Introduction

Cancer-related fatigue, a debilitating tiredness or loss of energy, is subjective, multi-factorial and adversely affects the quality of life among men receiving treatment for prostate cancer, including those undergoing radiotherapy [1–9]. Prostate cancer is the most commonly diagnosed non-cutaneous malignancy among Canadian men [10].

Prostate cancer is being diagnosed earlier because of the routine use of prostate-specific antigen (PSA). The combination of early diagnosis combined with technological advances has resulted in major changes in radiation therapy technique. Treatment by radiotherapy is based on a risk-adjusted protocol based on the presence of prognostic factors consisting of clinical stage, PSA level and Gleason score. Those with good prognostic factors are treated with a conformal technique delivering a high dose to a small volume while those with poor prognostic factors receive whole pelvic irradiation with a boost to the prostate and hormone therapy [11, 12]. Salvage radiotherapy for patients who have positive margins or rising PSA is delivered to an intermediate volume to cover the prostate bed. We routinely inform patients of the side effects they may expect from a course of radiotherapy. We tell them that they may experience fatigue during the course of their treatment. Patients usually want to know how fatigued they will be, whether they will be able to carry on with their regular activities and what they should do if they are fatigued?

It is our impression that patients receiving conformal radiotherapy experience less fatigue than those treated to the whole pelvis. Several fatigue assessment tools are available [13–19]. The majority of those tools are designed for research rather than for clinical use. A new clinical assessment tool: the fatigue pictogram (FP) was recently validated [20]. To evaluate the frequency and severity of fatigue during a curative course of radiotherapy, we therefore carried out a prospective study. We used two different instruments to assess fatigue to evaluate whether the simpler FP could provide similar information as an 11-item fatigue questionnaire (FQ).

We report our experience with radiation-induced fatigue among patients undergoing curative radiotherapy at the Toronto Sunnybrook Regional Cancer Centre.

Objectives

The objectives of the study were to determine the frequency, severity and change in fatigue during a course of fractionated radiation therapy and evaluate the clinical fatigue pictogram.

Materials and methods

Consecutive English speaking men with localised prostate cancer scheduled for curative radiation therapy were approached to participate in an ethics board approved prospective study. The study protocol was submitted to and approved by the Ethics Review Board of the Sunnybrook Health Sciences Centre, which is fully affiliated with the University of Toronto. Those who consented were asked to complete the FQ and shown the FP by the radiation oncologist or nurse at their planned weekly treatment review, which took place every Friday during the course of their irradiation treatment. The FQ was a modified FACT-AN questionnaire [21]. We did not include some of the questions such as “I have trouble walking” or I am too tired to eat” as they did not apply to our group of patients who were receiving curative radiotherapy as outpatients. We added the question “Does fatigue affect the quality of your life?” Furthermore, the initial FACT-AN questionnaire asks subjects to indicate “How true each statement was for you during the past 7 days”. In the modified questionnaire, we did not specify that the questions were about the preceding 7 days, so that patients responded to how they felt, when the questionnaire was administered during their treatment review, each Friday rather than how they felt during the preceding week. Each of the 11 items of the FQ had a score of 1 to 5 with 1=not at all, 2=a little bit, 3=somewhat, 4=quite a bit and 5=very much. The pictogram is a brief validated instrument consisting of two questions: FP1=“How tired have you felt over the last week?” and FP2=“How much does feeling tired prevent you from doing what you want?” Each one has five possible scores with a descriptive statement and a pictogram over a different colour background for each option (Fig. 1). The radiation oncologist or nurse recorded the weekly response on a study work sheet. We used the SAS statistical software version 9.1.3. (SAS Institute, Cary, NC) for data analysis.

Fig. 1
figure 1

Fatigue pictogram (FP1 and FP2)

Full size image

Statistical analysis

The raw item data were summarised for both the FQ and FP. The proportion (and 95% confidence interval) of patients reporting any level of fatigue was calculated at baseline, week 3 and week 6 for the FQ and the FP. The 11 FQ items were summed to calculate a fatigue scale (FS) value. The constant value of 11 was subtracted so that the final FS would range from 0 to 44. The coding for three items related to energy, motivation and activity level were reversed before calculating the FS value. Means and 95% confidence intervals for the FS were calculated at baseline, week 3 and the end of week 6 of radiotherapy.

