Fax: (626) 301-8865
Elderly patients with colon cancer have unique tumor characteristics and poor survival
Article first published online: 25 SEP 2012
Copyright © 2012 American Cancer Society
Volume 119, Issue 4, pages 739–747, 15 February 2013
How to Cite
Patel, S. S., Nelson, R., Sanchez, J., Lee, W., Uyeno, L., Garcia-Aguilar, J., Hurria, A. and Kim, J. (2013), Elderly patients with colon cancer have unique tumor characteristics and poor survival. Cancer, 119: 739–747. doi: 10.1002/cncr.27753
- Issue published online: 4 FEB 2013
- Article first published online: 25 SEP 2012
- Manuscript Accepted: 18 JUN 2012
- Manuscript Revised: 18 MAY 2012
- Manuscript Received: 3 APR 2012
- colon cancer;
- Los Angeles County
The incidence of colon cancer increases with age, and colon cancer predominantly affects individuals >65 years old. However, there are limited data regarding clinical and pathologic factors, treatment characteristics, and survival of older patients with colon cancer. The objective of this study was to determine the effects of increasing age on colon cancer.
Patients diagnosed with colon cancer between 1988 and 2006 were identified through the Los Angeles County Cancer Surveillance Program, in Southern California. Patients were stratified into 4 age groups: 18-49, 50-64, 65-79, and ≥80 years. Clinical and pathologic characteristics and disease-specific and overall survival were compared between patients from different age groups.
A total of 32,819 patients were assessed. Patients aged 18 to 49 and 65 to 79 years represented the smallest and largest groups, respectively. A near equal number of males and females were diagnosed with colon cancer in the 3 youngest age groups, whereas patients who were ≥80 years old were more commonly white and female. Tumor location was different between groups, and the frequency of larger tumors (>5 cm) was greatest in youngest patients (18-49 years). The oldest patients (≥80 years) were administered chemotherapy at the lowest frequency, and disease-specific and overall survival rates decreased with increasing age.
This investigation demonstrates that older age is associated with alterations in clinical and pathologic characteristics and decreased survival. This suggests that the phenotype of colon cancer and the efficacy of colon cancer therapies may be dependent on the age of patients. Cancer 2013. © 2012 American Cancer Society.
Colon cancer is the most common malignancy of the gastrointestinal tract and the second leading cause of death from cancer in the United States.1 It is predominantly associated with older age, with an increasing incidence in patients over the age of 50 years.2 As the elderly population continues to expand in the United States, with a projected estimate of 54,632,000 persons older than 65 years by 2020,3 the prevalence of colon cancer is expected to concomitantly rise. However, the impact of older age on tumor biology and outcomes remains somewhat unclear. Previous studies have suggested that older age may be associated with an increased propensity to epigenetic alterations as well as increased susceptibility to carcinogenesis.4, 5 Furthermore, when cancers develop in the elderly, they are associated with disproportionately high mortality rates compared with malignancies in younger patients.6 Given the rising burden of colon cancer in the elderly population, an understanding of the effects of increasing age on clinical and pathologic characteristics and survival of patients with this disease is thus becoming increasingly relevant.
Previous studies suggest that survival for older patients with colon cancer may be worse compared with that in younger patients.7-10 These reports suggest that the poorer outcomes for older patients (≥80 years old) may be secondary to higher rates of morbidity, mortality, and hospital readmission following surgery.11 An underlying etiology attributed to such outcomes has been an increasing rate of comorbidities with older age.7, 9, 12 However, other studies have indicated similar survival rates between older and younger patients undergoing curative surgery for colon cancer.13 In fact, when survival has been corrected for patient age, the prognosis of older and younger patients with colon cancer have appeared to be similar.14 Taken together, the current data are conflicting regarding the association of increasing age and survival of patients with colon cancer.
