Psychological Outcomes Associated With Anemia-Related Fatigue in Cancer Patients

ABSTRACT: This article examines the relationships between chemotherapy-induced anemia, fatigue, and psychological distress among anemic cancer patients with solid tumors. Patients participating in two randomized clinical trials evaluating the efficacy of darbepoetin alfa (Aranesp) completed a questionnaire at baseline, at the beginning of each chemotherapy cycle, and at the end of the 12-week treatment period. The questionnaire included four psychological distress outcomes: Brief Symptom Inventory (BSI) Depression and Anxiety, Functional Assessment of Cancer Therapy (FACT)-Emotional Well-Being, numeric rating scale of Overall Health, and the FACT-Fatigue subscale. Patients with a hemoglobin response of at least a 2 g/dL increase were more likely to experience meaningful improvements (at least 3 points) in FACT-Fatigue scores than nonresponders (55.0% vs 39.8%; P = .0004). Patients with meaningful improvements in FACT-Fatigue scores reported significantly greater improvements in each of the psychological outcomes relative to those without improved fatigue (P Recent studies report thattheprevalence of adjustment disorders[1] (ie, the development of clinically significant emotional or behavioralsymptoms) in patients with cancer ranges from 15% to 25%.[2,3] It is theorizedthat adjustment disorders arise in response to the diagnosis of cancer and thestress of coping with cancer treatment. These disorders are often serious enoughto warrant therapeutic intervention, but do not meet all of the criteria for amajor mental disorder.[1-3] Further, several studies have reported prevalencerates of clinical depression in oncology patients to be between 7% and 10%.[2-5]

Psychological Distressin Cancer Patients

Prevalence

An early study of the prevalence of psychiatric disorders incancer patients conducted by Derogatis et al[5] at three cancer centers in theearly 1980s found that 32% of nonterminal patients (n = 215) undergoing activetreatment had adjustment disorders, and that adjustment disorders comprised morethan two-thirds of all Diagnostic and Statistical Manual of Mental Disorders(DSM-III) diagnoses identified in the sample. In more recent studies,Morasso et al reported that 23.5% of breast cancer patients completingchemotherapy were found to have adjustment disorders.[3] Similarly, Skarstein etal found that 24.3% of hospitalized and ambulatory cancer patients with avariety of tumor types scored as borderline cases on the Hospital Anxiety andDepression Scale (HADS)-Anxiety scale, while for depression, borderlinedisorders represented 15.5% of patients.[2] Using the Brief Symptom Inventory(BSI), Zabora et al determined the prevalence of distress to range from 29.6% inpatients with gynecologic cancer to 43.4% in those with lung cancer.[6] Thus,one can conclude from this review of the literature that the prevalence ofpsychological disorders in the oncology population is approximately 15% to 32%.

Symptom-SpecificAssessments

Identification and treatment of patients with the symptoms ofdepression/anxiety or clinical depression is important, not only becausepatients’ functioning and well-being are compromised, but because thesesymptoms may also affect the ability to comply with and tolerate treatment, andthus could affect survival.[7] Clinicians face several challenges inidentifying, assessing, and treating the presence of psychological distress incancer patients: selecting a screening tool from the many available; having thespace and time necessary for valid and reliable assessment, and having trainedstaff available for the assessment and treatment of anxiety and depression.

First, one must choose from a wide variety of instrumentsthat have been shown to be approximately equivalent in their reliability andvalidity. Issues that vary widely across the instruments that the clinician mustgrapple with include the length of the instrument, the scoring, andinterpretation of the scores. Diagnosis of psychological distress, moreover, iscomplicated in patients with cancer because the diagnostic criteria for specificpsychological conditions include somatic symptoms that may also be associatedwith cancer or its treatment. For example, the symptoms of decreased appetite,insomnia, and fatigue (or loss of energy) can be either the direct result of thetoxic effects of chemotherapy or indicators of psychological distress. Thus, itis difficult for well-trained and highly experienced clinicians to diagnoseanxiety and depression in patients with cancer. Finally, many oncology clinicsare not appropriately staffed to diagnose and treat psychological distress.

In this article, we examine the relationships betweenanemia-related fatigue and patients’ emotional well-being and psychologicaldistress—specifically, depression and anxiety. Many cancer patients sufferanemia as consequences of the disease process and particularly of chemotherapy,with fatigue being the most common symptom of anemia.[8] Previous blinded,randomized controlled trials have shown that levels of fatigue experienced bypatients undergoing treatment for cancer can be significantly reduced byincreasing hemoglobin levels through the use of either epoetin alfa (Epogen,Procrit) or darbepoetin alfa (Aranesp).[9,10] However, little attention has been directedto the effects of fatigue on patients and the benefits of reducing fatigue.

