Fatigue in Long-Term Hodgkin's Disease Survivors: A Follow-Up Study
http://www.100md.com
《临床肿瘤学》
the Ullev?l University Hospital HF
Department of Behavioral Sciences in Medicine, University of Oslo
Departments of Medical Oncology and Clinical Research, the Norwegian Radium Hospital, Oslo
Department of Clinical and Molecular Medicine, Faculty of Medicine, the Norwegian University of Technology and Science, Trondheim, Norway
ABSTRACT
PURPOSE: To describe total fatigue (TF) and chronic fatigue (CF) in 476 long-term Hodgkin's disease survivors (HDSs). The development in CF over time was explored in 280 of the patients who had also been assessed 8 years earlier.
PATIENTS AND METHODS: In 2003, the Fatigue Questionnaire was mailed to 610 successfully treated HDSs at the Norwegian Radium Hospital from 1971 to 1997. Mean TF scores and occurrence of CF were compared with general population (GP) values.
RESULTS: Four hundred seventy-six complete forms (81%) were received (median age, 46 years; 56% males, median follow-up time, 195 months). There was a positive association between age and TF (P < .05), whereas presence of B symptoms at diagnosis and treatment before 1980 were associated with CF. Mean TF scores were elevated in HDSs compared with the GP (mean TF score, 14.6; 95% CI, 14.1 to 15.7 v 12.1; 95% CI, 11.9 to 12.3, respectively; P < .001), as was the proportion of persons with CF (30% v 11%, respectively; odds ratio = 3.6; P < .001). The 70 patients with CF 8 years earlier still reported higher TF at follow-up than the 210 patients without CF at the previous assessment (mean TF score, 17.0; 95% CI, 15.6 to 18.3 v 13.1; 95% CI, 12.9 to 14.1, respectively; P < .001). Significantly more patients with persisting CF had B symptoms at diagnosis compared with patients who had recovered (P = .05). No significant association with treatment modality and intensity was found.
CONCLUSION: Fatigue remains a major complaint in HDSs several years after treatment, but the association between treatment and fatigue still remains unclear. Many HDSs may recover from CF, particularly patients without B symptoms at diagnosis.
INTRODUCTION
The 5-year survival for all patients with Hodgkin's disease (HD) exceeds 80%, and for patients who are younger than age 40 years at diagnosis, more than 90% are expected to live free from HD for 5 years or more after diagnosis.1 Consequently, systematic assessment of late effects is of special relevance.
Fatigue is a frequently reported symptom among cancer patients in general2-5 and in survivors of HD in particular.6-13 Studies in HD survivors (HDSs) that have used fatigue-specific questionnaires, such as the Multidimensional Fatigue Inventory14 and the Fatigue Questionnaire (FQ),15 have shown that up to 27% of the patients met the criteria for chronic fatigue (CF),13,16 which was defined as elevated fatigue levels that lasted for 6 months or longer.17 This percentage was significantly higher compared with the general population.13,16
Among the few comparative studies, there are indications that fatigue may be more prevalent among HDSs than among survivors of other cancer types.9,18,19 One study comparing relatively large samples of HDSs (n = 249) and testicular cancer survivors (n = 791)19 showed a significantly higher prevalence of CF in HDSs (16% v 24%, respectively), which was also demonstrated in an earlier, smaller study by Bloom et al.9 Another of these comparative studies reported results from unselected HDSs compared with other unselected cancer survivors with similar follow-up time.19
Despite the indication from the literature of more fatigue in HDSs compared with other cancer survivors, little is known about the development in fatigue levels over time. Furthermore, the relationship between persisting fatigue and disease and treatment characteristics remains unclear, partly because the close connection between disease stage and type of treatment renders separate analyses difficult. The only prospective report on fatigue in HDSs showed that combination treatment with radiotherapy and chemotherapy resulted in higher levels of fatigue, as assessed by the quality-of-life (QOL) tool Short Form 36, during the first year compared with radiation therapy alone, but the levels were similar between the two groups after 2 years.10
In the last two decades, the treatment regimens for HD have been gradually modified to reduce the overall toxicity. However, there is no documentation on the association between treatment burden and the longitudinal course of fatigue, and we were not able to identify any studies on the development of fatigue in HDSs beyond the first year by the use of fatigue-specific instruments.
Thus, the present study was launched in 2002 to assess the level of total fatigue (TF) and the prevalence of CF by the FQ in HDSs treated at the Norwegian Radium Hospital (NRH) from 1971 to 1997 relative to age- and sex-adjusted data from the general population (GP). The second objective was to examine the relationship between TF and CF and medical and treatment-related variables. As a result, a new method for quantification of the total treatment was attempted to look for associations between treatment burden and fatigue outcomes. Because no reports have investigated the development in CF over time and because a subset of our sample had also completed the FQ 8 years earlier in 1994, the third and novel objective was to examine the development in TF and CF in patients who were suffering from CF at the previous assessment.
PATIENTS AND METHODS
Patients
The majority of Norwegian cancer patients (92%) in the age group between 15 and 39 years who were diagnosed with HD before 1980 received their treatment at the NRH. The corresponding percentages in the age groups of 40 to 59 years and 60 years were 80% and 53%, respectively.20 Although the treatment of HD gradually became more decentralized after 1980, the patients in this report constitute a nationally representative sample because the hospital has had treatment responsibility for more than 50% of the patient population.
Treatment Summary
For the patients included in this report, the therapeutic regimens for HD at the NRH can largely be divided into the following three time periods.
1971 to 1979. Stage IA to IIB patients received mantle-field (a selection of standard blocks was adapted for lung shielding) or inverted Y–field radiotherapy (2 Gy x 20). Patients with stage III to IV disease were treated with eight cycles of chemotherapy (mustine, vincristine, procarbazine, and prednisone; or chlorambucil, vinblastine, procarbazine, prednisone [ChlVPP]) that was supplemented by radiotherapy when indicated (2 Gy x 20). Some stage III patients received total nodal irradiation (mantle-field and inverted Y–field radiotherapy).
1980 to 1986. Stage I to IIB patients with risk factors for relapse (B symptoms, bulky tumor, histologic lymphocyte-depleted type, or four or more involved sites) received four cycles of ChlVPP or doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) or two of each of the same chemotherapy combinations before irradiation. Stage IA to IIA patients without risk factors received irradiation only (2 Gy x 20). Stage III to IV patients received chemotherapy (eight cycles of ChlVPP or alternating ChlVPP/ABVD) that was supplemented with radiotherapy when indicated (2 Gy x 20).
1987 to 1997. The radiotherapy was modified to reduce long-term adverse effects. The fractionation was altered to 1.8 Gy x 23 (total, 41.4 Gy), and a subcarinal block was adapted after 30.6 Gy if there was no sign of disease below carina at the time of diagnosis. Furthermore, to improve the lung shielding, individual blocks were made. An anterior or posterior field was treated every other day until 1990; thereafter, both fields were administered daily. Chemotherapy was largely administered with ABVD as the preferred regimen from 1993 for patients with extensive disease and epirubicin, bleomycin, vinblastine, and prednisone as the preferred regimen from 1988 for patients with limited disease with risk factors. Full details of the treatment regimens are described in detail elsewhere.21-23
Quantification of Total Treatment Burden
In line with the study objectives, we attempted to quantify the total treatment burden for each patient: The number of chemotherapy cycles administered for primary and salvage therapy was counted and summarized without weighting the agents. Accumulated toxicity of individual cytostatic agents was not taken into account by this approach. The 25 different types of single or combined radiation fields (dose range, 40 to 41.4 Gy) were evaluated by four experienced oncologists and a radiation physicist. The total radiation volume (in mL) was estimated for two patients based on the dose planning. Each field was assigned a score derived from the average of the estimates based on the assumption that large radiation volumes might be associated with fatigue. Mantle-field irradiation volume was weighted to 5 mL, and the volumes of the other fields were given numbers relative to this, as presented in Table 1.
For the quantification of total treatment burden, the patients were divided into the following four groups on the basis of radiation volume (Table 1) and number of chemotherapy cycles (Table 2): (1) less than the median dose of radiotherapy and less than the median number of chemotherapy cycles; (2) less than the median dose of radiotherapy and the median number of chemotherapy cycles; (3) more than the median dose of radiotherapy and less than the median number of chemotherapy cycles; and (4) more than the median dose of radiotherapy and the median number of chemotherapy cycles.
