- Email: [email protected]
Association Between Symptom Distress and Survival in Outpatients Seen in a Palliative Care Cancer Center
Vol. 29 No. 6 June 2005
Journal of Pain and Symptom Management
565
Original Article
Association Between Symptom Distress and Survival in Outpatients Seen in a Palliative Care Cancer Center J. Lynn Palmer, PhD and Michael J. Fisch, MD, MPH Departments of Palliative Care and Rehabilitation Medicine (J.L.P., M.J.F.) and Biostatistics (J.L.P.), The University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA
Abstract Clinical observation and preliminary reports suggest that higher scores for symptoms such as pain may be associated with shorter survival. We undertook a survival analysis to determine whether symptom expression in outpatients with complex cancer is related to the duration of their survival. Participants were 225 outpatients with cancer evaluated in our comprehensive cancer center for pain management or palliative care over a 10-week period ending June 2000. In addition to age and other clinical and demographic information, the patients completed the Anderson Symptom Assessment System (ASAS), which assesses pain, fatigue, nausea, depression, anxiety, drowsiness, shortness of breath (dyspnea), appetite, sleep, and feeling of well-being on a 0–10 scale. Univariate analyses showed that higher symptoms of dyspnea, drowsiness, problems with appetite, and nausea were significantly associated with shorter survival whereas pain, depression and other ASAS items were not. In multivariate analyses, only higher levels of dyspnea and drowsiness showed a significant association (P ⫽ 0.01 and P ⫽ 0.02, respectively) with shorter survival. Knowledge about these symptoms may be important in formulating adaptive randomization techniques for clinical trials and for research concerning estimates of survival. J Pain Symptom Manage 2005;29:565–571. 쑖 2005 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Palliative care, symptom distress, survival
Introduction Roughly 1.3 million people per year are diagnosed with cancer in the United States, and more than half of those patients live for five
Address reprint requests to: J. Lynn Palmer, PhD, Department of Palliative Care and Rehabilitation Medicine (Unit 8), The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030-0049, USA. Accepted for publication: November 2, 2004.
쑖 2005 U.S. Cancer Pain Relief Committee Published by Elsevier Inc. All rights reserved.
years or more at risk for bothersome symptoms that may diminish their quality of life. The symptom experience for patients with cancer may be related to the malignancy itself, cancer treatment, and procedures involved in cancer care.1 Bothersome symptoms are prevalent in roughly 50–84% of cancer patients, and they commonly include physical symptoms (pain, fatigue, shortness of breath), cognitive symptoms (memory problems, impaired concentration), and affective symptoms (especially depression and anxiety).2–6 Patients have a 0885-3924/05/$–see front matter doi:10.1016/j.jpainsymman.2004.11.007
566
Palmer and Fisch
median number of 11 symptoms, and even in outpatients and patients with good performance status, the median number of symptoms tends to be more than nine.4,5 Comorbidity has an influence on symptoms, and it is also an independent prognostic factor in cancer.7,8 In addition, symptoms and symptom biology may interact with both treatment outcome and compliance. Depressed cancer patients, for example, have been found to have higher than normal levels of interleukin-6, a proinflammatory cytokine,9 and antidepressants have been demonstrated in vitro to reduce the release of proinflammatory cytokines.10 This has led investigators to hypothesize that use of the antidepressant paroxetine before instituting biologic therapy with interferon-alpha for malignant melanoma may reduce the symptoms of depression and produce better outcomes. Indeed, a landmark study of depression11 demonstrated paroxetine was effective to prevent depression and reduce dropout from anti-cancer treatment. The study demonstrated both the intrinsic value of treating symptoms as well as the growing appreciation of the interaction of symptoms, cancer biology, and cancer treatment outcomes. Prospective data linking symptom expression to survival outcomes in cancer patients are rare. One intriguing exception is an evaluative study in Sweden12 in which 135 patients with advanced solid tumors and cancer-related cachexia were randomized to placebo, prednisolone 10 mg twice daily, or indomethacin 50 mg twice daily until death. Not only did the indomethacin-treated patients maintain superior Karnofsky indices, they experienced less pain and required fewer other analgesics. Perhaps most surprising was that patients maintained on indomethacin survived an average of 510 days compared to 250 days for patients who received placebo (P ⬍ 0.05). A handful of other provocative studies have indicated an association between global symptom burden or specific symptoms and prognosis for advanced cancer patients in various settings,13–20 but most of these data were collected in an era when cancer patients routinely received care in the hospital. Our cross-sectional study explored the hypothesis that higher symptom scores are associated with shorter survival in an outpatient
Vol. 29 No. 6 June 2005
population followed for pain management and palliative care in a tertiary cancer center.