Clinical Data

This study incorporated individual patient data from twomulticenter, randomized, clinical trials evaluating the efficacy of darbepoetinalfa among anemic cancer patients receiving multicycle chemotherapy. Thedose-finding phase II study of darbepoetin alfa was an active-controlled,open-label trial for anemia treatment among patients with solid tumors receivingchemotherapy conducted in the United States. The doses and schedules ofdarbepoetin alfa that were studied in this trial were 0.5, 1.0, 1.5, 2.25, 4.5,6.0, and 8.0 µg/kg/wk and 3.0, 5.0, 7.0, and 9.0 µg/kg every 2 weeks.

The pivotal phase III study was a double-blind,placebo-controlled study of darbepoetin alfa (2.25 µg/kg) for treatment ofanemia among lung cancer patients (either small-cell lung cancer or non-small-celllung cancer) receiving platinum-containing chemotherapy, conducted in Canada,Australia, and Europe. The two clinical trials were carried out between April1999 and January 2001.

Patients from these studies were included in these analysesif they completed at least 4 weeks of treatment, reported BSI depression andanxiety scale scores at baseline, and reported BSI depression or anxiety scoresat least one time after 4 weeks of treatment. Baseline and last-observed BSIscores were used to assess change over time, where the last observed valueoccurred any time after week 4 and before the end of the 12-week treatmentperiod. In the two trials, darbepoetin alfa, epoetin alfa, or placebo wereadministered for a maximum of 12 weeks. Written, signed informed consent wasobtained from each subject.

Eligibility criteria for the two trials were almost the same,with the exception of the phase III lung cancer trial that was limited to lungcancer patients receiving platinum chemotherapy. Inclusion and exclusioncriteria included anemia, as indicated by a hemoglobin concentration ≤11.0 g/dL; no evidence of iron, folate, or vitamin B₁₂deficiency; adequate renal and liver function; and Eastern Cooperative OncologyGroup (ECOG) performance status of 0 to 2. Patients were required to bereceiving cyclic chemotherapy, with no recombinant human erythropoietin therapywithin 8 weeks of randomization and no more than two red blood cell transfusionswithin 2 weeks of randomization.

Patient-reported health-related quality-of-life outcomes wereincluded as secondary end points to assess patient-perceived functioning andwell-being. Patients completed the quality-of-life questionnaire at thebeginning of each chemotherapy cycle and at the end of the treatment period.

Outcome Measures

Four measures of psychological distress were used in thisstudy:

(1) the depression scalefrom the BSI
(2) the anxiety scale fromthe BSI
(3) the emotionalwell-being scale from the Functional Assessment of Cancer Treatment-General(FACT)-G
(4) the 11-point numericrating scale of Overall Health status.

Fatigue was measured using the Fatigue subscale of theFACT-Anemia instrument. These measures are summarized in Table1.

Subjects completed the questionnaires at baseline,approximately every 3 to 4 weeks during treatment, and at the end-of-treatmentperiod. Since one of these studies was multinational, linguistically validatedversions of the BSI and FACT surveys were used for the required languages.

Brief SymptomInventory Anxiety and Depression Scales—TheBSI was developed as a shorter version of the Symptom Checklist-90 (SCL-90) tomeasure the psychological symptom patterns of both patients and nonpatients.[11]It hasbeen used extensively to assess the psychological distress of cancer patients[12-14], and Zabora found it to yield a specificity of .87 and a sensitivity of.89 in this population.[15] The BSI scales each include six symptoms;respondents report how much the symptom or problem distressed or bothered themduring the past 2 weeks on a 5-point scale: not at all (0), a little bit (1),moderately (2), quite a bit (3), or extremely (4). In this study, we computednorm-based scores by applying gender-specific general population means andstandard deviations for an adult, nonpatient sample. These normed scores have amean of 50 and standard deviation of 10 in the general population.