The eligibility criteria for the present study in 2002 included the following: a verified diagnosis of HD and treatment at the NRH in the years from 1971 to 1997; age of 15 years at the time of diagnosis; age from 18 to 74 years at the time of assessment; complete remission at the time of assessment; no treatment for HD the previous year; and no secondary cancer. A total of 611 adult patients met the inclusion criteria. Ten patients were no longer registered at the Norwegian Census Bureau and could not be contacted. Ten patients who had received the majority of their treatment in another health region and who had recently participated in another postal survey on psychosocial late effects were not contacted.
The 591 eligible patients were contacted by mail. They received a questionnaire packet consisting of the FQ and other questionnaires.15 A total of 479 patients returned the questionnaire packets. However, three patients failed to fill out the FQ and were excluded from analyses in this report. Thus, the response rate for the FQ overall was 81% (476 of 591 patients). The respondents had a median age of 46 years (range, 21 to 73 years), and 56% were males (n = 267).
Fifty-nine percent (n = 280) of the patients had participated in a similar study including the FQ 8 years earlier (1994).13 They were significantly older compared with the remaining 196 patients in this cross-sectional sample, with a median age of 49 years (range, 23 to 74 years) v 43 years (range, 21 to 74 years), respectively, and had a significantly longer observation time (223 v 113 months, respectively; P < .001) from diagnosis to assessment. No other demographic or medical differences were found across these groups.
Missing items on the completed forms were less than 0.05%. To avoid omitting patients with partially incomplete data, mean imputation was used for scale scores when patients completed at least 50% of the scale, according to the methods recommended by the European Organisation for Research and Treatment of Cancer Manual.24 All clinical variables were retrieved from the lymphoma database at the NRH.
All FQ scores were compared with reference values from a Norwegian GP survey.25 A nationwide representative sample of 3,500 Norwegian adults aged 19 to 80 years had received the FQ by mail, yielding a response rate of 67%. Women reported higher levels of TF than men, and the prevalence of CF was 11%. TF and age were weakly correlated (r < 0.2). Patients' TF scores and population data within the same age range were compared after adjusting for age, sex, and education.
Questionnaires
The FQ15,26,27 asks about fatigue symptoms experienced during the last month compared with how the subjects felt when they were last well. It consists of 11 items that were generated to reflect physical and mental fatigue. Judged by inspection, the items correspond well with six of the proposed criteria for cancer-related fatigue.28 The FQ was originally validated in primary care and has demonstrated good face and discriminant validity and good and stable psychometric properties across populations.13,15,25-27,29 Physical fatigue corresponds to the subjective feeling of being exhausted and lacking energy, whereas mental fatigue describes the subjective feeling of being mentally exhausted, encompassing items on concentration, memory, and speech. The seven items on physical fatigue and the four items on mental fatigue all have four response categories (0 = better than usual; 1 = no more than usual; 2 = worse than usual; and 3 = much worse than usual). Thus, higher scores imply more fatigue, and the sum score of all 11 items was designated TF, with a maximum scale score of 33. The FQ contains two additional items about the duration and extent of fatigue. All 11 items were also scored on a dichotomized scale (0 = better than usual and no more than usual; and 1 = worse than usual and much worse than usual) for the four response categories for the identification of CF.15 By comparing the FQ with the fatigue question in the Revised Clinical Interview Schedule, a relative operating characteristics analysis suggested that a cutoff point of 4 or higher on the dichotomized scale would be the optimum cutoff for a case definition. We have preferred to use the concept of CF instead of fatigue caseness in this article. In addition to a dichotomized score of 4 or higher, CF also implies a duration of 6 months or longer, according to a consensus definition.17
Ethical Considerations
Data collection was conducted according to the guidelines of the Helsinki Declaration. The Regional Committee for Medical Research Ethics, Health Region I, Norway, and the Institutional Review Board at the NRH approved the study. Appropriate informed consent was obtained from all respondents.
Statistical Analyses
No gold standard exists regarding the clinically significant numerical changes on the FQ. Thus, the magnitude of the differences in TF for the statistically significant results in the univariate analyses and for the differences between patients and the GP was presented as effect size (ie, the mean score difference divided by the mean difference of the standard deviation [SD] across groups). The effect size accounts for the variability of scores making statistically significant differences less important if they occur in a scale with a large range of values and, hence, a large SD.30 To interpret effect sizes, we followed the suggestion of Cohen31 and regarded effect sizes of 0.2 to 0.5 as small, 0.5 to 0.8 as moderate, and 0.8 as large.
Clinically significant changes may also be reported through a subjective approach, as frequently used in QOL studies.32 Here, mean score differences from 5 to 10 on 0- to 100-point scales were reported as small but perceptible changes by the patients, mean score differences from 10 to 20 were reported as moderate changes, and differences greater than 20 were regarded as large changes. As such, differences of approximately ± 3.3 points on the FQ representing a 10% change in TF was regarded as indicating clinical significance, whereas differences of ± 1.5 points (7%) were regarded as small but perceptible differences.
Standard descriptive statistics were used, including 2 and Kruskal-Wallis tests (nominal categoric variables), Wilcoxon's test (two tailed) for independent samples, and one-way analysis of variance for differences in medical and demographic characteristics and for the variables related to the specified study objectives, as appropriate, after checking for multicollinearity. Linear and logistic regression analyses were used to identify associations between background variables and CF by entering those variables that were significant at P = .05 in the univariate analyses. Because there were only two primary study outcomes (TF and CF), P .05 was defined as the level of significance. Analyses were performed using SPSS Statistical Software version 12.0 (SPSS Inc, Chicago, IL).
RESULTS
There were more males than females among the nonresponders (21% v 14.5%, respectively; P < .05), whereas no significant differences were found with age, observation time, primary treatment, and relapse between responders and nonresponders.
Background and Medical Variables
The patients were all white, with a median age of 46 years (range, 21 to 74 years), and the majority (69%, n = 328) was working or studying. A significantly higher proportion of the patients had higher education compared with the GP sample (37% v 29%, respectively; P < .01).
Sixty-three percent of the patients (n = 300) were diagnosed with stage I or II disease. Most patients (55%) had received combination treatment, whereas 149 patients (31%) were treated with radiotherapy only. Mantle-field irradiation alone, in combination with other fields or as part of the multimodality treatment, was most frequently administered (administered to 72% of the patients who were irradiated). Fifteen patients had received autologous stem-cell transplantation. Eleven percent of the patients (54 patients) had relapsed during follow-up.
TF and CF
Patient characteristics, the distribution of TF mean scores, and the proportion of patients with CF are listed in Table 2. Age greater than the median and treatment before 1980 were positively associated with higher TF mean scores (both P < .05; effect sizes = 0.2 and 0.3, respectively). Presence of CF was associated with the presence of B symptoms at diagnosis (P < .05) and treatment before 1980 (P < .05). No association between TF and treatment intensity or modality was found.
A multiple regression analysis, with TF as the dependent variable and including the variables that yielded statistically significant results in the univariate analyses (age at follow-up and period of treatment) as independent variables, revealed a significant association between age and high TF levels (P < .05). A logistic regression analysis was performed with CF as the dependent variable and presence of B symptoms and period of treatment as explanatory variables (Table 3). Both variables were significantly associated with CF (both P < .05).
Patients Versus the GP
The mean TF score was significantly elevated compared with values from the GP, adjusted for age, sex, and education (mean TF score, 14.6; 95% CI, 14.1 to 15.7 v 12.1; 95% CI, 11.9 to 12.3, respectively; P < .001; Table 4). The effect size of the difference was moderate (effect size = 0.6).31 The magnitude of the score difference of 2.5 might also be characterized as moderate based on the subjective method, representing a 7.6% change on the FQ.32 The age- and sex-adjusted proportion of patients with CF was significantly higher among patients than among the GP (30% v 11%, respectively; P < .001; odds ratio = 3.6).
Patients With CF at First Assessment
Among the 280 patients who had completed the FQ 8 years previously, the following four groups with different development of CF could be identified: patients with CF at both assessments (n = 35); patients with CF at first assessment but not at follow-up (n = 35); patients without CF at first assessment but who fulfilled the criteria at follow-up (n = 43); and patients with no CF at either assessment (n = 167). There were more patients with B symptoms at diagnosis among patients who met the criteria for CF at both assessments (P = .05) compared with patients who recovered (Table 5). No significant differences in sociodemographic or treatment-related variables were found. The TF mean score at the first assessment was significantly lower in the 167 patients with no CF at either assessment compared with the other three groups (mean TF score, 12.0; SD, 3.0; P < .001, data not tabulated).
