Surgical treatment of colon cancer in patients aged 80 years and older

Analysis of 31,574 patients in the SEER-Medicare database

Authors

  • Heather B. Neuman MD,

    Corresponding author
    1. Department of Surgery, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
    • Department of Surgery, University of Wisconsin School of Medicine and Public Health, Clinical Science Center H4/726, 600 Highland Ave, Madison, WI 53792-7375

    Search for more papers by this author
    • Fax: (608) 263-7652

  • Erin S. O'Connor MD,

    1. Department of Surgery, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
    Search for more papers by this author
  • Jennifer Weiss MD,

    1. Division of Gastroenterology and Hepatology, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
    Search for more papers by this author
  • Noelle K. LoConte MD,

    1. Division of Hematology and Oncology, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
    Search for more papers by this author
  • David Y. Greenblatt MD,

    1. Department of Surgery, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
    Search for more papers by this author
  • Caprice C. Greenberg MD,

    1. Department of Surgery, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
    Search for more papers by this author
  • Maureen A. Smith MD

    1. Department of Surgery, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
    2. Department of Population Health Sciences, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
    3. Department of Family Medicine, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
    Search for more papers by this author

  • The ideas and opinions expressed herein are those of the author(s) and endorsement by the State of California Department of Public Health, the National Cancer Institute, and the Centers for Disease Control and Prevention or their contractors and subcontractors is not intended nor should be inferred.

Abstract

BACKGROUND:

Age-related disparities in colon cancer treatment exist, with older patients being less likely to receive recommended therapy. However, to the authors' knowledge, few studies to date have focused on receipt of surgery. The objective of the current study was to describe patterns of surgery in patients aged ≥ 80 years with colon cancer and examine outcomes with and without colectomy.

METHODS:

Medicare beneficiaries aged ≥ 80 years with colon cancer who were diagnosed between 1992 and 2005 were identified from the Surveillance, Epidemiology, and End Results-Medicare database. Multivariable logistic regression analysis was used to assess factors associated with nonoperative management. Kaplan-Meier survival analysis determined 1-year overall and colon cancer-specific survival.

RESULTS:

Of 31,574 patients, 80% underwent colectomy. Approximately 46% were diagnosed during an urgent/emergent hospital admission, with decreased 1-year overall survival (70% vs 86% for patients diagnosed during an elective admission) noted among these individuals. Factors found to be most predictive of nonoperative management included older age, black race, more hospital admissions, use of home oxygen, use of a wheelchair, being frail, and having dementia. For both operative and nonoperative patients, the 1-year overall survival rate was lower than the colon cancer-specific survival rate (operative patients: 78% vs 89%; nonoperative patients: 58% vs 78%).

CONCLUSIONS:

The majority of older patients with colon cancer undergo surgery, with improved outcomes noted compared with nonoperative management. However, many patients who are not selected for surgery die of unrelated causes, reflecting good surgical selection. Patients undergoing surgery during an urgent/emergent admission have an increased short-term mortality risk. Because the earlier detection of colon cancer may increase the percentage of older patients undergoing elective surgery, the findings of the current study may have policy implications for colon cancer screening and suggest that age should not be the only factor driving cancer screening recommendations. Cancer 2013. © 2012 American Cancer Society.

INTRODUCTION

The US population is aging1 and in 2010 included more than 11 million individuals aged > 80 years. The growing numbers of the “oldest old” are at least partially a reflection of increasing life expectancy, because the average life expectancy for an 80-year-old alive today is > 8 years.2 These changing demographics have a significant impact on health care as a whole and cancer care in particular. For cancers such as colon cancer, which is largely a disease of the elderly, the impact may be especially significant.

Prior research has demonstrated that age-related disparities in the treatment of colon cancer exist, with older patients less likely to receive recommended therapy.3 The majority of studies examining age-related disparities have focused on adjuvant chemotherapy, with several studies demonstrating that older adults are less likely to be recommended or to receive adjuvant treatment.4-8 To the best of our knowledge, only 1 study to date, by O'Connell et al, has focused on the receipt of surgery.9 This study demonstrated that the majority of older patients with colon cancer undergo cancer-directed surgery. However, the study excluded patients with regional disease or those whose disease was unstaged. Considering that older patients are more likely to have unstaged disease,10 the conclusions from the study by O'Connell et al may not be reflective of surgical treatment for the majority of older Americans.9 In addition, the study did not examine factors associated with the selection of surgery in these older patients. The current research was initiated with the goal of examining current practices in the United States with regard to the surgical management of colon cancer in patients aged ≥ 80 years.

