The following institutions participated in this study: CALGB Statistical Center, Durham, NC (Stephen George, Ph.D. [supported by CA33601]); Dana Farber Cancer Institute, Boston, MA (George P. Canellos, M.D. [supported by CA32291]); Dartmouth Medical School-Norris Cotton Cancer Center, Lebanon, NH (Marc Ernstoff, M.D.; [supported by CA04326]); Duke University Medical Center, Durham, NC (Jeffrey Crawford, M.D. [supported by CA47577]); McGill Cancer Center, Montreal, Quebec, Canada (Brian Leyland-Jones, M.D. [supported by CA31809]); Mount Sinai School of Medicine, New York, NY (Lewis Silverman, M.D. [supported by CA04457]); North Shore-Long Island Jewish Medical Center, Manhasset, NY (Daniel R. Budman, M.D. [supported by CA35279]); Rhode Island Hospital, Providence, RI (William Sikov, M.D. [supported by CA08025]); Roswell Park Cancer Institute, Buffalo, NY (Ellis Levine, M.D. [supported by CA02599]); State University of New York-Upstate Medical University, Syracuse, NY (Stephen L. Graziano, M.D. [supported by CA21060]); University of Alabama Birmingham, Birmingham, AL (Robert Diasio, M.D. [supported by CA47545]); University of California at San Diego, San Diego, CA (Stephen Seagren, M.D. [supported by CA11789]); University of California at San Francisco, San Francisco, CA (Alan Venook, M.D. [supported by CA60138]); University of Chicago Medical Center, Chicago, IL (Gini Fleming, M.D. [supported by CA41287]); University of Iowa, Iowa City, IA (Gerald Clamon, M.D. [supported by CA47642]); University of Maryland Cancer Center, Baltimore, MD (David Van Echo, M.D. [supported by CA31983]); University of Massachusetts Medical Center, Worcester, MA (Pankaj Bhargava, M.D. [supported by CA37135]); University of Minnesota, Minneapolis, MN (Bruce A Peterson, M.D. [supported by CA16450]); University of North Carolina at Chapel Hill, Chapel Hill, NC (Thomas C. Shea, M.D. [supported by CA47559]); Wake Forest University School of Medicine, Winston-Salem, NC (David D. Hurd, M.D. [supported by CA03927]); Washington University School of Medicine, St. Louis, MO (Nancy Bartlett, M.D. [supported by CA77440]); and Weill Medical College of Cornell University, New York, NY (Scott Wadler, M.D. [supported by CA07968]).
Presented as Abstract 1593 at the annual meeting of the American Society of Clinical Oncology, Atlanta, Georgia, May 15–19, 1999.
The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute.
The objective of this study was to compare the quality of life (QOL) after treatment among patients who had breast carcinoma with multiple positive lymph nodes. The patients were randomized to receive either high-dose chemotherapy with autologous stem cell support (HDC) or intermediate-dose chemotherapy (IDC) in the adjuvant setting.
Two hundred forty-six patients with AJCC Stage IIA, IIB, or IIIA breast carcinoma who had ≥ 10 positive lymph nodes and who were participants in Cancer and Leukemia Group B (CALGB) 9082 were enrolled in this companion study, CALGB 9066. Patients were randomized to receive either high-dose cyclophosphamide, carmustine, and cisplatin (CPA/cDDP/BCNU) and autologous bone marrow transplantation (the HDC arm) or intermediate-dose CPA/cDDP/BCNU as consolidation to adjuvant chemotherapy (the IDC arm). QOL was assessed at baseline and at 3 months, 12 months, 24 months, and 36 months using the Functional Living Index-Cancer (FLIC), the Psychosocial Adjustment to Illness Scale (PAIS)-Self Report, and the McCorkle Symptom Distress Scale (SDS).
