Fatigue in ovarian carcinoma patients

A neglected issue?




Although fatigue is a commonly reported symptom in cancer patients it is rarely investigated, especially in patients with ovarian carcinoma. The main focus of the current study was to assess fatigue in these patients and to investigate the impact of fatigue and other clinical and psychosocial variables on their quality of life (QOL).


Ninety-eight ovarian carcinoma survivors (average age of 57.4 ± 12.5 years) were included in the study. All women had received cancer therapy but had not been treated for at least 6 months. The average time elapsed since first diagnosis was 5.7 ± 5.5 years. Fatigue was measured with the Multidimensional Fatigue Inventory (MFI-20) and QOL was measured with the Functional Assessment of Cancer Therapy (FACT)-ovarian carcinoma part and the European Organization for Research and Treatment of Cancer Care Questionnaire, including the ovarian carcinoma module.


Thirty-two of 98 ovarian carcinoma patients (32.7%, 95% confidence interval, 23.5–42.9%) reported MFI-20 General Fatigue scores ≥ 12.0 and therefore could be characterized as suffering from fatigue. This group of patients had a significantly lower QOL, had higher scores of anxiety and depression, and perceived that they had less social support. In a multiple regression model, mental adjustment, social support, anxiety, and depression as well as fatigue were significant predictors of QOL (FACT-generic part total score) whereas clinical and sociodemographic variables were not.


A remarkably high proportion of ovarian carcinoma survivors suffered from fatigue. Because this symptom is a key predictor of QOL, it should be given more attention in aftercare programs. Cancer 2003;97:1564–72. © 2003 American Cancer Society.

DOI 10.1002/cncr.11253

In Europe, 27,000–30,000 women are diagnosed with ovarian carcinoma annually.1 Medical treatment for ovarian carcinoma currently results in a 5-year survival rate of 57–89% for patients with early-stage ovarian carcinoma (International Federation of Gynecology and Obstetrics [FIGO] Stage I and II) and a rate of 11–24% for patients with advanced disease (FIGO Stage III and IV).2

Given the relative efficacy of treatment for ovarian carcinoma and the aversive nature of the treatment, especially the side effects of chemotherapy,3 it is argued that research should not focus solely on the immediate effects of treatment, disease-free intervals, and survival rates but also on the patients' quality of life (QOL). Within this area of research, the effects of ovarian carcinoma on long-term levels of effective functioning and on subjective well-being are of paramount interest. Both are mediated to a large extent by fatigue which is the most commonly reported symptom of cancer and cancer therapy4–6 and has a profound effect on patients' QOL.7

The specific etiology of cancer or treatment-induced fatigue is still poorly understood.8 Although it is well known that anemia is one of the main factors inducing fatigue9 in cancer patients, this syndrome may also be mediated by psychosocial factors such as depression or inadequate coping behavior.10 Despite the fact that depression and fatigue often occur concurrently in cancer patients, the causal relationship between these symptoms remains unclear.

Although there is a lack of specific studies investigating fatigue in ovarian carcinoma survivors in detail, a few groups have attempted to investigate more general QOL issues in this group of patients.

In a study of 200 ovarian carcinoma survivors without active disease for at least 2 years, Stewart et al.11 found that the patients reported better mental health and the same energy levels (i.e., level of fatigue) as the general population. Despite some impairments in the area of sexuality, the majority of the women reported that surviving ovarian carcinoma had a positive impact on their lives, especially with respect to their relationship with family and friends and their social roles.

Comparable results were observed by Ersek et al.12 in a survey of 152 women with ovarian carcinoma (54% had no evidence of active disease, 7% were receiving therapy, and 39% had active disease without treatment). The findings revealed that QOL is moderately high for this group of patients, despite having experienced some negative facets of both the illness and the associated treatment.

In contrast to the findings described above, Guidozzi3 reported that, in a study of 32 patients with ovarian carcinoma, treatment produced behavioral disruption and emotional distress, with lasting effects seen even 2 years later after commencing therapy in patients with complete response. Similarly, Bodurka-Bevers et al.13 found in a cohort of 246 ovarian carcinoma patients (49% were receiving therapy) that levels of depression and anxiety were higher in this group of patients than in the general population and that these levels were higher in patients with poor performance status.