When considering colon cancer phenotype, tumors in older patients may be different than tumors in younger patients. Specifically, older patients (≥65 years old) appear to have an increased frequency of right-sided tumors.15 However, this proximal shift in tumor location with age is not consistent for all patients. Wu et al demonstrated that whereas older white and black patients showed an increasing trend toward proximal tumors, older Asian and Pacific Islander patients showed a similar incidence of proximal and distal tumors.16 Indeed, this issue is relevant because location of disease has important prognostic implications. Approximately 20% of right-sided colon cancers are associated with microsatellite instability (MSI) due to defects in DNA mismatch repair.17 Although MSI tumors are generally associated with an improved prognosis,17 the impact on the outcomes of older patients with colon cancer remains unclear.
A disparity may also exist in the administration of systemic therapies to patients with colon cancer based on their age. Older patients may not receive the same systemic therapies that are offered to younger patients.18 In fact, patients >70 years of age are less likely to receive chemotherapy than their younger counterparts.7, 19 This appears to occur despite the established benefits of systemic chemotherapy in advanced disease,20-23 and older patients who receive chemotherapy derive benefits similar to those observed in younger patients, without an apparent increase in toxicity.20, 21, 23, 24 As such, the administration of systemic therapies and outcomes across colon cancer age groups requires further examination.
As the disease burden in older patients with colon cancer continues to rise, a more comprehensive understanding of the effects of increasing age on clinical and pathologic characteristics and patient survival is warranted. Therefore, we examined and compared patient characteristics, pathologic features, treatment patterns, and survival of patients with colon cancer in the Los Angeles County (LAC), California, metropolitan area.
MATERIALS AND METHODS
Los Angeles County Cancer Surveillance Program
We used the LAC Cancer Surveillance Program (CSP), a population-based cancer registry, to obtain data for this study. Data on patient demographics, tumor characteristics, staging, treatment modalities (chemotherapy and surgery), and follow-up are included in the registry and the data are continuously monitored for quality assurance purposes.25 Institutional review board approval from the City of Hope and the State of California was obtained.
Patients diagnosed with adenocarcinoma of the colon between 1988 and 2006 were identified, using the CSP database. In total, 32,819 patients with colon adenocarcinoma were evaluated. Colon cancer location, histology, and staging were coded based on the International Classification of Disease for Oncology (ICD-O); the second edition was used for cases diagnosed between 1988 and 2000 and the third edition was used for cases diagnosed between 2001 and 2006. The specific histology of interest was colon adenocarcinoma, which included histology codes 8140-8147, 8210, 8211, 8220, 8221, 8260-8263, 8480, 8481, and 8570-8576. Patients with overlapping tumors involving the colon and rectum, patients who received radiation therapy or intraoperative chemotherapy, and patients with rectal cancer were excluded from this study.
Patients were stratified by age into 4 age groups: 18-49, 50-64, 65-79, and ≥80 years. The CSP database was then queried for information pertaining to patient demographics, including sex and race.26 Pathologic tumor characteristics including tumor size, location, grade, and extent of disease were collected. Based on extent of disease, tumors were staged using the American Joint Committee on Cancer (AJCC) sixth edition guidelines, such that localized disease was categorized as AJCC stages I and II, regional disease as AJCC stage III, and distant disease as AJCC stage IV. Treatment data regarding surgery and chemotherapy were also tabulated. Five-year disease-specific survival (DSS) and overall survival (OS) were calculated for each patient. Patients were censored at the date of last follow-up (April 13, 2008) or the date they were last known to be alive.
Bivariate analysis was performed for patient, tumor, and treatment variables stratified by age group (18-49, 50-64, 65-79, and ≥80 years). Chi-square test for significance was performed for all categorical variables to examine age-related differences in clinical, pathologic, and treatment-related characteristics. Kaplan-Meier survival curves were constructed, and comparisons between age groups were made using the log-rank test. For survival analysis, age groups were further stratified by stage of disease, receipt of surgery, and receipt of chemotherapy. Multivariate Cox regression analysis was performed to determine the association of patient, tumor, and treatment-related variables with age. Statistical analysis was performed using SAS (version 9.1; SAS Institute, Cary, NC). P values were 2-sided and values <.05 were considered statistically significant.