At the 3-month assessment, patients in the HDC arm demonstrated significant worsening of QOL compared with the IDC arm in terms of their physical well being (FLIC, P = 0.023), social functioning (FLIC, P = 0.026; PAIS, P < 0.0001), symptom distress (SDS, P = 0.0002), and total QOL scores (FLIC, P = 0.042). At 12 months, the differences in QOL scores between the HDC arm and the IDC arm had resolved.
Patients with locally advanced breast carcinoma that involves ≥ 10 axillary lymph nodes face a particularly high rate of recurrence despite appropriate local and systemic therapy.1–3 Given the relatively poor outcomes noted in this patient subgroup, it has been viewed as an appropriate patient population for new and/or more aggressive treatment approaches. Initial exploration of high-dose chemotherapy with autologous bone marrow support (HDC) for women with breast carcinoma was based on the premise that increasing the intensity of chemotherapy would increase the percentage of tumor cells killed, thereby leading to improved outcomes. Early studies of HDC in patients with metastatic disease and in women who had breast carcinoma with multiple positive lymph nodes generated results that appeared to be superior to historic controls.4–7 Although the preliminary results seen with HDC were promising, it was recognized that the benefits suggested by uncontrolled studies may have been due to selection bias rather than the treatment itself.8 To determine whether HDC was effective in the high-risk setting, the Cancer and Leukemia Group B (CALGB) initiated a randomized control trial (CALGB 9082) of HDC versus intermediate-dose chemotherapy (IDC) as adjuvant therapy for patients who had breast carcinoma with ≥ 10 positive lymph nodes.9
Many of the toxicities from chemotherapy are dose dependent. In CALGB 9082, because it was understood that high-dose therapy requiring autologous stem cell transplantation support caused greater toxicity than lower dose regimens, it was expected that short-term quality of life (QOL) would be compromised. What was not clear when the study was initiated was whether any survival benefit due to HDC may be offset by a significant and sustained worsening of QOL among patients who received this treatment. The question was particularly important in the adjuvant setting, in which therapy would be given to otherwise healthy patients, albeit at high risk of recurrent disease. The current study, CALGB 9066, was initiated as a companion study to evaluate QOL after the completion of therapy among participants in CALGB 9082.
MATERIALS AND METHODS
All participants in the QOL component of the CALGB clinical trial 9082 had a diagnosis of AJCC Stage IIA, IIB, or IIIA breast carcinoma with ≥ 10 positive lymph nodes. Participants were age 18 years or older, had a Karnofsky performance status ≥ 80%, and had not received prior chemotherapy or radiation therapy. After informed consent was obtained, patients were randomized at the CALGB Statistical Center to receive either high-dose cyclophosphamide, carmustine, and cisplatin (CPA/cDDP/BCNU) and autologous bone marrow transplantation or intermediate-dose CPA/cDDP/BCNU as consolidation to adjuvant cyclophosphamide, doxorubicin, and 5-fluorouracil (CAF). All patients who participated in the QOL component were enrolled on CALGB 9082 and were required to speak English, because most of the measures had been developed and validated only in English. Initially, patients were not required to participate in the QOL study if they had been treated on CALGB 9082. However, after 3 months of accrual, this protocol was amended and, as of September 1992 until the closure of CALGB 9066 (on reaching the accrual goal), all English-speaking patients who were enrolled in the clinical trial (CALGB 9082) were required to participate in the current study, so that the QOL findings would be a valid reflection of the impact of treatment on the QOL of a consecutively enrolled cohort. The QOL protocol was approved by the Institutional Review Boards of the participating institutions, and all enrolled patients provided informed consent.
Timing and Method of Assessment
Five QOL assessments were scheduled over 3 years: at baseline and at 3 months, 12 months, 24 months, and 36 months after the completion of therapy. QOL assessments were discontinued when patients either developed recurrent disease or withdrew from the study.
