Palliative chemotherapy for patients 70 years of age and older with metastatic colorectal cancer : a single-centre experience

Background Metastatic colorectal cancer (mcrc) commonly affects elderly people, an understudied subset of patients. We analyzed the survival impact of the first and subsequent lines of chemotherapy in eligible non-trial patients 70 years of age and older with mcrc treated between 2004 and 2012. Methods This single-centre retrospective analysis estimated overall survival (os) and progression-free survival (pfs) using the Kaplan–Meier method. Multivariate analysis was used to adjust for age, sex, Eastern Cooperative Oncology Group performance status, score on the Charlson comorbidity index, dependency in activities of daily living, and exposure to 1 or more chemotherapy doublets, capecitabine alone, or best supportive care (bsc). Results Of 109 patients identified, 29 elected bsc, and 80 received chemotherapy. In multivariate analysis, age was not associated with os [hazard ratio (hr): 0.99; 95% confidence interval (ci): 0.92 to 1.05], but a performance status of 2 or higher was associated with a decreased likelihood of survival (hr: 3.12; 95% ci: 1.87 to 5.76), and exposure to 1 or more doublets was associated with improved survival (hr: 0.33; 95% ci: 0.17 to 0.66). In univariate analysis, a trend toward improved os was observed for first-line doublet chemotherapy compared with capecitabine (hr: 0.66; 95% ci: 0.41 to 1.07), and pfs was superior (hr: 0.46; 95% ci: 0.26 to 0.84). Compared with exposure to 1 doublet, exposure to the 3 potential cytotoxic chemotherapies was not associated with improved os (hr: 0.77; 95% ci: 0.41 to 1.43). The incidence of neutropenia with first-line folfiri was 40%; the incidences of bevacizumab-related arterial and venous thrombosis were both 8%. Conclusions Exposure to 1 or more doublet chemotherapies for mcrc was associated with better outcomes in non-trial patients 70 years of age and older. Elderly patients treated with palliative chemotherapy and bevacizumab should be monitored carefully for arterial and venous thrombotic events.


BACKGROUND
Colorectal cancer is the fourth most prevalent cancer in North America.Nearly 50% of all cases are diagnosed in patients 70 years of age and older, and 20% of cases have metastasized by the time of the diagnosis 1 .Recent advances in the treatment of metastatic colorectal cancer (mcrc) have led to significant improvements in survival, increasing average survival to more than 24 months from 11-12 months 2,3 .
Cytotoxic 5-fluorouracil-capecitabine or doublets using oxaliplatin (folfox, xelox) or irinotecan (folfiri) in combination with bevacizumab, an inhibitor of vascular endothelial growth factor, are established as the standard of care in fit adults, but their benefits are difficult to extrapolate to unselected elderly patients for many reasons, including inadequate evidence to support the use of those regimens because of the exclusion of elderly patients from large clinical trials 4 .Heterogeneity of functional status in elderly patients, age-related alterations in drug metabolism, and a greater number of comorbidities can affect both the expected efficacy and the adverse effects of the regimens 5,6 .
Considerable efforts have recently been made to promote the enrolment of older patients in clinical trials.Pooled analyses that regroup the elderly patients from relevant trials [7][8][9][10] , as well as a limited number of focused trials 11,12 , have suggested that older patients can gain appreciable benefit from chemotherapy and targeted therapy, with acceptable safety.Compared with their younger counterparts, older patients eligible to receive chemotherapy are more likely to be initially offered capecitabine alone instead of a doublet regimen and are less likely to receive a doublet chemotherapy regimen [13][14][15] .Chemotherapy safety, tolerance, and efficacy are clinical concerns when older patients with mcrc who do not fit the stringent inclusion criteria of clinical trials are treated.Given that background, we conducted a retrospective study to analyze survival and toxicity data in patients 70 years of age and older treated for mcrc in our academic tertiary care centre.