Criteria for defining low, moderate, and high levels ofdistress have been defined in previous cancer studies for normed BSI scores.[16]Specifically, low, moderate, and high distress have been defined as £ 0.5 standard deviations above the mean,0.5 to 1.49 standard deviations above the mean, and ³ 1.5 standard deviations above the mean,respectively. As such, low, moderate, and high levels of distress are associatedwith normed BSI scores of ≤55, 55 to 64.9, and ³65. A "case" of psychological distress for the BSI Depression andAnxiety scales was defined as a normed score ³ 65 on both scales.[16] This case definitionis a conservative one, given that Derogatis uses a normed score of 63 on twoscales to define a case.[11]

FunctionalAssessment of Cancer Therapy Scales—TheFACT-Emotional scale score ranges from 0 to 24. Responses to each question canvary from "not at all true" (0) to "very much true" (4). Ahigher score indicates better emotional well-being. The FACT-Emotional scale wasfound to have reliability and validity in a population of 545 oncology patientswith several different tumor types and chemotherapy regimens.[17] TheFACT-Fatigue subscale (Table 1) has 13 questions evaluating the impact of anemiaon cancer patients receiving chemotherapy. Responses to each question can rangefrom 0 to 4, with a total fatigue subscale score ranging from 0 to 52.[18] Ahigher score indicates less fatigue. The FACT-Fatigue subscale was found to havegood reliability and validity.[19] "Clinically meaningful improvement"in fatigue was defined as an increase of 3 points in the FACT-Fatigue subscalescore; a "case" of fatigue was defined as a FACT-Fatiguesubscale score £  36.5.[20]This cut-off is one standard deviation above the mean of an anemic cancerpatient population.[21]

Clinical efficacy outcomes included the proportion ofpatients achieving a hemoglobin response (³ 2 g/dL increase from baseline) and the changein the subject’s last hemoglobin in the treatment phase compared withbaseline.

Statistical Analysis

To analyze these two clinical trials, we used a meta-analyticapproach that pooled individual patient data.[22-25] Darbepoetin alfa at severaldoses and schedules, epoetin alfa at several doses and schedules, and placebowere combined. This meta-analysis was done to improve the power of the analysis.To describe the study sample, we computed means, standard deviations, and rangesof continuous variables and frequency distributions of categorical variables. Weconducted bivariate analyses evaluating the association of changes in normed BSIscores with changes in FACT-Fatigue subscale scores and with the othercovariates. For changes in FACT-Fatigue scores and continuous covariates, wecomputed Pearson correlation coefficients. For the fatigue case variable andcategorical covariates, we compared mean changes in normed BSI scores acrosslevels of the covariates using analysis of variance (ANOVA) models. Statisticalsignificance was evaluated at the .05 level.

Results

There were 357 evaluable subjects at baseline in the phase IIsolid tumor trial, and 250 evaluable subjects in the phase III lung cancertrial, for a total number of 607. Their baseline demographic characteristics aredisplayed in Table 2. Subjects ranged in age from 20 to 91, with a mean age of61.4 years. Patients aged 65 years and older represented 43.3% of theoverall sample, with 10.7% being 75 years of age and older. Slightly morethan half (52.9%) were women. The sample was predominantly non-Hispanic white(87.3%), with African-American and Hispanic patients comprising 5.1% each, andAsian and Pacific Islander Americans, 2.0%.

Table 3 shows the baseline clinical characteristics. Meanbaseline hemoglobin was 10.0 g/dL, with a range of 6.2 to 13.6 g/dL. There were275 subjects with hemoglobin levels below the mean (45.5%), while 330 (54.5%)had baseline hemoglobin at or above the mean. Slightly more than half of thesubjects (53.2%) had lung cancer, small-cell or non-small-cell. Breast andgynecologic cancer patients comprised 17.3% and 7.7%, respectively; patientswith gastrointestinal cancer made up 12.7%; and patients with other solidtumors, 9.1%. Most subjects had an ECOG status of 0 or 1 (26.5% and 62.2%,respectively), while 11.4% had an ECOG status of 2.

As shown in previous studies, increased hemoglobin issignificantly associated with improvement in fatigue.[9,29] Figure 1 displaysthe mean change in hemoglobin and the mean change in FACT-Fatigue subscale scoreover time. More than half (55.0%) of subjects whose hemoglobin increased 2.0 g/dLby the end of treatment also reported clinically meaningful improvement in theirFACT-Fatigue subscale score, compared with 39.8% who reported improvement infatigue without a hemoglobin response (P = .0004) (Table4).