Patients were given a QOL questionnaire at the time of entry to the clinical trial with a request to complete the baseline assessment in the clinic. Upon completion of the questionnaire, patients returned it to the clinical research associate. Data from the questionnaire were then mailed to the CALGB Statistical Center for data entry. For all subsequent assessments, the research interviewer, who was located at Duke University Medical Center, called the patient to schedule a time to conduct a QOL research interview over the telephone. The QOL questionnaire was then mailed to patients along with a request to complete it prior to the time of the interview. Patients' answers to the questionnaires were collected during the telephone interview. Interviews lasted approximately 30 minutes. Centralized telephone interviews to collect QOL data had been used successfully in numerous studies within the CALGB that involved patients in active treatment as well as cancer survivors.10
Functional Living Index-Cancer
The Functional Living Index-Cancer (FLIC)11, 12 is a widely used, disease-specific, patient-rated QOL scale that contains some of the most sensitively worded items related to cancer's impact on QOL of all current measures. It is a 22-item, linear analogue scale, with a 7-point Likert format superimposed on the rating line, and with each end of the item anchored specifically in relation to that item. The FLIC consists of five factors: Physical Functioning, Psychological Functioning, Social Functioning, Gastrointestinal Symptoms, and Current Well-Being.13 The FLIC total score and all factor scores were calculated by summing across all items, with higher scores indicating a better QOL. Norms for the FLIC are available through the World Health Organization Collaborating Center for Quality of Life in Cancer Care.
Psychosocial Adjustment to Illness Scale-Self Report
The Psychosocial Adjustment to Illness Scale (PAIS)-Self Report14 is a self-report, 45-item inventory of 7 dimensions of adaptation to illness: health beliefs and patient satisfaction with medical care; vocational, domestic, and social functioning; sexual relationships; extended family relationships; and psychologic distress. Items are rated along a four-point scale, with increasing scores denoting worse adaptation.
McCorkle Symptom Distress Scale
The McCorkle Symptom Distress Scale (SDS)15 is a 13-item scale that was designed to measure common physical and emotional symptoms experienced by patients with cancer who are receiving chemotherapy and radiation therapy. Each item is rated on a five-point Likert scale, with each point on the scale anchored specifically in relation to the symptom. The SDS total score is the sum across all items, ranging from 13 to 65, with higher scores indicative of greater symptom distress.
Basic sociodemographic information was obtained using the CALGB Background Information Form,16 which consists of eight questions concerning education, employment, occupation, marital status, household composition, health insurance, and the number of children who live with the patient. Data concerning age and race were obtained from the registration form at the time of accrual to the study. Menopausal status and history of toxicity ≥ NCI CDC Grade 3 from initial cycles of chemotherapy prior to randomization to HDC or IDC was obtained from data that were collected for CALGB 9082.
The current study was designed as a QOL outcomes assessment of a prospective, nonblinded, randomized controlled trial of HDC versus IDC for patients with high-risk breast carcinoma. Two primary treatment comparisons were of interest: a comparison of FLIC scores at 3 months and a comparison of the same scores at 12 months. Previous research at Duke University had shown that the standard deviation of FLIC scores is 19.1.17 To detect a treatment difference of 10 at each time point with 80% power using a test conducted at the 0.025 level of significance (2-tailed), 76 patients with follow-up measurements were needed in each treatment group. Assuming an attrition rate during the first year of 25%, an accrual objective of 204 patients was set.
Analyses of variance with repeated measures were used to test for differences in adjustment between the HDC group and the IDC group, changes in adjustment over time, and the presence of an interaction between treatment and time. The latter was an assessment of whether treatment differences in adjustment were consistent over time. The primary hypothesis was that the major differences in adjustment between treatment arms would occur at 3 months after treatment with no significant differences between arms by 12 months after treatment. Follow-up assessments also were conducted at 24 months and 36 months after treatment. Analyses focused on patients who had at least one postbaseline assessment.
