|Home | About | Journals | Submit | Contact Us | Français|
Multiple agents and combination therapies available to patients with advanced colorectal cancer have significantly improved survival and provided an opportunity for individualization of care, allowing clinicians and patients to prioritize risks and benefits of comparable regimens.
Patients with metastatic colorectal cancer (mCRC) are increasingly exposed to multiple chemotherapy regimens. Insight into patterns of care in mCRC is crucial to understanding physician and patient decision making.
Patients with mCRC diagnosed between June 2003 and June 2006 were identified from one academic and nine community oncology practices in the southeastern United States. Demographic, disease, treatment, and toxicity data were abstracted by retrospective medical record review.
After screening 738 medical records, 110 patients were determined eligible. Of these, mean age was 58 years (standard deviation, 12 years), 74% had stage IV disease at diagnosis, 39% were male, 53% were white, 26% were black, and 13% were age 70 years or older. As part of first-line mCRC chemotherapy, 100% of patients received regimens containing fluorouracil (FU), 87% received oxaliplatin (95% CI, 81% to 93%), 12% received irinotecan (95% CI, 6% to 18%), and 74% received bevacizumab (95% CI, 66% to 82%). The proportions of patients receiving subsequent lines of chemotherapy were: second line, 48% (n = 53); third line, 26% (n = 29); fourth line, 14% (n = 15); and fifth line, 5% (n = 5). From first- to third-line therapy, use of oxaliplatin and bevacizumab decreased, whereas irinotecan use increased. Among patients for whom therapy was discontinued, 29% experienced disease progression (PD), and 19% experienced toxicity; for 27%, no reason for discontinuation was documented. Of regimens containing oxaliplatin and irinotecan, 22% (n = 25 of 114) and 34% (n = 20 of 59) were discontinued because of PD, respectively. Of the same regimens, 19% (n = 21 of 114) and 20% (n = 12 of 59) were discontinued because of toxicity, respectively.
FU, oxaliplatin, and bevacizumab were most commonly used in first-line therapy for mCRC, despite data showing equivalency between regimens containing oxaliplatin and irinotecan. Use of oxaliplatin decreased and irinotecan use increased as treatment progressed beyond first-line therapy.
New combination therapies have significantly improved survival in patients with metastatic colorectal cancer (mCRC). The addition of oxaliplatin or irinotecan to flouoruracil (FU) has improved overall survival, progression-free survival, and response rates compared with FU alone.1–4 As survival with disease continues to increase, patients are exposed to an increasing number of chemotherapeutic agents and lines of chemotherapy.
Multiple agents and combination therapies available to patients with advanced CRC have provided an opportunity for individualization of care, whereby clinicians and patients can prioritize risks and benefits of comparable regimens.5 For example, in one study, two FU-based regimens were evaluated in the first and second lines of therapy for mCRC.6 Patients were randomly assigned to receive either FU, leucovorin, and irinotecan (FOLFIRI) or FU, leucovorin, and oxaliplatin (FOLFOX6) in the first line. On progression, patients were switched to the alternate regimen. Similar progression-free survival and response rates in the first and second lines of therapy were achieved with both regimens. However, toxicity profiles differed significantly between the first-line regimens. Patients receiving first-line FOLFIRI experienced greater frequencies mucositis, nausea/vomiting, and alopecia, whereas patients receiving first-line FOLFOX6 experienced more neutropenia and neurotoxicity. When regimens have equivalent efficacy but differing toxicity, patients and clinicians should have more freedom to select therapy on the basis of preference and patient circumstances. For example, a pianist could choose FOLFIRI to avoid sensory neuropathy.
With the addition of biologic agents such as bevacizumab—a monoclonal antibody directed against the vascular endothelial growth factor—and cetuximab and panitumumab—monoclonal antibodies targeting the endothelial growth factor receptor—various combinations and schedules have made management of mCRC extremely complex. Recent studies have continued to contrast efficacy between regimens containing these drugs, but to the best of our knowledge, no published study has examined community-based treatment patterns for mCRC in the United States. A better understanding of these treatment patterns could clarify how published findings from clinical trials are translated into practice. Given the proliferation of treatment options and lack of published data pertaining to community practice patterns, we undertook a longitudinal review to describe chemotherapeutic treatment patterns for mCRC.