METHODS
We retrospectively collected data for patients treated at the Centre hospitalier universitaire de Sherbrooke, Quebec, between January 2004 and December 2012.The inclusion criteria were histologically proven crc with radiologic metastasis, an age of 70 years or more at the time of diagnosis, oncology assessment at our centre, and candidacy for chemotherapy.To find charts for patients fulfilling the study criteria, the archive service at our institution used definitive mcrc-related key words to search admission summaries, radiology and pathology reports, and the oncology pharmacy's drug delivery registry.Upon approval by our institutional ethics committee, the identified charts were manually reviewed, and those not meeting the study criteria were excluded (Figure 1).Patients who received no chemotherapy despite being deemed eligible by an oncologist were kept in the study.
The patients were divided into two groups: best supportive care (bsc) and chemotherapy.The latter group was in turn subdivided in two age groups: 70-74 years and 75 years and older.The baseline characteristics extracted from the charts included primary tumour site, sites of metastasis, prior metastasectomy, prior adjuvant therapy, Eastern Cooperative Oncology Group (ecog) performance status, basic (adl) and instrumental (iadl) activities of daily living autonomy scales, and comorbidities.Autonomy was dichotomized into "autonomous" or "dependent," indicating that the patient was dependent with respect to 1 or more adls or iadls.Comorbidities were scored according to the Charlson comorbidity index (cci) 16 .Additionally, prognostic biochemical determinants such as serum lactate dehydrogenase, carcinoembryonic antigen, and platelet count were collected.Chemotherapy toxicities were retrospectively graded using the U.S. National Cancer Institute's Common Toxicity Criteria, version 2.0.The indexed treatments were capecitabine alone, doublets (folfox, xelox, folfiri), and bevacizumab.

Statistical Methods
The statistical analysis was performed using IBM SPSS Statistics (version 20.0: IBM, Armonk, NY, U.S.A.).Kaplan-Meier curves and log-rank tests were used to estimate median overall survival (os) and progression-free survival (pfs) with 95% confidence intervals (cis).Cox regression modelling was used to determine predictors of os and pfs; models included age, sex, cci score, ecog performance status, adl, doublet chemotherapy, capecitabine monotherapy, and bsc.Statistical significance was assumed at p ≤ 0.05.

RESULTS
Of the 109 patients who met the study criteria, 80 (73%) received chemotherapy, and 29 (27%) opted for bsc despite being deemed eligible for chemotherapy by an oncologist.Table i reports their baseline characteristics.In the chemotherapy group, 34 patients were 70-74 years of age, and 46 were 75 years of age or older.The age ranges were 70-87 years in the chemotherapy group and 70-94 years in the bsc group.Both groups contained 14 patients 80 years of age and older.The patients who selected bsc had higher ecog scores (p = 0.02).
Table ii presents the multivariate analysis of the 109 patients, which indicates that an ecog performance status of 2 or greater was associated with an increased risk of death [hazard ratio (hr): 3.12; 95% ci: 1.69 to 5.76].Other baseline characteristics such as age, sex, and cci score were not associated with an increased risk of death.Unfortunately, too many adl and iadl data were missing to permit the functional status of the patients to be studied.A sensitivity analysis by multiple imputations was conducted, and the missing adls were placed in the multivariate model.The addition of those adls to the model did not change the significance of the other factors.

Advanced Age and Chemotherapy
Median os was 19.7 months (95% ci: 12.6 to 26.7 months) for the 70-74 age group and 17.5 months (95% ci: 11.7 to 23.4 months) for the 75 and older group (log-rank p = 0.822).Univariate analysis revealed no significant difference in os between the age groups receiving chemotherapy [hr: 1.055; 95% ci: 0.66 to 1.67; Figure 2(A)].No significant differences in the os and pfs rates were found between the age groups for any of the first-line chemotherapy subgroups, including capecitabine and all doublets.In the multivariate analysis, age was not a factor influencing the risk of death (hr: 0.97; % ci: 0.91 to 1.04; p = 0.43).