MATERIALS AND METHODS

The current study was approved by the University of Wisconsin Institutional Review Board and given a waiver of consent.

Data Source

We examined data from the linked Surveillance, Epidemiology, and End Results (SEER)-Medicare database for patients diagnosed with colon cancer between 1992 and 2005. The SEER cancer registries include information regarding patient demographics, tumor characteristics, first course of treatment, and survival for persons newly diagnosed with cancer. For individuals who are eligible for Medicare services, the SEER-Medicare database includes claims for covered health care services, including hospital, physician, outpatient, home health, and hospice bills. The SEER-Medicare data set has successfully linked 93% of individuals aged > 65 years at the time of diagnosis with their Medicare record.11, 12 In 2000, SEER regions included approximately 26% of the US population.12

Patient Selection

All Medicare-enrolled patients aged ≥ 80 years who were diagnosed with primary colon cancer within a SEER region during the years 1992 through 2005 were considered for study inclusion. Patients with colon cancer were identified by SEER anatomic site (18.0-18.9 and 19.9) and histology (8140-8417, 8210-8211, 8220-8221, 8260-8263, 8480-8481, and 8490) codes. Patients were excluded if they had distant metastases (American Joint Committee on Cancer [AJCC] stage IV disease or SEER summary stage “distant” disease). Continuous enrollment in Medicare Parts A and B was required for the 3 years preceding diagnosis through the 3 years after discharge, death, or December 31, 2005 (whichever came first) to allow for the ascertainment of comorbidities and survival. Patients were excluded if they were enrolled in a health maintenance organization during the same time period. Patients were also excluded if they were diagnosed with another malignancy 1 year before or after the date of the colon cancer diagnosis, or if their first diagnosis of colon cancer was made after death (ie, on autopsy or death certificate). From an initial cohort of 42,873 patients aged ≥ 80 years with colon cancer, the following patients were sequentially excluded: those with distant metastases (6755 patients), those with discontinuous Medicare Part A and Part B coverage (1143 patients), and those with another cancer diagnosed within 1 year of the colon cancer diagnosis (3401 patients). The final sample size was 31,574 patients.

Outcome Variables

The primary outcome measure for the current study was the receipt of “curative surgery.” Surgery was considered to be of curative intent if it occurred within 90 days of the diagnosis of colon cancer and included 1 of the following International Classification of Diseases, Ninth Revision (ICD-9) procedure codes: 45.7X or 45.8X.

The 1-year overall and cancer-specific survival rates were examined based on dates of death recorded in the SEER Patient Entitlement and Diagnosis Summary File (PEDSF) according to Social Security administration data. Length of hospital stay, 30-day readmission rates (readmission to any acute care hospital), and in-hospital complications were also examined. In-hospital complications included complications that resulted in reoperation or other procedural interventions.13

Patient-Related Variables

Basic patient-related variables included date of birth, gender, race/ethnicity, marital status, SEER registry region, and census tract median level of household income and median level of education (which were used as proxies for socioeconomic status). Geographic region was represented by SEER registry and rural/urban county of residence, based on 2003 Rural/Urban Continuum Codes identified from the PEDSF.

We assessed patients' overall health by several mechanisms. The number of emergency room visits and hospital admissions within the year before diagnosis was examined to assess health care use. Specific comorbidities previously found to be associated with patient outcomes were identified.14 Given the importance of dementia in treatment decision-making for older adults, we assessed for dementia using a separate, validated algorithm based on the following ICD-9 codes: 331.0 to 331.1, 331.11, 331.19, 331.7, 290.0, 290.10 to 290.13, 290.20 to 290.21, 290.3, 290.40 to 290.43, 294.0, 294.1, 294.10, 294.11, 294.8, and 797.15 Home oxygen use was recorded if patients had 2 codes related to home oxygen usage within the 3 years before diagnosis (durable medical equipment code A7017). In addition, functional status was indirectly assessed by examining claims for mobility devices within the 3 years before diagnosis (assessment for device: 97755 and 97542; cane: E0100 and E0105; crutch: E0110-E0017; walker: E0130, E0135, E0140-E0141, E0143-E0144, and E0147-E0149; wheelchair: K0001-K0007, E0983, E0984, E1210-E1213, E2368-E2370, K0010-K0012, K0014, and K0800-K0899). Patient frailty (defined within this context as decreased physiologic reserve and resistance to stressors as a result of physiologic multisystem decline) was assessed using ICD-9 codes for 12 conditions as defined by the John Hopkins Adjusted Clinical Groups case-mix system (eg, difficulty walking, weight loss, frequent falls, malnutrition, impaired vision, decubitus ulcer, and incontinence).16 Patients who had billing codes for any of these conditions were considered to be frail. Of the 4233 patients meeting the criteria for frailty, 75% met the criteria with a single ICD-9 code. The most common single codes were difficulty walking (38%), weight loss (30%), frequent falls (16%), and decubitus ulcers (10%); the remaining codes occurred most frequently in combination.