Relationship of Fatigue to Psychological Distress

The 76.3% of subjects who were fatigued at baseline(FACT-Fatigue score ≤ 36.5)reported statistically significant greater levels of psychological distress(higher scores on the BSI Depression and Anxiety Scales) and lower levels ofoverall well-being than patients who were not fatigued (Table5). Patients withsubstantial fatigue reported mean BSI Depression scores that were 10 points (or1 standard deviation) higher than patients who did not report significantfatigue at baseline (61.0 vs 50.9, P < .001). Similarly, thedifference in BSI Anxiety scores between these two groups was 8 points (58.7 vs50.6, P < .001). BSI scores for patients who were not substantiallyfatigued at baseline (depression = 50.9; anxiety = 50.6) were at the mean of thegeneral population sample (50), whereas scores for those who were substantiallyfatigued indicated a much greater level of distress than the general population(depression = 61.0; anxiety = 58.7).

Furthermore, changes in fatigue were strongly associated withchanges in measures of psychological distress. As shown in Table6, subjects whoreported clinically meaningful improvement in the FACT-Fatigue subscale scorealso showed improvement in scores on the BSI Depression and Anxiety scales,FACT-Emotional Well-Being, and Overall Health status, whereas those who did notexperience improvement in fatigue reported worse psychological outcomes.

Differences between the groups were significant for all threepsychological outcome measures, as well as for the measure of overallhealth (P < .0001). When patients reported that theirfatigue levels had improved, they also reported improvements in BSI Depressionand Anxiety scores of 5.1 and 6.1, respectively. Conversely, subjects who didnot improve their self-assessed fatigue reported worsening BSI Depression andAnxiety scores of 3.1 and 1.6, respectively. Thus, the magnitude of thedifference in BSI scores among groups with varying levels of change in fatigueis approximately 8 points.

Patients who had substantial fatigue improvement alsoreported substantial improvements (11.2) in self-assessment of their overallhealth, while those with no improvement in their fatigue at end of treatmentreported a decline in overall health (-8.4).

At baseline, 76.3% of the 607 study subjects were classifiedas substantially fatigued (FACT-Fatigue ≤ 36.5).At the end of treatment, 20.7% of those who were fatigued at baseline (96 out of463) were no longer experiencing substantial fatigue (ie, reporting aFACT-Fatigue score > 36.5 at end of treatment). These subjects experiencedsignificantly greater reductions in psychological distress and improvements inemotional well-being and overall health, compared with those who did nottransition from being substantially fatigued (P < .001 for allmeasures), as illustrated in Table 7. Among patients who were substantiallyfatigued at baseline but no longer substantially fatigued at follow-up, BSIscores improved by approximately 7 points. This increase brings their means muchcloser to the general population mean of 50 and indicates a level ofpsychological distress that is comparable with the general population.

Further evidence of the strong relationship of fatigue topsychological distress is provided by the significant correlation between thesemeasures. The correlations between change in the FACT-Fatigue subscale score andthe change in the BSI Depression and Anxiety scores were -0.350 (P < .0001)and -0.353 (P< .0001), respectively. The negative correlationresults from the fact that an increase in the FACT-Fatigue score indicatesimprovement in fatigue, whereas an increase in the BSI measures indicatesgreater reporting of severity or number of depressive and anxiety symptoms.

Discussion

The effective treatment of cancer-related anemia witherythropoietin results in significant improvements in hemoglobin levels, andrising hemoglobin levels are associated with improvements in energy levels,activity levels, and quality of life.[26,27] Cancer-related anemia and theassociated fatigue have, therefore, gained clinical importance as clinicianshave effective interventions to reverse them.

Growing evidence suggests, however, that anemia may havebroader and deeper implications for cancer patients, including an impact oncancer response to radiation therapy[28,29] and chemotherapy,[30] and an adverseeffect on psychologic status. In fact, our data indicate that anxiety anddepression scores improve when anemia-related fatigue scores improve in cancerpatients. These findings provide background information that supports theinclusion of psychologic distress variables as outcome measures in prospectiveanemia treatment clinical trials.

Intuitively, it seems clinically reasonable to expect thatthe decrease in functional ability caused by anemia would result in anxiety andpossibly depression. In this context, anxiety and depression are reactions tothe physical dysfunction created by anemia. Cancer patients, in their dailylives, are concerned with practical issues, such as being able to carry outtheir usual activities of daily life, as well as being independent.Anemia-related fatigue can be devastating not only physically, but alsoemotionally. In turn, these effects may lead to additional adverse outcomes,such as willingness to continue treatment, increase in health-care resourceutilization, increased need for caregiver time,[31] and, where applicable,reduced productivity in the workplace by both the patient and the informalcaregiver(s).