First-Line Chemotherapy
First-line oxaliplatin and irinotecan doublets resulted in similar os (p = 0.804) and pfs (p = 0.450) rates for all of the elderly patients who received chemotherapy and were therefore pooled.Figure 3 presents the Kaplan-Meier curves for the patients who received one of the doublets, capecitabine, or bsc as initial management.Compared with patients who received capecitabine, those who received a doublet as first-line chemotherapy showed a trend toward increased os (hr: 0.66; 95% ci: 0.41 to 1.07).Median survival rates in those groups were 20.7 months (95% ci: 14.9 to 26.6 months) and 14.9 months (95% ci: 12.3 to 17.4 months) respectively (log-rank p = 0.073).Capecitabine and doublet regimens were both superior to bsc, which was associated with a median os of 7.6 months [95% ci: 3.9 to 11.2 months; Figure 3(A)].First-line pfs was longer with a doublet than with capecitabine [hr: 0.46; 95% ci: 0.26 to 0.84; p = 0.011; Figure 3(B)].Compared with the 70-74 age group, the 75 and older age group contained a greater proportion of patients who received first-line capecitabine rather than a doublet (57% vs. 35%, p = 0.05).Nevertheless, pfs after any first-line chemotherapy was similar in the two age groups [hr: 1.33; 95% ci: 0.77 to 2.31; Figure 2(B)].First-line bevacizumab was used in the same proportion in both groups.At the time of the study, no patients were given therapy targeting the epidermal growth factor receptor.

Treatment Intensity
Patients 75 years of age and older were less likely than those 70-74 years of age to be exposed to all three chemotherapy backbones (5-fluorouracil-capecitabine, irinotecan, oxaliplatin: 22% versus 53%, p = 0.02), and a smaller proportion of the older group received second-and third-line chemotherapy (Table i).Exposure to two doublets compared with only one was not associated with a better os (hr: 0.77; 95% ci: 0.41 to 1.43; Figure 4).In the multivariate analysis, receipt of capecitabine as the only line of treatment was not statistically associated with longer survival (hr: 0.57; 95% ci: 0.28 to 1.12; p = 0.13); exposure to at least one doublet was associated with longer survival (hr: 0.36; 95% ci: 0.17 to 0.66).The 46 patients who received bevacizumab experienced a median survival of 24 months (95% ci: 16.4 to 31.6 months).

Toxicity
Table iii reports the grade 3 and 4 toxicities related to chemotherapy.Oxaliplatin doublets were associated with a significant rate of peripheral neuropathy (29%), and dose reductions were instituted in 71% of patients during their course of treatment.The irinotecan doublet led to a 40% rate of neutropenia, with 7% having febrile neutropenia, and 11% being admitted with an underlying infection.Of the patients on capecitabine, 13% developed hand-foot syndrome, and 11% had an infection requiring admission.Adverse effects of bevacizumab included 8% rates of both arterial (ate) and venous (vte) thrombotic events in our cohort of elderly patients.One patient (2.2%) experienced a visceral perforation while on bevacizumab (Table iv).Proteinuria, hypertension, and poor wound healing were inconsistently tracked in our retrospective review and were therefore not analyzed.

DISCUSSION
Treating elderly patients with mcrc can sometimes be challenging, given considerable patient heterogeneity, poor representation of this group in clinical trials, and the various available treatment strategies.Additionally, elderly patients not eligible for clinical trials might be more susceptible to drug toxicity or might receive less-intensive chemotherapies.The present retrospective study reports outcomes in elderly real-life non-trial patients (≥70 years of age) with mcrc treated with chemotherapy, and it explores the survival impact of the first and subsequent lines of treatment.
Compared with their counterparts who chose bsc, elderly patients who received chemotherapy experienced prolonged survival.In accord with the findings of the fo-cus2 study of elderly and frail patients, the results of our multivariate analysis show that poor performance status was associated with worse outcomes, but that age and cci score were not 11 .In fact, number of comorbidities might be a better predictor of chemotherapy toxicity than of e352  mortality 17,18 .Too many data on adl and iadl were missing to evaluate functional status in our patients.Selection of the best first-line chemotherapy in elderly patients with mcrc has been poorly studied.In the focus2 trial, the addition of oxaliplatin to capecitabine failed to significantly improve pfs 11 .In comparison, in our study, doublets were associated with longer pfs and a trend toward better os that persisted beyond 2, 3, and 4 years (Figure 3).Given the lack of randomization, patients initially treated with capecitabine were likely to be more vulnerable or frail than the patients treated first with a doublet.Thus, the differences in functional and performance status could potentially account for the observed difference.Nevertheless, data from the U.S. Surveillance, Epidemiology, and End Results Medicare database suggest that choice of first-line of treatment could play an important role in older adults and could affect short-term and 5-year survival rates 19 .The retrospective nature of the data and the small number of patients limited interpretation of the results.However, it could be hypothesized that, although older and clinically fit patients clearly benefit from systemic therapy 14,21 , they might benefit not all from the same intensity and amount of chemotherapy.A prospective trial in elderly patients receiving comprehensive geriatric assessments would be necessary to make a more accurate determination.Among other factors, frailty and nutrition status could be particularly important in elderly patients with mcrc 17 .