Colon cancer disease stage was assessed using the SEER historic stage variable, which categorizes patients as having local, regional, distant, or unstaged disease. Within SEER, staging is generated using a combination of the most precise clinical and pathological documentation of the extent of disease. Because our patient cohort included patients who did not undergo surgery, there was a significant percentage of unstaged patients. This percentage was smaller when the SEER historic stage variable was used rather than AJCC staging. All patients with unstaged disease were kept in the study cohort.

Treatment-Related Variables

In addition to the receipt of “curative surgery,” we examined whether patients received a colectomy after the defined 90-day curative “ window.” Receipt of a diverting stoma without accompanying colectomy (ICD-9 codes 46.0, 46.01, 46.03, 46.1, 46.11, 46.13, 46.2, 46.20, 46.21, and 46.23 and Current Procedural Terminology [CPT] codes 44141, 44143, 44144, 44146, 44150, 44155, 44206, 44212, 44310, 44320, and 44322) or a colonic stent (ICD-9 codes 46.86, 46.87, and 97.04 and CPT codes 44397, 45327, and 45387) any time after diagnosis was noted. We also examined whether patients had an intestinal obstruction (ICD-9 code 560.89 or 560.90) or an intestinal perforation (ICD-9 code 569.83) at the time of surgery, or were admitted under urgent/emergent conditions.

The question of whether patients received a surgical evaluation after diagnosis was examined by recording the date of service from a surgical specialist (defined as Centers for Medicare and Medicaid Services provider specialty codes 2 [general surgery], 28 [colorectal surgery], and 91 [surgical oncology]) with an appropriate Evaluation and Management service code (inpatient: 99025, 99218-99223, 99234-99236, 99251-99255, 99261-99263, 99281-99285, 99356, and 99357; and outpatient: 99058, 99201-99205, 99211-99215, 99241-99245, 99271-99275, 99354-99355, 99381-99387, 99391-99397, 99401-99404, 99411-99412, 99420, 99429, 99431-99432, G0101, G0245-G0250, G0344, G0375, and G0376).

Statistical Analysis

Descriptive statistics of patient characteristics were generated. The frequency of all patient-related and treatment variables by receipt of surgery with curative intent were compared using chi-square tests for categorical variables and 2-way analysis of variance (ANOVA) tests for continuous variables. Multivariable logistic regression analysis was used to assess factors associated with not undergoing surgery. Additional surgical interventions (delayed colectomy, diverting stoma, or colonic stent) for the group not undergoing curative surgery were examined. The 1-year overall and cancer-specific survival rates were assessed using Kaplan-Meier analysis. Cox regression analysis was used to identify factors associated with overall survival. Given the high percentage of unstaged patients in the “no surgery” group, sensitivity analyses were performed to evaluate the impact of cancer stage on survival. Because the addition of cancer stage was found to have only minimal impact on the model, it was not included in the final models.

RESULTS

Of the 31,574 patients aged ≥ 80 years who were diagnosed with colon cancer, 80% underwent colectomy within 90 days of their diagnosis (Fig. 1). Patient characteristics are described in Table 1. As would be expected, several differences existed between those patients who underwent surgery and those who did not, with those patients treated nonoperatively more likely to be older, male, nonwhite, and unmarried. In addition, nonoperative patients were less “healthy,” with more hospital admissions noted within the prior year, more use of home oxygen, more claims for mobility assist devices, and greater comorbidities (Table 1). In adjusted analyses, the factors found to be most predictive of being managed nonoperatively after a diagnosis of colon cancer included older age, black race, more hospital admissions within the prior year, use of home oxygen, use of a wheelchair, being frail, and having dementia (Table 2).