When comparing the 70 patients who were classified as having CF at the first assessment with patients who were not (n = 210, 75%), there was a significant decline (P .001) in TF score in patients with CF, from 19.6 (95% CI, 18.6 to 20.6) to 17.0 (95% CI, 15.6 to 18.3) at follow-up, compared with an increase in TF score in patients without CF at first assessment, from 12.5 (95% CI, 12.0 to 13.0) to 13.4 (95% CI, 12.9 to 14.1; P < .01; Fig 1). The TF scores were significantly elevated relative to GP values adjusted for age, sex, and education (P < .001), with a clinical effect size of 0.6.
DISCUSSION
This study confirms that fatigue is prevalent several years after successful treatment for HD and that HDSs report elevated mean levels compared with GP values decades after diagnosis, which is in line with previous reports.8,10,13,16 This is the first report that has examined the course of CF over time. Our results might be interpreted positively because more than 50% of patients with a previous condition of CF did not report CF 8 years later. From a clinical point of view, this is new knowledge that can be used in patient information about what to expect after successful treatment. The only significant association for persisting CF at follow-up was the presence of B symptoms at diagnosis. The mean TF in patients who suffered from CF 8 years earlier also showed a significant decline over time in contrast to the mean level in patients who did not meet the criteria for CF at that time. It cannot be ruled out that this decline was a result of a regression to the mean effect because of the high initial level. However, both mean levels were statistically and clinically significantly elevated relative to age- and sex-adjusted population data.
Our tentative quantification of treatment burden did not seem to influence the results; thus, we were unable to demonstrate associations between treatment and fatigue in line with other reports.8,13,19,33 Patients who were treated before 1980 reported higher TF, and adverse effects from extended radiation fields administered in the 1970s cannot be ruled out. However, the elevated TF levels in the univariate analyses were more likely related to the higher age of the survivors.25 The changes in treatment regimens were gradually introduced, and subgroup analyses comparing treatment protocols could not be performed. However, the changes were consistently in the direction of less toxic regimens, and as such, one could expect less fatigue in patients who were treated with milder regimens (adjusted for age). This was not found. The lack of relationship between treatment and fatigue is consistent with a new report that found no difference in fatigue associated with treatment modality beyond the first year.10 This might suggest that the fatigue induced by treatment might persist for some time before it subsides.
We anticipated that extensive radiation fields or mantle-field irradiation would be related to increased levels of fatigue,34 but this was not supported by our data. Seventy-two percent of patients who were irradiated received mantle-field irradiation, which might have obscured the findings. Pulmonary complications have been documented after mantle-field irradiation,35,36 leading to a treatment modification with individualized lung-shielding fields from 1971 to 1997, followed by a further reduction in radiotherapy volume and dose from 1997.
Clinical importance is not equivalent to statistically significant differences across groups. Thus, to make interpretation easier, we presented the widely used clinical approach by Cohen,31 as well as the subjective significance based on responses from cancer patients.32 Despite certain limitations to the use of the two methods, there was a high agreement between the two methods and the statistically significant analyses in this report. The differences in TF between age and treatment groups, as presented in Table 2, were small according to both approaches. When patients' TF scores were compared with scores from the GP, effect sizes were moderate and, thus, important to patients. The 20% difference in the prevalence of CF between HDSs and the reference GP supports the assertion that fatigue is a major problem several years after treatment.
Presence of B symptoms, regardless of stage of disease, was the only disease characteristic that was significantly associated with TF and CF, as documented by others.6,13,16 B symptoms were significantly more prevalent in patients who met the criteria for CF at both assessments compared with patients who did not. In our opinion, this lends support to the hypothesis that fatigue is part of the HD symptomatology, as set forth in previous studies.6,10
There is a lack of studies exploring possible etiologic explanations for fatigue. Hypotheses about Epstein-Barr virus,37 altered immuno- or neuroendocrinologic responses,38 and increased cytokine circulation during radiotherapy39 have been launched, but to our knowledge, no evidence-based knowledge about the etiology of fatigue in HDSs exists.
There were no significant differences between responders and nonresponders with respect to background variables. However, a response bias can never be ruled out. If only patients with major problems respond, the fatigue problem in HDSs would seem worse than reality, relative to the population at large. However, if only patients with few symptoms respond, there could be an underestimation of the true prevalence. In this respect, it should be remembered that the reference population consisted of individuals who suffered from the normal spectrum of physical and psychological conditions and that fatigue is the final common end point of a variety of conditions and life events for all people and not only for cancer survivors. This was the reason for performing comparisons that were adjusted for age, sex, and education. The compliance in the reference population might be regarded as low (67%). Still, these data represent the best estimate of fatigue in the Norwegian population. The response rate is well in line with similar surveys presenting data on QOL questionnaires, such as the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire C30 and the Short Form 36.40,41
It might be regarded as a limitation of our study that we did not present data on various physical conditions and psychological symptoms that might have impacted the reported levels of fatigue or on the possible interventions and treatments on the part of the patients for the relief of fatigue in the period between assessments. The perception of fatigue may be different in former cancer patients compared with the GP, particularly after treatment for a life-threatening disease. The items of the FQ refer to fatigue symptoms experienced during the last month compared with how the patients felt when they were last well. With this somewhat imprecise time format, it cannot be ruled out that the sensation of fatigue might be related to conditions other than HD. If this were the case, elevated levels of fatigue might be erroneously related to HD. However, various comorbid conditions that are related to fatigue are prevalent also in the GP. Thus, this should be controlled for by performing age- and sex-adjusted comparisons with the reference population (supported by the statistically and clinically significant differences in line with other reports).
Psychiatric morbidity was not assessed in the present study. One study demonstrated an association between CF, anxiety, and depression in HDSs, but no associations between CF and self-reported psychiatric symptoms before diagnosis or during treatment were reported.33 However, the level of anxiety and depression was significantly higher in CF individuals in the GP than in HDSs.33,42 CF is distressing but might also be part of a psychiatric condition such as depression. Better understanding of the association between psychiatric conditions and fatigue is warranted but requires another study design other than the present design.
Few studies specifically examine medical late effects in HDSs (overt or more subtle) and their association with fatigue. A recent study in disease-free HDSs treated with chest radiotherapy revealed subclinical cardiac abnormalities in 47 of 48 patients at screening,43 which might suggest a relationship between asymptomatic heart problems and fatigue. However, our clinical experience leans more in the direction of a hypothesis that fatigue is associated with reduced oxygen uptake after mantle-field irradiation.34 A possible relationship is further substantiated by a pilot study in HDSs with CF who participated in an aerobic exercise program for 20 weeks.44 Fatigue was significantly reduced, and the aerobic capacity was significantly improved, which is in line with results from a randomized trial of aerobic exercise versus standard care in a heterogeneous sample of cancer patients.45
The major strengths of this study are related to the high compliance rate among patients treated at a single institution, the use of representative population data for comparison, the use of a fatigue-specific questionnaire, and a design that enabled us to assess the course of CF. To date, no evidence-based data for recommendation of treatment of fatigue in HDSs exist, other than treatment of the underlying cause if identified and symptomatic treatment if necessary. The results from the pilot study with aerobic training are interesting but need to be replicated in larger samples. However, physical training might be effective only in a subset of HDSs with CF, as discussed in a recent review.46
In conclusion, our results, combined with previous studies, support the hypothesis of a multifactorial etiology of fatigue after HD. Although fatigue in HDSs is significantly higher relative to values from the GP, it does not seem to increase over time. A large proportion of the patients with CF can expect to improve their condition, which is promising news that can be used in patient information and counseling. Because fatigue is the final common end point for several physical and psychological conditions, large, prospective, comparative studies with fatigue and medical late effects as separate study outcomes are warranted.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Professor Thorbj?rn Moum, PhD, University of Oslo, for valuable advice and help regarding the statistical analyses and radiation physicist Eva Stabell Bergstrand, PhD, the Norwegian Radium Hospital (NRH), for her efforts in the estimation of radiation volumes. We also thank the Office of Clinical Research, NRH, for retrieving data and preparing files for analysis and the National Cancer Society for supporting this work.