Methods Participants were 225 outpatients with cancer evaluated in our comprehensive cancer center for pain management or palliative care over a 10-week period that ended in June 2000. The Edmonton Symptom Assessment Scale (ESAS) was noted to be one of the reliable and valid multidimensional instruments that are well-suited for measurement of symptom clusters.21–24 We used a version of this instrument adapted for ease-of-use by implementing numerical rating scales (0–10, where 10 represents the worst imaginable severity) rather than visual analog scales on a 10-cm line. This instrument is called the Anderson Symptom Assessment System (ASAS). Like the ESAS, the ASAS allows the patient to assess their physical symptoms of pain, fatigue, nausea, drowsiness, shortness of breath, appetite, and sleep and their psychological symptoms of depression and anxiety, in addition to their overall feeling of well-being. According to a factor analysis of the ASAS,25 the ASAS instrument can be conceptualized as having two factors: physical symptoms (pain, fatigue, nausea, drowsiness, shortness of breath, appetite, sleep, and feeling of well-being) and psychological symptoms (depression, anxiety, and feeling of well-being), with well-being associated with both factors. The earlier study25 also found that physical symptoms measured by the ASAS contribute uniquely to understanding the patient’s symptom profile, while psychological variables are more closely related to the overall ASAS score. We therefore included both factors and the overall ASAS score as potentially being related to survival. Another variable studied was the patient’s score on the Two-Question Screening Survey (TQSS) as an indicator of depression. This easy-to-use tool was found reasonably predictive of depression in patients, and its depression scores were in moderate agreement with those of a 0–10 Numerical Rating Scale (NRS) item for depression.26 The data were analyzed using descriptive statistics and survival analysis. Initially, univariate analyses were utilized to determine associations between survival and the individual variables.
Vol. 29 No. 6 June 2005
Symptom Distress and Survival in Outpatients
Then all variables with significance levels of ⬍0.20 were included in a multivariate analysis, and individual variables were removed from the model one at a time (backward selection method) until all variables remaining in the model were significant at a significance level of ⱕ0.05. All data were analyzed using SAS software version 8.01 (SAS Institute, Cary, NC).
Results The characteristics of our study population are shown in Table 1. The median age was 52 years, with a range of 17–83. A large proportion of patients had advanced disease (42%), although many patients had locoregional disease with no evidence of active disease (35%). Lower proportions of patients had advanced disease with no evidence of active disease (14%) or locoregional disease (9%). The most common primary cancers were those of breast (17%), lung (11%), sarcoma (10%), and head and neck (10%). Median follow-up time was 1.9 years with 78 patients known to have died, and 147 patients were censored or still alive. Sixty-eight percent of the 147 censored dates were updated in May 2003.
ASAS Questionnaire and Survival Based on individual items on the ASAS scale, univariate analyses showed significant associations between shorter survival and the symptoms of dyspnea (P ⫽ 0.002), drowsiness (P ⫽ 0.005), problems with appetite (P ⫽ 0.02), and nausea (P ⫽ 0.045) (Table 2). Pain, depression, and other ASAS items were not independently associated with survival (each P ⬎ 0.4). Table 1 Characteristics of Study Population (n ⫽ 225) Median Age, years (range) Sex Disease status,a n (%) locoregional locoregional-NED advanced advanced-NED Primary cancer, % Breast Lung Sarcoma Head and neck Other
52 (17–83) 41% Male 20 79 95 31
(9) (35) (42) (14)
17 11 10 10 52
a Locoregional ⫽ stages 1 or 2; advanced ⫽ stages 3 or 4; NED ⫽ no evidence of active disease.
567
Table 2 Associations Among Survival and Other Variables Significance Level (P) Variable
All patients Local-NED Advanced (n ⫽ 225) (n ⫽ 79) (n ⫽ 95)
Results of univariate analyses, variables in order significancea Dyspnea 0.002 0.035 Drowsiness 0.005 0.020 Physical factor 0.021 0.242 (8 items) Appetite 0.024 0.274 Nausea 0.045 0.942 Total ASAS score 0.096 0.500 (10 items) Age 0.114 0.147 TQSS depression 0.141 0.337 score Pain 0.420 0.453 Anxiety 0.496 0.590 Depression 0.512 0.202 Fatigue 0.526 0.743 Well-being 0.554 0.944 Psychological factor 0.765 0.500 (3 items) Problems with sleep 0.956 0.946 Results of multivariate analyses, based on above variablesa Dyspnea 0.010 Drowsiness 0.022 0.020
of 0.022 0.959 0.252 0.097 0.261 0.426 0.162 0.322 0.541 0.802 0.739 0.587 0.965 0.841 0.262 0.022
a
All variables from ASAS except for Age and TQSS depression score.
Two survival curves are shown in Figures 1 and 2, each grouping the variables dyspnea and drowsiness by the categories of 0, 1–3, and 4–10 on a 10-point scale. The categorization of no dyspnea appeared to be more closely related to longer survival than the other two categories of dyspnea. As to drowsiness, the highest categorization appeared to be more closely associated with shorter survival than the two lower categorizations. When the variables were categorized into three groups, the significance levels of the associations between survival and the two variables of dyspnea and drowsiness were P ⫽ 0.005 and P ⫽ 0.015 using the log-rank test.