Figure 1.

Surgical treatment of patients aged ≥ 80 years with colon cancer is shown. yo indicates years old.

Table 1. Characteristics of Patients Aged ≥80 Years With Colon Cancer Treated With and Without Colectomy
CharacteristicsColectomy N=25,358No Colectomy N=6216P
  1. Abbreviations: ER, emergency room; SD, standard deviation.

Demographics   
Mean age at diagnosis (SD), y85.0 (4.0)86.0 (4.6)<.005
Male sex36.0%38.3%.001
Race   
 White88.7%83.8%<.005
 Black5.1%9.3%
 Other6.2%7.0%
Marital status   
 Married36.5%31.6%<.005
 Widowed50.4%51.2%
 Single/separated/  divorced9.7%11.5%
Cancer stage   
 Localized46.4%46.7%<.005
 Regional52.3%23.2%
 Unstaged1.3%30.1%
Health factors   
Mean no. of hospital admissions with prior y (SD)1.49 (0.90)1.70 (1.2)<.0005
Mean no. of ER visits within prior y (SD)3.62 (4.1)3.53 (4.1).280
Home oxygen use2.4%3.7%<.005
Mobility assist device   
 No assessment91.7%87.2%<.005
 Crutch/cane4.6%6.1%
 Wheelchair3.7%6.8%
Frail6.8%14.6%<.005
Comorbidities   
 Dementia1.8%6.1%<.005
 Congestive heart  failure38.6%42.8%<.005
 Valvular disease14.2%13.7%<.005
 Hypertension75.1%71.6%<.005
 Peripheral vascular  disease31.0%33.3%<.005
 Paralysis3.7%5.1%<.005
 Other neurologic disorders14.2%16.8%<.005
 Chronic pulmonary  disease25.6%26.9%<.005
 Diabetes  (uncomplicated)15.1%15.4%<.005
 Diabetes (with  complications)8.7%9.9%<.005
 Renal failure5.7%6.2%<.005
 Weight loss6.7%7.1%<.005
Table 2. Multivariable Analysis of Factors Associated With Not Undergoing Colectomy for Colon Cancer in Patients Aged ≥80 Years
FactorAdjusted OR95% CIP
  1. Abbreviations: 95% CI, 95% confidence interval; ER, emergency room; OR, odds ratio; SD, standard deviation.

Demographics   
Mean age at diagnosis (SD)1.041.03-1.05<.005
Male sex1.040.93-1.16.46
Race   
 WhiteReference  
 Black1.421.18-1.70<.005
 Other1.070.86-1.34.52
Marital status   
 MarriedReference  
 Widowed1.070.95-1.21.23
 Single/separated/  divorced1.191.0-1.41.05
Health factors   
Mean no. of hospital admissions within prior y (SD)1.101.05-1.14<.005
Mean no. of ER visits within prior y (SD)0.990.98-1.01.37
Home oxygen use1.441.15-1.78.001
Mobility assist device   
 No assessmentReference  
 Crutch/cane1.160.98-1.37.08
 Wheelchair1.241.05-1.48.01
Frail1.791.56-2.05<.005
Comorbidities   
 Dementia2.221.80-2.73<.005
 Congestive heart failure1.161.05-1.28.01
 Valvular disease1.070.94-1.21.32
 Hypertension0.910.81-1.03.13
 Peripheral vascular  disease1.050.95-1.16.32
 Paralysis1.371.12-1.67.002
 Other neurologic  disorders1.070.95-1.22.28
 Chronic pulmonary  disease1.090.98-1.21.11
 Diabetes  (uncomplicated)1.070.94-1.22.27
 Diabetes (with  complications)1.231.06-1.42.01
 Renal failure1.090.91-1.31.34
 Weight loss0.800.67-0.96.02

For patients undergoing curative colectomy, the 1-year overall survival rates were 78% versus 89% (Fig. 2). The colon cancer-specific survival rate was similarly higher for those patients who did not undergo surgery (58% vs 76%). Factors found to be associated with decreased overall and colon cancer-specific survival included older age, being widowed, having more hospitalizations or emergency room visits within the prior year, use of home oxygen, being frail, having dementia, and not receiving elective surgery (Table 3). In addition, male gender, being single/separated/divorced, and use of a wheelchair were associated with decreased overall survival. Some specific comorbidities (such as hypertension, peripheral vascular disease, and chronic pulmonary disease) were found to be associated with improved overall and cancer-specific survival.