The repeated-measures analysis initially was stratified by the point of the last QOL assessment to assess the impact of dropout patterns on efficacy outcome. Because the analyses focused on the first year posttreatment, there were 3 stratification groups: Subgroup 1 involved patients who were assessed only at baseline (excluded from analyses), Subgroup 2 involved patients who were assessed at baseline and at 3 months after treatment, and Subgroup 3 involved patients who were assessed from baseline through at least 12 months after treatment. It was found that, qualitatively, the inferences associated with the stratified and unstratified analyses were the same. Therefore, this report focuses on the experiences of all patients who provided follow-up measurements using repeated-measures analysis without stratification.
Univariate and multiple regression analyses were conducted to determine whether patient characteristics, including QOL variables, were predictors of QOL at 3 months and 6 months. All data were entered into the CALGB data base. The statistical analysis was performed at the CALGB Statistical Center.
Between December 1992 and September 1994, 246 of 885 patients (28%) who were enrolled in the treatment trial (CALGB 9082) were enrolled in the QOL study (CALGB 9066). Thirty-six patients were excluded from the final analysis, including 24 patients who dropped out of the associated clinical study prior to randomization, 2 patients who withdrew from study participation without receiving any protocol therapy, 2 patients in the HDC group who were not participating in the randomized component of CALGB 9082 and were enrolled in CALGB 9066 in error, and 8 patients who did not complete any of the QOL assessments. Two hundred ten patients were evaluable and completed at least 1 of the QOL assessments in this study. Baseline characteristics of patients from CALGB 9082 who participated in CALGB 9066 were comparable in terms of age, race, disease stage, receptor status, menopausal status, number of positive lymph nodes, and tumor size to patients who did not participate. Characteristics of the evaluable patients enrolled in CALGB 9066 are presented in Table 1 by treatment arm. Baseline sociodemographic and medical characteristics were similar for the two treatment groups.
Fifteen patients were enrolled in CALGB 9066 after they had started treatment in CALGB 9082 and therefore were excluded from the baseline assessment. The baseline assessment was completed for 93% of evaluable patients. The numbers and percentages of patients withdrawn or deceased prior to each assessment are presented in Table 2. QOL assessments at 3 months, 1 year, 2 years, and 3 years were completed by 90%, 78%, 52%, and 69% of surviving patients, respectively. The completion of assessments was similar between treatment groups.
Table 2. Compliance with Quality of Life Assessmentsa
There were 106 patients in the high-dose chemotherapy group and 104 patients in the intermediate-dose chemotherapy group.
Withdrawn and alive
Pooled data from patients who completed at least 1 postbaseline assessment are summarized in Table 3. At the 3-month assessment, patients in the HDC arm demonstrated significant worsening of QOL compared with patients in the IDC arm in terms of physical well-being (FLIC; P = 0.023) (Fig. 1), social functioning (FLIC; P = 0.026) (Fig. 2) (PAIS; P < 0.0001) (Fig. 3), symptom distress (SDS; P = 0.0002) (Fig. 4), and total QOL scores (FLIC; P = 0.042) (Fig. 5).
Table 3. Quality of Life Scores According to Treatment Arm and Time of Assessmenta
Assessment: mean score (SE)
SE: standard error; HDC: high-dose chemotherapy arm; IDC: intermediate-dose chemotherapy arm; FLIC: the Functional Living Index-Cancer; SDS: the McCorkle Symptom Distress Scale; PAIS: the Psychosocial Adjustment to Illness Scale.
Mean scores and standard errors are reported for patients who had at least one postbaseline assessment of quality of life.
An increase in the Functional Living Index-Cancer score is associated with improved quality of life.
Significantly worse for the high-dose group versus the intermediate-dose group at 3 months (P < 0.05).
Significantly improved from baseline scores for both treatment arms (P < 0.001).
An increase in the McCorkle Symptom Distress Scale score is associated with worse symptoms.
Significantly worse for the high-dose group versus the intermediate-dose group at 3 months (P < 0.0005).