Other Variables and Survival Depression measured by the Two-Question Screening Survey focusing on depressed mood and anhedonia27 and age were marginally associated (P ⫽ 0.14, P ⫽ 0.11). The physical factor of the ASAS25 was significant (P ⫽ 0.02), but the psychological factor was not (P ⫽ 0.77). The total ASAS score and the total of 9 symptoms were only marginally related (P ⫽ 0.10, P ⫽ 0.08). In multivariate analyses, only increased symptoms of dyspnea and drowsiness were
568
Palmer and Fisch
Vol. 29 No. 6 June 2005
Fig. 1. Survival in years by dyspnea score category.
significantly associated with shorter survival (P ⫽ 0.01, P ⫽ 0.02). Controlling for drowsiness, an 11% decrease in survival time is expected for each one-unit increase in dyspnea. Controlling for dyspnea, a 10% decrease in sur-
vival time is expected for each one-unit increase in dyspnea. Fairly similar, although less significant results were found when only advanced cancer patients were included in the model (n ⫽ 95) and
Fig. 2. Survival in years by drowsiness score category.
Vol. 29 No. 6 June 2005
Symptom Distress and Survival in Outpatients
when patients with locoregional disease with no evidence of active disease (n ⫽ 79) were included (Table 2). The other two categories of disease status had too few observations (n ⫽ 20 and n ⫽ 31) to provide stable estimates or statistical significance. When a combined category of local (with and without evidence of disease, n ⫽ 99) was used, similar findings were found to those of locoregional disease with no evidence of active disease (n ⫽ 79); dyspnea was found to be significant at 0.028 and drowsiness was significant at 0.017. When the two advanced categories were combined (with and without evidence of disease, n ⫽ 126), no variables were found to be significant at P ⬍ 0.05, although dyspnea was closest to significance at P ⫽ 0.066. In advanced cancer patients with evidence of disease (n ⫽ 95), only the variable dyspnea remained in the multivariate model at a significance level less than 0.05 (P ⫽ 0.02). When drowsiness was forced into this model, its significance level was ⬎ 0.6. For patients with locoregional disease, only drowsiness with no evidence of active disease remained a significant variable (P ⫽ 0.02) in multivariate analysis. When dyspnea was forced into this model, dyspnea was nearly significant (P ⫽ 0.06).
Discussion The terms advanced cancer, terminal cancer, and end-of-life malignancy are often difficult to define. Some research has focused on selfreported health and symptom status and their relationship to survival in ambulatory patients seen in oncology clinics,28,29 but little research has been related to the survival outcomes of cancer patients seen in ambulatory palliative care clinics. Our tertiary cancer center follows the trend toward earlier integration of experts in pain and palliative care into the multidisciplinary treatment team. As such, we described a cohort of patients who achieved longer survival than was observed in several other studies of the relationship of clinical factors to survival of patients with advanced cancer.30–32 These kind of outpatient palliative care clinics are a new phenomenon and are becoming increasingly prevalent as the demand for palliative care services escalates. With a population that is both expanding and aging, the burden of disease
569
and suffering attributed to malignancy is expected to grow significantly,33 and a better understanding of outcomes in this emerging outpatient palliative care venue is important. The pain and palliative care outpatient practice in our tertiary cancer center includes patients with advanced and non-advanced cancer, and their survival distribution is far different from that typically associated with palliative and/or hospice care. Our palliative care population is based on internal referrals and includes a wider variety of primary malignancies compared to a typical outpatient practice. However, we feel that our data is typical of an academic outpatient palliative care practice and that our data can be generalized to such practices. The finding that increased physical symptoms other than pain were associated with shorter survival is consistent with data obtained from other settings of advanced cancer care. Sloan and colleagues,13 for example, examined data concerning outpatients with advanced cancer receiving oncology care in the North Central Cancer Treatment Group clinical trials and found that the only self-reported symptom that influenced their Cox proportional hazards model of survival was patient-reported appetite. Morita and colleagues19 have also documented the prognostic significance of appetite. Moreover, in a cohort of ambulatory lung cancer patients presenting to our cancer center emergency department, dyspnea was the presenting symptom associated with the worst prognosis.20 In our study, appetite was a significant predictor along with drowsiness and dyspnea. This confirmed the importance of physical symptoms derived from concrete biological changes associated with disease compared to the significance of more subjective and variable symptoms (pain, depression, fatigue) which sometimes are expressions of global suffering as much as indicators of changing biological conditions.34 Knowledge of the importance of non-pain physical symptom expression may be important when the survival of these patients is estimated, and for design of adaptive randomization techniques in clinical trials that address estimates of survival of ambulatory cancer patients referred for palliative care.
References 1. McGuire DB. Occurence of cancer pain. J Natl Cancer Inst Monogr 2004;32:51–56.
570
Palmer and Fisch
2. Cleeland CS. Cancer-related symptoms. Semin Radiat Oncol 2000;10:175–190. 3. Chang VT, Hwang SS, Feuerman M, et al. Symptom and quality of life survey of medical oncology patients at a Veteran Affairs Medical Center. Cancer 2000;88:1175–1183.
Vol. 29 No. 6 June 2005
17. Rana A, Chisholm GD, Rashwan HM, et al. Symptomatology of metastatic prostate cancer: prognostic significance. Br J Urol 1994;73:683–686. 18. Graf W, Glimelius B, Pahlman L, et al. Determinants of prognosis in advanced colorectal cancer. Eur J Cancer 1991;27:1119–1123.
4. Walsh D, Donnelly S, Rybicki L. The symptoms of advanced cancer: relationship to age, gender, and performance status in 1,000 patients. Support Care Cancer 2000;8:175–179.
