Abstract
Background
Previous studies have shown that summary measures of comorbid conditions are associated with decreased overall survival in breast cancer patients. However, less is known about associations between specific comorbid conditions on the survival of breast cancer patients.
Methods
The Surveillance, Epidemiology, and End Results–Medicare database was used to identify primary breast cancers diagnosed from 1992 to 2000 among women aged 66 years or older. Inpatient, outpatient, and physician visits within the Medicare system were searched to determine the presence of 13 comorbid conditions present at the time of diagnosis. Overall survival was estimated using age-specific Kaplan–Meier curves, and mortality was estimated using Cox proportional hazards models adjusted for age, race and/or ethnicity, tumor stage, cancer prognostic markers, and treatment. All statistical tests were two-sided.
Results
The study population included 64 034 patients with breast cancer diagnosed at a median age of 75 years. None of the selected comorbid conditions were identified in 37 306 (58%) of the 64 034 patients in the study population. Each of the 13 comorbid conditions examined was associated with decreased overall survival and increased mortality (from prior myocardial infarction, adjusted hazard ratio [HR] of death = 1.11, 95% CI = 1.03 to 1.19, P = .006; to liver disease, adjusted HR of death = 2.32, 95% CI = 1.97 to 2.73, P < .001). When patients of age 66–74 years were stratified by stage and individual comorbidity status, patients with each comorbid condition and a stage I tumor had similar or poorer overall survival compared with patients who had no comorbid conditions and stage II tumors.
Conclusions
In a US population of older breast cancer patients, 13 individual comorbid conditions were associated with decreased overall survival and increased mortality.
CONTEXT AND CAVEATS
Prior knowledge
Previous reports have combined comorbidities into a summary measure and found that they are associated with poor overall survival and increased all-cause mortality in breast cancer patients.
Study design
The individual associations between 13 comorbidities and overall survival and all-cause mortality were investigated among breast cancer patients aged 65 years or older using data from the Surveillance, Epidemiology, and End Results cancer registry that had been linked with Medicare data for each patient.
Contribution
The large population size allowed the investigators to study the association between rare conditions and overall survival in older breast cancer patients in the United States and to assess the impact of age and tumor stage. All 13 comorbidities were associated with decreased overall survival and increased all-cause mortality in the study population. The age at diagnosis modified these relationships. Patients with comorbidities diagnosed with early-stage breast cancer had survival outcomes similar to or worse than that of patients with no comorbiditites diagnosed with later-stage tumors.
Implication
Comorbidities should be considered when determining prognosis for older breast cancer patients.
Limitations
The Medicare data, created for billing purposes, may not have included data about possible confounding factors such as smoking status, obesity, and physical activity level. Also, the severity of the comorbidity and the time since diagnosis were unknown and therefore not considered in the analyses.
From the Editors
One in eight women will be diagnosed with breast cancer during her lifetime (1). Breast cancer is more common among older women, yet cancer survivors diagnosed at age 65 years or older still represent an understudied population in cancer research, and older cancer patients and patients with comorbid chronic conditions are often systematically excluded from clinical trials (2–6).
Older persons diagnosed with cancer often have comorbid conditions in addition to their primary cancer diagnosis (7–9). These conditions can be highly prevalent, such as hypertension and heart-related conditions, or less common, such as diabetes. Comorbid conditions may be considered when deciding treatment regimens and otherwise increase the patient’s risk of death (10–12). Gaining a better understanding of the relationship between comorbid conditions and cancer survival and mortality may assist with the assessment of a patient’s prognosis and treatment management.
The Surveillance, Epidemiology, and End Results (SEER) is a national population-based cancer registry that has been linked with Medicare data at the patient level. SEER provides detailed information on patient demographics and tumor characteristics, whereas the Medicare database offers information on health-care utilization leading up to and following cancer diagnosis. This linkage provides a large unified data source that enables analysis of the relationships between clinical information that is not routinely included in cancer registries and includes cancer-related outcomes.
Previous studies of breast cancer patients have found that the presence of comorbid conditions is statistically significantly associated with overall survival and all-cause mortality (11–18). However, most published reports on this subject have combined comorbid conditions into one summary measure (a comorbidity index), and few studies have assessed the individual associations between specific comorbid conditions and prognosis (11,12,17). In addition, no previous studies have assessed whether the relationships between comorbid conditions and survival after breast cancer are modified by other prognostic factors, such as age at diagnosis and tumor characteristics. Determining the specific influence of individual comorbidities on mortality can enable a better assessment of prognosis for breast cancer patients with these conditions.
The current study aims to measure the individual associations between specific comorbidities of interest and overall survival and all-cause mortality among older women diagnosed with breast cancer. We hypothesize that some comorbidities will be more strongly associated with survival and mortality and also that these associations may be modified by age.
Methods
Patient Population
The SEER program is a population-based tumor registry of selected geographic areas within the United States designed to be representative of the country’s general population. Demographic data from SEER include sex, race, and dates of birth and death; tumor-specific data include diagnosis date, site, stage at diagnosis, and grade; and estrogen and progesterone receptor status of the tumor is reported for breast cancer patients. Surgical and radiation treatment within 4 months after cancer diagnosis are also included in the SEER database.
The following types of Medicare files from women aged 65 years or older were also included in this study: 1) the Medicare Provider Analysis and Review file containing inpatient hospital claims, 2) the Hospital Outpatient Standard Analytic file for outpatient claims, and 3) the Physician/Supplier file containing claims for physician visits. Medicare files were searched for diagnostic and procedure codes related to the comorbidities and treatments of interest. Codes for diagnoses and procedures are derived from the International Classification of Diseases, Ninth Revision, Clinical Modification; the Health Care Financing Administration Common Procedure Coding System; and Revenue Center codes (19,20). Medicare patients can also be enrolled in either a Fee-for-Service (FFS) or Health Maintenance Organization (HMO) plan. Because Medicare payments for HMO plans are not tied to specific clinical services, the claims files are not available for the 17% of Medicare recipients enrolled in such plans (21).
The National Cancer Institute, the SEER program, and the Centers for Medicare and Medicaid Services have collaborated to link the SEER cancer registries and Medicare enrollment and claims files to produce the de-identified SEER–Medicare dataset. These two databases are periodically linked based on an algorithm consisting of an individual social security number, first and last name, sex, and dates of birth and death (22).