Characteristic
Grade 3 and greater hematologic adverse effects of the 3 active standard chemotherapies were limited, but the irinotecan doublet was associated with 40% neutropenia and 7% febrile neutropenia, despite one third of the patients receiving dose-reduced chemotherapy.The rates of diarrhea, neuropathy, and hand-foot syndrome might have been underreported given the retrospective nature of the study.
Bevacizumab was associated with an 8% rate of ate and an additional 8% rate of vte.Although those rates are high, similar rates of thrombotic events have been reported in older adults 10,22 .In a pooled analysis, Cassidy et al. 10 reported a 6.7% incidence of ate in adults 70 years of age and older with mcrc treated with bevacizumab.Likewise, those authors reported an incidence of vte of nearly 13%.Patients in the control groups who did not receive bevacizumab had incidences of ate and vte of 3% and 10% respectively.In the analysis, the proportion of bevacizumab-related ate increased with age.Other studies have proposed that bevacizumab-related thrombotic events are more common in colon cancer and in people with vascular comorbidities 23,24 , which emphasizes the need to carefully select elderly patients based on their comorbidities before adding bevacizumab to their chemotherapy.In our study, 29 of 46 patients who received bevacizumab had a cci score of 1 or higher, and it is therefore likely that some had one or more cardiovascular risk factors beyond advanced age.Of the 46 patients, one experienced a visceral perforation, for a rate comparable to that in the brite observational cohort study 14 .
Neutropenia associated with first-line folfiri was non-negligible, at an incidence of 40%, with a 7% rate of febrile neutropenia.Similar rates were reported in the bicc-c trial 25 , but the incidence of neutropenia remained high, given that up to 32% of the patients in that trial received initial dose-adjusted folfiri.Souglakos et al. 26 reported a 20% incidence of grade 3 or greater neutropenia in patients 70 and more years of age treated with first-line folfiri.
The limitations of our study include its retrospective nature, and accordingly, the results should be carefully interpreted.The small number of patients, combined with their great heterogeneity, hindered our ability to adjust for all prognostic and predictive determinants in our analyses.Finally, an evaluation of the functional status of our patients, an important prognostic factor in elderly patients, could not be adequately performed given the amount of missing data.

CONCLUSIONS
Despite its limitations, our study emphasizes that elderly (70 and more years of age) non-trial patients with mcrc seem to derive benefit from a first-line doublet, with improvement in pfs compared w ith that achieved on capecitabine alone.Patients 75 and more years of age fared as well as those 70-74 years of age despite less-intense chemotherapy treatments.The high incidence of folfirirelated neutropenia and bevacizumab-related thrombotic events emphasizes the need to use adjusted-dose irinotecan in this population and to carefully select patients suitable for treatment with bevacizumab.Future prospective studies, with comprehensive systematic and geriatric assessment, are needed to evaluate the best treatment strategy in elderly patients and to better individualized the intensity and the type of chemotherapy delivered.

FIGURE 1
FIGURE 1 Flow diagram of patient selection for the study.
p ≤ 0.05) compared with (B).b Significant difference (p ≤ 0.05) compared with (D).c Patient groups could not be compared because of missing data.d Statistical tests not applicable.CCI = Charlson comorbidity index; ECOG = Eastern Cooperative Oncology Group; NA = not applicable.

aA
sensitivity analysis by multiple imputations was performed, and the factor of dependency for activities of daily living was added.b For activities of daily living, 35% of the data were missing.HR = hazard ratio; CI = confidence interval; CCI = Charlson comorbidity index; ECOG = Eastern Cooperative Oncology Group; ADL = activities of daily living.

FIGURE 2
FIGURE 2Kaplan-Meier estimates for probability of (A) survival and (B) progression-free survival, by age group, for elderly patients who received chemotherapy.HR = hazard ratio; CI = confidence interval.

TABLE II
Results of multivariate analysis a