19. Morita T, Tsunoda J, Inoue S, et al. Prediction of survival in terminally ill cancer patients—a prospective study. Gan To Kagaku Ryoho 1998;25: 1203–1211.
5. Portenoy RK, Thaler HT, Kornblith AB. Symptom prevalence, characteristics and distress in a cancer population. Qual Life Res 1994;3:183–189.
20. Escalante CP, Martin CG, Elting LS, et al. Dyspnea in cancer patients: etiology, resource utilization, and survival—implications in a managed care world. Cancer 1996;78:1314–1319.
6. Schuit KW, Sleijfer DT, Meijler WJ, et al. Symptoms and functional status of patients with disseminated cancer visiting outpatient departments. J Pain Symptom Manage 1998;16:290–297.
21. Chang VT, Hwang SS, Feuerman M. Validation of the Edmonton Symptom Assessment Scale. Cancer 2000;88:2164–2171.
7. Piccirillo JF, Feinstein AR. Clinical symptoms and comorbidity: significance for the prognostic classification of cancer. Cancer 1996;77:834–842. 8. Piccirillo JF, Tierney RM, Costas I, et al. Prognostic importance of comorbidity in a hospital-based cancer registry. JAMA 2004;291:2441–2447. 9. Musselman DL, Miller AH, Porter MR, et al. Higher than normal plasma interleukin-6 concentrations in cancer patients with depression: preliminary findings. Am J Psychiatry 2001;158:1252–1257. 10. Maes M, Song C, Lin AH, et al. Negative immunoregulatory effects of antidepressants: inhibition of interferon-gamma and stimulation of interleukin-10 secretion. Neuropsychopharmacology 1999;20:370–379. 11. Musselman DL, Lawson DH, Gumnick JF, et al. Paroxetine for the prevention of depression induced by high dose interferon alfa. N Engl J Med 2001;344:961–966. 12. Lundholm K, Gelin J, Hyltander A, et al. Antiinflammatory treatment may prolong survival in undernourished patients with metastatic solid tumors. Cancer Res 1994;54:5602–5606. 13. Chang VT, Thaler HT, Polyak TA, et al. Quality of life and survival: the role of multidimensional symptom assessment. Cancer 1998;83:173–179. 14. Lassauniere JM, Vinant P. Prognostic factors, survival, and advanced cancer. J Palliat Care 1992;8: 52–54. 15. Stanley KE. Prognostic factors for survival in patients with inoperable lung cancer. J Natl Cancer Inst 1980;65:25–32. 16. Reuben DB, Mor V, Hiris J. Clinical symptoms and length of survival in patients with terminal cancer. Arch Intern Med 1988;148:1586–1591.
22. Dudgeon DJ, Harlos M, Clinch JJ. The Edmonton Symptom Assessment Scale (ESAS) as an audit tool. J Palliative Care 1999;15:14–19. 23. Bruera E, Kuehn N, Miller MJ, et al. The Edmonton Symptom Assessment System (ESAS): a simple method for the assessment of palliative care patients. J Palliat Care 1991;7:6–9. 24. Paice JA. Assessment of symptom clusters in people with cancer. J Natl Cancer Inst Monogr 2004; 32:98–102. 25. Palmer JL, Sweeney C, Fisch MJ. Factor analysis of the Anderson Symptom Assessment System. American Society of Clinical Oncology, 2002, Abstract 1516, 380a. 26. Fisch MJ, Palmer JL, Zhukovsky DS, et al. Screening for depressive symptoms in cancer outpatients: association between a numerical rating scale (NRS) and a two-question screening survey (TQSS). American Society of Clinical Oncology, 2002, Abstract 1499, 375a. 27. Fisch MJ, Loehrer PJ, Kristeller J, et al. Fluoxetine versus placebo in advanced cancer outpatients: a double-blinded trial of the Hoosier Oncology Group. J Clin Oncol 2003;21:1937–1943. 28. Sloan JA, Loprinzi CL, Laurine JA, et al. A simple stratification factor prognostic for survival in advanced cancer: The Good/Bad/Uncertain Index. J Clin Oncol 2001;19:3539–3546. 29. Coates A, Porzsolt F, Osoba D. Quality of life in oncology practice: prognostic value of EROTC QLQC30 scores in patients with advanced malignancy. Eur J Cancer 1997;33:1025–1030. 30. Vigano A, Bruera E, Jhangri GS, et al. Clinical survival predictors in patients with advanced cancer. Arch Intern Med 2000;160:861–868. 31. Maltoni M, Nanni O, Pirovano M, et al. Successful validation of the palliative prognostic score in
Vol. 29 No. 6 June 2005
Symptom Distress and Survival in Outpatients
terminally ill cancer patients. J Pain Symptom Manage 1999;17:240–247. 32. Bruera E, Miller MJ, Kuehn N, et al. Estimate of survival of patients admitted to a palliative care unit: a prospective study. J Pain Symptom Manage 1992;7: 82–86.
571
33. Carr DB, Goudas LC, Balk EM, et al. Evidence report on the treatment of pain in cancer patients. J Natl Cancer Inst Monogr 2004;32:23–31. 34. Cohen MZ, Williams L, Knight P, et al. Symptom masquerade: understanding the meaning of symptoms. Support Care Cancer 2004;12:184–190.