Female primary malignant cancers of the breast diagnosed at age 66 years or older and reported to SEER from January 1, 1992, through December 31, 2000, were eligible for study inclusion. Women aged 65 years were not included to allow for a 1-year period after Medicare enrollment during which comorbidities could be recorded in claims files for the period before breast cancer diagnosis. Women were excluded if they: 1) lacked full coverage of both Medicare Part A (covers hospital, skilled-nursing facility, hospice, and some home health care) and Part B (covers physician and outpatient services), 2) were enrolled in a Medicare HMO, 3) had an unknown month of cancer diagnosis, 4) had the same month of diagnosis of breast cancer and death, or 5) had records considered to have unreliable Medicare coding for a comorbid condition (e.g. bills that are not encoded by a clinician). These exclusions were to ensure that claims files were available for accurate detection of comorbidities and accurate calculation of survival time from diagnosis to death.
The outcomes of interest were overall survival and all-cause mortality. Survival time was measured from the date of diagnosis of breast cancer until the date of SEER-recorded death or the censor date of December 15, 2005. Because SEER does not include the exact day for date of diagnosis or death, the 15th of the month was arbitrarily assigned to the reported month and year.
Measurement of Comorbidity
The SEER database identifies whether another primary cancer had been diagnosed before the cancer of interest in this analysis. This field was used to determine whether a previous cancer constituted a comorbidity. All other comorbidities were searched in Medicare files for the period of 1 year before and 30 days after the cancer diagnosis. This long period before cancer diagnosis and short period after cancer diagnosis allows for substantial time to identify diagnoses representing comorbidities without capturing conditions that may result from the cancer treatment. Comorbidity codes had to be recorded in the Medicare files but not necessarily first diagnosed in the patient during this period. Diagnostic coding for comorbidities had to be included in the Medicare inpatient files or had to occur at least two times 30 days apart in the outpatient and/or physician files.
The comorbidities of interest in this analysis are the 19 conditions included in the Charlson index (23,24), adapted by Deyo and Klabunde (25,26) to be used with administrative claims data. The Charlson index has been validated as a predictor of short- and long-term mortality and specifically for use with breast cancer (27). The Charlson index has five conditions that have two categories of severity that were combined for the purposes of our analysis. The conditions assessed in this study include: cerebrovascular disease (stroke), chronic obstructive pulmonary disease, chronic renal failure, congestive heart failure, dementia, diabetes, liver disease, myocardial infarction, paralysis (hemiplegia, hemiparesis, or paraplegia), peripheral vascular disease, previous cancer (except for nonmelanoma skin cancer), rheumatoid arthritis, and ulcers (chronic gastric or duodenal ulcers). We also assessed HIV status to determine if an individual had any comorbid conditions, but because there were only three patients diagnosed with HIV, separate data are not presented for this comorbidity, leaving 13 conditions specifically examined in this analysis. In addition, the Charlson Comorbidity Index Score was used as a summary measure of comorbidity in which each comorbidity is assigned a weight based on the association with survival, and the weights were then summed. SAS (SAS Institute, Cary, NC) programming to identify diagnostic and procedure codes for these comorbidities is available on the SEER–Medicare website (28).
Cancer Treatments
Fields for surgical treatment of breast cancer and radiation provided within 4 months after breast cancer diagnosis are available in the SEER database, but chemotherapy is not coded. Medicare files were therefore searched for chemotherapy within 6 months of diagnosis and for radiation codes within 9 months of diagnosis. These periods were chosen based on published articles of chemotherapy (29–31) and radiation therapy (30,32,33) to capture treatment for primary and not recurrent disease. Radiation treatment was counted if it occurred in either SEER or Medicare (33). The International Classification of Diseases, Ninth Revision, Clinical Modification diagnostic and procedure codes; Health Care Financing Administration Common Procedure Coding System codes; and Revenue Center codes used to identify chemotherapy and radiation treatment in Medicare are included in the Appendix.
Statistical Methods
Proportions, 5-year overall survival rates, and hazard ratios (HRs) for all-cause mortality according to specific patient characteristics were calculated. Multivariable Cox proportional hazards models were used to estimate the association between selected comorbidities and all-cause mortality in the form of crude and adjusted hazard ratios and 95% confidence intervals (CIs). Partially adjusted models controlled for age (66–74 years, 75–84 years, and 85 years or more), race and/or ethnicity (white, black, Hispanic, other, or unknown), stage (I, II, III and IV, or unknown), grade (1, 2, 3 and 4, or unknown), estrogen receptor status (yes, no, or unknown), surgery (yes, no, or unknown), chemotherapy (yes or no), and radiation treatment (yes or no). Fully adjusted models also controlled for the presence of all other comorbidities (13 yes or no conditions). Stratified analyses were conducted to determine if the relationships between comorbidities and survival were modified by age. Unadjusted Kaplan–Meier survival curves were generated for selected comorbidities and stratified by presence of comorbidity and stage category (stage I, II, and III and IV) for women diagnosed with breast cancer. Missing or unknown data values were grouped into their own category and included in Cox proportional hazards models. Two approaches were used to determine whether missing data on predictors contributed to the results. Records with missing or unknown values for any variable of interest were excluded from the analyses, and the R package NestedCohort was used to weight records with known values of age, race and/or ethnicity, and death status to account for records with missing data (34). SAS version 9.1 and 9.2 were used for analysis. All statistical tests were two-sided and a P value less than .05 was considered to be statistically significant.
Results
A total of 96 954 women aged 66 years or older and diagnosed with malignant breast cancer between January 1, 1992, and December 31, 2000, were potentially eligible for the study. One-third of women (32 920 women or 34.0%) were excluded because of the following (some patients were excluded for more than one reason): being in a Medicare HMO (22 106 women or 22.8%), unreliable diagnosis coding (3878 women or 4.0%), not having both Medicare Part A and B (5332 women or 5.5%), month and year of death were the same as that of diagnosis (2521 women or 2.6%) and/or unknown month of diagnosis (679 women or 0.7%). Women who were excluded from the study were younger than patients chosen for the study (median age = 74 vs 75 years), were more likely to be Hispanic (6.8% vs 3.0%), less likely to be white (78.4% vs 87.6%), more likely to have higher tumor stage (12.2% had stage III or IV vs 9.3%) or unknown tumor stage (10.7% vs 7.0%), and less likely to have undergone surgery (87.4% vs 93.9%). A total of 64 034 patients met the study inclusion criteria. Excluding records with unknown or missing values and using R to account for missing and unknown data values did not influence the study findings.
The median follow-up time was 104 months and the median age at diagnosis was 75 years. A breast cancer diagnosis was made for 28 812 study participants (45.0%) aged 66–74 years, 26 496 (41.1%) aged 75–84 years, and 8726 (13.6%) aged 85 years or older. The majority of patients were white (56 116 women or 87.6%), 33 017 tumors (51.6%) were stage I, 20 559 tumors (32.1%) were stage II, 5965 tumors (9.3%) were stage III and IV, and the stage was unknown for 4493 tumors (7.0%) (Table 1). Most patients (60 153 or 93.9%) underwent some type of surgery, 21 826 (34.1%) were treated with radiation, and 9459 (14.8%) were treated with chemotherapy.
