been substantially greater in the Medicare population than those encountered in the efficacy studies. Notably, toxicities detected in our effectiveness cohort are only marginally higher for bevacizumab-treated patients and are similar to the rates reported in clinical trials. Administrative claims data are not optimized for identification of certain bevacizumab-associated adverse effects such as hypertension, proteinuria, and delayed wound healing. However, major toxicities such as stroke, perforation, and myocardial infarction are well captured in billing claims. Bevacizumab did not result in a substantially higher rate of myocardial infarction or venous thrombosis in this cohort, although there was a modestly higher rate of strokes and GI perforations.

The advantage of using the SEER-Medicare linked database is that the choice of chemotherapy regimens is not limited to the stringent treatment program required in clinical trials and analysis by chemotherapy type is possible. Such analyses also capture the outcomes of real-world usage of these medications, outside of protocol-prescribed dose modifications and treatment holds.

Several limitations of this analysis must be noted. The potential for inaccurate coding exists for any claims-based analysis, clinical information available from billing records is not as detailed as from chart reviews or clinical trials, and the details of dose modification could not be captured.¹¹ Our results may not generalize to the non-Medicare population or the approximately 16% of Medicare beneficiaries who receive care in a health maintenance organization setting, where patterns of care may be different.²³ However, because two thirds of patients with colorectal cancer are over age 65 years and Medicare is the primary insurer for the vast majority, our study represents the care received by typical US patients with colorectal cancer. Because identification of recurrent cancer is challenging using administrative data, we restricted our cohort to patients with stage IV colorectal cancer. Most clinical trials do not identify substantial differences in prognosis for patients with stage IV disease and patients with recurrent metastatic disease. However, some caution is nonetheless warranted because our results may not generalize to patients with recurrent metastatic disease. In the context of an observational cohort study, we cannot be certain that the survival differences we observed are a result of bevacizumab or alternatively a result of baseline differences in patient attributes. Our baseline data do not demonstrate appreciable differences in baseline characteristics, and we used propensity score modeling to adjust for potential biases. Nonetheless, 60-day mortality is greater in patients not receiving bevacizumab. This difference may be a result of chance or an actual early benefit of therapy that includes bevacizumab. Alternatively, patients receiving bevacizumab may be different from patients not receiving bevacizumab, and this difference may be associated with early mortality risk.

In conclusion, the addition of bevacizumab to first-line therapy for metastatic colorectal cancer significantly improved overall survival in a cohort derived from the linked SEER-Medicare database. The data from this analysis suggest the benefit is derived from irinotecan-based regimens, with marginal benefit in oxaliplatin-treated patients. Furthermore, the benefit of bevacizumab was more attenuated in the latter years when there was an increasing shift away from bolus to infusional fluoropyrimidine regimens, increased availability of latter-line epidermal growth factor inhibitors, and more aggressive surgical management of limited stage IV disease. An ideal trial to answer the question of whether bevacizumab benefit is truly limited to irinotecan-based regimens would be a 2 × 2 randomization between FOLFOX and FOLFIRI plus or minus bevacizumab. However, because multiple prior FOLFOX versus FOLFIRI trials have been reported and suggest noninferiority between the two regimens (albeit none have included bevacizumab),^24,25 enthusiasm for such a trial would be low. In summary, the routine use of bevacizumab in combination with cytotoxic chemotherapy for Medicare beneficiaries with stage IV colorectal cancer is associated with a modest survival advantage and a modest excess risk of harms from perforation and stroke. On balance, elderly patients with colorectal cancer can be counseled that including bevacizumab in first-line therapy regimens for metastatic colorectal cancer seems to be no more than marginally effective.

This study used the linked Surveillance, Epidemiology, and End Results (SEER) -Medicare database. This resource has been made available to the research community through collaborative efforts of the National Cancer Institute (NCI) and Centers for Medicare and Medicaid Services (CMS). We acknowledge the efforts of the Applied Research Program, NCI; the Office of Research, Development and Information, CMS; Information Management Services; and the SEER Program tumor registries in the creation of the SEER-Medicare database.

Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.

Employment or Leadership Position: None Consultant or Advisory Role: Jeffrey A. Meyerhardt, sanofi-aventis (C), Bayer Pharmaceuticals (C) Stock Ownership: None Honoraria: None Research Funding: None Expert Testimony: None Other Remuneration: None