NOTES
Supported by the National Cancer Society.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Aleman BM, van den Belt-Dusebout AW, Klokman WJ, et al: Long-term cause-specific mortality of patients treated for Hodgkin's disease. J Clin Oncol 21:3431-3439, 2003
Smets EM, Garssen B, Schuster-Uitterhoeve AL, et al: Fatigue in cancer patients. Br J Cancer 68:220-224, 1993
Curt GA: The impact of fatigue on patients with cancer: Overview of FATIGUE 1 and 2. Oncologist 5:9-12, 2000 (suppl 2)
Stone P, Richards M, Hardy J: Fatigue in patients with cancer. Eur J Cancer 34:1670-1676, 1998
Loge JH: Unpacking fatigue. Eur J Pall Care 10:14-20, 2003
Fobair P, Hoppe RT, Bloom J, et al: Psychosocial problems among survivors of Hodgkin's disease. J Clin Oncol 4:805-814, 1986
Flechtner H, Ruffer JU, Henry-Amar M, et al: Quality of life assessment in Hodgkin's disease: A new comprehensive approach—First experiences from the EORTC/GELA and GHSG trials: EORTC Lymphoma Cooperative Group, Groupe D'Etude des Lymphomes de L'Adulte, and German Hodgkin Study Group. Ann Oncol 9:S147-S154, 1998 (suppl 5)
Joly F, Henry-Amar M, Arveux P, et al: Late psychosocial sequelae in Hodgkin's disease survivors: A French population-based case-control study. J Clin Oncol 14:2444-2453, 1996
Bloom JR, Fobair P, Gritz E, et al: Psychosocial outcomes of cancer: A comparative analysis of Hodgkin's disease and testicular cancer. J Clin Oncol 11:979-988, 1993
Ganz PA, Moinpour CM, Pauler DK, et al: Health status and quality of life in patients with early-stage Hodgkin's disease treated on Southwest Oncology Group Study 9133. J Clin Oncol 21:3512-3519, 2003
Wettergren L, Bjorkholm M, Axdorph U, et al: Individual quality of life in long-term survivors of Hodgkin's lymphoma: A comparative study. Qual Life Res 12:545-554, 2003
van Tulder MW, Aaronson NK, Bruning PF: The quality of life of long-term survivors of Hodgkin's disease. Ann Oncol 5:153-158, 1994
Loge JH, Abrahamsen AF, Ekeberg O, et al: Hodgkin's disease survivors more fatigued than the general population. J Clin Oncol 17:253-261, 1998
Smets EM, Garssen B, Bonke B, et al: The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res 39:315-325, 1995
Chalder T, Berelowitz G, Pawlikowska T, et al: Development of a fatigue scale. J Psychosom Res 37:147-153, 1993
Ruffer JU, Flechtner H, Tralls P, et al: Fatigue in long-term survivors of Hodgkin's lymphoma: A report from the German Hodgkin Lymphoma Study Group (GHSG). Eur J Cancer 39:2179-2186, 2003
Fukuda K, Straus SE, Hickie I, et al: The chronic fatigue syndrome: A comprehensive approach to its definition and study: International Chronic Fatigue Syndrome Study Group. Ann Intern Med 121:953-959, 1994
Kornblith AB, Herndon JE, Zuckerman E, et al: Comparison of psychosocial adaptation of advanced stage Hodgkin's disease and acute leukemia survivors: Cancer and Leukemia Group B. Ann Oncol 9:297-306, 1998
Fossa SD, Dahl AA, Loge JH: Fatigue, anxiety, and depression in long-term survivors of testicular cancer. J Clin Oncol 21:1249-1254, 2003
Abrahamsen AF, Egeland T, Hansen S, et al: Hodgkin's disease in a national and hospital population: Trends over 20 years. Eur J Cancer 33:2380-2383, 1997
Holte H, Mella O, Wist E, et al: ChlVPP is as effective as alternating ChlVPP/ABOD in advanced stage Hodgkin's disease. Acta Oncol 35:73-80, 1996 (suppl 8)
Abrahamsen AF, Hannisdal E, Nome O, et al: Clinical stage I and II Hodgkin's disease: Long-term results of therapy without laparotomy—Experience at one institution. Ann Oncol 7:145-150, 1996
Abrahamsen AF, Host H: Mantle field irradiation for stages IA and IIA Hodgkin's disease. Scand J Haematol 26:306-310, 1981
Fayers PM, Aaronson N, Bjordal K, et al: EORTC QLQ-C30: Scoring Manual. Brussels, Belgium, European Organisation for Research and Treatment of Cancer Quality of Life Study Group, 1995
Loge JH, Ekeberg O, Kaasa S: Fatigue in the general Norwegian population: Normative data and associations. J Psychosom Res 45:53-65, 1998
Pawlikowska T, Chalder T, Hirsch SR, et al: Population based study of fatigue and psychological distress. BMJ 308:763-766, 1994
Wessely S, Powell R: Fatigue syndromes: A comparison of chronic "postviral" fatigue with neuromuscular and affective disorders. J Neurol Neurosurg Psychiatry 52:940-948, 1989
Cella D, Davis K, Breitbart W, et al: Cancer-related fatigue: Prevalence of proposed diagnostic criteria in a United States sample of cancer survivors. J Clin Oncol 19:3385-3391, 2001
Stone P, Hardy J, Huddart R, et al: Fatigue in patients with prostate cancer receiving hormone therapy. Eur J Cancer 36:1134-1141, 2000
Fayers PM, Machin D: Quality of life: Assessment, analysis and interpretation. Chichester, United Kingdom, John Wiley and Sons Ltd, 2000
Cohen J: Statistical Power Analysis for the Behavioral Sciences. New York, NY, Lawrence Erlbaum Associates, 1977
Osoba D, Rodrigues G, Myles J, et al: Interpreting the significance of changes in health-related quality-of-life scores. J Clin Oncol 16:139-144, 1998
Loge JH, Abrahamsen AF, Ekeberg, et al: Fatigue and psychiatric morbidity among Hodgkin's disease survivors. J Pain Symptom Manage 19:91-99, 2000
Knobel H, Loge JH, Lund M, et al: Late medical complications and fatigue in Hodgkin's disease survivors. J Clin Oncol 19:3226-3233, 2001
Lund MB, Kongerud J, Nome O, et al: Lung function impairment in long-term survivors of Hodgkin's disease. Ann Oncol 6:495-501, 1995
Loyer E, Fuller L, Libshitz HI, et al: Radiographic appearance of the chest following therapy for Hodgkin disease. Eur J Radiol 35:136-148, 2000
Wessely S: The epidemiology of chronic fatigue syndrome. Epidemiol Psichiatr Soc 7:10-24, 1998
Greenberg DB, Kornblith AB, Herndon JE, et al: Quality of life of adult leukemia survivors treated on clinical trials of Cancer and Leukemia Group B during the period 1971-1988. Cancer 80:1936-1944, 1997
Cleare AJ, Bearn J, Allain T, et al: Contrasting neuroendocrine responses in depression and chronic fatigue syndrome. J Affect Disord 34:283-289, 1995
Hjermstad MJ, Fayers PM, Bjordal K, et al: Health-related quality of life in the general Norwegian population assessed by the EORTC Core Quality of Life Questionnaire: The QLQ-C30 (+3). J Clin Oncol 16:1188-1196, 1998
Jenkinson C, Coulter A, Wright L: Short Form 36 (SF36) health survey questionnaire: Normative data for adults of working age. BMJ 306:1437-1440, 1993
Wessely S, Chalder T, Hirsch S, et al: Psychological symptoms, somatic symptoms, and psychiatric disorder in chronic fatigue and chronic fatigue syndrome: A prospective study in the primary care setting. Am J Psychiatry 153:1050-1059, 1996
Adams MJ, Lipsitz SR, Colan SD, et al: Cardiovascular status in long-term survivors of Hodgkin's disease treated with chest radiotherapy. J Clin Oncol 22:3139-3148, 2004
Oldervoll LM, Kaasa S, Knobel H, et al: Exercise reduces fatigue in chronic fatigued Hodgkin's disease survivors: Results from a pilot study. Eur J Cancer 39:57-63, 2003
Thorsen L, Skovlund E, Stromme SB, et al: Effectiveness of physical activity on cardiorespiratory fitness and health-related quality of life in young and middle-aged cancer patients shortly after chemotherapy. J Clin Oncol 23:2378-2388, 2005
Oldervoll LM, Kaasa S, Hjermstad MJ, et al: Physical exercise results in the improved subjective well-being of a few or is effective rehabilitation for all cancer patients? Eur J Cancer 40:951-962, 2004(Marianne J. Hjermstad, So)
Department of Behavioral Sciences in Medicine, University of Oslo
Departments of Medical Oncology and Clinical Research, the Norwegian Radium Hospital, Oslo
Department of Clinical and Molecular Medicine, Faculty of Medicine, the Norwegian University of Technology and Science, Trondheim, Norway
ABSTRACT
PURPOSE: To describe total fatigue (TF) and chronic fatigue (CF) in 476 long-term Hodgkin's disease survivors (HDSs). The development in CF over time was explored in 280 of the patients who had also been assessed 8 years earlier.
PATIENTS AND METHODS: In 2003, the Fatigue Questionnaire was mailed to 610 successfully treated HDSs at the Norwegian Radium Hospital from 1971 to 1997. Mean TF scores and occurrence of CF were compared with general population (GP) values.
RESULTS: Four hundred seventy-six complete forms (81%) were received (median age, 46 years; 56% males, median follow-up time, 195 months). There was a positive association between age and TF (P < .05), whereas presence of B symptoms at diagnosis and treatment before 1980 were associated with CF. Mean TF scores were elevated in HDSs compared with the GP (mean TF score, 14.6; 95% CI, 14.1 to 15.7 v 12.1; 95% CI, 11.9 to 12.3, respectively; P < .001), as was the proportion of persons with CF (30% v 11%, respectively; odds ratio = 3.6; P < .001). The 70 patients with CF 8 years earlier still reported higher TF at follow-up than the 210 patients without CF at the previous assessment (mean TF score, 17.0; 95% CI, 15.6 to 18.3 v 13.1; 95% CI, 12.9 to 14.1, respectively; P < .001). Significantly more patients with persisting CF had B symptoms at diagnosis compared with patients who had recovered (P = .05). No significant association with treatment modality and intensity was found.
CONCLUSION: Fatigue remains a major complaint in HDSs several years after treatment, but the association between treatment and fatigue still remains unclear. Many HDSs may recover from CF, particularly patients without B symptoms at diagnosis.
INTRODUCTION
The 5-year survival for all patients with Hodgkin's disease (HD) exceeds 80%, and for patients who are younger than age 40 years at diagnosis, more than 90% are expected to live free from HD for 5 years or more after diagnosis.1 Consequently, systematic assessment of late effects is of special relevance.
Fatigue is a frequently reported symptom among cancer patients in general2-5 and in survivors of HD in particular.6-13 Studies in HD survivors (HDSs) that have used fatigue-specific questionnaires, such as the Multidimensional Fatigue Inventory14 and the Fatigue Questionnaire (FQ),15 have shown that up to 27% of the patients met the criteria for chronic fatigue (CF),13,16 which was defined as elevated fatigue levels that lasted for 6 months or longer.17 This percentage was significantly higher compared with the general population.13,16
Among the few comparative studies, there are indications that fatigue may be more prevalent among HDSs than among survivors of other cancer types.9,18,19 One study comparing relatively large samples of HDSs (n = 249) and testicular cancer survivors (n = 791)19 showed a significantly higher prevalence of CF in HDSs (16% v 24%, respectively), which was also demonstrated in an earlier, smaller study by Bloom et al.9 Another of these comparative studies reported results from unselected HDSs compared with other unselected cancer survivors with similar follow-up time.19
Despite the indication from the literature of more fatigue in HDSs compared with other cancer survivors, little is known about the development in fatigue levels over time. Furthermore, the relationship between persisting fatigue and disease and treatment characteristics remains unclear, partly because the close connection between disease stage and type of treatment renders separate analyses difficult. The only prospective report on fatigue in HDSs showed that combination treatment with radiotherapy and chemotherapy resulted in higher levels of fatigue, as assessed by the quality-of-life (QOL) tool Short Form 36, during the first year compared with radiation therapy alone, but the levels were similar between the two groups after 2 years.10
In the last two decades, the treatment regimens for HD have been gradually modified to reduce the overall toxicity. However, there is no documentation on the association between treatment burden and the longitudinal course of fatigue, and we were not able to identify any studies on the development of fatigue in HDSs beyond the first year by the use of fatigue-specific instruments.
Thus, the present study was launched in 2002 to assess the level of total fatigue (TF) and the prevalence of CF by the FQ in HDSs treated at the Norwegian Radium Hospital (NRH) from 1971 to 1997 relative to age- and sex-adjusted data from the general population (GP). The second objective was to examine the relationship between TF and CF and medical and treatment-related variables. As a result, a new method for quantification of the total treatment was attempted to look for associations between treatment burden and fatigue outcomes. Because no reports have investigated the development in CF over time and because a subset of our sample had also completed the FQ 8 years earlier in 1994, the third and novel objective was to examine the development in TF and CF in patients who were suffering from CF at the previous assessment.
PATIENTS AND METHODS
Patients
The majority of Norwegian cancer patients (92%) in the age group between 15 and 39 years who were diagnosed with HD before 1980 received their treatment at the NRH. The corresponding percentages in the age groups of 40 to 59 years and 60 years were 80% and 53%, respectively.20 Although the treatment of HD gradually became more decentralized after 1980, the patients in this report constitute a nationally representative sample because the hospital has had treatment responsibility for more than 50% of the patient population.
Treatment Summary
For the patients included in this report, the therapeutic regimens for HD at the NRH can largely be divided into the following three time periods.
1971 to 1979. Stage IA to IIB patients received mantle-field (a selection of standard blocks was adapted for lung shielding) or inverted Y–field radiotherapy (2 Gy x 20). Patients with stage III to IV disease were treated with eight cycles of chemotherapy (mustine, vincristine, procarbazine, and prednisone; or chlorambucil, vinblastine, procarbazine, prednisone [ChlVPP]) that was supplemented by radiotherapy when indicated (2 Gy x 20). Some stage III patients received total nodal irradiation (mantle-field and inverted Y–field radiotherapy).
1980 to 1986. Stage I to IIB patients with risk factors for relapse (B symptoms, bulky tumor, histologic lymphocyte-depleted type, or four or more involved sites) received four cycles of ChlVPP or doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) or two of each of the same chemotherapy combinations before irradiation. Stage IA to IIA patients without risk factors received irradiation only (2 Gy x 20). Stage III to IV patients received chemotherapy (eight cycles of ChlVPP or alternating ChlVPP/ABVD) that was supplemented with radiotherapy when indicated (2 Gy x 20).
1987 to 1997. The radiotherapy was modified to reduce long-term adverse effects. The fractionation was altered to 1.8 Gy x 23 (total, 41.4 Gy), and a subcarinal block was adapted after 30.6 Gy if there was no sign of disease below carina at the time of diagnosis. Furthermore, to improve the lung shielding, individual blocks were made. An anterior or posterior field was treated every other day until 1990; thereafter, both fields were administered daily. Chemotherapy was largely administered with ABVD as the preferred regimen from 1993 for patients with extensive disease and epirubicin, bleomycin, vinblastine, and prednisone as the preferred regimen from 1988 for patients with limited disease with risk factors. Full details of the treatment regimens are described in detail elsewhere.21-23
Quantification of Total Treatment Burden
In line with the study objectives, we attempted to quantify the total treatment burden for each patient: The number of chemotherapy cycles administered for primary and salvage therapy was counted and summarized without weighting the agents. Accumulated toxicity of individual cytostatic agents was not taken into account by this approach. The 25 different types of single or combined radiation fields (dose range, 40 to 41.4 Gy) were evaluated by four experienced oncologists and a radiation physicist. The total radiation volume (in mL) was estimated for two patients based on the dose planning. Each field was assigned a score derived from the average of the estimates based on the assumption that large radiation volumes might be associated with fatigue. Mantle-field irradiation volume was weighted to 5 mL, and the volumes of the other fields were given numbers relative to this, as presented in Table 1.
For the quantification of total treatment burden, the patients were divided into the following four groups on the basis of radiation volume (Table 1) and number of chemotherapy cycles (Table 2): (1) less than the median dose of radiotherapy and less than the median number of chemotherapy cycles; (2) less than the median dose of radiotherapy and the median number of chemotherapy cycles; (3) more than the median dose of radiotherapy and less than the median number of chemotherapy cycles; and (4) more than the median dose of radiotherapy and the median number of chemotherapy cycles.
The eligibility criteria for the present study in 2002 included the following: a verified diagnosis of HD and treatment at the NRH in the years from 1971 to 1997; age of 15 years at the time of diagnosis; age from 18 to 74 years at the time of assessment; complete remission at the time of assessment; no treatment for HD the previous year; and no secondary cancer. A total of 611 adult patients met the inclusion criteria. Ten patients were no longer registered at the Norwegian Census Bureau and could not be contacted. Ten patients who had received the majority of their treatment in another health region and who had recently participated in another postal survey on psychosocial late effects were not contacted.
The 591 eligible patients were contacted by mail. They received a questionnaire packet consisting of the FQ and other questionnaires.15 A total of 479 patients returned the questionnaire packets. However, three patients failed to fill out the FQ and were excluded from analyses in this report. Thus, the response rate for the FQ overall was 81% (476 of 591 patients). The respondents had a median age of 46 years (range, 21 to 73 years), and 56% were males (n = 267).
Fifty-nine percent (n = 280) of the patients had participated in a similar study including the FQ 8 years earlier (1994).13 They were significantly older compared with the remaining 196 patients in this cross-sectional sample, with a median age of 49 years (range, 23 to 74 years) v 43 years (range, 21 to 74 years), respectively, and had a significantly longer observation time (223 v 113 months, respectively; P < .001) from diagnosis to assessment. No other demographic or medical differences were found across these groups.
Missing items on the completed forms were less than 0.05%. To avoid omitting patients with partially incomplete data, mean imputation was used for scale scores when patients completed at least 50% of the scale, according to the methods recommended by the European Organisation for Research and Treatment of Cancer Manual.24 All clinical variables were retrieved from the lymphoma database at the NRH.
All FQ scores were compared with reference values from a Norwegian GP survey.25 A nationwide representative sample of 3,500 Norwegian adults aged 19 to 80 years had received the FQ by mail, yielding a response rate of 67%. Women reported higher levels of TF than men, and the prevalence of CF was 11%. TF and age were weakly correlated (r < 0.2). Patients' TF scores and population data within the same age range were compared after adjusting for age, sex, and education.
Questionnaires
The FQ15,26,27 asks about fatigue symptoms experienced during the last month compared with how the subjects felt when they were last well. It consists of 11 items that were generated to reflect physical and mental fatigue. Judged by inspection, the items correspond well with six of the proposed criteria for cancer-related fatigue.28 The FQ was originally validated in primary care and has demonstrated good face and discriminant validity and good and stable psychometric properties across populations.13,15,25-27,29 Physical fatigue corresponds to the subjective feeling of being exhausted and lacking energy, whereas mental fatigue describes the subjective feeling of being mentally exhausted, encompassing items on concentration, memory, and speech. The seven items on physical fatigue and the four items on mental fatigue all have four response categories (0 = better than usual; 1 = no more than usual; 2 = worse than usual; and 3 = much worse than usual). Thus, higher scores imply more fatigue, and the sum score of all 11 items was designated TF, with a maximum scale score of 33. The FQ contains two additional items about the duration and extent of fatigue. All 11 items were also scored on a dichotomized scale (0 = better than usual and no more than usual; and 1 = worse than usual and much worse than usual) for the four response categories for the identification of CF.15 By comparing the FQ with the fatigue question in the Revised Clinical Interview Schedule, a relative operating characteristics analysis suggested that a cutoff point of 4 or higher on the dichotomized scale would be the optimum cutoff for a case definition. We have preferred to use the concept of CF instead of fatigue caseness in this article. In addition to a dichotomized score of 4 or higher, CF also implies a duration of 6 months or longer, according to a consensus definition.17
Ethical Considerations
Data collection was conducted according to the guidelines of the Helsinki Declaration. The Regional Committee for Medical Research Ethics, Health Region I, Norway, and the Institutional Review Board at the NRH approved the study. Appropriate informed consent was obtained from all respondents.
Statistical Analyses
No gold standard exists regarding the clinically significant numerical changes on the FQ. Thus, the magnitude of the differences in TF for the statistically significant results in the univariate analyses and for the differences between patients and the GP was presented as effect size (ie, the mean score difference divided by the mean difference of the standard deviation [SD] across groups). The effect size accounts for the variability of scores making statistically significant differences less important if they occur in a scale with a large range of values and, hence, a large SD.30 To interpret effect sizes, we followed the suggestion of Cohen31 and regarded effect sizes of 0.2 to 0.5 as small, 0.5 to 0.8 as moderate, and 0.8 as large.
Clinically significant changes may also be reported through a subjective approach, as frequently used in QOL studies.32 Here, mean score differences from 5 to 10 on 0- to 100-point scales were reported as small but perceptible changes by the patients, mean score differences from 10 to 20 were reported as moderate changes, and differences greater than 20 were regarded as large changes. As such, differences of approximately ± 3.3 points on the FQ representing a 10% change in TF was regarded as indicating clinical significance, whereas differences of ± 1.5 points (7%) were regarded as small but perceptible differences.
Standard descriptive statistics were used, including 2 and Kruskal-Wallis tests (nominal categoric variables), Wilcoxon's test (two tailed) for independent samples, and one-way analysis of variance for differences in medical and demographic characteristics and for the variables related to the specified study objectives, as appropriate, after checking for multicollinearity. Linear and logistic regression analyses were used to identify associations between background variables and CF by entering those variables that were significant at P = .05 in the univariate analyses. Because there were only two primary study outcomes (TF and CF), P .05 was defined as the level of significance. Analyses were performed using SPSS Statistical Software version 12.0 (SPSS Inc, Chicago, IL).
RESULTS
There were more males than females among the nonresponders (21% v 14.5%, respectively; P < .05), whereas no significant differences were found with age, observation time, primary treatment, and relapse between responders and nonresponders.
Background and Medical Variables
The patients were all white, with a median age of 46 years (range, 21 to 74 years), and the majority (69%, n = 328) was working or studying. A significantly higher proportion of the patients had higher education compared with the GP sample (37% v 29%, respectively; P < .01).
Sixty-three percent of the patients (n = 300) were diagnosed with stage I or II disease. Most patients (55%) had received combination treatment, whereas 149 patients (31%) were treated with radiotherapy only. Mantle-field irradiation alone, in combination with other fields or as part of the multimodality treatment, was most frequently administered (administered to 72% of the patients who were irradiated). Fifteen patients had received autologous stem-cell transplantation. Eleven percent of the patients (54 patients) had relapsed during follow-up.
TF and CF
Patient characteristics, the distribution of TF mean scores, and the proportion of patients with CF are listed in Table 2. Age greater than the median and treatment before 1980 were positively associated with higher TF mean scores (both P < .05; effect sizes = 0.2 and 0.3, respectively). Presence of CF was associated with the presence of B symptoms at diagnosis (P < .05) and treatment before 1980 (P < .05). No association between TF and treatment intensity or modality was found.
A multiple regression analysis, with TF as the dependent variable and including the variables that yielded statistically significant results in the univariate analyses (age at follow-up and period of treatment) as independent variables, revealed a significant association between age and high TF levels (P < .05). A logistic regression analysis was performed with CF as the dependent variable and presence of B symptoms and period of treatment as explanatory variables (Table 3). Both variables were significantly associated with CF (both P < .05).
Patients Versus the GP
The mean TF score was significantly elevated compared with values from the GP, adjusted for age, sex, and education (mean TF score, 14.6; 95% CI, 14.1 to 15.7 v 12.1; 95% CI, 11.9 to 12.3, respectively; P < .001; Table 4). The effect size of the difference was moderate (effect size = 0.6).31 The magnitude of the score difference of 2.5 might also be characterized as moderate based on the subjective method, representing a 7.6% change on the FQ.32 The age- and sex-adjusted proportion of patients with CF was significantly higher among patients than among the GP (30% v 11%, respectively; P < .001; odds ratio = 3.6).
Patients With CF at First Assessment
Among the 280 patients who had completed the FQ 8 years previously, the following four groups with different development of CF could be identified: patients with CF at both assessments (n = 35); patients with CF at first assessment but not at follow-up (n = 35); patients without CF at first assessment but who fulfilled the criteria at follow-up (n = 43); and patients with no CF at either assessment (n = 167). There were more patients with B symptoms at diagnosis among patients who met the criteria for CF at both assessments (P = .05) compared with patients who recovered (Table 5). No significant differences in sociodemographic or treatment-related variables were found. The TF mean score at the first assessment was significantly lower in the 167 patients with no CF at either assessment compared with the other three groups (mean TF score, 12.0; SD, 3.0; P < .001, data not tabulated).
When comparing the 70 patients who were classified as having CF at the first assessment with patients who were not (n = 210, 75%), there was a significant decline (P .001) in TF score in patients with CF, from 19.6 (95% CI, 18.6 to 20.6) to 17.0 (95% CI, 15.6 to 18.3) at follow-up, compared with an increase in TF score in patients without CF at first assessment, from 12.5 (95% CI, 12.0 to 13.0) to 13.4 (95% CI, 12.9 to 14.1; P < .01; Fig 1). The TF scores were significantly elevated relative to GP values adjusted for age, sex, and education (P < .001), with a clinical effect size of 0.6.
DISCUSSION
This study confirms that fatigue is prevalent several years after successful treatment for HD and that HDSs report elevated mean levels compared with GP values decades after diagnosis, which is in line with previous reports.8,10,13,16 This is the first report that has examined the course of CF over time. Our results might be interpreted positively because more than 50% of patients with a previous condition of CF did not report CF 8 years later. From a clinical point of view, this is new knowledge that can be used in patient information about what to expect after successful treatment. The only significant association for persisting CF at follow-up was the presence of B symptoms at diagnosis. The mean TF in patients who suffered from CF 8 years earlier also showed a significant decline over time in contrast to the mean level in patients who did not meet the criteria for CF at that time. It cannot be ruled out that this decline was a result of a regression to the mean effect because of the high initial level. However, both mean levels were statistically and clinically significantly elevated relative to age- and sex-adjusted population data.
Our tentative quantification of treatment burden did not seem to influence the results; thus, we were unable to demonstrate associations between treatment and fatigue in line with other reports.8,13,19,33 Patients who were treated before 1980 reported higher TF, and adverse effects from extended radiation fields administered in the 1970s cannot be ruled out. However, the elevated TF levels in the univariate analyses were more likely related to the higher age of the survivors.25 The changes in treatment regimens were gradually introduced, and subgroup analyses comparing treatment protocols could not be performed. However, the changes were consistently in the direction of less toxic regimens, and as such, one could expect less fatigue in patients who were treated with milder regimens (adjusted for age). This was not found. The lack of relationship between treatment and fatigue is consistent with a new report that found no difference in fatigue associated with treatment modality beyond the first year.10 This might suggest that the fatigue induced by treatment might persist for some time before it subsides.
We anticipated that extensive radiation fields or mantle-field irradiation would be related to increased levels of fatigue,34 but this was not supported by our data. Seventy-two percent of patients who were irradiated received mantle-field irradiation, which might have obscured the findings. Pulmonary complications have been documented after mantle-field irradiation,35,36 leading to a treatment modification with individualized lung-shielding fields from 1971 to 1997, followed by a further reduction in radiotherapy volume and dose from 1997.
Clinical importance is not equivalent to statistically significant differences across groups. Thus, to make interpretation easier, we presented the widely used clinical approach by Cohen,31 as well as the subjective significance based on responses from cancer patients.32 Despite certain limitations to the use of the two methods, there was a high agreement between the two methods and the statistically significant analyses in this report. The differences in TF between age and treatment groups, as presented in Table 2, were small according to both approaches. When patients' TF scores were compared with scores from the GP, effect sizes were moderate and, thus, important to patients. The 20% difference in the prevalence of CF between HDSs and the reference GP supports the assertion that fatigue is a major problem several years after treatment.
Presence of B symptoms, regardless of stage of disease, was the only disease characteristic that was significantly associated with TF and CF, as documented by others.6,13,16 B symptoms were significantly more prevalent in patients who met the criteria for CF at both assessments compared with patients who did not. In our opinion, this lends support to the hypothesis that fatigue is part of the HD symptomatology, as set forth in previous studies.6,10
There is a lack of studies exploring possible etiologic explanations for fatigue. Hypotheses about Epstein-Barr virus,37 altered immuno- or neuroendocrinologic responses,38 and increased cytokine circulation during radiotherapy39 have been launched, but to our knowledge, no evidence-based knowledge about the etiology of fatigue in HDSs exists.
There were no significant differences between responders and nonresponders with respect to background variables. However, a response bias can never be ruled out. If only patients with major problems respond, the fatigue problem in HDSs would seem worse than reality, relative to the population at large. However, if only patients with few symptoms respond, there could be an underestimation of the true prevalence. In this respect, it should be remembered that the reference population consisted of individuals who suffered from the normal spectrum of physical and psychological conditions and that fatigue is the final common end point of a variety of conditions and life events for all people and not only for cancer survivors. This was the reason for performing comparisons that were adjusted for age, sex, and education. The compliance in the reference population might be regarded as low (67%). Still, these data represent the best estimate of fatigue in the Norwegian population. The response rate is well in line with similar surveys presenting data on QOL questionnaires, such as the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire C30 and the Short Form 36.40,41
It might be regarded as a limitation of our study that we did not present data on various physical conditions and psychological symptoms that might have impacted the reported levels of fatigue or on the possible interventions and treatments on the part of the patients for the relief of fatigue in the period between assessments. The perception of fatigue may be different in former cancer patients compared with the GP, particularly after treatment for a life-threatening disease. The items of the FQ refer to fatigue symptoms experienced during the last month compared with how the patients felt when they were last well. With this somewhat imprecise time format, it cannot be ruled out that the sensation of fatigue might be related to conditions other than HD. If this were the case, elevated levels of fatigue might be erroneously related to HD. However, various comorbid conditions that are related to fatigue are prevalent also in the GP. Thus, this should be controlled for by performing age- and sex-adjusted comparisons with the reference population (supported by the statistically and clinically significant differences in line with other reports).
Psychiatric morbidity was not assessed in the present study. One study demonstrated an association between CF, anxiety, and depression in HDSs, but no associations between CF and self-reported psychiatric symptoms before diagnosis or during treatment were reported.33 However, the level of anxiety and depression was significantly higher in CF individuals in the GP than in HDSs.33,42 CF is distressing but might also be part of a psychiatric condition such as depression. Better understanding of the association between psychiatric conditions and fatigue is warranted but requires another study design other than the present design.
Few studies specifically examine medical late effects in HDSs (overt or more subtle) and their association with fatigue. A recent study in disease-free HDSs treated with chest radiotherapy revealed subclinical cardiac abnormalities in 47 of 48 patients at screening,43 which might suggest a relationship between asymptomatic heart problems and fatigue. However, our clinical experience leans more in the direction of a hypothesis that fatigue is associated with reduced oxygen uptake after mantle-field irradiation.34 A possible relationship is further substantiated by a pilot study in HDSs with CF who participated in an aerobic exercise program for 20 weeks.44 Fatigue was significantly reduced, and the aerobic capacity was significantly improved, which is in line with results from a randomized trial of aerobic exercise versus standard care in a heterogeneous sample of cancer patients.45
The major strengths of this study are related to the high compliance rate among patients treated at a single institution, the use of representative population data for comparison, the use of a fatigue-specific questionnaire, and a design that enabled us to assess the course of CF. To date, no evidence-based data for recommendation of treatment of fatigue in HDSs exist, other than treatment of the underlying cause if identified and symptomatic treatment if necessary. The results from the pilot study with aerobic training are interesting but need to be replicated in larger samples. However, physical training might be effective only in a subset of HDSs with CF, as discussed in a recent review.46
In conclusion, our results, combined with previous studies, support the hypothesis of a multifactorial etiology of fatigue after HD. Although fatigue in HDSs is significantly higher relative to values from the GP, it does not seem to increase over time. A large proportion of the patients with CF can expect to improve their condition, which is promising news that can be used in patient information and counseling. Because fatigue is the final common end point for several physical and psychological conditions, large, prospective, comparative studies with fatigue and medical late effects as separate study outcomes are warranted.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Professor Thorbj?rn Moum, PhD, University of Oslo, for valuable advice and help regarding the statistical analyses and radiation physicist Eva Stabell Bergstrand, PhD, the Norwegian Radium Hospital (NRH), for her efforts in the estimation of radiation volumes. We also thank the Office of Clinical Research, NRH, for retrieving data and preparing files for analysis and the National Cancer Society for supporting this work.