Using a population-based strategy, patients with CRC were retrospectively identified from one academic (Duke University Medical Center, Durham, NC) and nine Duke University–affiliated community oncology practices. Eligible patients were adults with CRC diagnosed between June 2003 and June 2006 who received chemotherapy at a participating site. Their medical records had to contain at least 1 year of data on chemotherapeutic treatment plans and outcomes of treatments for mCRC. Patients had to have received as first-line chemotherapy either the combination of bolus FU, infusional FU, leucovorin, and oxaliplatin (FOLFOX) or FOLFIRI. Patients were excluded if they had been treated on a research protocol during the first or second lines of treatment. Informed consent was waived by the institutional review board. The research presented was conducted in accordance with the principles of the Declaration of Helsinki.7
Lists of potential participants generated using the local tumor registry were prepared for each participating site. Tumor registries were searched for analytic and nonanalytic cases in which patients potentially met eligibility criteria. At sites that did not have tumor registries or where tumor registries included only analytic cases, billing or other available onsite databases were searched for potential participants. Case report forms were developed for medical record abstraction by trained abstractors.
Data were abstracted in two phases. Phase I was devoted to evaluation of CRC according to relevant quality metrics issued by national oversight agencies (eg, the National Comprehensive Cancer Network and ASCO). These data will be presented separately. Phase II focused on evaluation of longitudinal management of mCRC. These data are reported here. Data were abstracted with regard to patient sociodemographics; disease stage; chemotherapeutic regimens, doses, modifications, and treatment frequency; clinicians' reasons for choosing or discontinuing therapy; and adverse events. Data regarding toxicity and toxicity-related decision making will be presented separately. Data were double-entered from the paper questionnaire forms. Built-in validation parameters and electronic forms that mimicked the paper forms were used to reduce risk of data entry errors. A 10% visual data inspection, with 100% inspection of critical data elements, was completed. Analysis focused on descriptive statistics using SAS 9.1 (SAS Institute, Cary, NC).
Of the 738 patients initially screened, 400 were eligible overall (Fig 1). Of these, 110 were eligible for phase II (Table 1). Of the 110 eligible patients, 39% were male, 53% were white, 29% had Medicare/Medicaid as a payer source, and 66% were younger than age 70 years. Mean age was 57.9 years. Men versus women were approximately equally likely to be treated with regimens containing FU (48% v 52%), oxaliplatin (48% v 52%), bevacizumab (45% v 55%), or irinotecan (41% v 59%). A majority of patients receiving any chemotherapy were younger than age 70 years (FU, 81%; bevacizumab, 82%; oxaliplatin, 85%; irinotecan, 79%).
A majority of patients (64%) were treated with FOLFOX-bevacizumab in the first line of treatment (Fig 2). Only six distinct regimens were used. In the second line of treatment, FOLFIRI-bevacizumab was the most common regimen used (13% of patients), but much greater regimen variety was noted; 19 distinct regimens were used. In the third line of treatment, irinotecan plus cetuximab was the most common regimen (21% of patients); 14 distinct regimens were used. As treatment progressed from first to third line, use of bevacizumab, FU, and oxaliplatin decreased, whereas use of irinotecan increased (Table 2).
In the first line of treatment, regimens were infrequently used before approval by the US Food and Drug Administration (Table 3). Regimens were often used before data from phase III studies supporting use of these regimens were published, suggesting that oncologists were willing to treat patients on the basis of preliminary or phase II data. Bevacizumab was used in combination with chemotherapy after US Food and Drug Administration approval but before phase III study publication and recommendation by the National Comprehensive Cancer Network.
One patient was abstracted as having received FOLFOX plus bevacizumab and capecitabine. Because it is unlikely that this patient received FU and capecitabine simultaneously, one could infer that the patient was switched from FU to capecitabine, but this was not clear from medical record abstraction.
Reason for physician choice of regimen was abstracted from medical records when that information was available. Physician reason for choosing a particular regimen was poorly documented and therefore missing from the majority of medical records (65%). Of those records in which reason for therapy choice was documented, 10% attributed choice to treatment guidelines, 12% to perceived patient tolerability, 1% to patient preference, fewer than 1% to comfort or experience with the particular regimen, and fewer than 1% to patient comorbidities.
Reasons for ending therapy were documented by physicians more often than reasons for choice of therapy; only 27% of medical records were missing reason for chemotherapy discontinuation. At time of abstraction, 13% of patients were still receiving therapy. Of those records in which reason for therapy discontinuation was documented, 18% attributed ending of treatment to tolerability or toxicity, 29% to disease progression, 10% to completion of predetermined treatment course, 2% to patient preference, and fewer than 1% to physician preference. For one patient, death was given as reason for discontinuation of therapy.
This community-based study reveals general treatment patterns for patients with mCRC. Oxaliplatin-based regimens were more commonly used in the first line of treatment than irinotecan-based regimens, and age played an important role in delivery of chemotherapy.