Significantly improved from baseline scores for both treatment arms (P ≤ 0.05).
An increase in the Psychosocial Adjustment to Illness Scale score is associated with worse functioning.
At 12 months, the differences in QOL scores between the HDC and IDC groups had resolved. In addition, scores for physical well-being (FLIC; P < 0.0001), social functioning (FLIC; P = 0.001), symptom distress (SDS; P = 0.037), and total QOL (FLIC; P < 0.0001) indicated significant improvement for patients in both groups at the 12-month assessment compared with baseline scores.
Family disruption scores (FLIC) improved significantly in both treatment groups between the baseline assessment and the 3-month assessment (P < 0.0001) and continued to improve between the 3-month and 12-month assessments. By contrast, sexual subscale scores (PAIS) were significantly worse in both treatment groups at 3 months compared with baseline scores (P < 0.002) and remained worse than baseline despite a slight improvement from nadir in both groups at 12 months. Although the differences in emotional functioning scores (FLIC) and vocational subscale scores (PAIS) were not significant between baseline and 3 months or between 3 months and 12 months, there was a significant improvement in these areas when comparing baseline scores with the 12-month assessment (FLIC, P < 0.0001; PAIS, P < 0.0001).
The degree to which particular symptoms distressed patients in both treatment arms at baseline, 3 month, and 12 month assessments is summarized in Table 4. Based on the number of patients who scored ≥ 3 on the 5-point SDS, the most common symptom in both treatment groups was fatigue. Moderate-to-severe fatigue (defined as a score ≥ 3) was reported by 61% of patients in the HDC group at 3 months and by 47% of patients in the IDC group. By 12 months, 34% of patients in the HDC group reported frequent fatigue compared with 32% of patients in the IDC group. The only other symptoms with scores ≥ 3 among > 33% of patients were appetite (at 3 mos for the HDC arm; improved by 12 mos), insomnia (in both arms at 3 mos; improved at 12 mos), and outlook (in both arms at baseline; improved by 12 mos).
Table 4. Number and Percentage of Patients with Scores ≥ 3 on the McCorkle Symptom Distress Scale
Table 2 shows that the proportion of patients completing assessments at 24 months and 36 months declined to < 60% of the initial sample, making these results subject to greater bias as a result of drop-out. However, in general, the trends established at 12 months persisted at the 2-year and 3-year assessments. In the majority of patients, differences seen between treatment arms at 3 months had resolved by 12 months, and the QOL in the 2 arms remained similar over time. In a comparison of the last QOL assessments among patients who participated in the QOL survey for at least 24 months, the only significant differences between treatment arms were noted in the PAIS vocational subscale (HDC vs. IDC, 3.46 vs. 2.17; P = 0.028), the PAIS social subscale (HDC vs. IDC, 3.63 vs. 2.05; P = 0.015), and the PAIS total score (HDC vs. IDC, 29.55 vs. 22.50; P = 0.029). This finding suggests that patients in the HDC arm who survived for > 2 years after treatment had not adjusted as well, particularly in terms of vocational and social functioning, compared with patients in the IDC arm. At no assessment for any QOL subscale did patients in the HDC arm report significantly better QOL than patients in the IDC arm.
Predictors of QOL
A multivariate analysis was performed to determine whether, after they were adjustment for standard clinical prognostic factors, including CALGB 9082 treatment arm, baseline characteristics or QOL scores were predictive of QOL scores after treatment. Predictors included age, marital status, education, race, employment, income, menopausal status, FLIC total score at baseline, SDS total score at baseline, and history of toxicity ≥ Grade 3 from initial cycles of chemotherapy prior to randomization to HDC or IDC.
In general, QOL at baseline predicted QOL at 3 months and 12 months after treatment. The strongest predictor of QOL (FLIC or PAIS) at 3 months and 12 months (FLIC only) was QOL at baseline, as measured by the FLIC total score (FLIC, P < 0.0001; PAIS, P < 0.001). A backward elimination procedure was used to identify a multivariate model that predicted PAIS scores at 12 months. Included in the final model and associated with lower PAIS scores were white race (P = 0.007), postmenopausal status (P = 0.022), and the baseline measure of FLIC (P = 0.017).