Journal of Pain and Symptom Management
565
Original Article
Association Between Symptom Distress and Survival in Outpatients Seen in a Palliative Care Cancer Center J. Lynn Palmer, PhD and Michael J. Fisch, MD, MPH Departments of Palliative Care and Rehabilitation Medicine (J.L.P., M.J.F.) and Biostatistics (J.L.P.), The University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA
Abstract Clinical observation and preliminary reports suggest that higher scores for symptoms such as pain may be associated with shorter survival. We undertook a survival analysis to determine whether symptom expression in outpatients with complex cancer is related to the duration of their survival. Participants were 225 outpatients with cancer evaluated in our comprehensive cancer center for pain management or palliative care over a 10-week period ending June 2000. In addition to age and other clinical and demographic information, the patients completed the Anderson Symptom Assessment System (ASAS), which assesses pain, fatigue, nausea, depression, anxiety, drowsiness, shortness of breath (dyspnea), appetite, sleep, and feeling of well-being on a 0–10 scale. Univariate analyses showed that higher symptoms of dyspnea, drowsiness, problems with appetite, and nausea were significantly associated with shorter survival whereas pain, depression and other ASAS items were not. In multivariate analyses, only higher levels of dyspnea and drowsiness showed a significant association (P ⫽ 0.01 and P ⫽ 0.02, respectively) with shorter survival. Knowledge about these symptoms may be important in formulating adaptive randomization techniques for clinical trials and for research concerning estimates of survival. J Pain Symptom Manage 2005;29:565–571. 쑖 2005 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Palliative care, symptom distress, survival
Introduction Roughly 1.3 million people per year are diagnosed with cancer in the United States, and more than half of those patients live for five
Address reprint requests to: J. Lynn Palmer, PhD, Department of Palliative Care and Rehabilitation Medicine (Unit 8), The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030-0049, USA. Accepted for publication: November 2, 2004.
쑖 2005 U.S. Cancer Pain Relief Committee Published by Elsevier Inc. All rights reserved.
years or more at risk for bothersome symptoms that may diminish their quality of life. The symptom experience for patients with cancer may be related to the malignancy itself, cancer treatment, and procedures involved in cancer care.1 Bothersome symptoms are prevalent in roughly 50–84% of cancer patients, and they commonly include physical symptoms (pain, fatigue, shortness of breath), cognitive symptoms (memory problems, impaired concentration), and affective symptoms (especially depression and anxiety).2–6 Patients have a 0885-3924/05/$–see front matter doi:10.1016/j.jpainsymman.2004.11.007
566
Palmer and Fisch
median number of 11 symptoms, and even in outpatients and patients with good performance status, the median number of symptoms tends to be more than nine.4,5 Comorbidity has an influence on symptoms, and it is also an independent prognostic factor in cancer.7,8 In addition, symptoms and symptom biology may interact with both treatment outcome and compliance. Depressed cancer patients, for example, have been found to have higher than normal levels of interleukin-6, a proinflammatory cytokine,9 and antidepressants have been demonstrated in vitro to reduce the release of proinflammatory cytokines.10 This has led investigators to hypothesize that use of the antidepressant paroxetine before instituting biologic therapy with interferon-alpha for malignant melanoma may reduce the symptoms of depression and produce better outcomes. Indeed, a landmark study of depression11 demonstrated paroxetine was effective to prevent depression and reduce dropout from anti-cancer treatment. The study demonstrated both the intrinsic value of treating symptoms as well as the growing appreciation of the interaction of symptoms, cancer biology, and cancer treatment outcomes. Prospective data linking symptom expression to survival outcomes in cancer patients are rare. One intriguing exception is an evaluative study in Sweden12 in which 135 patients with advanced solid tumors and cancer-related cachexia were randomized to placebo, prednisolone 10 mg twice daily, or indomethacin 50 mg twice daily until death. Not only did the indomethacin-treated patients maintain superior Karnofsky indices, they experienced less pain and required fewer other analgesics. Perhaps most surprising was that patients maintained on indomethacin survived an average of 510 days compared to 250 days for patients who received placebo (P ⬍ 0.05). A handful of other provocative studies have indicated an association between global symptom burden or specific symptoms and prognosis for advanced cancer patients in various settings,13–20 but most of these data were collected in an era when cancer patients routinely received care in the hospital. Our cross-sectional study explored the hypothesis that higher symptom scores are associated with shorter survival in an outpatient
Vol. 29 No. 6 June 2005
population followed for pain management and palliative care in a tertiary cancer center.