NOTES
Supported by the National Cancer Society.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Aleman BM, van den Belt-Dusebout AW, Klokman WJ, et al: Long-term cause-specific mortality of patients treated for Hodgkin's disease. J Clin Oncol 21:3431-3439, 2003
Smets EM, Garssen B, Schuster-Uitterhoeve AL, et al: Fatigue in cancer patients. Br J Cancer 68:220-224, 1993
Curt GA: The impact of fatigue on patients with cancer: Overview of FATIGUE 1 and 2. Oncologist 5:9-12, 2000 (suppl 2)
Stone P, Richards M, Hardy J: Fatigue in patients with cancer. Eur J Cancer 34:1670-1676, 1998
Loge JH: Unpacking fatigue. Eur J Pall Care 10:14-20, 2003
Fobair P, Hoppe RT, Bloom J, et al: Psychosocial problems among survivors of Hodgkin's disease. J Clin Oncol 4:805-814, 1986
Flechtner H, Ruffer JU, Henry-Amar M, et al: Quality of life assessment in Hodgkin's disease: A new comprehensive approach—First experiences from the EORTC/GELA and GHSG trials: EORTC Lymphoma Cooperative Group, Groupe D'Etude des Lymphomes de L'Adulte, and German Hodgkin Study Group. Ann Oncol 9:S147-S154, 1998 (suppl 5)
Joly F, Henry-Amar M, Arveux P, et al: Late psychosocial sequelae in Hodgkin's disease survivors: A French population-based case-control study. J Clin Oncol 14:2444-2453, 1996
Bloom JR, Fobair P, Gritz E, et al: Psychosocial outcomes of cancer: A comparative analysis of Hodgkin's disease and testicular cancer. J Clin Oncol 11:979-988, 1993
Ganz PA, Moinpour CM, Pauler DK, et al: Health status and quality of life in patients with early-stage Hodgkin's disease treated on Southwest Oncology Group Study 9133. J Clin Oncol 21:3512-3519, 2003
Wettergren L, Bjorkholm M, Axdorph U, et al: Individual quality of life in long-term survivors of Hodgkin's lymphoma: A comparative study. Qual Life Res 12:545-554, 2003
van Tulder MW, Aaronson NK, Bruning PF: The quality of life of long-term survivors of Hodgkin's disease. Ann Oncol 5:153-158, 1994
Loge JH, Abrahamsen AF, Ekeberg O, et al: Hodgkin's disease survivors more fatigued than the general population. J Clin Oncol 17:253-261, 1998
Smets EM, Garssen B, Bonke B, et al: The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res 39:315-325, 1995
Chalder T, Berelowitz G, Pawlikowska T, et al: Development of a fatigue scale. J Psychosom Res 37:147-153, 1993
Ruffer JU, Flechtner H, Tralls P, et al: Fatigue in long-term survivors of Hodgkin's lymphoma: A report from the German Hodgkin Lymphoma Study Group (GHSG). Eur J Cancer 39:2179-2186, 2003
Fukuda K, Straus SE, Hickie I, et al: The chronic fatigue syndrome: A comprehensive approach to its definition and study: International Chronic Fatigue Syndrome Study Group. Ann Intern Med 121:953-959, 1994
Kornblith AB, Herndon JE, Zuckerman E, et al: Comparison of psychosocial adaptation of advanced stage Hodgkin's disease and acute leukemia survivors: Cancer and Leukemia Group B. Ann Oncol 9:297-306, 1998
Fossa SD, Dahl AA, Loge JH: Fatigue, anxiety, and depression in long-term survivors of testicular cancer. J Clin Oncol 21:1249-1254, 2003
Abrahamsen AF, Egeland T, Hansen S, et al: Hodgkin's disease in a national and hospital population: Trends over 20 years. Eur J Cancer 33:2380-2383, 1997
Holte H, Mella O, Wist E, et al: ChlVPP is as effective as alternating ChlVPP/ABOD in advanced stage Hodgkin's disease. Acta Oncol 35:73-80, 1996 (suppl 8)
Abrahamsen AF, Hannisdal E, Nome O, et al: Clinical stage I and II Hodgkin's disease: Long-term results of therapy without laparotomy—Experience at one institution. Ann Oncol 7:145-150, 1996
Abrahamsen AF, Host H: Mantle field irradiation for stages IA and IIA Hodgkin's disease. Scand J Haematol 26:306-310, 1981
Fayers PM, Aaronson N, Bjordal K, et al: EORTC QLQ-C30: Scoring Manual. Brussels, Belgium, European Organisation for Research and Treatment of Cancer Quality of Life Study Group, 1995
Loge JH, Ekeberg O, Kaasa S: Fatigue in the general Norwegian population: Normative data and associations. J Psychosom Res 45:53-65, 1998
Pawlikowska T, Chalder T, Hirsch SR, et al: Population based study of fatigue and psychological distress. BMJ 308:763-766, 1994
Wessely S, Powell R: Fatigue syndromes: A comparison of chronic "postviral" fatigue with neuromuscular and affective disorders. J Neurol Neurosurg Psychiatry 52:940-948, 1989
Cella D, Davis K, Breitbart W, et al: Cancer-related fatigue: Prevalence of proposed diagnostic criteria in a United States sample of cancer survivors. J Clin Oncol 19:3385-3391, 2001
Stone P, Hardy J, Huddart R, et al: Fatigue in patients with prostate cancer receiving hormone therapy. Eur J Cancer 36:1134-1141, 2000
Fayers PM, Machin D: Quality of life: Assessment, analysis and interpretation. Chichester, United Kingdom, John Wiley and Sons Ltd, 2000
Cohen J: Statistical Power Analysis for the Behavioral Sciences. New York, NY, Lawrence Erlbaum Associates, 1977
Osoba D, Rodrigues G, Myles J, et al: Interpreting the significance of changes in health-related quality-of-life scores. J Clin Oncol 16:139-144, 1998
Loge JH, Abrahamsen AF, Ekeberg, et al: Fatigue and psychiatric morbidity among Hodgkin's disease survivors. J Pain Symptom Manage 19:91-99, 2000
Knobel H, Loge JH, Lund M, et al: Late medical complications and fatigue in Hodgkin's disease survivors. J Clin Oncol 19:3226-3233, 2001
Lund MB, Kongerud J, Nome O, et al: Lung function impairment in long-term survivors of Hodgkin's disease. Ann Oncol 6:495-501, 1995
Loyer E, Fuller L, Libshitz HI, et al: Radiographic appearance of the chest following therapy for Hodgkin disease. Eur J Radiol 35:136-148, 2000
Wessely S: The epidemiology of chronic fatigue syndrome. Epidemiol Psichiatr Soc 7:10-24, 1998
Greenberg DB, Kornblith AB, Herndon JE, et al: Quality of life of adult leukemia survivors treated on clinical trials of Cancer and Leukemia Group B during the period 1971-1988. Cancer 80:1936-1944, 1997
Cleare AJ, Bearn J, Allain T, et al: Contrasting neuroendocrine responses in depression and chronic fatigue syndrome. J Affect Disord 34:283-289, 1995
Hjermstad MJ, Fayers PM, Bjordal K, et al: Health-related quality of life in the general Norwegian population assessed by the EORTC Core Quality of Life Questionnaire: The QLQ-C30 (+3). J Clin Oncol 16:1188-1196, 1998
Jenkinson C, Coulter A, Wright L: Short Form 36 (SF36) health survey questionnaire: Normative data for adults of working age. BMJ 306:1437-1440, 1993
Wessely S, Chalder T, Hirsch S, et al: Psychological symptoms, somatic symptoms, and psychiatric disorder in chronic fatigue and chronic fatigue syndrome: A prospective study in the primary care setting. Am J Psychiatry 153:1050-1059, 1996
Adams MJ, Lipsitz SR, Colan SD, et al: Cardiovascular status in long-term survivors of Hodgkin's disease treated with chest radiotherapy. J Clin Oncol 22:3139-3148, 2004
Oldervoll LM, Kaasa S, Knobel H, et al: Exercise reduces fatigue in chronic fatigued Hodgkin's disease survivors: Results from a pilot study. Eur J Cancer 39:57-63, 2003
Thorsen L, Skovlund E, Stromme SB, et al: Effectiveness of physical activity on cardiorespiratory fitness and health-related quality of life in young and middle-aged cancer patients shortly after chemotherapy. J Clin Oncol 23:2378-2388, 2005
Oldervoll LM, Kaasa S, Hjermstad MJ, et al: Physical exercise results in the improved subjective well-being of a few or is effective rehabilitation for all cancer patients? Eur J Cancer 40:951-962, 2004(Marianne J. Hjermstad, So)