Among those who were treated in our study, oxaliplatin-based regimens (eg, FOLFOX) were preferred over irinotecan-based regimens (eg, FOLFIRI). This pattern was noted despite data available during the study interval that demonstrated equivalent efficacy between the two regimens. Tournigand et al6 randomly assigned patients previously untreated for mCRC to receive either FOLFOX6 or FOLFIRI. On disease progression, oxaliplatin was replaced with irinotecan or vice versa. Similar response rates, progression-free survival, and overall survival in the first line of therapy were achieved with both regimens. Despite similar efficacy results, toxicity profiles were different for the two regimens. Patients experienced more neurotoxicity and hematologic toxicities with oxaliplatin and more febrile neutropenia, nausea/vomiting, mucositis, and fatigue with irinotecan.
Because a majority of medical records in our study did not contain documentation on why certain regimens were chosen, we can, for the most part, only speculate as to why oxaliplatin-based regimens were preferred over irinotecan-based regimens. First, oncologists might have been more familiar with oxaliplatin because of its success in the adjuvant setting.8 Irinotecan-based therapy, on the other hand, has been shown to be an ineffective adjuvant therapy when compared with FU alone.9 Second, the perceived toxicity profile of oxaliplatin-based regimens might be more acceptable to oncologists than that of irinotecan-based regimens. Along these lines, oncologists might be disinclined to use irinotecan because of worse outcome of irinotecan combined with bolus FU in comparison with FOLFOX.10 Ease of use should not have been a significant factor, because in both regimens, infusional FU is used, and 2-week cycles can be established.2,6 Other influences, including pharmaceutical industry marketing, cost, and reimbursement, might have also played a role.
The findings regarding physician treatment intent are incomplete, because 65% of medical records did not indicate why specific chemotherapeutic regimens were chosen over others. Of those in which treatment intent was documented, a majority indicated that oncologists made their choices on the basis of either perceived patient tolerability or treatment guidelines. Patients discontinued treatment most commonly as a result of disease progression and, secondly, as a result of treatment-related toxicity. Although these findings are limited by missing data, they indicate that as patients live longer with mCRC, and as the number of drugs available to patients grows, toxicity plays an increasingly important role in selection of regimen.
A growing body of evidence supports an age bias against treating older patients, and this bias is additionally supported by our findings. A study of 6,959 patients with CRC identified through the Surveillance, Epidemiology, and End-Results–Medicare linked database demonstrated that older patients were significantly less likely to receive adjuvant chemotherapy, even after adjusting for presence of comorbid conditions.11 In a study of California Cancer Registry data including 1,956 patients, Ayanian et al12 found that patients age 85 years or older were significantly less likely than younger patients to receive adjuvant radiotherapy for rectal cancers. Schrag et al13 reported similar results in 1,670 patients with stage II and III rectal cancer identified through the Surveillance, Epidemiology, and End-Results–Medicare linked database. In this study, patients age 85 years or older were significantly less likely to receive radiotherapy or chemoradiotherapy, even after adjusting for degree of comorbidity.
Despite this evident bias, older patients who receive appropriate therapy tend to do as well as their younger counterparts. Goldberg et al14 conducted a retrospective, pooled analysis of four clinical trials evaluating adjuvant chemotherapy for CRC. The relative benefit of chemotherapy as measured by response rate and survival did not significantly differ in patients younger or older than age 70 years. Although this pooled analysis only included patients healthy enough for clinical trial participation, these important data suggest that numeric age by itself might not be as useful a tool in treatment-related decision making. Older age is also a negative predictor of whether a patient receives stage-appropriate chemotherapy in the usual care setting, even when controlling for comorbidity.15 In a single-institution study of 844 patients, Popescu et al16 found that elderly patients who received chemotherapy for mCRC did not experience worse toxicity or longer inpatient hospital stays compared with their younger counterparts.
The findings of our study are limited primarily by the design of retrospective medical record review. Understanding physician intent in selection of chemotherapeutic regimen was confounded by missing data, in that reasons for selection or discontinuation of therapies were infrequently and inconsistently documented. Furthermore, because of agreements to maintain site-level anonymity of data, we were unable to compare treatment patterns between academic and community practices. Because the vast majority of our patient sample was treated in the community setting, even if such a comparison were to be presented, the data would likely be less than informative.
Despite these limitations, our study indicates a preference for oxaliplatin-based first-line therapy among community medical oncologists. Furthermore, our findings support data from prior studies that demonstrate an age bias against treating older patients, regardless of type of chemotherapy. Prospective studies are needed for better understanding of physician intent in regimen selection for treatment of mCRC.
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: Michael A. Morse, Genentech (C); Amy P. Abernethy, Pfizer (C) Stock Ownership: None Honoraria: S. Yousuf Zafar, Genentech Research Funding: Amy P. Abernethy, Pfizer Expert Testimony: None Other Remuneration: None
Presented at the 44th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, May 30-June 3, 2008, and the 2008 Duke Comprehensive Cancer Center Annual Meeting, Durham, NC, March 10, 2008. Funded through an Outcomes Research service agreement with Pfizer. Pfizer has no access to individual data or authority to prohibit publication results.