Figure 2.

Overall and colon cancer-specific survival rates are shown for patients aged ≥ 80 years who were treated with and without curative colectomy.

Table 3. Predictors of Overall Survival and Cancer-Specific Survival in Patients Aged ≥80 Years With Colon Cancer
PredictorOverall SurvivalCancer-Specific Survival
 Adjusted HR95% CIPAdjusted HR95% CIP
  1. Abbreviations: 95% CI, 95% confidence interval; ER, emergency room; HR, hazard odds ratio; SD, standard deviation.

Demographics      
Mean age at diagnosis (SD)1.041.03-1.05<.0051.041.03-1.05<.005
Male sex1.261.16-1.36<.0051.080.96-1.21.23
Race      
 WhiteReference     
 Black1.130.99-1.29.071.120.93-1.37.23
 Other0.970.83-1.14.741.050.83-1.33.67
Marital status      
 MarriedReference     
 Widowed1.211.11-1.32<.0051.201.05-1.36.01
 Single/separated/divorced1.171.03-1.32.011.090.91-1.31.37
Health factors      
Mean no. of hospital admissions within prior y (SD)1.131.10-1.16<.0051.081.03-1.12.001
Mean no. of ER visits within prior y (SD)1.021.01-1.02<.0051.021.01-1.03.001
Home oxygen use1.521.32-1.76<.0051.281.01-1.61.04
Mobility assist device      
 No assessmentReference     
 Crutch/cane0.900.79-1.02.110.910.75-1.09.31
 Wheelchair1.181.03-1.33.011.100.91-1.33.34
Frail4.744.33-5.18<.0055.244.6-5.97<.005
Comorbidities      
 Dementia1.901.64-2.19<.0052.051.70-2.48<.005
 Congestive heart failure1.000.93-1.08.920.850.76-0.94.003
 Valvular disease0.890.81-0.97.010.810.70-0.94.01
 Hypertension0.660.60-0.71<.0050.600.53-0.67<.005
 Peripheral vascular disease0.770.72-0.83<.0050.710.64-0.80<.005
 Paralysis0.930.81-1.08.370.870.69-1.10.25
 Other neurologic disorders0.800.73-0.88<.0050.750.66-0.87<.005
 Chronic pulmonary disease0.900.83-0.97.010.790.70-0.89<.005
 Diabetes (uncomplicated)0.990.90-1.09.880.950.82-1.10.48
 Diabetes (with complications)0.990.88-1.12.930.820.67-0.99.04
 Renal failure1.000.88-1.15.920.920.74-1.15.49
 Weight loss0.880.78-1.00.050.850.70-1.02.09
Receipt of surgery      
 ElectiveReference     
 Urgent/emergent1.791.65-1.95<.0052.091.83-2.38<.005
 No surgery2.151.95-2.37<.0052.762.38-3.21<.005

Of the 6216 patients who did not undergo colectomy within 90 days, 8.9% required a delayed colectomy (Fig. 1). An additional 87 patients (1.4%) required a diverting stoma and 24 patients (0.4%) received a colonic stent. Overall, 5554 (89.3%) of the patients not selected to undergo immediate colectomy did not require any procedural intervention for their colon cancer. Only 31.1% of nonoperative patients had an evaluation and management code for a surgeon visit recorded. Outcomes between nonoperative patients who did and did not see a surgeon were equivalent (data not shown).

The majority of patients undergoing surgery did so during an elective admission (53.6%). The characteristics of patients undergoing elective versus urgent/emergent surgery differed significantly, with patients undergoing urgent/emergent surgery more likely to be older (P < .001), nonwhite (P < .001), and married (P < .001). In addition, patients undergoing urgent/emergent surgery were found to have had more hospitalizations within the prior year (P < .001); a higher stage of disease (P < .001); more claims for mobility assist devices (P < .001); more use of home oxygen (P < .001); and, in general, greater comorbidities (P < .001). Although the overall rate of postoperative complications for the overall cohort was relatively low at 4.3%, the complication rate in patients undergoing surgery during an urgent/emergent admission was higher (5.4% vs 3.4%; P < .0001). Similarly, the length of hospital stay (median, 13 days vs 9 days; P < .0001) and readmission rates (14.3% vs 10.9%; P < .0001) were increased. The overall 1-year survival rate for patients undergoing elective surgery was 86% compared with 70% after an urgent/emergent admission (Fig. 3). Of the 1851 patients who died within 30 days of surgery, the majority of these deaths (70%) occurred after an urgent/emergent admission.