In a study of 235 women with gynecologic carcinoma (24% had ovarian carcinoma), Carlsson et al.14 showed that patients previously treated with chemotherapy had poorer role and cognitive functioning and more problems with fatigue, nausea/emesis, dyspnea, constipation, and financial issues than those not treated with chemotherapy. Comparable results were found by Lutgendorf et al.15 in an investigation of 48 gynaecologic carcinoma patients (70.9% had ovarian carcinoma). Patients treated extensively with chemotherapy reported lower QOL, more fatigue, and less vigor compared with less intensively treated early-stage disease patients. Depression and anxiety scores did not differ between the groups.

All available evidence suggests a relation between impairments in QOL and the syndrome of fatigue. To test this hypothesis, we performed a cross-sectional study on fatigue in ovarian carcinoma patients to investigate the impact of fatigue and other clinical and psychosocial variables on QOL.

The study specifically addressed the following questions: 1) What is the prevalence of subjective fatigue in ovarian carcinoma patients? 2) In which domains of QOL do ovarian carcinoma patients with and without fatigue differ? 3) What impact does fatigue have on anxiety/depression, mental adjustment, and social support in ovarian carcinoma patients? 4) Can the effect of fatigue on QOL be explained by other clinical or psychosocial variables?

Based on the findings, appropriate medical and psychooncologic measures that improve the QOL of ovarian carcinoma survivors will be discussed. In answering these questions, implications for possible treatment strategies are also considered.



We investigated patients with a diagnosis of ovarian carcinoma who attended the outpatient unit of the Department of Gynaecology, University of Innsbruck. Subjects were recruited consecutively between February and October 2000. The inclusion criteria were a diagnosis of ovarian carcinoma, age younger than 75 years, no treatment for cancer in the last 6 months, without known active disease, German speaking, no cognitive impairments, expected survival time of at least 3 months, and obtainment of informed consent. Exclusion criteria were the presence of other severe mental or physical diseases at the time of investigation.

All patients were treated initially after 1980. Treatment included surgery in all cases; first-line chemotherapy (before 1997: FIGO stage above Stage IA/B, G > 1 patients were treated with cisplatin/cyclophosphomid; after 1997: FIGO stage > IA/B, G > 1 patients were treated with paclitaxel/carboplatin); second-line chemotherapy (recurrent-free survival > 6 months: patients were treated with carboplatin or carbotaxol; < 6 months: [patients were treated with] etoposide or topotecan [since 1997]); third-line chemotherapy (gemcitabine or etoposide or caelyx); radiotherapy in cases of solitary tumors; and regular medical aftercare visits every 3 months for 0–2 years, every 6 months for 2–5 years, and every 12 months 5 years or more after initial treatment.

During one of their aftercare visits at the outpatient unit of the Department of Gynaecology, subjects were asked to fill in questionnaires concerning fatigue (Multidimensional Fatigue Inventory [MFI-20]), QOL (European Organization for Research and Treatment of Cancer Care Questionnaire, including the ovarian carcinoma module [EORTC QLQ-C30/OV28], Functional Assessment of Cancer Therapy-ovarian carcinoma part [FACT-O]), anxiety and depression (Hospital Anxiety and Depression Scale [HADS]), mental adjustment (Mental Adjustment to Cancer Scale [MAC]), and social support (social support questionnaire [F-SOZU]). At the same visit, blood samples were collected and sociodemographic and clinical data, including hormone replacement and pharmacologic therapy, were obtained from the medical charts.

Assessment of Fatigue

The MFI-20 is a 20-item self-report instrument designed to measure fatigue8, 16–20 and includes the subscales General Fatigue, Physical Fatigue, Mental Fatigue, Reduced Motivation, and Reduced Activity. The MFI-20 has good psychometric properties.20

Physical Fatigue measures the physical sensations related to the feeling of tiredness. Possible somatic symptoms of fatigue such as light-headedness or sore muscles are not included in this subscale to exclude contamination with the symptoms of somatic illness that are independent of fatigue. Reduction in activities and lack of motivation are covered in the Reduced Activity and Reduced Motivation Subscales. Finally, cognitive symptoms, such as having difficulties concentrating, are included in the Mental Fatigue Subscale. Each subscale contains four items that are rated on a five-point Likert scale (higher scores indicate more fatigue). The use of a total score over all 20 items is not recommended. The General Fatigue Subscale score measures comprehensive aspects of fatigue.

Assessment of QOL

Quality of life was assessed using two different instruments, the European Organization for Research and Treatment of Cancer Core Questionnaire (EORTC QLC-C30), including the ovarian cancer module OV28,21 and the FACT-O,22 consisting of a generic part (FACT-G) and a module specific for ovarian cancer patients. Both questionnaires have good validity and reliability, both for the English original and for the translations into various languages including German.23, 24 Detailed descriptions [the EORTC QLQ-C30/OV28] were published earlier.25–31

The core questionnaires of the instruments are similar with respect to their length, type of response scale, time frame and, to a lesser extent, their subscale structure. Both have subscales that measure the physical, emotional, functional, and social aspects of QOL. In addition, the EORTC QLQ-C30 has additional subscales that assess cognitive functions, symptomatology, and the financial impact of the disease.

In spite of similar nomenclature, the two core instruments emphasize somewhat different aspects of QOL. Whereas the EORTC QLQ-C30 concentrates largely on physical functioning and clinical symptoms, the FACT-G has a stronger focus on social and emotional aspects.32 For this reason, both instruments were used in the study.

Assessment of Anxiety and Depression, Mental Adjustment, and Social Support

The HADS,33 which is a widely used self-rating scale for detecting states of depression (HADS-D) and anxiety (HADS-A), has good psychometric properties.34–36 The HADS-A and HADS-D subscales (each with seven items, with scales ranging from 0 [no distress] to 21 [maximum distress]) are also valid measures of the severity of the emotional disorders in clinical populations with symptoms of physical disease.36

The MAC37 is a 40-item self-rating scale that measures adjustment and coping style. The inventory consists of five subscales: Fighting Spirit, Anxious Preoccupation, Fatalism, Helplessness/Hopelessness, and Avoidance. Each item is rated on a four-point Likert scale, ranging from 1 (definitely does not apply to me) to 4 (definitely applies to me). A higher score denotes a greater tendency to adopt a specific coping style. Previous studies have found that the MAC is reliable and valid.37–39

The F-SOZU40 enables a differentiated assessment of the main aspects of social support, namely, emotional support, social integration, practical support, and social strain. It is a 54-item questionnaire that consists of four subscales corresponding to the dimensions of social support described above. The subscales Emotional Support, Social Integration, and Practical Support can be reduced to a single global scale, Perceived Social Support. The internal consistency of the subscales is above .80.

Partitioning of the Sample

To differentiate between ovarian carcinoma patients with and without fatigue, the sample was split into two subgroups according to the MFI-20 General Fatigue. Patients were allocated to the fatigue subgroup, if their General Fatigue subscore was greater than or equal to 12.0. The rationale for this cutoff score was that in a previous study,41 we found that the mean + 1 SD of the MFI-General Fatigue subscore was 12.0 in a healthy population.

Statistical Methods

MFI-20, FACT-O, EORTC QLQ-C30/OV28, HADS, MAC, and F-SOZU subscores were calculated according to the instructions of the developers. Comparisons between fatigued and nonfatigued groups used the Mann–Whitney U test for independent samples for continuous/ordinal variables and the chi-square test for dichotomous variables. To quantify differences between groups, effect sizes were determined (i.e., the difference between group means was divided by the average of the SD of the two groups).

Stepwise multiple linear regression was used to assess the degree to which the FACT total QOL score was determined by the various clinical and psychologic parameters measured in the study. Five regression analyses was performed, starting with only MFI-General Fatigue as an independent variable with increasing numbers of the remaining independent variables added in sequence. The number of independent variables eventually used in any of these regression analyses was limited to 10 by using only those independent variables from earlier analyses that had been statistically significant.


Patient Characteristics

One hundred twelve ovarian carcinoma patients originally consented to complete data sheets and assessment instruments. Six patients (5.4%) failed to complete the questionnaires, citing a lack of time, and eight (7.1%) were excluded because they were suffering from other diseases at the time of investigation. Therefore, 98 patients were analyzed in this study (age, 57.4 ± 12.5; time since initial diagnosis, 5.7 ± 5.5 years). Hormone replacement therapy was used by 27 (27.5%), antidepressants by 9 (9.2%), and benzodiazepines by 8 (8.2%) patients. All subjects had hemoglobin levels above 10 g/dL and therefore did not receive any treatment for anemia. A detailed description of the sociodemographic and clinical data is given in Table 1.

Table 1. Sociodemographic and Clinical Data for 98 Patients with Ovarian Carcinoma
  • FIGO: International Federation of Gynecology and Obstetrics.

  • a

    Mean (standard deviation).

  • b

    At the time of diagnosis.

Age (yrs)  
 57.4 (12.5)a  
Marital status  
 Married/with partner70(71.4)
 Compulsory school66(67.3)
 Completed apprenticeship28(28.6)
Time since first diagnosis (yrs)  
 5.7 (5.5)a  
FIGO stages  
 Stage I49(50.0)
  Stage IA/IB/IC21/4/24(21.4/4.1/24.5)
 Stage II9(9.2)
  Stage IIA/IIB/IIC2/3/4(2.0/3.1/4.1)
 Stage III37(37.8)
  Stage IIIA/IIIB/IIIC8/7/22(8.2/7.1/22.4)
 Stage IV3(3.1)
Hormone replacement therapy  
 Conjugated estrogens17(17.3)
 Micronized estradiol10(10.2)
Herbal supplements10(10.2)
Pharmacologic treatment  
Hemoglobin level (g/dL)  
 13.52 (1.87)a  
Hematocrit (%)  
 40.25 (5.57)a  

Table 2 provides an overview of the MFI-20 fatigue scores, namely, means and percentiles of the MFI-20 subscales. A more detailed analysis of the fatigue scores is given below.

Table 2. MFI-20 Subscales of the Total Sample of Ovarian Carcinoma Patients (n = 98)
MFI-20 subscalesaOvarian carcinoma (range, 4–20)Percentiles
  • MFI-20: Multidimensional Fatigue Inventory-20.

  • a

    Higher values indicate more fatigue.

General Fatigue9.4 ±
Physical Fatigue8.1 ±
Reduced Activity8.4 ±
Reduced Motivation7.4 ±
Mental Fatigue8.3 ±

Prevalence of Subjective Fatigue in Ovarian Carcinoma Patients

When the sample was divided according to the criterion described in the Materials Methods section (MFI-20 General Fatigue ≥ 12.0 vs. < 12.0), 32 of 98 patients suffered from fatigue. This corresponds to a prevalence of 32.7% (95% confidence interval, 23.5–42.9%).

The two groups, fatigue versus nonfatigue ovarian carcinoma survivors, did not differ significantly with regard to sociodemographic variables (age, marital status, education). The only statistically significant difference with respect to clinical parameters (time since diagnosis, FIGO stage, menopausal status, chemotherapy, radiotherapy, disease recurrence, blood data) was observed for the hemoglobin level (fatigue patients: mean, 13.4 mg/dL, standard deviation [SD] of 1.4; nonfatigue patients: mean, 13.8 mg/dL, SD of 1.0 mg/dL, P = 0.033).

Differences in QOL in Patients with and without Fatigue

As shown in Table 3, the ovarian carcinoma patients with fatigue report a significantly lower QOL in all domains assessed by the FACT-O and the EORTC QLQ-C30. With effect sizes ranging from 0.55 (FACT-Social well-being) to 1.35 (FACT-total score), QOL differences between fatigue and nonfatigue patients were considerable. Significant differences between groups were also observed on the EORTC QLQ-C30/OV28 Symptom subscales that assess nausea/emesis, pain, sleep disturbance, financial impact, peripheral neuropathy, and other chemotherapy side effects.

Table 3. Comparison of Quality of Life (FACT-O, EORTC QLQ-C30/OV-28) between Fatigued (MFI-GF ≥ 12, n = 32) and Nonfatigued (MFI-GF < 12, n = 66) Ovarian Carcinoma Patients
 Fatigued (n = 32)aNonfatigued (n = 66)aEffect sizeP value
  • NS: not significant.

  • FACT-O: Functional Assessment to Cancer Therapy—Ovarian; EORTC QLQ-C30: European Organization for Research and Treatment of Cancer Quality of Life Core Questionnaire; OV-28: ovarian module; MFI-GF: Multidimensional Fatigue Inventory-General Fatigue (higher scores indicate more fatigue).

  • a

    Values are mean ± the standard deviation.

  • b

    Higher values indicate a better quality of life.

  • c

    Higher values indicate higher severity.

FACT-O subscalesb
 Physical well-being21.2 ± 5.026.2 ± 3.21.21< 0.001
 Emotional well-being17.0 ± 4.620.5 ± 3.50.85< 0.001
 Social well-being20.8 ± 6.323.7 ± 4.30.550.030
 Functional well-being19.3 ± 3.523.3 ± 3.71.11< 0.001
 Total score78.1 ± 13.293.8 ± 10.01.35< 0.001
 Ovarian-modul33.3 ± 5.737.8 ± 3.31.00< 0.001
EORTC QLQ-C30 functioning scalesb
 Physical functioning73.3 ± 16.786.8 ± 18.50.76< 0.001
 Emotional functioning49.7 ± 25.573.7 ± 16.01.15< 0.001
 Social functioning65.5 ± 26.785.1 ± 22.50.80< 0.001
 Role functioning58.5 ± 28.385.1 ± 22.11.06< 0.001
 Cognitive functioning68.3 ± 29.088.5 ± 17.50.87< 0.001
 Global quality of life49.7 ± 19.576.9 ± 19.21.41< 0.001
EORTC QLQ-C30 symptom scalesc
 Fatigue55.6 ± 27.823.4 ± 19.6−1.35< 0.001
 Nausea/emesis16.1 ± 28.07.1 ± 20.7−0.370.047
 Pain32.8 ± 29.717.4 ± 21.6−0.600.011
 Dyspnea21.5 ± 23.614.4 ± 24.3−0.30NS
 Sleep disturbance50.5 ± 37.433.3 ± 31.7−0.510.032
 Appetite loss16.1 ± 28.48.6 ± 19.7−0.31NS
 Constipation21.5 ± 32.88.7 ± 18.9−0.49NS
 Diarrhea11.8 ± 28.05.1 ± 14.7−0.31NS
 Financial impact42.2 ± 37.110.8 ± 22.9−1.05< 0.001
EORTC OV-28 functioning scalesb
 Body image58.6 ± 26.763.8 ± 19.80.22NS
 Attention to disease/treatment34.8 ± 27.544.6 ± 27.80.35NS
EORTC OV-28 symptom scalesc
 Abdominal/gastrointestinal27.7 ± 29.116.2 ± 16.8−0.50NS
 Peripheral neuropathy44.1 ± 39.325.7 ± 28.6−0.540.030
 Chemotherapy side effects55.0 ± 20.835.4 ± 20.7−0.950.038
 Hormonal44.1 ± 40.530.8 ± 32.6−0.36NS

Impact of Fatigue on Anxiety/Depression, Mental Adjustment, and Social Support

Significantly higher values of depression and anxiety (HADS), as well as increased helpless coping behavior, were observed in the ovarian carcinoma patients with fatigue. Furthermore, this group reported significantly lower social support and experienced more social strain (F-SOZU) than ovarian carcinoma patients who were not suffering from fatigue (Table 4).

Table 4. Comparison of Anxiety, Depression (HADS), Mental Adjustment (MAC), and Social Support (F-SOZU) between Fatigued (MFI-GF ≥ 12) and Nonfatigued (MFI-GF < 12) Ovarian Carcinoma Patients
 Fatigued (n = 32)Nonfatigued (n = 66)Effect sizeP value
  1. NS: not significant.

  2. HADS: Hospital Anxiety and Depression Scale (higher scores indicate more anxiety and depression); MAC: Mental Adjustment to Cancer Scale (higher scores indicate a higher degree); F-SOZU: Social Support Questionnaire (higher scores indicate a higher degree); MFI-GF: Multidimensional Fatigue Inventory-General Fatigue (higher scores indicate more fatigue).

 Anxiety7.38 ± 3.824.44 ± 2.520.93< 0.001
 Depression6.13 ± 4.282.37 ± 2.291.14< 0.001
 Fighting Spirit50.07 ± 5.1351.83 ± 4.57−0.36NS
 Anxious Preoccupation21.18 ± 5.4419.68 ± 3.650.33NS
 Fatalism21.17 ± 3.6520.10 ± 3.790.29NS
 Helplessness/Hopelessness11.24 ± 3.117.85 ± 2.291.26< 0.001
 Avoidance2.63 ± 1.132.20 ± 1.120.38NS
 Emotional Support4.11 ± 0.754.55 ± 0.65−0.620.002
 Practical/Useful Support4.21 ± 0.724.58 ± 0.52−0.600.007
 Social Strain2.30 ± 0.941.65 ± 0.490.910.002
 Social Integration3.79 ± 0.694.41 ± 0.51−1.02< 0.001
 Perceived Support4.06 ± 0.594.57 ± 0.45−0.98< 0.001

Effects of Fatigue on QOL: Influence of Clinical or Psychosocial Variables

The stepwise regression results are reported in Table 5. The effect of the MFI-General Fatigue subscale on the FACT total score cannot be accounted for by clinical and/or sociodemographic variables. The clinical variables considered were tumor stage, time since diagnosis, chemotherapy, radiotherapy, disease recurrence, hormone replacement therapy, herbal supplements, pharmacologic therapy, and hemoglobin level. The sociodemographic variables investigated were age, marital status, and education (Models 2,3). When adding symptoms assessed by the EORTC QLQ-C30, pain and constipation had a significant impact on QOL (FACT total score), but MFI-General Fatigue was retained as a significant predictor (Model 4).

Table 5. Regression Model, Dependent Variable: FACT-Global Score
ModelVariables retainedRegression coefficientSEP valueAdjusted R2
  1. FACT: Functional Assessment of Cancer Therapy; SE: standard error; MFI: Multidimensional Fatigue Inventory-20; clinical variables: stages, time since diagnoses, chemotherapy, radiotherapy, recurrence, hormone replacement therapy, herbal supplements, pharmacologic therapy, hemoglobin level; sociodemographic variables: age, marital status, education; MAC: Mental Adjustment to Cancer; F-SOZU: Social Support Questionnaire; EORTC symptom: European Organization for Research and Treatment of Cancer symptom subscales; HADS: Hospital Anxiety and Depression Scale.

1 MFI-General FatigueMFI-General Fatigue−1.7480.250< 0.0010.337
2 MFI-General Fatigue, clinical variablesMFI-General Fatigue−1.7480.250< 0.0010.337
3 MFI-General Fatigue, clinical variables, sociodemographic variablesMFI-General Fatigue−1.7480.250< 0.0010.337
4 MFI-General Fatigue, clinical variables, sociodemographic variables, EORTC symptom subscalesMFI-General Fatigue−1.0590.245< 0.0010.474
EORTC pain−0.1500.0430.001 
5 MFI-General Fatigue, clinical variables, sociodemographic variables, EORTC symptom subscales, MAC subscales, F-SOZU subscalesMFI-General Fatigue−0.9730.220< 0.0010.608
MAC-Fighting Spirit0.5130.1810.006 
MAC-Anxious Preoccupation−0.5450.2130.013 
 F-SOZU-Emotional Support6.4081.417< 0.001 
6 MFI-General Fatigue, clinical variables, sociodemographic variables, EORTC symptom subscales, MAC subscales, F-SOZU subscales, HADSMFI-General Fatigue−0.6410.2450.0110.611
MAC-Fighting Spirit0.4930.1770.007 
MAC-Anxious Preoccupation−0.4340.1990.033 
 F-SOZU-Emotional Support3.9241.2630.003 

Extension of the model by including psychosocial parameters (subscales of the MAC and the F-SOZU led to an inclusion of three of the MAC subscales (Fighting Spirit, Avoidance/Preoccupation, Avoidance) and Emotional Support (F-SOZU) as significant variables (Model 5). Again, MFI-General Fatigue was retained in the model, whereas the EORTC symptoms pain and constipation were deleted.

In the last step (Model 6), when adding anxiety and depression assessed by the HADS, all variables of the previous model stayed in the model in addition to anxiety and depression. The percentage of explained variance (adjusted R2) increased from 33.7% in the first step (MFI-General Fatigue alone) to 61.1% in the final model.


Fatigue is the most frequently reported syndrome accompanying cancer and cancer treatment.7, 42 It can significantly impair a cancer patient's QOL during all phases of the illness.43 Unfortunately, there is limited information and research on the causes and patterns of fatigue in cancer patients and the multiple factors that may improve or worsen feelings of fatigue.6

The aim of this study was to assess fatigue in ovarian carcinoma survivors and to investigate the impact of fatigue and other clinical and psychosocial variables on the QOL of this group of patients.

All patients included in the study had hemoglobin levels in the normal or in the mild anemic range (> 10 g/dL) and therefore did not receive any treatment for anemia at the time of investigation. However, our results show that a remarkably high proportion (32.7%) of these patients suffered from fatigue. This is of interest also when considering that we only investigated patients currently in complete clinical disease remission.

With regard to sociodemographic and clinical variables, the only difference between fatigued and nonfatigued ovarian carcinoma survivors was found for hemoglobin level. A relationship between hemoglobin values in the normal range and subjective fatigue has been described for different groups of cancer patients in previous studies.41

Given the current result, it is suggested that medical treatment to increase hemoglobin levels be offered to certain patients suffering from the syndrome of fatigue even if their hemoglobin levels are in the normal or subnormal range.44

Women suffering from fatigue differed substantially from women without these symptoms with respect to QOL. This difference extended to all QOL domains investigated. A relation between fatigue and low QOL has also been reported for other groups of cancer patients.45

Ovarian carcinoma patients with fatigue also differ from nonfatigue patients in several other respects, with the former showing more anxiety and depression as well as an increased helpless coping behavior. In addition, this group reported significantly lower social support and higher social stress. These results are in accordance with findings in other groups of cancer patients.46

Another primary aim of the study was to investigate whether QOL is also mediated by other clinical and psychosocial variables. We analyzed the relative impact of fatigue on QOL in comparison to other factors in a regression analysis and found no effect of clinical variables such as stage, time since diagnosis, chemotherapy, radiotherapy, disease recurrence, and hemoglobin level or sociodemographic factors, such as, age, marital status, and education. In particular, no association between QOL and the use of hormone replacement therapy, herbal supplements, and psychotropic drugs was observed.

One could hypothesize that in the aftercare period, the development of fatigue is fairly independent of the primary disease and its treatment and that other factors play an important role. Even when including coping styles, social support as well as anxiety and depression in the regression model fatigue remained an independent factor in predicting QOL.

Two main conclusions can be drawn from our results. First, fatigue can be regarded as a key variable for QOL in ovarian carcinoma survivors. Given this finding, it can be argued that attention should be paid to the patients not only during their hospital stay and when they are receiving chemotherapy, but also after disease remission and completion of therapy.47–49 For long-term aftercare, this means that patients should be screened for fatigue regularly, ideally by clinical interview, supplemented with short questionnaires such as the MFI-General Fatigue subscale.

Second, our multiple regression results indicate that to improve the QOL of ovarian carcinoma survivors, the treatment of fatigue should follow a multidimensional concept that focuses not only on somatic issues but especially on the psychologic and psychosocial aspects of the disease. Psychologic treatment could encompass the whole spectrum of therapy, extending from providing adequate information on fatigue to specific psychotherapeutic interventions. The aim is to increase the patients' autonomy by strengthening coping mechanisms and helping them to develop new perspectives.50, 51 As some patients will benefit from increased social support, significant others should also be involved in the treatment processes if necessary. In some cases, pharmacologic intervention with antidepressants may also be indicated in addition to psychologic therapy.9, 17

Fatigue is a widespread and highly significant symptom in ovarian carcinoma survivors. It is a key mediator of the patients' QOL. Ovarian carcinoma survivors suffering from fatigue reported low QOL, high rates of anxiety and depression, and low social support. Our data suggest that to improve the QOL of ovarian carcinoma patients, special attention should be focused on the detection and treatment of fatigue. It is also suggested that as the relationship between fatigue and other psychologic variables (e.g., depression and anxiety) is relatively complex, this area would benefit from further more detailed investigation, possibly using long-term prospective intervention studies.