Anemia, therefore, is no longer an expected and ignoredbyproduct of a cancer diagnosis and its treatment, but an important cause ofphysical and emotional morbidity that must be recognized and addressed byclinicians. The relationship between hemoglobin change, fatigue, andpsychological outcomes needs further investigation in which other covariates—includinganalgesic/opioid use, antidepressive medication use, red blood cell transfusion,and neutropenia—are explored.

These analyses were based on data pooled from two trials. Theinclusion and exclusion criteria were virtually identical. However, the US trialincluded multiple types of cancer, whereas the other trial, conducted in Canada,Australia, and Europe, only enrolled lung cancer patients. We conducted theanalyses on each trial separately, and our findings were consistent with theoverall results. Therefore, we believe pooling these data is appropriate forevaluating the association between fatigue and psychological outcomes. We note,however, that confounding factors could result from the pooling.

These analyses were exploratory and not prespecified, andstatistical tests were not adjusted for multiple comparisons. Furtherprospective trials should be conducted to confirm these findings. Given theconsistent and highly significant data, the results should be viewed asproviding relatively strong evidence of a relationship between fatigue andpsychological outcomes in patients being treated for cancer. Previous studieshave shown a single point-in-time relationship between fatigue and psychologicaloutcomes, further supporting our findings here.

Conclusions

These data indicate that improvements in anemia-relatedfatigue in cancer patients are associated with reduced symptoms of anxiety anddepression. Improving hemoglobin has significant patient benefits beyond areduction in the chances of receiving a red blood transfusion. These includealleviation of fatigue (which cancer patients consistently rate as the mostdeleterious effect of cancer and its treatment),[32] improvements in physicalwell-being, improvements in depressive symptoms, reductions in feelings ofanxiety and panic, and improvements in overall health. Fatigue, depression, andother psychological outcomes are increasingly being recognized as critical forthe comprehensive and appropriate treatment of patients with cancer. Patientsbeing treated for cancer often are burdened by significant physical andemotional symptoms that affect their quality of life. These issues have recentlybeen raised at the National Institutes of Health conference on SymptomManagement in Cancer: Pain, Depression and Fatigue, where many experts calledfor increased attention to be paid to the identification, treatment, andmonitoring of these symptoms.[33]

References:

1. Strain J, Newcorn JH: Adjustment disorder, in Tasman A(ed): Psychiatry, 1st ed, p 1280. Philadelphia, WB Saunders, 1997.

2. Skarstein J, Aass N, Fossa SD, et al: Anxiety anddepression in cancer patients: Relations between the Hospital Anxiety andDepression Scale and the European Organization for Research and Treatment ofCancer Core Quality of Life Questionnaire. J Psychosom Res 49:27-34,2000.

3. Morasso G, Constantini M, Viterbori P, et al: Predictingmood disorders in breast cancer patients. Eur J Cancer 37:216-223, 2001.

4. Pascoe S, Edelman S, Kidman A: Prevalence of psychologicaldistress and use of support services by cancer patients at Sydney hospitals. AustN Z J Psychiatry 34:785-791, 2000.

5. Derogatis LR, Morrow GR, Fetting J, et al: The prevalenceof psychiatric disorders among cancer patients. JAMA 249:751-757, 1983.

6. Zabora J, Brintzenhofeszoc K, Curbow B, et al: Theprevalence of psychological distress by cancer site. Psychooncology10:19-28, 2001.

7. Colleoni M, Mandala M, Peruzzotti G, et al: Depression anddegree of acceptance of adjuvant cytotoxic drugs. Lancet 356:1326-1327,2000.

8. Groopman JE, Itri LM: Chemotherapy-induced anemia inadults: Incidence and treatment. J Natl Cancer Inst 91:1616-1634,1999.

9. Abels RI: Recombinant human erythropoietin in thetreatment of the anaemia of cancer. Acta Haematologica 87(suppl 1):4-11,1992.

10. Vansteenkiste J, Pirker R, Massuti B, et al:Double-blind, placebo-controlled, randomized, phase III trial of darbepoetinalfa in lung cancer patients receiving chemotherapy. J Natl Cancer Inst.In press.

11. Derogatis LR: Brief Symptom Inventory: Administration,Scoring, and Procedures Manual, 4th ed. Minneapolis, National ComputerSystems, 1993.

12. Stefanek ME, Derogatis LR, Shaw A: Psychological distressamong oncology outpatients: Prevalence and severity as measured with the BriefSymptom Inventory. Psychosomatics; 28:530-539, 1987.

13. Schain WS, Wellisch DK, Pasnau RO, et al: The sooner thebetter: A study of psychological factors in women undergoing immediate versusdelayed breast reconstruction. Am J Psychiatry 142:40-46, 1983.

14. Schover LR, Fife M, Gershenson DM: Sexual dysfunction andtreatment for early-stage cervical cancer. Cancer 63:202-212, 1989.

15. Zabora JR: Pragmatic approaches in the screeningprocedures for psychosocial distress, in Holland JC (ed): The Handbook ofPsycho-Oncology, 2nd ed, pp 652-661. New York, Oxford University Press,1998.

16. Kornblith AB, Herndon JE, Zuckerman E, et al: Comparisonof psychosocial adaptation of advanced stage Hodgkin’s disease and acuteleukemia survivors. Cancer and Leukemia Group B. Ann Oncol 9:297-306,1998.

17. Cella D, Tulsky DS, Gray G, et al: TheFunctional Assessment of Cancer Therapy scale: Development and validation of thegeneral measure. J Clin Oncol 11:570-579,1993.

18. Cella D: The Functional Assessment of Cancer-Anemia(FACT-An scale): A new tool for the assessment of outcomes in cancer anemia andfatigue. Semin Hematol 34(3 suppl 2):13-19 1997.

19. Yellen S, Cella D, Webster K, et al: Measuringfatigue and other anemia-related symptoms with the Functional Assessment ofCancer Therapy (FACT) measurement system. J Pain Symptom Manage 13:63-74,1997.

20. Cella D, Eton DT, Lai JS, et al: Combining anchor anddistribution based methods to derive clinically meaningful differences on theFunctional Assessment of Cancer Therapy (FACT) Anemia and Fatigue Scales. JPain Symptom Manage. In press.

21. Cella D, Lai J, Chang C, et al: Fatigue in cancerpatients compared with fatigue in the general United States population. Cancer94:528-538, 2002.

22. Duchateau L, Sylvester R: Meta-analysis, in Crowley J(ed): Handbook of Statistics in Clinical Oncology, pp 525-543. New York,Marcel Dekker, 2001.

23. Early Breast Cancer Trials’ Collaborative Group:Effects of radiotherapy and surgery in early breast cancer: An overview of therandomized trials. N Engl J Med 333:1444-1455, 1995.

24. Non-small Cell Cancer Collaborative Group: Chemotherapyin non-small cell lung cancer: A meta-analysis using updated data on individualpatients from 52 randomized clinical trials. BMJ 311:899-909, 1995.

25. Clarke M, Stewart L, Pignon JP, et al: Individual patientdata meta-analyses in cancer. Br J Cancer 77:2036-2044, 1998.

26. Glaspy J, Bukowski R, Steinberg D, et al: Impact oftherapy with epoetin alfa on clinical outcomes in patients with nonmyeloidmalignancies during cancer chemotherapy in community oncology practice. JClin Oncol 15:1218-1234, 1997.

27. Littlewood TJ, Bajetta E, Nortier JWR, et al: Effects ofepoetin alfa on hematologic parameters and quality of life in cancer patientsreceiving nonplatinum chemotherapy: Results of a randomized, double-blind,placebo-controlled trial. J Clin Oncol 19:2865-2874, 2001.

28. Smaniooto D, Luzi S, Morganti A: Prognostic significanceof anemia and the role of erythropoietin in radiation therapy. Tumori86:17-23, 2000.

29. Pirker R, Vansteenkiste J, Gately J, et al: A phase III,double-blind, placebo-controlled, randomized study of novel erythropoiesisstimulating protein (NESP) in patients undergoing platinum treatment for lungcancer (abstract 1772). Proc Am Soc Clin Oncol 20:394a, 2001.

30. MacRae R, Choy H, Shyr Y, et al.: Significance ofhemoglobin changes during chemotherapy in non-small cell lung cancer (NSCLC): Ananalysis of three sequential trials (abstract 1336). Proc Am Soc Clin Oncol20:335a, 2001.

31. Kallich J, Erder H, Glaspy J, et al: Darbepoetin alfa hashigher observed improvements in fatigue and physical well-being than epoetinalfa (abstract 1466). Proc Am Soc Clin Oncol 21:367a, 2002.

32. Curt G, Breitbart W, Cella D, et al: Impact ofcancer-related fatigue on the lives of patients: New findings from the FatigueCoalition. Oncologist 5:353-60, 2000.

33. Symptom Management in Cancer: Pain, Depression andFatigue. Presented at the state-of-the-science conference of the US Departmentof Health and Human Services, National Institutes of Health; July 15-17, 2002;Bethesda, Md.