Figure 3.

Overall survival of colon cancer is shown for patients aged ≥ 80 years by urgency of surgery. Admissi indicates admission.

DISCUSSION

The results of the current study demonstrate that the majority of the “oldest old” patients with colon cancer in the United States are undergoing surgical resection. For both surgical and nonsurgical patients, colon cancer-specific survival was higher than overall survival,17, 18 demonstrating that many older patients who are diagnosed with colon cancer ultimately die of unrelated causes. The majority of patients selected for surgery do well, despite their advanced age and numerous comorbidities. This is especially true for older patients who undergo surgery under elective circumstances, with a 30-day mortality rate of only 3%. This suggests that for those “oldest old” patients believed to be good candidates for elective surgical intervention, surgery should be considered as a standard of care.

Conversely, surgical outcomes during an urgent/emergent admission were significantly poorer, with a 30-day mortality rate of 10%. These outcomes mirror the findings of prior retrospective studies.19-21 Several factors may be contributing to the poorer outcomes observed. In the current study cohort, patients who underwent urgent/emergent surgery were overall “less healthy” compared with those patients selected for elective surgery. Therefore, the poorer outcomes observed may occur as a result of delays in intervention when these higher risk older adults present with an urgent/emergent condition. Alternatively, a conscious decision may have been made to avoid elective procedures in these patients. If this included cessation of screening colonoscopy, delays in diagnosis may have occurred, necessitating urgent/emergent surgery because of serious complications such as bowel perforation or obstruction. Finally, a decision may have been made by the patient and their care provider to avoid an elective colectomy itself due to the individual's health status. It may be these higher risk patients who avoided an elective colectomy at the time of diagnosis but then presented emergently who are contributing to the poorer outcomes observed. The limitations of the SEER-Medicare data set do not allow us to explore these potential factors driving the outcomes after urgent/emergent surgery further, but highlight the importance of clinical judgment in the management of older patients with colon cancer.

It is important to note that only a few patients who were not selected for initial surgical resection required a delayed surgical intervention for symptom management. This is likely a result of the high mortality rate in this patient population and reflects the good judgment of clinicians in selecting patients for surgical resection. Patients selected for nonoperative management were overall “less healthy” than surgical patients, as reflected by more frequent hospital admissions, higher rates of home oxygen and wheelchair use, and increased rates of frailty and dementia. Although only 31% of nonoperative patients saw a surgeon for a preoperative consultation, the outcomes for these patients were equivalent to those for patients who did not have a surgical evaluation. This suggests that primary care providers and oncologists are effectively choosing patients for a surgical referral. Overall, this observation highlights the finding that for select older adults believed to be poor operative candidates by their health care providers, the risk of death from a competing cause may be higher than the risk of developing colon cancer-related symptoms. In these highly selected patients, a selective, symptom-directed approach to the management of their colon cancer may be reasonable.

There are some limitations to the current study. First, our use of Medicare claims data limits our assessment of patients' overall health to those factors that are associated with a billing claim, such as the presence of specific comorbidities, claims for mobility assist devices such as walkers or wheelchairs, and the use of home oxygen. However, our claims data cannot provide insight into the chronicity or severity of these factors. This is especially relevant when assessing comorbidities in patients aged ≥ 80 years. Although the past medical history for many patients in this age group will include a significant list of comorbidities, there is a selection bias associated with living to be an octogenarian or nonagenarian that suggests that the severity of an individual's comorbidities may be relatively low. This selection bias may explain our observation that several specific comorbidities were associated with improved survival. In addition, clinical factors not captured by billing claims were not included in the current analysis. Factors such as independent living22 and frailty22-25 have been shown to be significantly associated with outcomes in the older adult population. Although we were able to assess for frailty, patients identified as “frail” through billing codes likely represent only the most severely impaired, and it is very possible we are underestimating the degree of frailty in this “oldest old” patient cohort. Next, a significant percentage of the patients in the current study who were not selected to undergo surgery had unstaged disease (30% vs 1.3% in the operative cohort); the remaining patients underwent clinical or radiologic staging alone. Because of this, we were unable to fully assess the relation between cancer stage and survival. Finally, we were restricted to using survival as our primary outcome. Older adults may place relatively greater importance on other outcomes such as quality of life; maintaining the ability to live independently; or surviving to a particular milestone, such as a birth or wedding.26 Our claims-based data set cannot provide insight into these values-dependent outcomes.