Patient Demographics Change With Increasing Age
We stratified patients into 4 groups and assessed patient demographics and tumor characteristics (Table 1). Patients who were 18 to 49 years old (n = 2358; 7%) and 65 to 79 years old (n = 14,651; 45%) represented the smallest and largest groups, respectively. A near-equal number of males and females were diagnosed with colon cancer in the 3 youngest groups, whereas patients who were ≥80 years old were more commonly female (60%). The majority of patients (41%-72%) across all age groups were white.
|Characteristic||18-49 n = 2358 (7%)||50-64 n = 7167 (22%)||65-79 n = 14,651 (45%)||≥80 n = 8643 (26%)|
|Mean age (y) ± SD||42.3 ± 6||58 ± 4.2||72.4 ± 4.2||85 ± 4.1|
|Male||1225 (52%)||3796 (53%)||7586 (51.8%)||3429 (39.7%)|
|Female||1133 (48%)||3371 (47%)||7065 (48.2%)||5214 (60.3%)|
|White||975 (41.3%)||3570 (49.8%)||9094 (62.1%)||6203 (71.8%)|
|Black||363 (15.4%)||1268 (17.7%)||1857 (12.7%)||784 (9.1%)|
|Hispanic||664 (28.2%)||1400 (19.5%)||2020 (13.8%)||857 (9.9%)|
|Asian||356 (15.1%)||929 (13%)||1680 (11.5%)||799 (9.2%)|
|I||408 (17.3%)||1671 (23.3%)||3730 (25.5%)||1952 (22.6%)|
|II||674 (28.6%)||2012 (28.1%)||4703 (32.1%)||3258 (37.7%)|
|III||711 (30.2%)||1990 (27.8%)||3840 (26.2%)||2094 (24.2%)|
|IV||565 (24%)||1494 (20.8%)||2378 (16.2%)||1339 (15.5%)|
|I||182 (8.2%)||607 (9%)||1214 (8.8%)||669 (8.1%)|
|II||1453 (65.6%)||4759 (70.9%)||9503 (69.1%)||5558 (67.6%)|
|III||564 (25.5%)||1302 (19.4%)||2943 (21.4%)||1924 (23.4%)|
|IV||16 (0.7%)||42 (0.6%)||101 (0.7%)||69 (0.8%)|
|Ascending||932 (39.5%)||2903 (40.5%)||7061 (48.2%)||4750 (55%)|
|Transverse||239 (10.1%)||631 (8.8%)||1418 (9.7%)||941 (10.9%)|
|Descending/sigmoid||1187 (50.3%)||3633 (50.7%)||6172 (42.1%)||2952 (34.2%)|
|0-5 cm||1108 (55.2%)||3853 (63.9%)||8398 (66.4%)||5013 (65.5%)|
|>5 cm||901 (44.8%)||2173 (36.1%)||4241 (33.6%)||2638 (34.5%)|
Tumor Characteristics Change With Increasing Age
Tumor location was different among the age groups. The 2 youngest age groups (18-49 years and 50-64 years) presented most commonly with descending tumors (50% and 51%, respectively), whereas the majority of the oldest patients (≥80) presented with ascending tumors (55%) (Fig. 1A). Interestingly, the frequency of larger tumors (>5 cm) was greatest in the youngest patients (aged 18-49 years; 45%). Across all age groups, most patients presented with local and regional disease (Fig. 1B). However, a decreasing rate of metastatic disease was observed with increasing patient age (24%, 18-49 age group vs 16%, ≥80 age group; P < .001).
Treatment Patterns Change With Increasing Age
Treatment patterns were analyzed for the 4 age groups (Table 2). Across all groups, the majority of patients underwent curative surgery (91%-92%). The youngest patients (age 18-49) had the highest frequency (52%) of ≥12 lymph nodes examined, but they also had the highest rate of lymph node–positive disease (45%). There was also a clear decrease in chemotherapy administration with increasing age; half of the patients in the youngest age group (age 18-49; 50%) had chemotherapy, whereas only 8% of the oldest patients (≥80) received chemotherapy (Fig. 1C).
