Age-Specific Administration of Chemotherapy and Long-Term Quality of Life in Stage II and III Colorectal Cancer Patients: A Population-Based Prospective Cohort

Study Design and Study Population

This analysis is based on patients with a colorectal cancer diagnosis in 2003 or 2004 from a population-based case–control study carried out in southwest Germany (Rhine-Neckar-Odenwald region) including patients with a first histologically confirmed diagnosis of colorectal cancer [Darmkrebs–Chancen der Verhütung durch Screening (DACHS)]. Further details of the original study have been described elsewhere [12, 13].

Eligible patients who were mentally and physically able to participate in a personal interview of ∼1 hour were informed about the study by their clinicians, usually a few days after surgery, and the study center was notified upon receipt of informed consent. Additionally, patients who could not be recruited during their in-patient stay were contacted by mail after discharge by clinicians and clinical cancer registries. All participants gave written informed consent. The study was approved by the ethics committees of the University of Heidelberg and the state medical boards of Baden-Wuerttemberg and Rhineland-Palatinate.

After an average of 3 years since diagnosis, detailed information on colorectal cancer treatment was collected from attending physicians in a standardized questionnaire. The 5-year follow-up was conducted in 2009. After ascertainment of vital status and dates of death through population registries, a QoL questionnaire was sent to all patients alive who had not refused further contact. Nonresponders were mailed up to two reminder letters, which were followed by one reminder phone call in cases of nonresponse to mailings. Patients who did not return the QoL questionnaire were offered a phone interview on the basis of a shortened version of the QoL questionnaire (referred to as short questionnaire; for details, see below). Both questionnaires entailed questions concerning new concomitant diseases or recurrences after the colorectal cancer diagnosis. In cases of reported new diagnoses, including cancer recurrence, attending physicians were contacted to validate the diagnoses. Attending physicians were also contacted when information on recurrence within 5 years after diagnosis was lacking or uncertain.

QoL Questionnaire

QoL was assessed with the Quality of Life Questionnaire Core 30 Items (QLQ-C30) [14] and the tumor specific module QLQ-CR29 [15, 16], both developed by the European Organization for Research and Treatment of Cancer (EORTC). The short questionnaire, which was used in cases when a patient was not willing to complete the full questionnaire, was restricted to the global QLQ-C30 item as the sole QoL measure.

Assessment of Treatment

Information on treatment was collected during the 3-year follow-up and during the 5-year follow-up. For the 3-year follow-up, attending physicians were asked for information on initial treatment of the tumor and, separately, for information on treatment of any local recurrence or distant metastasis after diagnosis. Only information on the initial treatment of the tumor was used to define whether or not a patient received chemotherapy, because the aim of our study was to investigate associations between primary chemotherapy treatment and long-term survival and QoL. For the 5-year follow-up, information on recurrence after diagnosis and, in cases of recurrence, information on treatment was requested from the attending physician. This information was only used to assess how much time had elapsed between the last chemotherapy and the reply to the questionnaire during the 5-year follow-up.

Statistical Methods

The distributions of patient and disease characteristics were compared among responders, nonresponders, short-questionnaire responders, and deceased patients to assess the potential impact of nonrandom missing data. The allocation of chemotherapy with respect to gender, age (≤59 years, 60–69 years, 70–79 years, and ≥80 years), stage, cancer site (colon and rectum), and self-reported medical conditions (myocardial infarction, coronary heart disease, heart failure, stroke, and diabetes) at baseline was descriptively compared. This analysis was done in the cohort of all patients with a diagnosis in 2003 or 2004 to avoid survival bias. In addition, the association between chemotherapy allocation and each of the aforementioned factors was tested in a logistic regression model with adjustment for all other factors.

Cox proportional-hazards regression was employed to estimate the hazard ratio (HR) for chemotherapy-associated survival adjusting for sex, age at follow-up, stage at diagnosis, cancer site, radiotherapy, and presence (yes/no) of any self-reported comorbidities (myocardial infarction, coronary heart disease, heart failure, stroke, and diabetes) at baseline. The same analysis was done in the subgroups of stage II and stage III patients.

For the QoL analysis, scoring of the EORTC QLQ-C30 and EORTC QLQ-CR29 was performed according to the EORTC scoring manual [16, 17]. All scores were linearly transformed to a 0–100 scale. In cases of missing items, multi-item scores were calculated as the mean of the nonmissing items if at least half of the items from the corresponding scale had been completed. Differences in mean QoL scores >10 points were considered clinically meaningful [18].

The QoL of patients with and without chemotherapy was compared using mean values and standard errors. Analyses of covariance with the factors sex, age at follow-up, stage at diagnosis, cancer site, radiotherapy, chemotherapy, and presence of any self-reported comorbidities at baseline were carried out to test for an independent effect of chemotherapy on QoL. In sensitivity analyses, additional separate adjustments for living with a partner, highest level of education (three levels), recurrence, second cancer, physical activity (assessed by a modified version of the international physical activity questionnaire [19] with the categories low, moderate, and high), and body mass index at follow-up were conducted, but results are not reported because no substantial deviations were found. The same QoL analysis was performed separately for two age groups, using the median age at follow-up (71 years) as the cutoff, and for stage II and stage III patients.

All analyses were performed with SAS software, version 9.2 (SAS Institute Inc., Cary, NC). Statistical significance was defined by p < .05. No multiple comparison corrections were made.

Administration of Chemotherapy

The pattern of chemotherapy administration according to sex, age at diagnosis, stage, cancer type, and comorbidities is shown in Table 2. Adjuvant or neoadjuvant chemotherapy was more often administered for stage III cancer at diagnosis (77%, compared with 25% of stage II patients). The three most frequently administered cytostatics were 5-fluorouracil (n = 257, 93%), folinic acid (n = 214, 77%), and oxaliplatin (n = 36, 13%). Chemotherapy was administered in 71% of patients aged <60 years but in only 20% of patients aged ≥80 years. For both stage II and stage III patients, chemotherapy administration was strongly associated with age, even after adjustment for sex, cancer site, stage, and baseline comorbidities (p = .0007, p < .0001). For stage II patients, the odds of receiving chemotherapy was 4.92 times higher for rectal cancer patients than for colon cancer patients, whereas for stage III cancer, the odds of receiving chemotherapy was 2.50 times higher for colon cancer patients. Sample sizes for the analysis of the association between chemotherapy and comorbidities were small and no significant associations were found. However, the odds for chemotherapy administration were generally lower for patients who had heart failure, stroke, and diabetes but higher for patients with a previous myocardial infarction or prevalent coronary heart disease.

Primary chemotherapy was started, on average, 2 months after diagnosis (standard deviation, 3 months) and was stopped after an average of 6 months (standard deviation, 4 months). Chemotherapeutic treatment was started after >12 month for six patients. Among the patients who provided information during the 5-year follow-up, only one patient with stage II disease and three patients with stage III disease were reported to have received chemotherapy during the last 12 months. In the previous 2 years, two and five responders had received chemotherapy, respectively.

Association Between Survival and Chemotherapy

The vital status could be determined for all except one patient who moved abroad. Of the 561 patients with a known vital status, 178 (32%) patients died prior to or during the 5-year follow-up. After adjustment for confounders, chemotherapy was nonsignificantly associated with a longer overall survival time (HR, 0.80; 95% confidence interval [CI], 0.55–1.17). This trend resulted from a significant survival benefit for stage III patients (HR=0.63, 95% CI: 0.39–0.99) and lack of survival benefit for stage II patients (HR, 1.22; 95% CI, 0.62–2.40). Exclusion of six patients who received chemotherapy in the second year after diagnosis or later did not materially alter the findings.

Follow-up with Respect to QoL After 5 Years

Of the 383 patients who were alive through the 5-year follow-up, 20 were excluded from the survey because they were physically or mentally not able to participate (e.g., because of dementia or hospitalization), 12 patients did not respond to mailings and could not be contacted by phone, eight agreed to participate but did not return the questionnaire, and 16 actively refused to participate. The remaining patients returned a full (n = 312) or short (n = 15) questionnaire, resulting in a response rate of 90%. Full responders were significantly more often male, were younger, were more likely to have had rectal cancer or stage III cancer, were more likely to have received chemotherapy and radiotherapy, and were less likely to have had a recurrence than nonresponders, including short-questionnaire responders (Table 1). The overall QoL of responders who completed a full questionnaire was higher than the QoL of short-questionnaire responders (Δ = 11.06), but after adjustment for age and sex the difference was no longer significant (p = .13).

Association Between QoL and Chemotherapy

Table 3 presents the raw QLQ-C30 and QLQ-CR29 QoL mean scores of survivors stratified by chemotherapy and the difference between the adjusted mean scores of survivors who did and did not receive chemotherapy. After adjustment for age at follow-up, sex, cancer site, stage, radiotherapy, and comorbidities, all functioning scores except the global QoL score were lower and all QLQ-C30 symptom scores except diarrhea were higher in survivors who received chemotherapy, but none of these associations was significant. Of the QLQ-CR29 scores, anxiety, hair loss, and trouble with taste symptom levels were significantly and meaningfully higher in survivors who received chemotherapy.

Results from an age-stratified analysis are presented in Table 4. Survivors who were aged ≤70 years at the 5-year follow-up and received chemotherapy reported significantly lower adjusted physical, role, and cognitive functioning and significantly higher pain, appetite loss, hair loss, and trouble with taste symptom levels than survivors of the same age group who did not receive chemotherapy. Differences were clinically meaningful for all these scales except appetite loss. Differences in QoL according to chemotherapy status were much smaller for survivors who were >70 years old. Among the elderly, only hair loss and dry mouth were significantly and meaningfully associated with chemotherapy.

Results from a stage-stratified analysis suggest that chemotherapy was associated with better QoL in stage III colorectal cancer patients, whereas an opposite association was observed for stage II patients (Table 5). For example, concern about weight was significantly stronger in stage II patients who received chemotherapy, whereas stage III patients who received chemotherapy had tentatively fewer concerns than patients who did not receive chemotherapy. This pattern (i.e., higher QoL for stage III patients who underwent chemotherapy but higher QoL for stage II patients who did not) was stronger for younger patients, patients with colon cancer, and patients not undergoing radiotherapy (data not shown). In contrast, little or no difference in QoL outcomes between stage II and stage III patients with and without chemotherapy administration was observed for older patients, patients with rectal cancer, and patients undergoing radiotherapy (data not shown). However, because of small sample sizes, results of this subgroup analysis must be interpreted with caution.