Table 1.
Prevalence, 5-year survival rates, and hazard ratios of all-cause mortality by demographic and clinical characteristics in breast cancer patients aged 66 years or older, SEER-Medicare 1992–2000*
| Characteristic | No. (%) | 5-year survival rate, No. (%) |
HR† (95% CI) | |
| No. (%) | 95% CI | |||
| Total | 64 034 (100) | 43 151 (67.4) | 67.0 to 67.7 | NA |
| Age, y | ||||
| 66–74 | 28 812 (45.0) | 22 618 (78.5) | 78.0 to 79.0 | 1.0 (referent) |
| 75–84 | 26 496 (41.4) | 17 402 (65.7) | 65.1 to 66.2 | 1.86 (1.81 to 1.91) |
| ≥85 | 8726 (13.6) | 3131 (35.9) | 34.9 to 36.9 | 4.45 (4.32 to 4.58) |
| Race/ethnicity | ||||
| White | 56 116 (87.6) | 38 034 (67.8) | 67.4 to 68.2 | 1.0 (referent) |
| Black | 3901 (6.1) | 2192 (56.2) | 54.6 to 57.7 | 1.41 (1.35 to 1.46) |
| Hispanic | 1927 (3.0) | 1308 (68.4) | 65.7 to 69.9 | 0.97 (0.91 to 1.04) |
| Other | 1928 (3.0) | 1493 (77.4) | 75.5 to 79.2 | 0.66 (0.61 to 0.71) |
| Unknown | 162 (0.2) | 124 (76.5) | 69.2 to 82.3 | 0.64 (0.49 to 0.82) |
| Stage | ||||
| I | 33 017 (51.6) | 26 025 (78.8) | 78.4 to 79.3 | 1.0 (referent) |
| II | 20 559 (32.1) | 13 057 (63.5) | 62.8 to 64.2 | 1.63 (1.59 to 1.67) |
| III and IV | 5965 (9.3) | 1664 (27.9) | 26.8 to 29.0 | 4.53 (4.38 to 4.68) |
| Unknown | 4493 (7.0) | 2405 (53.5) | 52.1 to 55.0 | 2.02 (1.94 to 2.11) |
| Grade | ||||
| 1 | 10 519 (16.4) | 8171 (77.7) | 76.9 to 78.5 | 1.0 (referent) |
| 2 | 22 729 (35.5) | 16 314 (71.8) | 71.2 to 72.4 | 1.25 (1.20 to 1.29) |
| 3 and 4 | 16 859 (26.3) | 10 113 (60.0) | 59.2 to 60.7 | 1.72 (1.66 to 1.78) |
| Unknown | 13 927 (21.8) | 8553 (61.4) | 60.6 to 62.2 | 1.63 (1.57 to 1.69) |
| ER status | ||||
| Positive | 41 018 (64.1) | 29 508 (71.9) | 71.5 to 72.4 | 1.0 (referent) |
| Negative | 7981 (12.5) | 4742 (59.4) | 58.3 to 60.5 | 1.33 (1.29 to 1.38) |
| Unknown/other | 15 035 (23.5) | 8901 (59.2) | 58.4 to 60.0 | 1.41 (1.30 to 1.44) |
| Surgery | ||||
| Yes | 60 153 (93.9) | 42374 (70.4) | 70.1 to 70.8 | 1.0 (referent) |
| No | 2385 (3.7) | 326 (13.7) | 12.3 to 15.1 | 6.16 (5.89 to 6.43) |
| Unknown | 1496 (2.3) | 451 (30.2) | 27.8 to 32.5 | 3.27 (3.09 to 3.47) |
| Chemotherapy | ||||
| Yes | 9459 (14.8) | 5948 (62.9) | 61.9 to 63.8 | 1.0 (referent) |
| No | 54 575 (85.2) | 37 203 (68.2) | 67.8 to 68.6 | 0.90 (0.88 to 0.93) |
| Radiation | ||||
| Yes | 21 826 (34.1) | 17 005 (77.9) | 77.4 to 78.5 | 1.0 (referent) |
| No | 42 208 (65.9) | 26 146 (62.0) | 61.5 to 62.4 | 1.72 (1.68 to 1.77) |
CI = confidence interval, ER = estrogen receptor, HR = hazard ratio. Calculations were based on log–log transformation.
Cox proportional hazards model was used to calculate values.
Specific comorbidities with the highest prevalence (Table 2) were previous cancer (16.3%), diabetes (13.0%), chronic obstructive pulmonary disease (8.8%), congestive heart failure (6.7%), and stroke (4.3%). Liver disease had the lowest prevalence (0.3%) among the study population. Fifty-eight percent of the study population had none of the selected comorbidities, 28.0% had one comorbidity, 8.8% had two comorbidities, and 4.9% had three or more of the conditions.
Table 2.
Comorbidities among breast cancer patients aged 66 years or older: prevalence, 5-year survival rates, and hazard ratios of all-cause mortality of the SEER–Medicare 1992–2000 population*
| Comorbidities | No. (%) | 5-year survival rate |
Crude HR (95% CI) | Age-adjusted HR (95% CI) | Partially adjusted HR† (95% CI) | Fully adjusted HR‡ (95% CI) | |
| No. (%) | 95% CI | ||||||
| Total patients | 64 034 (100) | 43 151 (67.4) | 67.0 to 67.7 | NA | NA | NA | NA |
| Comorbidities | |||||||
| None | 37 306 (58.3) | 27 956 (74.9) | 74.5 to 75.4 | 1.0 (referent) | 1.0 (referent) | 1.0 (referent) | 1.0 (referent) |
| Previous cancer | 10 422 (16.3) | 6386 (61.3) | 60.3 to 62.2 | 1.61 (1.56 to 1.66)§ | 1.48 (1.44 to 1.52)§ | 1.55 (1.50 to 1.60)§ | 1.27 (1.23 to 1.30)§ |
| Myocardial infarction | 1091 (1.7) | 525 (48.1) | 45.1 to 51.0 | 2.38 (2.21 to 2.55)§ | 2.14 (2.00 to 2.31)§ | 2.16 (2.01 to 2.32)§ | 1.11 (1.03 to 1.19)‖ |
| Congestive heart failure | 4280 (6.7) | 1405 (32.8) | 31.4 to 34.2 | 3.68 (3.55 to 3.82)§ | 2.76 (2.66 to 2.87)§ | 2.55 (2.46 to 2.65)§ | 1.70 (1.64 to 1.76)§ |
| Peripheral vascular disease | 1638 (2.6) | 718 (43.8) | 41.4 to 46.2 | 2.69 (2.54 to 2.86)§ | 2.16 (2.04 to 2.29)§ | 2.18 (2.05 to 2.31)§ | 1.36 (1.28 to 1.44)§ |
| Cerebrovascular disease | 2742 (4.3) | 1232 (44.9) | 43.1 to 46.8 | 2.65 (2.53 to 2.78)§ | 2.16 (2.06 to 2.26)§ | 2.08 (1.99 to 2.18)§ | 1.35 (1.28 to 1.42)§ |
| COPD | 5669 (8.8) | 2987 (52.7) | 51.4 to 54.0 | 2.13 (2.06 to 2.21)§ | 2.11 (2.04 to 2.19)§ | 2.14 (2.07 to 2.22)§ | 1.52 (1.47 to 1.58)§ |
| Dementia | 887 (1.4) | 168 (18.9) | 16.4 to 21.6 | 5.72 (5.33 to 6.14)§ | 3.47 (3.23 to 3.73)§ | 2.79 (2.59 to 3.00)§ | 1.96 (1.82 to 2.10)§ |
| Paralysis | 388 (0.6) | 138 (35.6) | 30.8 to 40.3 | 3.52 (3.16 to 3.93)§ | 2.73 (2.44 to 3.04)§ | 2.58 (2.31 to 2.88)§ | 1.23 (1.09 to 1.38)§ |
| Diabetes | 8332 (13.0) | 4631 (55.6) | 54.5 to 56.6 | 1.97 (1.91 to 2.03)§ | 1.97 (1.91 to 2.03)§ | 1.90 (1.85 to 1.96)§ | 1.41 (1.36 to 1.45)§ |
| Chronic renal failure | 590 (0.9) | 139 (23.6) | 20.2 to 27.1 | 4.90 (4.49 to 5.35)§ | 4.52 (4.14 to 4.94)§ | 4.42 (4.04 to 4.83)§ | 2.20 (2.02 to 2.41)§ |
| Liver disease | 186 (0.3) | 59 (31.7) | 25.2 to 38.5 | 3.46 (2.94 to 4.06)§ | 3.78 (3.22 to 4.44)§ | 4.04 (3.44 to 4.76)§ | 2.32 (1.97 to 2.73)§ |
| Stomach ulcer | 705 (1.1) | 387 (54.9) | 51.1 to 58.5 | 1.91 (1.74 to 2.10)§ | 1.73 (1.57 to 1.89)§ | 1.71 (1.56 to 1.88)§ | 1.12 (1.02 to 1.23)# |
| Rheumatoid arthritis | 1246 (2.0) | 747 (60.0) | 57.2 to 62.6 | 1.70 (1.58 to 1.83)§ | 1.65 (1.53 to 1.78)§ | 1.71 (1.58 to 1.84)§ | 1.27 (1.18 to 1.37)§ |
| Charlson comorbidity index score | |||||||
| 0 | 37 306 (58.3) | 27 956 (74.9) | 74.5 to 75.4 | 1.0 (referent) | 1.0 (referent) | 1.0 (referent) | NA |
| 1 | 17946 (28.0) | 11 392 (63.5) | 62.8 to 64.2 | 1.54 (1.50 to 1.57)§ | 1.45 (1.41 to 1.48)§ | 1.45 (1.41 to 1.48)§ | NA |
| 2 | 5639 (8.8) | 2787 (49.4) | 48.1 to 50.7 | 2.34 (2.25 to 2.42)§ | 2.13 (2.06 to 2.21)§ | 2.12 (2.05 to 2.20)§ | NA |
| ≥3 | 3143 (4.9) | 1016 (32.3) | 30.7 to 34.0 | 3.69 (3.54 to 3.84)§ | 3.24 (3.11 to 3.38)§ | 3.19 (3.06 to 3.32)§ | NA |
CI = confidence interval; COPD = chronic obstructive pulmonary disease, HR = hazard ratio, NA = not applicable.
Adjusted for age, race and/or ethnicity, stage, grade, estrogen receptor status, surgery, chemotherapy, and radiation with Cox proportional hazards modeling.
Adjusted for age, race and/or ethnicity, stage, grade, estrogen receptor status, surgery, chemotherapy, radiation, and all other individual comorbid conditions with Cox proportional hazards modeling.
Indicates P < .001, determined by a two-sided χ2 test.
Indicates P < .01, determined by a two-sided χ2 test.
Indicates P < .05, determined by a two-sided χ2 test.
In univariate analysis, 5-year overall survival rates were highest for those patients of younger age, earlier tumor stage, lower tumor grade, positive estrogen receptor status (Table 1), and having no comorbidities (Table 2). Age was the largest confounding factor. Adjustment for age as a continuous variable (data not shown) produced results virtually identical to the age group-adjusted analyses shown in Table 2. All mortality hazard ratios for comorbidities in the fully adjusted models, which adjusted for other comorbid conditions, were smaller than those calculated by partially adjusted models (Table 2). Partially and fully adjusted hazard ratios of comorbidities were all positive and statistically significant (eg, fully adjusted hazard ratios: for myocardial infarction, HR of death = 1.11, 95% CI = 1.03 to 1.19, P = .006; for liver disease, HR of death = 2.32, 95% CI = 1.97 to 2.73, P < .001). Other comorbidities with the highest fully adjusted hazard ratios include chronic renal failure (HR of death = 2.20, 95% CI = 2.02 to 2.41, P = .001), dementia (HR of death = 1.96, 95% CI = 1.82 to 2.10, P < .001), and congestive heart failure (HR of death = 1.70, 95% CI = 1.64 to 1.76, P < .001). Analysis of patients diagnosed with only a single comorbidity (ie, excluding those with multiple comorbidities) resulted in hazard ratios for all-cause mortality that were generally lower than the partially adjusted model but higher than the fully adjusted model for each comorbidity (data not shown). Compared with patients who had no comorbidities, hazard ratios of death increased as the Charlson Comorbidity Index Score increased (adjusted HR: for a score of 1, HR = 1.45, 95% CI = 1.41 to 1.48, P < .001; for a score of 2, HR = 2.12, 95% CI = 2.05 to 2.20, P = .001; and for a score of ≥3, HR = 3.19, 95% CI = 3.06 to 3.32, P = .001).
Younger patients were more likely to have no comorbidities; 18 310 (63.6%) patients aged 66–74 years had no comorbidities compared with 14 637 (55.2%) and 4359 (50.0%) patients aged 75–84 yearsand 85 years or older, respectively. In general, the prevalence of each type of comorbidity increased with age. Consistently across most of the comorbidities, the adjusted hazard ratios of the comorbidities decreased as age increased (Figure 1) statistically significantly for all comorbidities except ulcers and rheumatoid arthritis. Omitting records with missing values from multivariable analysis did not change the study findings (data not shown).
Figure 1.
Adjusted hazard ratios of comorbidities on all-cause mortality, stratified by age category, among breast cancer patients aged 66 years or older, Surveillance, Epidemiology, and End Results–Medicare 1992–2000. The hazard ratios were fully adjusted for race and/or ethnicity, stage category, grade category, estrogen receptor status, surgery, chemotherapy, radiation, and all other comorbid conditions. COPD = chronic obstructive pulmonary disease, PVD = peripheral vascular disease.
Across all of the comorbid conditions, patients aged 66−74 years with stage I tumors who had these conditions had survival curves similar to patients with stage II tumors who had no comorbidities (Figure 2). Similar trends were observed in the Kaplan–Meier survival curves for patients aged 75–84 years and 85 years or older (data not shown). In all of these graphs, the curves representing patients with no comorbidities for stage I, stage II, and stage III and IV tumors are the same. The most common pattern of survival, shown for myocardial infarction, peripheral vascular disease, stroke, chronic obstructive pulmonary disease, and diabetes, for patients with earlier stage cancers and comorbidities was to have survival outcomes that were shifted about one stage higher compared with patients without comorbidities. For more severe comorbidities (eg, dementia, chronic renal failure, and liver disease), the survival outcomes of patients with stage I and stage II tumors who had the specific comorbid condition resembled that of patients with stage III and IV tumors who did not have the specific conditions. The number of patients at risk at different time points after breast cancer diagnosis is presented in Table 3.
Figure 2.
Kaplan–Meier survival curves by stage and each of 13 selected comorbid conditions (A–M) among women diagnosed with breast cancer at age 66–74 years, from the Surveillance, Epidemiology, and End Results–Medicare 1992–2000 database. CHF = congestive heart failure, COPD = chronic obstructive pulmonary disease, CRF = chronic renal failure, MI = myocardial infarction, PVD = peripheral vascular disease, RA = rheumatoid arthritis.
Table 3.
Number of patients at risk by stage and selected comorbid condition among women diagnosed with breast cancer at ages 66–74 years, SEER–Medicare 1992–2000
| Comorbidity | No. of patients at risk* |
|||||
| 0 mo | 20 mo | 40 mo | 60 mo | 80 mo | 100 mo | |
| No comorbidities | ||||||
| Stage I | 10 058 | 9874 | 9609 | 9280 | 6811 | 5057 |
| Stage II | 5750 | 5529 | 5128 | 4723 | 3220 | 2313 |
| Stage III and IV | 1501 | 1110 | 806 | 626 | 399 | 261 |
| Previous cancer | ||||||
| Stage I | 2420 | 2267 | 2095 | 1941 | 1347 | 979 |
| Stage II | 1137 | 1014 | 873 | 767 | 508 | 352 |
| Stage III and IV | 313 | 193 | 120 | 86 | 48 | 30 |
| Myocardial infarction | ||||||
| Stage I | 196 | 171 | 158 | 143 | 92 | 58 |
| Stage II | 136 | 111 | 92 | 75 | 48 | 33 |
| Stage III and IV | 36 | 20 | 12 | 9 | 5 | 3 |
| Congestive heart failure | ||||||
| Stage I | 467 | 383 | 315 | 257 | 145 | 85 |
| Stage II | 375 | 287 | 217 | 157 | 94 | 59 |
| Stage III and IV | 120 | 52 | 28 | 19 | 11 | 4 |
| Peripheral vascular disease | ||||||
| Stage I | 254 | 220 | 189 | 165 | 94 | 65 |
| Stage II | 144 | 120 | 92 | 73 | 40 | 26 |
| Stage III and IV | 38 | 15 | 6 | 4 | 0 | 0 |
| Cerebrovascular disease | ||||||
| Stage I | 389 | 344 | 305 | 273 | 184 | 134 |
| Stage II | 288 | 243 | 191 | 151 | 88 | 57 |
| Stage III and IV | 90 | 44 | 22 | 11 | 8 | 3 |
| COPD | ||||||
| Stage I | 1333 | 1221 | 1094 | 965 | 630 | 415 |
| Stage II | 767 | 658 | 564 | 466 | 279 | 160 |
| Stage III and IV | 206 | 110 | 70 | 49 | 28 | 19 |
| Dementia | ||||||
| Stage I | 45 | 35 | 26 | 20 | 10 | 5 |
| Stage II | 36 | 28 | 16 | 11 | 7 | 1 |
| Stage III and IV | 15 | 5 | 4 | 2 | 1 | 0 |
| Paralysis | ||||||
| Stage I | 44 | 39 | 31 | 26 | 18 | 14 |
| Stage II | 49 | 42 | 32 | 25 | 17 | 12 |
| Stage III and IV | 16 | 9 | 4 | 2 | 0 | 0 |
| Diabetes | ||||||
| Stage I | 1807 | 1672 | 1518 | 1375 | 873 | 571 |
| Stage II | 1351 | 1189 | 1006 | 863 | 519 | 330 |
| Stage III and IV | 358 | 205 | 119 | 84 | 37 | 24 |
| Chronic renal failure | ||||||
| Stage I | 96 | 61 | 45 | 33 | 17 | 9 |
| Stage II | 92 | 61 | 36 | 25 | 12 | 5 |
| Stage III and IV | 34 | 8 | 3 | 3 | 2 | 1 |
| Liver disease | ||||||
| Stage I | 41 | 38 | 29 | 19 | 10 | 6 |
| Stage II | 37 | 25 | 20 | 14 | 8 | 5 |
| Stage III and IV | 7 | 1 | 1 | 0 | 0 | 0 |
| Stomach ulcer | ||||||
| Stage I | 140 | 138 | 128 | 115 | 82 | 61 |
| Stage II | 85 | 66 | 55 | 50 | 39 | 24 |
| Stage III and IV | 12 | 9 | 4 | 4 | 3 | 3 |
| Rheumatoid arthritis | ||||||
| Stage I | 290 | 276 | 262 | 239 | 153 | 97 |
| Stage II | 159 | 138 | 116 | 103 | 70 | 44 |
| Stage III and IV | 35 | 22 | 13 | 10 | 7 | 3 |
The number of patients at risk at different times after breast cancer diagnosis.
Discussion
Our analysis found an association between thirteen individual comorbidities and the overall survival and all-cause mortality of older female breast cancer patients. The presence of comorbid conditions in this population is substantially associated with decreased survival and is quantitatively similar to that of breast cancer stage. The hazard ratios of all-cause mortality depend on the specific comorbidity and statistically significantly decline with increasing age for 11 of the 13 comorbidities included in this study.
It is well established that breast cancer patients with comorbidities have a poorer prognosis than patients without comorbidities (11–18), but previous studies have used only a summary measure of comorbidities such as the Charlson Comorbidity Index Score. Although a comorbidity index can be useful for research purposes, measuring the specific hazard ratio of comorbidities can help delineate clinically meaningful differences that exist between cancer patients. For example, we found that conditions such as liver disease, chronic renal failure, dementia, and congestive heart failure produce the highest hazard ratios of all-cause mortality and that patients with these conditions diagnosed at stage I have survival experiences similar to those of patients with stage III and IV tumors lacking any coexisting disease. Interestingly, in this study, the hazard ratios of all-cause mortality for individual comorbidities differed somewhat from the weights assigned to each comorbidity in the Charlson Comorbidity Index Score. This observation agrees with other publications that have suggested that comorbidity indices may not always be adequate predictors of survival, as the influence of comorbid conditions may vary depending on the population and disease of interest (35–38).
Only three other studies have assessed the individual effects of specific comorbidities on breast cancer mortality (11,12,17), with two of these studies having a substantial overlap in study populations. Louwman et al. (12) analyzed patients in the Eindhoven Cancer Registry who were diagnosed from 1995 to 2001 at any age and Janssen-Heijnen et al. (17) studied patients in the Eindhoven Cancer Registry with stage I–III tumors diagnosed from 1995 to 2004 among women aged 50 years or older and included other cancer sites as well. Our study findings were similar to these previous studies in terms of which comorbidities had the highest hazard ratios. However, we identified more comorbidities that had a statistically significant association with overall mortality. The largest of these previous studies included approximately 9000 breast cancer patients, whereas our study population is almost seven times larger. The large size of our study population allowed us to assess the impact of rarer conditions and to see whether the relationship between comorbidities and all-cause mortality differed by age and tumor stage.
Our study population is large because SEER was used to identify cancer patients. SEER participation has expanded greatly since its inception. The population covered by SEER is intended to be comparable to the general US population and is true in regards to measures of poverty and education. However, the SEER population is somewhat more urban and has a higher proportion of foreign-born people than the general population. SEER continually monitors and evaluates their data to ensure high quality. The program’s standard for registry completeness for incident cancers is 98% (21).
We determined that age at diagnosis often modifies the relationships between comorbidities and all-cause mortality. This interaction has not been previously studied among breast cancer patients. Three separate studies among patients with prostate, colorectal, and head and neck cancers found that comorbidity indices modified overall survival more for younger age-groups (39–41). Froehner et al. (42) also determined that age modifies the relationship between a comorbidity index and prostate cancer survival with a comorbidity index having the greatest effect among men aged 63–69 years. Only one previous study, reported by Gross et al. (43), measured comorbidities at the individual level and assessed interactions with age. Similar to our results, age modified the relationship between survival among colon cancer patients and three specific comorbidities including heart disease, chronic obstructive pulmonary disease, and diabetes.
Our study has several potential limitations. Medicare data were created for billing purposes and may lack the precision needed for clinical research. For example, data on possible confounding factors such as smoking, body fatness (body mass index or percentage body fat), physical activity, and functional status are not available in the database. Recipients of Medicare can seek care from outside providers, such as long-term care or an oral prescription for Tamoxifen, which was not in the past covered by Medicare. In addition, both cancer treatment and treatment for comorbid conditions have improved since the time this study began. Data on medical history before enrolling in Medicare are also not available. The measurement of comorbidities does not consider the severity of diagnosis or how long the patient has had each condition. In addition, if a person has a comorbid condition that is not listed in Medicare, it would result in misclassification of the comorbidity status. However, comorbidity prevalence was defined using inpatient, outpatient, and physician files with restriction criteria that have been shown to be comparable to hospital medical records (44).
Approximately one-third (34%) of the potentially eligible study participants were excluded, most often because they were enrolled in an HMO and Medicare files were not available. Differences do exist between the population characteristics of HMO and FFS Medicare health-care delivery system subscribers—patients who are enrolled in HMOs are generally younger, healthier, diagnosed at an earlier stage, and have better overall survival compared with FFS patients (21,45). Although these differences between HMO patients and FFS patients could potentially influence the relationships between comorbidity and overall survival, there is no data to support this hypothesis. Our study findings are therefore most applicable to the population of Medicare FFS patients because this population is used in the present analysis. In addition, the number of patients in the study receiving chemotherapy was low (14.8%) and consistent with other Medicare patient series (46,47). The relatively low number of patients receiving chemotherapy in our study population may be because of the substantial drop-off in the use of adjuvant therapies that occurs with advancing age (46–49). Finally, our study population included a number of records with missing or unknown values for stage, grade, and estrogen receptor status. However, the use of two different methods to account for missing or unknown values indicated that missing records had no influence on the study findings.
In this study, analyses of Kaplan–Meier survival curves demonstrate that comorbidities can be as important as stage in predicting breast cancer survival. Comorbidities are associated with survival to the extent that patients with these comorbid conditions diagnosed with early-stage breast cancer have survival similar to or worse than that of patients with no comorbidities diagnosed with later-stage tumors. This finding suggests that including comorbid conditions in the assessment of prognosis in both observational research and patient care may be important and that careful attention to the effective management of comorbid conditions, as well as to the management of a patient’s cancer, may result in longer overall survival for older breast cancer patients.
Funding
This work was supported, in part, by the National Cancer Institute and the National Institute on Aging at the National Institutes of Health (P20 CA103680 to DD and JLP).
Notes
The funders did not have any involvement in the design of the study; the collection, analysis, and interpretation of the data; the writing of the article; or the decision to submit the article for publication.
Appendix
Appendix Table 1.
Codes used to identify chemotherapy and radiation treatment in Medicare files*
| Treatment | ICD-9 diagnostic codes | ICD-9 procedure codes | HCPCS codes | Revenue center codes |
| Chemotherapy | V58.1, V66.2, V67 | 99.25 | 96400–96499, 96500–96599, Q0083–Q0085, 51720, J8510, J8520, J8521, J8530-J8999, J9000-J9999 | 331, 332, 335 |
| Radiation | V58.0, V66.1, V67.1 | 92.21–92.33, 92.39 | 77401–77499, 77520, 77523, 77750–77799, G0256, G0261 | 330, 333, 339 |
* ICD-9 = International Classification of Diseases, Ninth Revision; HCPCS = the Health Care Financing Administration Common Procedure Coding System.
References
- 1.Horner MJ, Ries LAG, Krapcho M, et al. SEER Cancer Statistics Review, 1975–2006. Bethesda, MD: National Cancer Institute; 2009. http://seer.cancer.gov/csr/1975_2006/. Accessed May 16, 2011. [Google Scholar]
- 2.Lewis JH, Kilgore ML, Goldman DP, et al. Participation of patient 65 years of age or older in cancer clinical trials. J Clin Oncol. 2003;21(7):1383–1389. doi: 10.1200/JCO.2003.08.010. [DOI] [PubMed] [Google Scholar]
- 3.Trimble EL, Carter CL, Cain D, Freidlin B, Ungerleider RS, Friedman MA. Representation of older patients in cancer treatment trials. Cancer. 1994;74(7):2208–2214. doi: 10.1002/1097-0142(19941001)74:7+<2208::aid-cncr2820741737>3.0.co;2-#. [DOI] [PubMed] [Google Scholar]
- 4.Hutchins LF, Unger JM, Crowley JJ, Coltman CA, Albain KS. Underrepresentation of patients 65 years of age or older in cancer treatment trials. N Engl J Med. 1999;341(27):2061–2067. doi: 10.1056/NEJM199912303412706. [DOI] [PubMed] [Google Scholar]
- 5.Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical trials, race, sex and age-based disparities. JAMA. 2004;291(22):2720–2726. doi: 10.1001/jama.291.22.2720. [DOI] [PubMed] [Google Scholar]
- 6.Unger JM, Coltman CA, Crowley JJ, et al. Impact of the year 2000 Medicare policy change on older patient enrollment to cancer clinical trials. J Clin Oncol. 2006;24(1):141–144. doi: 10.1200/JCO.2005.02.8928. [DOI] [PubMed] [Google Scholar]
- 7.Guralnik JM. Assessing the impact of comorbidity in older populations. Ann Epidemiol. 1996;6(5):376–380. doi: 10.1016/s1047-2797(96)00060-9. [DOI] [PubMed] [Google Scholar]
- 8.Yancik R. Cancer burden in the aged: an epidemiologic and demographic overview. Cancer. 1997;80(7):1273–1283. [PubMed] [Google Scholar]
- 9.Ogle KS, Swanson GM, Woods N, Azzouz F. Cancer and comorbidity: redefining chronic diseases. Cancer. 2000;88(3):653–663. doi: 10.1002/(sici)1097-0142(20000201)88:3<653::aid-cncr24>3.0.co;2-1. [DOI] [PubMed] [Google Scholar]
- 10.Klabunde CN, Warren JL, Legler JM. Assessing comorbidity using claims data: an overview. Med Care. 2002;40(suppl 8):IV26–IV35. doi: 10.1097/00005650-200208001-00004. [DOI] [PubMed] [Google Scholar]
- 11.Yancik R, Wesley MN, Ries LAG, Havlik RJ, Edwards BK, Yates JW. Effect of age and comorbidity in postmenopausal breast cancer patients aged 55 years and older. JAMA. 2001;285(7):885–892. doi: 10.1001/jama.285.7.885. [DOI] [PubMed] [Google Scholar]
- 12.Louwman WJ, Janssen-Heijnen MLG, Houterman S, et al. Less extensive treatment and inferior prognosis for breast cancer patients with comorbidity: a population-based study. Eur J Cancer. 2005;41(5):779–785. doi: 10.1016/j.ejca.2004.12.025. [DOI] [PubMed] [Google Scholar]
- 13.Satariano WA, Ragland DR. The effect of comorbidity on 3-year survival of women with primary breast cancer. Ann Intern Med. 1994;120(2):104–110. doi: 10.7326/0003-4819-120-2-199401150-00002. [DOI] [PubMed] [Google Scholar]
- 14.Houterman S, Janssen-Heijnen MLG, Verheij CD, et al. Comorbidity has negligible impact on treatment and complications but influences survival in breast cancer patients. Br J Cancer. 2004;90(12):2232–2237. doi: 10.1038/sj.bjc.6601844. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Nagel G, Wedding U, Hoyer H, Rohrig B, Katenkamp D. The impact of comorbidity on the survival of postmenopausal women with breast cancer. J Cancer Res Clin Oncol. 2004;130(11):664–670. doi: 10.1007/s00432-004-0594-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Cronin-Fenton DP, Norgaard M, Jacobsen J, et al. Comorbidity and survival of Danish breast cancer patients from 1995 to 2005. Brit J Cancer. 2007;96(9):1462–1468. doi: 10.1038/sj.bjc.6603717. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Janssen-Heijnen MLG, Maas HAAM, Houterman S, Lemmens VE, Rutten HJ, Coebergh JW. Comorbidity in older surgical cancer patients: influence on patient care and outcome. Eur J Cancer. 2007;43(15):2179–2193. doi: 10.1016/j.ejca.2007.06.008. [DOI] [PubMed] [Google Scholar]
- 18.West DW, Satariano WA, Ragland DR, Hiatt RA. Comorbidity and breast cancer survival: a comparison between black and white women. Ann Epidemiol. 1996;6(5):413–419. doi: 10.1016/s1047-2797(96)00096-8. [DOI] [PubMed] [Google Scholar]
- 19.US Public Health Service. International Classification of Diseases, 9th Revision. Washington, DC: US GPO; 1996. [Google Scholar]
- 20.American Medical Association. Physicians’ Current Procedural Terminology. CPT 95. 5th ed. Chicago, IL: American Medical Association; 1994. [Google Scholar]
- 21.Warren JL, Klabunde CN, Schrag D, Bach PB, Riley GF. Overview of the SEER-Medicare data: content, research applications, and generalizability to the United States elderly population. Med Care. 2002;40(8 suppl):3–18. doi: 10.1097/01.MLR.0000020942.47004.03. [DOI] [PubMed] [Google Scholar]
- 22.Potosky AL, Riley GF, Lubitz JD, Mentnech RM, Kessler LG. Potential for cancer related health services research using a linked Medicare-tumor registry database. Med Care. 1993;31(8):732–748. [PubMed] [Google Scholar]
- 23.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis. 1987;40(5):373–383. doi: 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]
- 24.Charlson ME, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol. 1994;47(11):1245–1251. doi: 10.1016/0895-4356(94)90129-5. [DOI] [PubMed] [Google Scholar]
- 25.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613–619. doi: 10.1016/0895-4356(92)90133-8. [DOI] [PubMed] [Google Scholar]
- 26.Klabunde CN, Potosky AL, Legler JM, Warren JL. Development of a comorbidity index using physician claims data. J Clin Epidemiol. 2000;53(12):1258–1267. doi: 10.1016/s0895-4356(00)00256-0. [DOI] [PubMed] [Google Scholar]
- 27.Newschaffer CJ, Bush TL, Penberthy LT. Comorbidity measurement in elderly female breast cancer patients with administrative and medical records data. J Clin Epidemiol. 1997;50(6):725–733. doi: 10.1016/s0895-4356(97)00050-4. [DOI] [PubMed] [Google Scholar]
- 28.National Cancer Institute. SEER-Medicare Database. Bethesda, MD: National Cancer Institute; http://healthservices.cancer.gov/seermedicare/. Accessed December 7, 2009. [Google Scholar]
- 29.Du Xl, Key CR, Dickie L, Darling R, Geraci JM, Zhang D. External validation of Medicare claims for breast cancer chemotherapy compared with medical chart reviews. Med Care. 2006;44(2):124–131. doi: 10.1097/01.mlr.0000196978.34283.a6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Smith Gl, Smith BD, Giordano SH, et al. Risk of hypothyroidism in older breast cancer patients treated with radiation. Cancer. 2008;112(6):1371–1379. doi: 10.1002/cncr.23307. [DOI] [PubMed] [Google Scholar]
- 31.Warren Jl, Harlan LC, Fahey A, et al. Utility of the SEER-Medicare data to identify chemotherapy use. Med Care. 2002;40(suppl 8):IV-55–IV-61. doi: 10.1097/01.MLR.0000020944.17670.D7. [DOI] [PubMed] [Google Scholar]
- 32.Srokowski TP, Fang S, Duan Z, et al. Completion of adjuvant radiation therapy among women with breast cancer. Cancer. 2008;113(1):22–29. doi: 10.1002/cncr.23513. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Virnig BA, Warren JL, Cooper GS, Klabunde CN, Schussler N, Freeman J. Studying radiation therapy using SEER-Medicare linked data. Med Care. 2002;40(suppl 8):IV49–IV54. doi: 10.1097/00005650-200208001-00007. [DOI] [PubMed] [Google Scholar]
- 34.Katki HA, Mark S. Survival analysis for cohorts with missing covariate information. R News. 2008;8(1):14–19. [Google Scholar]
- 35.Fleming ST, Rastogi A, Dmitrienko A, Johnson KD. A comprehensive prognostic index to predict survival based on multiple comorbidities: a focus on breast cancer. Med Care. 1999;37(6):601–614. doi: 10.1097/00005650-199906000-00009. [DOI] [PubMed] [Google Scholar]
- 36.Elixhauser A, Claudia S, Harris RD, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8–27. doi: 10.1097/00005650-199801000-00004. [DOI] [PubMed] [Google Scholar]
- 37.Extermann M. Measuring comorbidity in older cancer patients. Eur J Cancer. 2000;36(4):453–471. doi: 10.1016/s0959-8049(99)00319-6. [DOI] [PubMed] [Google Scholar]
- 38.Havlik RJ, Yancik R, Long S, Ries L, Edwards B. The National Institute on Aging and the National Cancer Institute SEER collaborative study on comorbidity and early diagnosis of cancer in the elderly. Cancer. 1994;74(7):2101–2106. doi: 10.1002/1097-0142(19941001)74:7+<2101::aid-cncr2820741718>3.0.co;2-m. [DOI] [PubMed] [Google Scholar]
- 39.Kahi CJ, Azzouz F, Juliar BE, Imperiale TF. Survival of elderly persons undergoing colonoscopy: implications for colorectal cancer screening and surveillance. Gastrointest Endosc. 2007;66(3):544–550. doi: 10.1016/j.gie.2007.01.008. [DOI] [PubMed] [Google Scholar]
- 40.Alho OP, Hannula K, Luokkala A, Teppo H, Koivunen P, Kantola S. Differential prognostic impact of comorbidity in head and neck cancer. Head Neck. 2007;29(10):913–918. doi: 10.1002/hed.20608. [DOI] [PubMed] [Google Scholar]
- 41.Post PN, Hansen BE, Kil PJ, Janssen-Heijnen ML, Coebergh JW. The independent prognostic value of comorbidity among men aged < 75 with localized prostate cancer: a population-based study. BJU Int. 2001;87(9):821–826. doi: 10.1046/j.1464-410x.2001.02189.x. [DOI] [PubMed] [Google Scholar]
- 42.Froehner M, Koch R, Litz RJ, Hakenberg OW, Oehlschlaeger S, Wirth MP. Interaction between age and comorbidity as predictors of mortality after radical prostatectomy. J Urol. 2008;179(5):1823–1829. doi: 10.1016/j.juro.2008.01.023. [DOI] [PubMed] [Google Scholar]
- 43.Gross CP, Guo Z, McAvay GJ, Allore HG, Young M, Tinetti ME. Multimorbidity and survival in older persons with colorectal cancer. J Am Geriatr Soc. 2006;54(12):1898–1904. doi: 10.1111/j.1532-5415.2006.00973.x. [DOI] [PubMed] [Google Scholar]
- 44.Klabunde CN, Harlan LC, Warren JL. Data sources for measuring comorbidity: a comparison of hospital records and Medicare claims for cancer patients. Med Care. 2006;44(10):921–928. doi: 10.1097/01.mlr.0000223480.52713.b9. [DOI] [PubMed] [Google Scholar]
- 45.Kirsner RS, Ma F, Fleming L, et al. The effect of Medicare health care delivery systems on survival for patients with breast and colorectal cancer. Cancer Epidemiol Biomarkers Prev. 2006;15(4):769–773. doi: 10.1158/1055-9965.EPI-05-0838. [DOI] [PubMed] [Google Scholar]
- 46.Du X, Goodwin JS. Patterns of use of chemotherapy for breast cancer in older women: findings from Medicare claims data. J Clin Oncol. 2001;19(5):1455–1461. doi: 10.1200/JCO.2001.19.5.1455. [DOI] [PubMed] [Google Scholar]
- 47.Baxter NN, Durham SB, Phillips KA, Habermann EB, Virning BA. Risk of dementia in older breast cancer survivors: a population-based cohort study of the association with adjuvant chemotherapy. J Am Geriatrics Soc. 2009;57(3):403–411. doi: 10.1111/j.1532-5415.2008.02130.x. [DOI] [PubMed] [Google Scholar]
- 48.Kirkpatrick HM, Aitelli CL, Qin H, et al. Referral patterns and adjuvant chemotherapy use in patients with stage II colon cancer. Clin Colorectal Cancer. 2010;9(3):150–156. doi: 10.3816/CCC.2010.n.020. [DOI] [PubMed] [Google Scholar]
- 49.DeMichele A, Putt M, Zhang Y, Glick JH, Norman S. Older age predicts a decline in adjuvant chemotherapy recomendations for patients with breast carcinoma. Cancer. 2003;97(9):2150–2159. doi: 10.1002/cncr.11338. [DOI] [PubMed] [Google Scholar]