|Treatment Characteristic||18-49 n = 2358 (7%)||50-64 n = 7167 (22%)||65-79 n = 14,651 (45%)||≥80 n = 8643 (26%)|
|None||149 (6.3%)||409 (5.7%)||684 (4.7%)||531 (6.1%)|
|Curative||2140 (90.8%)||6490 (90.6%)||13402 (91.5%)||7820 (90.5%)|
|Noncurative||69 (2.9%)||268 (3.7%)||565 (3.9%)||292 (3.4%)|
|Number of lymph nodes|
|0||276 (12.1%)||890 (12.9%)||1685 (12%)||1078 (12.9%)|
|1-11||827 (36.4%)||3163 (45.9%)||7302 (52.1%)||4344 (52.1%)|
|12+||1169 (51.5%)||2832 (41.1%)||5040 (35.9%)||2912 (34.9%)|
|Lymph node status|
|N0||1186 (52.7%)||3960 (57.6%)||8922 (63%)||5493 (66.1%)|
|N+||1064 (47.3%)||2916 (42.4%)||5236 (37%)||2815 (33.9%)|
|No||1133 (50%)||4187 (60.9%)||10519 (75%)||7742 (92%)|
|Yes||1134 (50%)||2691 (39.1%)||3498 (25%)||677 (8%)|
|No||2298 (97.5%)||6951 (97%)||13994 (95.5%)||7837 (90.7%)|
|Yes||60 (2.5%)||216 (3%)||657 (4.5%)||806 (9.3%)|
|History of cancer|
|No||2186 (92.7%)||6279 (87.6%)||11778 (80.4%)||6459 (74.7%)|
|Yes||172 (7.3%)||888 (12.4%)||2873 (19.6%)||2184 (25.3%)|
To assess survival for the 4 age groups, we matched patients according to stage and treatment (surgery or chemotherapy). First, we examined patients in the different age groups who underwent curative intent surgery for early-stage disease (ie, AJCC stage I and II). With increasing age, 5-year DSS in patients with AJCC stage I disease decreased from 94% (18-49 years old) to 88% (≥80 years old) (Fig. 2A). Decreased OS (88% to 50%) was also observed with increasing age (Fig. 2A). Similarly, poorer DSS and OS were observed with increasing age in patients with AJCC stage II and III disease (Fig. 2B,C). Finally, we examined DSS and OS for patients with stage IV disease who received chemotherapy with or without any type of surgery. For this analysis, we also observed decreased 5-year DSS or OS across the age groups (Fig. 2D). Finally, there was a clear increase in 30-day mortality from 2.5% (18-49 years) to 9.3% (≥80 years) across all stages of disease with increasing age.
We performed univariate and multivariate analysis to identify predictors of OS. On univariate analysis, all of the clinical and pathologic factors, except sex, were associated with survival differences (Table 3). To determine whether age group was independently prognostic, we performed a multivariate Cox regression analysis that identified age group as an independent predictor of OS. Specifically, the oldest age group (≥80) predicted worse OS (hazard ratio = 3.38, 95% confidence interval 3.14-3.64; P < .001). Additional factors that were prognostic for OS on multivariate analysis included sex, race, stage, tumor location, tumor size, and treatment modality.
|Factors||N||Univariate Analysis||Stepwise Multivariate Analysis|
|HR (95% CI)||P||HR (95% CI)||P|
|Age group, y||<.001||<.001|
|50-64||7,167||1.20 (1.00-1.20)||.014||1.21 (1.12-1.30)||<.001|
|65-79||14,651||1.60 (1.60-1.80)||<.001||2.03 (1.89-2.18)||<.001|
|80+||8,642||3.00 (2.80-3.20)||<.001||4.04 (3.76-4.34)||<.001|
|Female||16,783||1.00 (1.00-1.00)||.131||0.89 (0.86-0.91)||<.001|
|Black||4,272||1.20 (1.00-1.20)||<.001||1.20 (1.15-1.25)||<.001|
|Hispanic||4,941||0.80 (0.80-0.80)||<.001||0.95 (0.91-1.00)||.042|
|Asian||3,764||0.80 (0.80-0.80)||<.001||0.84 (0.80-0.88)||<.001|
|II||10,647||1.20 (1.20-1.40)||<.001||1.29 (1.23-1.35)||<.001|
|III||8,635||1.80 (1.60-1.80)||<.001||1.39 (1.28-1.51)||<.001|
|IV||5,776||7.20 (6.80-7.40)||<.001||6.14 (5.70-6.61)||<.001|
|Ascending colon||15,646||1.20 (1.20-1.20)||<.001||1.05 (1.02-1.09)||.002|
|Transverse colon||3,229||1.20 (1.00-1.20)||<.001||1.09 (1.04-1.15)||.001|
|>5 cm||9,953||1.20 (1.20-1.20)||<.001||1.06 (1.03-1.10)||.001|
|Unknown||4,494||1.20 (1.20-1.20)||<.001||1.20 (1.14-1.26)||<.001|
|N+||12,031||1.80 (1.80-2.00)||<.001||1.36 (1.27-1.47)||<.001|
|NX||1,227||8.40 (8.00-9.00)||<.001||2.09 (1.90-2.28)||<.001|
As the population in the United States continues to grow and the number of persons greater than age 65 years expands, there will be a growing population of older individuals with colon cancer.9 However, much remains unknown about the specific changes in colon cancer presentation, treatment patterns, and outcomes associated with increasing patient age. In this study, we observed that increasing age was generally associated with worse outcomes. Tumor characteristics and treatment biases that vary with patient age may have affected these outcomes to some degree, but they do not fully account for the clear differences in survival for the patients of the 4 age groups.
First, our study demonstrated a marked shift in tumor location from left side to right side with increasing age. This shift in tumor location has been inconsistently observed, but the explanation for the shift with increasing age remains unclear.10, 16, 27-30 Furthermore, the implication of tumor location on prognosis is conflicting. In our investigation, we observed that patients with left-sided tumors had improved survival over patients with right-sided disease. Previously published data have suggested that right-sided cancers tend more often to be poorly differentiated and locally advanced than left-sided tumors, and are more frequently associated with worse survival outcomes.31
It has also been shown that right-sided tumors are more frequently associated with MSI-H (high levels of MSI), and these tumors are associated with improved prognosis and decreased metastatic potential compared with microsatellite-stable tumors.17, 32, 33 However, some studies have suggested that MSI-H tumors may derive little benefit from 5-fluorouracil–based therapies, whereas other studies have shown the opposite.17, 32-34 In this investigation, we were unable to evaluate MSI secondary to the inherent limitations of the LAC tumor registry. Nevertheless, the older patients in our study cohort, despite having a greater percentage of right-sided disease, frequently had more favorable tumor profiles, including earlier disease stage, small tumors (<5 cm), and lymph node–negative disease. This clinical phenotype may also be secondary to the effect of use of screening colonoscopy in elderly patients, which may facilitate detection of earlier stage disease.35, 36 Nonetheless, when we compared outcomes between older patients (<80 years) with left-sided versus right-sided disease, we observed no difference in OS (data not shown).
The mainstay of curative colon cancer treatment remains surgical resection.11, 37 Consistent with previously published literature, we observed that curative surgical resection was performed at a near equal rate among the 4 age groups.19 However, we observed that DSS and OS of older patients were worse than that of younger patients. In fact, our investigation showed a progressive decrease in DSS and OS across the age groups, and this trend was observed across all stages of disease. In contrast, Mulcahy et al demonstrated 5-year relative survivals of 59% in patients <70 years and 68% in patients ≥70 years old who underwent surgical resection, and concluded that age does not play a significant role in the long-term survival of patients with colon cancer.13 Widdison et al and Devon et al were unable to demonstrate significant differences in DSS with increasing patient age, whereas McMillan et al and Latkauskas et al both noted decreased survival in older patients with early-stage colorectal cancer.38-41
Our investigation revealed significant differences in both OS (reduction from 66% to 34%) and DSS (reduction from 73% to 64%), with increasing patient age among patients with all stages of disease. Even in patients with stage I disease, large differences in survival were noted with increasing age; patients >80 years old had DSS and OS that were 6% and 38% lower, respectively, than patients <50 years old. These differences persisted with advanced disease, and were especially pronounced in patients with stage III disease, where the DSS and OS survivals of the cohort of those aged ≥80 were 22% and 41%, respectively, less than the <50-year-old cohort. There are potential etiologies for the worse outcomes in older patients. Kunitake et al observed that patients ≥80 years old had higher readmission rates, and higher in-hospital and 1-year mortality rates.11 Indeed, we observed that 30-day mortality was significantly higher in the oldest patients, which may contribute to the overall poorer survival in older patient groups. The shorter DSS in the older patients, however, suggests that the worse outcomes are not simply a result of patient comorbidities or postoperative complications, but rather that the cancer is recurring sooner, possibly secondary to differences in tumor characteristics or treatment patterns. Finally, given the lower number of lymph nodes examined and decreased rate of lymph node positivity in the older patients of this study, the possibility of confounding via stage migration exists. Moreover, there are prior reports suggesting that the number of examined lymph nodes correlates with lymph node status and survival.42
The mainstay of treatment for metastatic colon cancer is systemic chemotherapy. Although the majority of patients in our cohort did not receive chemotherapy, the administration of chemotherapy, nevertheless, decreased with increasing age. This may partially be attributed to patient preference to defer recommended therapies. Landrum et al demonstrated that of those patients who did not receive chemotherapy for stage III colon cancer, the patient declined chemotherapy in 58% of cases.43 The magnitude of decreased rate of chemotherapy receipt with increasing patient age was far greater than any of the other changes in clinical or pathologic factors. Perhaps the disparity in chemotherapy treatment patterns may partially be attributed to the insufficient investigation of chemotherapeutic regimens in older patients.18 Indeed, older patients are frequently excluded from clinical trials, although the majority of cancer diagnoses occur in older individuals.18, 19, 44 Multiple reasons have been cited to account for this underrepresentation, including the presence of debilitating comorbidities, history of prior malignancy, lack of education regarding trial participation, and skepticism among caregivers regarding utility of cancer therapies and ability of older patients to tolerate them.45 Kahn et al provided data to substantiate these concerns, showing that even when older patients received adjuvant chemotherapy, they frequently had a shorter duration of treatment than did younger patients.20 However, despite these impediments, several studies have suggested that elderly patients can derive the benefits of adjuvant therapy observed in younger patients, without a significant increase in toxicity.21-23 In our study, we observed worse outcomes in older patients who had metastatic disease. Even when chemotherapy was added to the treatment regimen in patients with stage III and IV disease, there was a clear difference in outcomes among the 4 age groups, thus suggesting that older patients may not derive the same degree of benefit from systemic therapies than younger patients. However, given the inherent limitations of CSP, we were unable to assess comorbidities and performance status, or to determine the exact agents, dosages, and completion rates of chemotherapy for patients in this study.
In summary, our study demonstrates that colon cancer may be a different disease in older patients. It is a disease with a different phenotype and pattern of management, dependent on the age of the patient. Even when patients were matched by extent of disease and receipt of treatment, differences in outcome persist. Age influences outcomes. As the elderly population continues to increase and the prevalence of colon cancer continues to rise, our investigation provides a better understanding of age-related disparities in colon cancer presentation, care, and outcomes.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURE
The authors made no disclosure.
- 2Colorectal cancer incidence in the United States, 1999-2004: an updated analysis of data from the National Program of Cancer Registries and the Surveillance, Epidemiology, and End Results Program. Cancer. 2009; 115: 1967-1976., , , , .
- 3United States Census Bureau. U.S. Interim Projections by Age, Sex, Race, and Hispanic Origin: 2000-2050. http://www.census.gov/population/www/projections/usinterimproj/ Accessed July 6, 2011.
- 4Cancer in the elderly. Trans Am Clin Climatol Assoc. 2006; 117: 147-155., , , et al.
- 16Comparative Analysis of Incidence Rates Subcommittee, Data Evaluation and Publication Committee, North American Association of Central Cancer Registries. Subsite-specific colorectal cancer incidence rates and stage distributions among Asians and Pacific Islanders in the United States, 1995 to 1999. Cancer Epidemiol Biomarkers Prev. 2004; 13: 1215-1222., , , et al;
- 28Variations in demography and prognosis by colon cancer location. Anticancer Res. 2011; 31: 2347-2350., .