The primary objective of this study was to determine whether QOL was compromised in the months and years after treatment for those patients who received HDC versus less intensive chemotherapy. We found that QOL was compromised only transiently among patients in the HDC arm, with no difference between groups by 12 months after therapy. Most of the differences at 3 months were due to declines in social functioning, physical well being, and increases in symptom distress in the HDC arm. These declines likely were due to increased toxicity in the HDC arm. Despite these transient declines in QOL, patients in both treatment arms reported improved total QOL and total symptom distress at 12 months compared with baseline. Follow-up at 2 years and 3 years after treatment generally confirmed these findings.
Patients who participated in this QOL analysis were representative of those who participated in CALGB 9082, the randomized trial of HDC versus IDC, and baseline characteristics between the patients in both arms were well balanced. Completion of the study by 90% of living patients at 3 months and by 78% of living patients at 12 months was comparable to percentages reported by other studies of QOL in patients with cancer.18, 19
Since the completion of CALGB 9082 and CALGB 9066, several large randomized trials have demonstrated no benefit to HDC in the treatment of patients with high-risk breast carcinoma.20–23 Randomized trials also have shown no survival benefit to HDC in the treatment of patients with advanced and metastatic disease.24, 25 This body of evidence suggests there is no role for HDC in the treatment of breast carcinoma outside of a clinical trial. Although enthusiasm for HDC in women with breast carcinoma has fallen off sharply, it remains possible that HDC may have value in selected patients.26, 27 For this reason, and to augment our understanding of the impact of this therapy on the thousands of patients who were treated within and outside of clinical trials,6 and because high-dose chemotherapy is used in the treatment of other malignancies, a better understanding of the long-term effects of this treatment on QOL remains important.
Brandberg et al. have published what to our knowledge is the only other experience evaluating changes in QOL over time among women who were treated in a randomized controlled trial of HDC for high-risk breast carcinoma. The Scandinavian Breast Group Study 9401 randomized women with high-risk breast carcinoma to receive either HDC or 5-fluorouracil, epirubicin, and cyclophosphamide (FEC) tailored to hematologic effects. QOL was assessed using the European Organization for Research and Treatment of Cancer (EORTC) QLQ-C30 instrument and the EORTC breast cancer subscale (BR-23), which measure dimensions of QOL similar to those measured by the instruments used in the current study. Unlike CALGB 9066, assessments were performed during treatment as well as at subsequent intervals of 4–12 weeks, with the eighth and final assessment performed at 1 year after randomization. Not surprisingly, those authors reported a significant decline in many aspects of QOL during therapy, with a more rapid decline and earlier recovery among patients who were treated with the more intensive, but briefer, HDC. Similar to our findings, at 1 year after randomization, there were no significant differences between the 2 treatment arms for any QOL parameter. Brandberg et al. found that emotional functioning improved in both groups over time, as it did among our patients between baseline and 12 months. In addition, they found that fatigue was pronounced during treatment but approached baseline levels by 1 year, as it did among both treatment groups in our cohort.28
The predominant symptom reported by patients in both arms of our study was fatigue. This symptom peaked in both groups at 3 months and essentially returned to baseline by 12 months. Hann et al. compared QOL and the experience of fatigue among 31 patients with high-risk or advanced breast who were undergoing HDC with concurrent survey data from a group of 49 healthy controls. Patients who were undergoing HDC, as suspected, experienced significant fatigue and worsening of QOL during treatment compared with the healthy control group. This study documented that fatigue was correlated strongly with depression (correlation coefficient [r] = 0.77; P < 0.001), with the time to engraftment (r = 0.41; P < 0.05), and with the length of hospital stay (r = 0.5; P < 0.01). Those authors did not examine longitudinal changes in fatigue and QOL after HDC.29 Our results, along with others,28, 30 provide evidence that these declines may be transient. However, it is important to note that, even at baseline, nearly 25% of patients in both treatment groups reported significant fatigue. Future research should be directed at efforts to alleviate this symptom and to distinguish fatigue from insomnia, which may have different treatment implications.