However, the current study does have several strengths. By using the SEER-Medicare database, we were able to examine population-level outcomes for all patients aged > 80 years who were diagnosed with colon cancer. Severity of illness was examined using several different surrogates, including hospital admission, emergency room visits, and the use of home oxygen. We were able to indirectly assess functional status through billing claims for mobility assist devices. Finally, by using the SEER data, we were able to examine the cause of death and provide insight into competing mortalities in this population.

Conclusions

To the best of our knowledge, the current study represents the first population-based study to comprehensively examine outcomes for the “oldest old” patients with colon cancer. The results of the current study demonstrated that the majority of patients aged > 80 years with colon cancer are undergoing surgical resection, with good short-term outcomes. Outcomes for the 20% of patients who are not selected for surgical intervention are markedly poorer; however, because many of the deaths are unrelated to the colon cancer diagnosis, these findings likely reflect good clinical judgment in selecting appropriate patients for surgical treatment. For the patients who did undergo surgical resection, nearly 50% of operations occurred during an ugent/emergent admission, with an observed increase in short-term mortality.

The findings of the current study lead to several directions for future research. First, it is apparent that patients who undergo surgery under urgent/emergent conditions have poorer outcomes. Many different factors may be driving this observation. Some may represent an appropriate reflection of patients' preferences and values, such as a decision reached by a patient and their care provider to avoid elective surgery given the patient's poor overall health status. Other factors, such as a delay in diagnosis due to the cessation of screening colonoscopies, may be less appropriate. The earlier diagnosis of colon cancer in older adults represents a potentially modifiable means of decreasing high-risk emergent surgery by conversion to an elective procedure. This observation has important policy implications for colon cancer screening. Based on the currently available data, the US Preventive Services Task Force recommends against routine colon cancer screening after the age of 75 years except in patients at an increased risk of developing colon cancer.27 However, the “oldest old” adults have largely been excluded from clinical trials and only limited data are available to guide screening recommendations for this age group. In addition, the US aging population is diverse and a patient's age may not accurately reflect their overall health status. Therefore, chronologic age should not be the only factor driving cancer screening recommendations. Future research should focus on how to account for increasing life expectancy in the US population in colon cancer screening recommendations.

Finally, the results of the current study demonstrate the importance of patient selection. Although significant perioperative risk is associated with the “oldest; old” patients presenting with urgent/emergent problems, these patients often represent less of a clinical decision-making dilemma to surgeons than do patients who are evaluated in the clinic for an elective surgical intervention. This is especially true when outcomes other than survival (such as quality of life and independent living) are considered. Future research should focus on patients being considered for elective surgery to better identify those patients at an increased risk of poor short-term outcomes who may be better managed with a nonoperative approach.

Acknowledgements

We acknowledge the efforts of the Applied Research Program of the National Cancer Institute; the Office of Research, Development and Information, Centers for Medicare and Medicaid Services; Information Management Services, Inc; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database.

FUNDING SUPPORT

Support provided by the Health Innovation Program and the Community-Academic Partnerships Core of the University of Wisconsin Institute for Clinical and Translational Research (UW ICTR); grant 1UL1RR025011 from the Clinical and Translational Science Award program of the National Center for Research Resources, National Institutes of Health; the University of Wisconsin Carbone Cancer Center; and grant P30CA014520-34 from the National Cancer Institute, National Institutes of Health. Additional funding for this project was provided by the University of Wisconsin School of Medicine and Public Health from The Wisconsin Partnership Program.

The collection of the California cancer incidence data used in this study was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the National Cancer Institute's Surveillance, Epidemiology, and End Results Program under contract N01-PC-35136 awarded to the Northern California Cancer Center; contract N01-PC-35139 awarded to the University of Southern California; contract N02-PC-15105 awarded to the Public Health Institute; and the Centers for Disease Control and Prevention's National Program of Cancer Registries under agreement #U55/CCR921930-02 awarded to the Public Health Institute.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary