Multiple myeloma (MM) is the most common indication for high-dose chemotherapy with autologous stem cell transplantation (ASCT) in the U.S. and can be associated with substantial morbidity. Thorough assessment and understanding of symptoms and risk factors for symptom development after ASCT are logical first steps toward developing strategies aimed at reducing the symptom burden associated with this procedure.
The authors performed a prospective evaluation of symptom burden among 64 patients with myeloma who underwent ASCT. Symptom data were collected using the M. D. Anderson Symptom Inventory (MDASI) at 4 time points: baseline, the day of stem cell infusion (Day 0), nadir of counts, and Day 30. Univariate analysis was performed to correlate pretransplantation variables with post-transplantation symptom burden at these time points.
MDASI scores increased significantly throughout transplantation, with most patients returning to baseline by Day 30 after the procedure. Patients with the highest MDASI scores at baseline had the highest MDASI scores at nadir (P = .02). Patients with prolonged time to transplantation and women had a trend toward higher nadir global symptom severity scores. These groups, as well as patients aged >60 years, had a trend toward higher nadir interference scores.
Multiple myeloma (MM) is a malignant plasma cell disorder that comprises approximately 1% of all malignant diseases in Caucasian Americans and 2% of all malignant diseases in African Americans.1 Although several studies have demonstrated an improvement in complete response rates and overall survival with the use of high-dose chemotherapy with autologous stem cell transplantation (ASCT), both the disease and its treatments still are associated with substantial morbidity and mortality.2–6 Toxicities are common during the acute phase of transplantation, and many patients experience pain, fever, mucositis, nausea, vomiting, diarrhea, organ toxicity, psychiatric sleep disturbance, and infectious complications.7–9
Fermand et al. reported that, despite the toxicities of ASCT, health-related quality of life, as measured by time without symptoms, treatment, and treatment toxicity (TwiSST), was significantly better in patients who underwent ASCT than in patients who received conventional chemotherapy (25.1 months vs 16.6 months; P = .03).10 Their results were corroborated in a meta-analysis by Levy et al. of 3 randomized French trials (French Myeloma Intergroup 90, Myeloma Autograph Group 90 [MAG 90], and MAG 91) that compared high-dose therapy (HDT) and ASCT with conventional chemotherapy. In that meta-analysis, a mean gain of 14.5 months (95% confidence interval [95% CI], 9.9–19.1 months) in TwiSST was observed in patients who received HDT plus transplantation compared with patients who received conventional chemotherapy.11 Notwithstanding those results, patients who undergo ASCT suffer significant and prolonged increases in multiple symptoms after transplantation, including sleep disturbance, sexual difficulties, and poorer physical, psychological, and social functioning, that interfere with activities of daily living and result in significant morbidity.12–21 Thus, to minimize the morbidity associated with transplantation and to optimize the gain in health-related quality of life that transplantation offers, interventions designed at reducing the symptom burden associated with ASCT need to be investigated aggressively.
Designing and developing clinical trials with the objective of reducing symptom burden during the acute phase of ASCT would benefit from the ability to identify the patient population most at risk for developing severe symptoms after ASCT. To our knowledge, this study is one of the first attempts to prospectively examine symptom burden during the first 30 days after transplantation in patients with myeloma who underwent ASCT and to evaluate possible pretransplantation predictors of symptom burden, including patient demographics, disease status, treatment, and comorbidity.
MATERIALS AND METHODS
Patients who received high-dose chemotherapy followed by ASCT for myeloma, non-Hodgkin lymphoma, and breast cancer at the University of Texas M. D. Anderson Cancer Center between June 2000 and May 2003 were considered eligible for, and were offered enrollment on, a study that would prospectively evaluate symptom burden in patients undergoing ASCT. Patients who were considered eligible for the study were aged ≥18 years, English-speaking, had hearing and vision adequate enough to use the interactive voice response (IVR) system, and had cognitive ability to complete the assessment measures. One hundred fifty-seven eligible patients were identified and approached regarding the study, and 109 patients (69%) agreed to participate and provided written informed consent. Of these, 64 patients were undergoing transplantation for myeloma and are the subject of this analysis. Basic demographic data were collected on all eligible patients; no statistically significant differences in sex, race, age, or diagnosis were observed between those who consented and those who did not consent to the study. All patients were treated on institutionally approved protocols and had signed separate written informed consent for their treatment and for this study. Approval for both the prospective and retrospective data collection and analysis for this study was obtained from the institutional review board.
The M. D. Anderson Symptom Inventory
The M. D. Anderson Symptom Inventory (MDASI) was developed as a brief but comprehensive means of measuring the severity of symptom burden in cancer patients and the impact that those symptoms have on various aspects of patients' lives.22 The MDASI is a questionnaire composed of symptoms and interference items and can be administered in paper format, by personal interview, or by telephone (IVR system). Patients rate the severity of each symptom and interference item on a standard scale from 0 to 10, with 0 indicating absence of the symptom and 10 indicating that the symptom is “as bad as you can imagine.” Symptoms can be analyzed individually or as a group, with the mean of the symptoms reflecting a global symptom severity (GSS) score. The interference items indicate how much a patient's symptoms interfered with various aspects of daily life in the previous 24 hours. The mean of the interference items is reported as an interference score.
Specific MDASI modules have been created to address symptoms particular to some diseases and their treatments; the MDASI-Blood and Marrow Transplantation (BMT) module was used for the current study. The MDASI-BMT assesses 14 symptoms (pain, fatigue, weakness, nausea, diarrhea, mouth sores, shortness of breath, lack of appetite, feeling physically sick, disturbed sleep, difficulty paying attention, sadness, distress, and bleeding) and 5 interference factors (activity, mood, walking, relationships with other individuals, and enjoyment of life). In previous studies, the MDASI-BMT demonstrated good internal reliability, with coefficient α values of .86 and .82 at baseline assessment for the symptom and interference scales, respectively, and with coefficient α values of .91 and .90 at nadir assessment for the symptom and interference scales, respectively.22–25
We performed prospective evaluation of symptom burden using the MDASI-BMT paper questionnaire at 4 time points: baseline (from Day −36 to Day −3), Day 0 (the day of stem cell infusion), at the nadir of counts (individualized, as indicated by the treating team), and at Day +30 after transplantation. We performed a retrospective review of the M. D. Anderson Cancer Center Department of Blood and Marrow Transplantation clinical database and a chart review to obtain pretransplantation variables, including patient demographics, performance status, preparative regimen, disease status, and comorbidities. The variables that were chosen for this analysis are summarized in Table 1. The Charlson Comorbidity Index (CCI) was calculated for all patients. The CCI was described originally by Charlson et al. in 1987, has been used extensively and validated in cancer populations, and has been used successfully in helping to predict outcomes in stem cell transplantation populations.26–30
Other disease status included recurrent sensitive (6 patients), recurrent untreated (1 patient), and recurrent refractory (1 patient).
Other preparative regimens included combined thiotepa, melphalan, and cyclophosphamide (6 patients); holmium and melphalan (2 patients); and combined thiotepa, busulfan, and cyclophosphamide (1 patient).
Other race included Hispanic (3 patients), Native American (1 patient), and other (1 patient).
Complete MDASI questionnaire data from all 4 time points were available for analysis on 56 of the 64 patients. Reasons for lack of response included patient illness and inability to contact patients who already had returned home before the time of the fourth assessment.
Distribution of MDASI scores at baseline, on Day 0, at nadir, and on Day 30 after transplantation were compared by using the Wilcoxon's rank-sum paired test. Predictors of nadir MDASI scores higher than the median (>3.5) were evaluated by using logistic regression analysis. Actuarial survival was estimated by using the Kaplan-Meier method and was compared on univariate analysis by using a Cox proportional hazards model. Statistical significance was determined at the .05 level. Analysis was performed using STATA software (version 7.0).
Patient and Treatment Characteristics
Sixty-four patients were enrolled on study between June 2000 and May 2003. Patient characteristics and demographics are summarized in Table 1. The median follow-up for the patients who remained alive at time of analysis was 37 months (range, 5–60 months), and 65% of patients still were alive at the time of analysis. In brief, the mean age of the patients at the time of transplantation was 55.2 years (range, 30–74 years), and the mean time from diagnosis of myeloma to transplantation was 7.5 months (range, 2.5–73.3 months). Ninety-four percent of patients had Durie-Salmon stage II or III disease. At the time of transplantation, 65.6% of patients were in their first remission (complete or partial), 21.9% of patients had primary refractory disease, and 13.6% of patients had recurrent disease (6 had recurrent sensitive disease, 1 had recurrent untreated disease, and 1 had recurrent refractory disease). Most patients (85.9%) received melphalan at a dose of 200 mg/m2 as their preparative regimen. The remaining patients (14.1%) were treated with combination high-dose chemotherapy regimens, including 6 patients who received topotecan, melphalan, and cyclophosphamide31; 2 patients who received holmium and melphalan32; and 1 patient who received thiotepa, busulfan, and cyclophosphamide.33 All patients received peripheral blood stem cell rescue with a median CD34-positive infused cell dose of 5.32 × 106 cells (range, 2.34–12.73 × 106). Other patient characteristics are summarized in Table 1. All patients received supportive care after ASCT in accordance with departmental guidelines. Consolidation or maintenance therapy was not started during the confines of this study.
Patients' symptom burden, as measured by the GSS and interference scores, increased from baseline (median GSS score, 1.25; range, 0–5.8), to Day 0 (median GSS score, 2.25; range, 0–7.8), to nadir of counts (median GSS score, 3.5; range, 0.6–8.5), with the majority of patients returning to their baseline symptom burden by Day 30 after transplantation (median GSS score, 1.4; range, 0–4.7) (see Figs. 1 and 2). Differences between symptom severity scores and interference scores between baseline and Day 0 and between baseline and nadir were significant despite broad ranges in patients' reported symptom scores (P < .001 for both GSS and interference scores). There was no statistical difference between patients' baseline GSS and interference scores and their Day 30 scores.
Mean GSS scores peaked at nadir of counts. The individual symptoms that were reported as most severe were fatigue (median score, 5.5), weakness (median score, 5.3), lack of appetite (median score, 5.2), feeling sick (median score, 4.5), nausea (median score, 4.1), difficulty sleeping (median score, 3.9), and diarrhea (median score, 3.7). Interference scores demonstrated that these symptoms significantly affected patients' activity level (median score, 5.3), ability to walk (median score, 4.2), and overall enjoyment of life (median score, 4.5). Responses ranged from 0 to 10 in all categories. The mean individual symptom intensity scores and interference intensity scores are reported in Table 2.
Table 2. Trends in M. D. Anderson Symptom Inventory Individual Symptom and Interference Scores Over the First 30 Days After Transplantation
Although we noted that the overall trend for symptoms and interference scores worsened with HDT until nadir and then return to baseline at 30 days post-transplantation, this pattern was not consistent across the individual symptoms and interference scores. Instead, we noted that trends in individual symptoms over the first 30 days clustered into 4 groups: 1) symptoms most patients reported as the same or better at Day 30 compared with baseline, 2) symptoms most patients reported as the same or worse at Day 30 compared with baseline, 3) symptoms reported equally as better or worse at Day 30 compared with baseline, and 4) symptoms infrequently reported over the first 30 days of transplantation. For symptoms and interference scores that were described previously as the most distressing at nadir (with the exception of diarrhea and difficulty sleeping), mean Day 30 scores of those symptoms returned to baseline for only 31% of patients on average, whereas they increased for 43% of patients, reaching a difference of 3 points on the MDASI scale, for 20% of patients. Similarly, Day 30 scores associated with mood, distress, feeling sad, and pain were equal to baseline scores for 33% of patients but were lower than baseline for 45% of patients. Scores associated with diarrhea, difficulty sleeping, attention, and relationships were unchanged for almost half of the patients (46%), with equivalent proportions of patients experiencing higher (28%) and lower (26%) scores on Day 30 compared with baseline. In summary, symptoms that worsened over the course of transplantation were those that reflected common transplantation toxicities, whereas symptoms that improved with transplantation were largely psychosocial factors (mood, distress, and sadness). We also noted that pain improved with transplantation (see Table 2).
Predictive Factors for Increased Symptom Burden
MDASI GSS and interference scores at baseline before transplantation were the only significant predictors of MDASI scores at the nadir of counts. Stratified analysis indicated that the proportion of patients with nadir scores >3.5 increased with increasing baseline scores. The strongest correlation was for patients with baseline GSS scores ≥3, who were 8 times as likely to have high scores at nadir (odds ratio, 8; P = .02) than patients with a GSS scores <1 (see Table 3). Women had a greater tendency than men to have GSS and interference scores higher than the median at nadir of counts, with 56% of women reporting symptoms greater than the median for the group compared with only 45% of man reporting MDASI scores above the median (P = .4). Other patient demographics, including age, race, marital status, and educational status, had no apparent correlation with symptom burden. Prolonged time from diagnosis to transplantation also resulted in higher nadir GSS and interference scores (P = .2 and P = .08, respectively). Disease status before transplantation, as measured by remission status, Durie-Salmon staging, and levels of lactate dehydrogenase, β2 microglobulin, albumin, and hemoglobin, had no apparent correlation with higher nadir GSS or interference scores. There was no significant effect of comorbidity as measured by the CCI. The stem cell dose, both as a continuous variable and as a dichotomous variable ≥5 × 106 cells/kg versus <5 × 106 cells/kg, had no significant effect on symptom burden. The impact of performance status could not be assessed, because the vast majority of patients (63 of 64 patients) were assessed with an Eastern Cooperative Oncology Group performance status of 1. Mouth sores and bleeding were excluded from these analyses, because very few patients on study reported worsening of these symptoms with treatment. Because none achieved significance, data for patient demographics, disease, and treatment variables are not shown.
Table 3. Impact of Baseline M. D. Anderson Symptom Inventory Global Symptom Severityand Interference Scores on Nadir Scores
Percent with score >3.5
OR indicates odds ratio; 95% CI, 95% confidence interval; GSS, Global Symptom Severity; IS, Interference Score.
Baseline GSS a continuous variable
≥1 and <2
≥2 and <3
Baseline IS as continuous variable
≥1 and <2
≥2 and <3
≥3 and <4
≥4 and <5
Symptom Burden and Survival
There was no effect of increased symptom burden, as measured by MDASI symptom severity scores >1.25 (median at baseline; hazards ratio [HR], 1.3; 95% CI, 0.6–3.2 [P = .5]) or scores >3.5 (median at nadir; HR, 0.8; 95% CI, 0.3–1.9 [P = .6]), on survival at a median follow-up of 37 months.
High-dose chemotherapy with ASCT is a standard therapy for MM based on several studies that produced improved response rates, event free survival, and overall survival.2–4 Although it generally is considered to be a safe therapy, it still can be associated with substantial morbidity and mortality.5, 9 Greater understanding of the causes of treatment-related symptom burden as well as identification of predictive factors would allow for rational development of interventions that could prevent or ameliorate post-transplantation symptoms.
Our current analysis provides a detailed description of symptom burden in patients undergoing high-dose chemotherapy with ASCT for myeloma during the time of greatest toxicity, the first 30 days of transplantation. To our knowledge, the current study represents the most comprehensive analysis of symptom burden in patients undergoing ASCT for myeloma. We observed that symptom burden, measured as the mean MDASI score, increased from baseline before ASCT until the nadir of counts then returned to baseline in the majority of patients by Day 30 after ASCT. Although the mean MDASI scores returned to baseline, the distribution of individual symptom scores changed to reflect improvements in psychosocial parameters and persistent worsening in symptoms that represent common toxicities of ASCT. High symptom burden at baseline before ASCT was the strongest predictor of severe symptoms after ASCT.
Although our study clearly demonstrated that patients with a high baseline symptom burden had a higher risk for increased transplantation toxicity, the reasons for this correlation were not defined by this study. Further investigation is warranted, keeping in mind that baseline symptom burden in patients with MM is reflective of multiple factors, including disease burden, prior treatment toxicity, comorbidity, and subjective factors, such as a patient's “scale” when reporting symptoms. The variables that were investigated in the current study, including patient demographic factors, disease burden, advanced age, and comorbidity, which are thought to be linked to increased symptom burden, did not have any apparent effect on transplantation-related symptom burden. This suggests that there may be other variables contributing to disease- and treatment-related symptom burden.
Hypotheses on other potential causes for increased symptom burden include variations in cytokine levels, cytokine gene polymorphisms, and drug metabolism. Inflammatory cytokines, such as interleukin 6 (IL-6), interferon γ, tumor necrosis factor-α, and IL-12, have been implicated in the pathogenesis of symptoms such as fatigue, fever, and cachexia.34–36 In a previous study, we demonstrated the relation between symptom burden, as measured by the MDASI questionnaire, and biologic variables such as white blood cell counts, albumin levels, and IL-6 levels.37 Studies using IL-6 antagonists in patients with refractory myeloma and in patients undergoing ASCT for myeloma incidentally demonstrated improvement in symptoms, including reduction in fever, weight loss, and mucositis.38, 39 To further investigate the effect of cytokine gene polymorphisms and cytokine blockade on symptom burden in patients undergoing either conventional or high-dose chemotherapy with ASCT, we are planning a prospective evaluation of cytokine gene polymorphisms in these patients as well as prospective, randomized trials using either nonspecific cytokine blockade, specific inhibition of IL-6, or specific inhibition of tumor necrosis factor-α in this patient population.
Our current study was limited by its small sample size and by some variability in treatment regimen. These factors limited the statistical power of the study and our ability to adjust for confounding factors. Comorbidity was measured with the CCI.28 This tool has been criticized in the past as not sensitive or specific enough to predict outcome in transplantation patients, because the comorbidities measured often exclude patients from transplantation.26, 30 To this end, Sorror et al. have created a transplantation-specific comorbidity index.30 Unfortunately, we were unable to apply that tool in our study because of the retrospective nature of the study and incompleteness of available comorbidity data. In the future, analysis of the effect of comorbidity on symptom burden may benefit from use of their hematopoietic cell transplantation-specific comorbidity index.
Symptom burden was measured with the MDASI-BMT module. Although there are many symptom and quality-of-life indices to choose from, the current study highlights the utility of the MDASI-BMT questionnaire in measuring symptom burden in this patient population. Other common symptom indices, such as the European Organization of Research and Treatment of Cancer Quality of Life Questionnaire 30 and the Functional Assessment of Cancer Therapy, measure symptom burden over the past week.40, 41 Patients who undergo transplantation will have more variability in their symptom burden, and the brevity of the form, the ease of administration (multiple formats, including paper and the IVR system), and the time frame of symptoms assessed all offer advantages in this patient population.
The usefulness of understanding symptom burden is clear. Early identification of patients with high symptom burden will allow for early and aggressive symptom management. Furthermore, the results of this study suggest that longitudinal measurement of symptom burden using the MDASI-BMT is useful both for targeting patients who are at high risk of severe symptom burden from treatment and for monitoring the efficacy of interventions aimed at reducing symptom burden. More effective management of symptom burden easily translates to improved quality of life, eventually leading to improved performance status, tolerance to therapy, and ability to deliver effective therapy. With treatment options for MM rapidly expanding, resulting in improved complete and partial response rates, event-free survival, and overall survival, more focus will need to be placed on treatment toxicity and overall quality of life to help in the determination of best treatment choices. Identifying severe symptoms and which patients are at the highest risk for increased symptom burden with transplantation will help target interventions aimed at reducing the symptom burden incurred with transplantation.