Several recent studies have evaluated QOL in the context of intensive treatment programs for breast carcinoma other than HDC. Therasse et al. evaluated QOL among patients with locally advanced breast carcinoma who were randomized to receive either FEC or higher dose epirubicin and cyclophosphamide in a dose-dense fashion. QOL (according to the EORTC QLQ C30) was worse among the higher intensity arm during the first 3 months of treatment but had returned to baseline by 6 months. At 1-year follow-up, there was no difference in QOL between treatment arms.31 Similarly, Del Mastro et al. documented a transient increase in psychologic distress (Psychological Distress Inventory)32 during treatment among patients who were randomized to receive either dose-dense FEC or standard FEC, but those authors observed no difference between the groups at 6 months and 12 months after treatment; to date, the clinical outcomes of that randomized trial have not been published.18 The results from those studies support our finding that more intensive therapy may result in transient declines in QOL, but minimal differences are seen in long-term follow-up compared with less intensive therapies.
The current study had several limitations. Although many well validated instruments to assess QOL have been established, the variability between instruments continues to make comparisons difficult across studies, rendering the interpretation of results more difficult. The rate of compliance with assessments through 12 months in the current study compared favorably with similar studies; however, a nonresponse rate of 27% at 1 year may have introduced significant bias if patients differentially were likely to respond based on their treatment arm. However, response rates were comparable between treatment arms. Finally, the current analysis involved multiple comparisons, increasing the chance for random correlations or differences. The general conclusions of the overall QOL analysis based on the FLIC were supported by results from the SDS, which demonstrated a high degree of statistical significance.
HDC continues to be used widely in the treatment of hematologic malignancies, such as lymphoma and multiple myeloma, and is the subject of ongoing clinical trials in patients with brain tumors, germ cell tumors, and sarcomas. Understanding the QOL among patients with breast carcinoma after HDC may serve as a predictor and comparison point for QOL after HDC in these other settings as well. The extent to which our data may be generalized to the QOL experiences of patients with other tumors undergoing similarly intensive but different regimens is not known. Our model suggests that baseline QOL may be the primary determinant of QOL after intensive treatment. This finding should be explored among patients undergoing HDC for other indications, because it may guide interventions that are designed to enhance QOL in these settings.
The findings of the current study indicate that adjuvant treatment with HDC leads to a transient decline in some aspects of QOL compared with IDC but results in no significant difference in QOL between therapies by 12 months. Many aspects of QOL improve significantly by 1 year after therapy. These findings are consistent with those reported in a similar QOL assessment of patients in a randomized controlled trial of HDC for patients with high-risk breast carcinoma.28 Furthermore, the current study results are consistent with studies to date comparing QOL among patients in dose-dense therapy trials, because those studies have shown transient declines in short-term QOL that are more pronounced with more intensive therapies.18, 31
Although the current study would have greater clinical relevance if HDC had proven beneficial for breast carcinoma, it remains important to document this experience both as a contribution to our general understanding of the trade-offs involved in more toxic therapies and out of an obligation to share the knowledge gained from the patients who generously volunteered to become research participants to contribute to the care of future patients. This study also demonstrates the feasibility of collecting longitudinal QOL data in large, cooperative group trials, even with intensive therapies. Longitudinal data can aid greatly in our understanding of the way novel treatments impact patient's lives and the durability of any changes in QOL.
The authors are grateful to the patients who participated in this study and to the participating institutions that made this work possible. They thank Carol Brunatti and K. Warren for their assistance with conducting this study.