Methods Participants were 225 outpatients with cancer evaluated in our comprehensive cancer center for pain management or palliative care over a 10-week period that ended in June 2000. The Edmonton Symptom Assessment Scale (ESAS) was noted to be one of the reliable and valid multidimensional instruments that are well-suited for measurement of symptom clusters.21–24 We used a version of this instrument adapted for ease-of-use by implementing numerical rating scales (0–10, where 10 represents the worst imaginable severity) rather than visual analog scales on a 10-cm line. This instrument is called the Anderson Symptom Assessment System (ASAS). Like the ESAS, the ASAS allows the patient to assess their physical symptoms of pain, fatigue, nausea, drowsiness, shortness of breath, appetite, and sleep and their psychological symptoms of depression and anxiety, in addition to their overall feeling of well-being. According to a factor analysis of the ASAS,25 the ASAS instrument can be conceptualized as having two factors: physical symptoms (pain, fatigue, nausea, drowsiness, shortness of breath, appetite, sleep, and feeling of well-being) and psychological symptoms (depression, anxiety, and feeling of well-being), with well-being associated with both factors. The earlier study25 also found that physical symptoms measured by the ASAS contribute uniquely to understanding the patient’s symptom profile, while psychological variables are more closely related to the overall ASAS score. We therefore included both factors and the overall ASAS score as potentially being related to survival. Another variable studied was the patient’s score on the Two-Question Screening Survey (TQSS) as an indicator of depression. This easy-to-use tool was found reasonably predictive of depression in patients, and its depression scores were in moderate agreement with those of a 0–10 Numerical Rating Scale (NRS) item for depression.26 The data were analyzed using descriptive statistics and survival analysis. Initially, univariate analyses were utilized to determine associations between survival and the individual variables.
Vol. 29 No. 6 June 2005
Symptom Distress and Survival in Outpatients
Then all variables with significance levels of ⬍0.20 were included in a multivariate analysis, and individual variables were removed from the model one at a time (backward selection method) until all variables remaining in the model were significant at a significance level of ⱕ0.05. All data were analyzed using SAS software version 8.01 (SAS Institute, Cary, NC).
Results The characteristics of our study population are shown in Table 1. The median age was 52 years, with a range of 17–83. A large proportion of patients had advanced disease (42%), although many patients had locoregional disease with no evidence of active disease (35%). Lower proportions of patients had advanced disease with no evidence of active disease (14%) or locoregional disease (9%). The most common primary cancers were those of breast (17%), lung (11%), sarcoma (10%), and head and neck (10%). Median follow-up time was 1.9 years with 78 patients known to have died, and 147 patients were censored or still alive. Sixty-eight percent of the 147 censored dates were updated in May 2003.
ASAS Questionnaire and Survival Based on individual items on the ASAS scale, univariate analyses showed significant associations between shorter survival and the symptoms of dyspnea (P ⫽ 0.002), drowsiness (P ⫽ 0.005), problems with appetite (P ⫽ 0.02), and nausea (P ⫽ 0.045) (Table 2). Pain, depression, and other ASAS items were not independently associated with survival (each P ⬎ 0.4). Table 1 Characteristics of Study Population (n ⫽ 225) Median Age, years (range) Sex Disease status,a n (%) locoregional locoregional-NED advanced advanced-NED Primary cancer, % Breast Lung Sarcoma Head and neck Other
52 (17–83) 41% Male 20 79 95 31
(9) (35) (42) (14)
17 11 10 10 52
a Locoregional ⫽ stages 1 or 2; advanced ⫽ stages 3 or 4; NED ⫽ no evidence of active disease.
567
Table 2 Associations Among Survival and Other Variables Significance Level (P) Variable
All patients Local-NED Advanced (n ⫽ 225) (n ⫽ 79) (n ⫽ 95)
Results of univariate analyses, variables in order significancea Dyspnea 0.002 0.035 Drowsiness 0.005 0.020 Physical factor 0.021 0.242 (8 items) Appetite 0.024 0.274 Nausea 0.045 0.942 Total ASAS score 0.096 0.500 (10 items) Age 0.114 0.147 TQSS depression 0.141 0.337 score Pain 0.420 0.453 Anxiety 0.496 0.590 Depression 0.512 0.202 Fatigue 0.526 0.743 Well-being 0.554 0.944 Psychological factor 0.765 0.500 (3 items) Problems with sleep 0.956 0.946 Results of multivariate analyses, based on above variablesa Dyspnea 0.010 Drowsiness 0.022 0.020
of 0.022 0.959 0.252 0.097 0.261 0.426 0.162 0.322 0.541 0.802 0.739 0.587 0.965 0.841 0.262 0.022
a
All variables from ASAS except for Age and TQSS depression score.
Two survival curves are shown in Figures 1 and 2, each grouping the variables dyspnea and drowsiness by the categories of 0, 1–3, and 4–10 on a 10-point scale. The categorization of no dyspnea appeared to be more closely related to longer survival than the other two categories of dyspnea. As to drowsiness, the highest categorization appeared to be more closely associated with shorter survival than the two lower categorizations. When the variables were categorized into three groups, the significance levels of the associations between survival and the two variables of dyspnea and drowsiness were P ⫽ 0.005 and P ⫽ 0.015 using the log-rank test.
Other Variables and Survival Depression measured by the Two-Question Screening Survey focusing on depressed mood and anhedonia27 and age were marginally associated (P ⫽ 0.14, P ⫽ 0.11). The physical factor of the ASAS25 was significant (P ⫽ 0.02), but the psychological factor was not (P ⫽ 0.77). The total ASAS score and the total of 9 symptoms were only marginally related (P ⫽ 0.10, P ⫽ 0.08). In multivariate analyses, only increased symptoms of dyspnea and drowsiness were
568
Palmer and Fisch
Vol. 29 No. 6 June 2005
Fig. 1. Survival in years by dyspnea score category.
significantly associated with shorter survival (P ⫽ 0.01, P ⫽ 0.02). Controlling for drowsiness, an 11% decrease in survival time is expected for each one-unit increase in dyspnea. Controlling for dyspnea, a 10% decrease in sur-
vival time is expected for each one-unit increase in dyspnea. Fairly similar, although less significant results were found when only advanced cancer patients were included in the model (n ⫽ 95) and
Fig. 2. Survival in years by drowsiness score category.
Vol. 29 No. 6 June 2005
Symptom Distress and Survival in Outpatients
when patients with locoregional disease with no evidence of active disease (n ⫽ 79) were included (Table 2). The other two categories of disease status had too few observations (n ⫽ 20 and n ⫽ 31) to provide stable estimates or statistical significance. When a combined category of local (with and without evidence of disease, n ⫽ 99) was used, similar findings were found to those of locoregional disease with no evidence of active disease (n ⫽ 79); dyspnea was found to be significant at 0.028 and drowsiness was significant at 0.017. When the two advanced categories were combined (with and without evidence of disease, n ⫽ 126), no variables were found to be significant at P ⬍ 0.05, although dyspnea was closest to significance at P ⫽ 0.066. In advanced cancer patients with evidence of disease (n ⫽ 95), only the variable dyspnea remained in the multivariate model at a significance level less than 0.05 (P ⫽ 0.02). When drowsiness was forced into this model, its significance level was ⬎ 0.6. For patients with locoregional disease, only drowsiness with no evidence of active disease remained a significant variable (P ⫽ 0.02) in multivariate analysis. When dyspnea was forced into this model, dyspnea was nearly significant (P ⫽ 0.06).
Discussion The terms advanced cancer, terminal cancer, and end-of-life malignancy are often difficult to define. Some research has focused on selfreported health and symptom status and their relationship to survival in ambulatory patients seen in oncology clinics,28,29 but little research has been related to the survival outcomes of cancer patients seen in ambulatory palliative care clinics. Our tertiary cancer center follows the trend toward earlier integration of experts in pain and palliative care into the multidisciplinary treatment team. As such, we described a cohort of patients who achieved longer survival than was observed in several other studies of the relationship of clinical factors to survival of patients with advanced cancer.30–32 These kind of outpatient palliative care clinics are a new phenomenon and are becoming increasingly prevalent as the demand for palliative care services escalates. With a population that is both expanding and aging, the burden of disease
569
and suffering attributed to malignancy is expected to grow significantly,33 and a better understanding of outcomes in this emerging outpatient palliative care venue is important. The pain and palliative care outpatient practice in our tertiary cancer center includes patients with advanced and non-advanced cancer, and their survival distribution is far different from that typically associated with palliative and/or hospice care. Our palliative care population is based on internal referrals and includes a wider variety of primary malignancies compared to a typical outpatient practice. However, we feel that our data is typical of an academic outpatient palliative care practice and that our data can be generalized to such practices. The finding that increased physical symptoms other than pain were associated with shorter survival is consistent with data obtained from other settings of advanced cancer care. Sloan and colleagues,13 for example, examined data concerning outpatients with advanced cancer receiving oncology care in the North Central Cancer Treatment Group clinical trials and found that the only self-reported symptom that influenced their Cox proportional hazards model of survival was patient-reported appetite. Morita and colleagues19 have also documented the prognostic significance of appetite. Moreover, in a cohort of ambulatory lung cancer patients presenting to our cancer center emergency department, dyspnea was the presenting symptom associated with the worst prognosis.20 In our study, appetite was a significant predictor along with drowsiness and dyspnea. This confirmed the importance of physical symptoms derived from concrete biological changes associated with disease compared to the significance of more subjective and variable symptoms (pain, depression, fatigue) which sometimes are expressions of global suffering as much as indicators of changing biological conditions.34 Knowledge of the importance of non-pain physical symptom expression may be important when the survival of these patients is estimated, and for design of adaptive randomization techniques in clinical trials that address estimates of survival of ambulatory cancer patients referred for palliative care.
References 1. McGuire DB. Occurence of cancer pain. J Natl Cancer Inst Monogr 2004;32:51–56.
570
Palmer and Fisch
2. Cleeland CS. Cancer-related symptoms. Semin Radiat Oncol 2000;10:175–190. 3. Chang VT, Hwang SS, Feuerman M, et al. Symptom and quality of life survey of medical oncology patients at a Veteran Affairs Medical Center. Cancer 2000;88:1175–1183.
Vol. 29 No. 6 June 2005
17. Rana A, Chisholm GD, Rashwan HM, et al. Symptomatology of metastatic prostate cancer: prognostic significance. Br J Urol 1994;73:683–686. 18. Graf W, Glimelius B, Pahlman L, et al. Determinants of prognosis in advanced colorectal cancer. Eur J Cancer 1991;27:1119–1123.
4. Walsh D, Donnelly S, Rybicki L. The symptoms of advanced cancer: relationship to age, gender, and performance status in 1,000 patients. Support Care Cancer 2000;8:175–179.
19. Morita T, Tsunoda J, Inoue S, et al. Prediction of survival in terminally ill cancer patients—a prospective study. Gan To Kagaku Ryoho 1998;25: 1203–1211.
5. Portenoy RK, Thaler HT, Kornblith AB. Symptom prevalence, characteristics and distress in a cancer population. Qual Life Res 1994;3:183–189.
20. Escalante CP, Martin CG, Elting LS, et al. Dyspnea in cancer patients: etiology, resource utilization, and survival—implications in a managed care world. Cancer 1996;78:1314–1319.
6. Schuit KW, Sleijfer DT, Meijler WJ, et al. Symptoms and functional status of patients with disseminated cancer visiting outpatient departments. J Pain Symptom Manage 1998;16:290–297.
21. Chang VT, Hwang SS, Feuerman M. Validation of the Edmonton Symptom Assessment Scale. Cancer 2000;88:2164–2171.
7. Piccirillo JF, Feinstein AR. Clinical symptoms and comorbidity: significance for the prognostic classification of cancer. Cancer 1996;77:834–842. 8. Piccirillo JF, Tierney RM, Costas I, et al. Prognostic importance of comorbidity in a hospital-based cancer registry. JAMA 2004;291:2441–2447. 9. Musselman DL, Miller AH, Porter MR, et al. Higher than normal plasma interleukin-6 concentrations in cancer patients with depression: preliminary findings. Am J Psychiatry 2001;158:1252–1257. 10. Maes M, Song C, Lin AH, et al. Negative immunoregulatory effects of antidepressants: inhibition of interferon-gamma and stimulation of interleukin-10 secretion. Neuropsychopharmacology 1999;20:370–379. 11. Musselman DL, Lawson DH, Gumnick JF, et al. Paroxetine for the prevention of depression induced by high dose interferon alfa. N Engl J Med 2001;344:961–966. 12. Lundholm K, Gelin J, Hyltander A, et al. Antiinflammatory treatment may prolong survival in undernourished patients with metastatic solid tumors. Cancer Res 1994;54:5602–5606. 13. Chang VT, Thaler HT, Polyak TA, et al. Quality of life and survival: the role of multidimensional symptom assessment. Cancer 1998;83:173–179. 14. Lassauniere JM, Vinant P. Prognostic factors, survival, and advanced cancer. J Palliat Care 1992;8: 52–54. 15. Stanley KE. Prognostic factors for survival in patients with inoperable lung cancer. J Natl Cancer Inst 1980;65:25–32. 16. Reuben DB, Mor V, Hiris J. Clinical symptoms and length of survival in patients with terminal cancer. Arch Intern Med 1988;148:1586–1591.
22. Dudgeon DJ, Harlos M, Clinch JJ. The Edmonton Symptom Assessment Scale (ESAS) as an audit tool. J Palliative Care 1999;15:14–19. 23. Bruera E, Kuehn N, Miller MJ, et al. The Edmonton Symptom Assessment System (ESAS): a simple method for the assessment of palliative care patients. J Palliat Care 1991;7:6–9. 24. Paice JA. Assessment of symptom clusters in people with cancer. J Natl Cancer Inst Monogr 2004; 32:98–102. 25. Palmer JL, Sweeney C, Fisch MJ. Factor analysis of the Anderson Symptom Assessment System. American Society of Clinical Oncology, 2002, Abstract 1516, 380a. 26. Fisch MJ, Palmer JL, Zhukovsky DS, et al. Screening for depressive symptoms in cancer outpatients: association between a numerical rating scale (NRS) and a two-question screening survey (TQSS). American Society of Clinical Oncology, 2002, Abstract 1499, 375a. 27. Fisch MJ, Loehrer PJ, Kristeller J, et al. Fluoxetine versus placebo in advanced cancer outpatients: a double-blinded trial of the Hoosier Oncology Group. J Clin Oncol 2003;21:1937–1943. 28. Sloan JA, Loprinzi CL, Laurine JA, et al. A simple stratification factor prognostic for survival in advanced cancer: The Good/Bad/Uncertain Index. J Clin Oncol 2001;19:3539–3546. 29. Coates A, Porzsolt F, Osoba D. Quality of life in oncology practice: prognostic value of EROTC QLQC30 scores in patients with advanced malignancy. Eur J Cancer 1997;33:1025–1030. 30. Vigano A, Bruera E, Jhangri GS, et al. Clinical survival predictors in patients with advanced cancer. Arch Intern Med 2000;160:861–868. 31. Maltoni M, Nanni O, Pirovano M, et al. Successful validation of the palliative prognostic score in
Vol. 29 No. 6 June 2005
Symptom Distress and Survival in Outpatients
terminally ill cancer patients. J Pain Symptom Manage 1999;17:240–247. 32. Bruera E, Miller MJ, Kuehn N, et al. Estimate of survival of patients admitted to a palliative care unit: a prospective study. J Pain Symptom Manage 1992;7: 82–86.
571
33. Carr DB, Goudas LC, Balk EM, et al. Evidence report on the treatment of pain in cancer patients. J Natl Cancer Inst Monogr 2004;32:23–31. 34. Cohen MZ, Williams L, Knight P, et al. Symptom masquerade: understanding the meaning of symptoms. Support Care Cancer 2004;12:184–190.