Univariate summaries of patient characteristics within treatment groups were compared by chi-squared test or one-way analysis of variance (ANOVA) f-tests. The proportion reporting fatigue (FQ item 1 or FQ1, and FP1) and mean FS values were also summarised by treatment group (at baseline, week 3 and week 6) and compared using chi-squared or ANOVA f-tests. Comparisons between baseline and week 6 were made using paired t-tests (including FQ and FP items). A linear mixed model was used to compare the FS for treatment groups using all weeks from baseline through week 6.

In addition, a detailed comparison of the two fatigue tools was carried out using several forms of reliability analysis. The fatigue scale was assessed for internal reliability and the two pictogram items were also compared. The questionnaire (tired) item was compared to the pictogram (tired), the questionnaire (social) item was compared to the pictogram (activity) and finally, the fatigue scale was compared to the pictogram (tired). The following statistics were calculated pooling data across time and at each time point: Cronbach’s alpha, Pearson’s correlation, intraclass correlation (ICC, both Shrout and Fleiss ICC[1,1] and ICC[1,k]). These various statistics are different approaches to evaluating how consistent responses are across patients and in this analysis, across time.

The detailed comparison of the two fatigue tools and reliability of the modified FQ is being finalised for a separate report.

Only Cronbach’s alpha will be reported in this paper.

Results

From June 2000 to June 2003, we enrolled a total of 130 consecutive patients scheduled to undergo a curative course of radiotherapy. Fifty-one received conformal radiotherapy (CRT) (7,000–7,600 cGy in 35–38 fractions), 46 received whole pelvis (4,500 cGy in 25 fractions) and prostate boost radiotherapy (WP+PB) (2,600 cGy in 13 fractions) and 33 had post prostatectomy radiotherapy to the prostate bed (PBRT) (6,000–6,600 cGy in 30–33 fractions). All treatment was delivered using a four-field conformal technique with an 18 MV photon beam 5 days a week.

The age, PSA, stage, Gleason score and hormonal treatment for those three groups are listed in Table 1. The CRT group tends to be older men with stage T1 or T2 disease with mid-range PSA values, lower Gleason scores and largely without hormone treatment. The WP+PB group tends to have mid-range ages, stage T2 or T3 disease, higher PSA values, higher Gleason scores and most received hormone treatment. The PBRT group tends to be younger with stage pT3 disease, lower PSA, mid-range Gleason scores and largely without hormone treatment.

Table 1 Patient characteristics by treatment group
Full size table

Table 2 summarises the proportion (and 95% confidence intervals) of patients who reported anything except “not at all” for each FQ, FP item at baseline, week 3 and week 6. Table 2 also summarises mean FS scores (and 95% confidence intervals). Table 3 lists items with significant mean difference from baseline to week 6 and shows a significant increase at week 6 for the following statements in the FQ: “I feel tired/fatigue” (proportion 39% vs 56%, mean increase of 0.24, p = 0.003), “I am able to do my usual activities (reversed)” (proportion 18% vs 27%, mean increase of 0.12, p = 0.011),“I need to sleep during the day” (proportion 42% vs 51% mean increase of 0.11, p = 0.018), “I am motivated to do my usual activities (reversed)” (proportion 29% vs 38% mean increase of 0.14, p = 0.015) and “Does fatigue affect the quality of your life?” (proportion 12% vs 17% mean increase of 0.09, p = 0.049).

Table 2 Proportions greater than ‘not at all’
Full size table
Table 3 Comparison of baseline scores to week 6
Full size table

The largest increase in proportion (or mean score) in the FQ was for “I feel tired/fatigued”. This can mostly be attributed to an increase in “a little bit” tired across time as shown in Fig. 2a. All other significant mean score increases are smaller and therefore less important clinically.

Fig. 2
figure 2

a Response frequency by week of RT to FQ1. b Response frequency by week of RT to FP1

Full size image

For the pictogram FP1 “How tired have you felt during the past week?”, 60% (52,68) reported some fatigue at baseline, increasing to 68% (60,76) at week 3 (mean increase of 0.18, p = 0.002) and a return to baseline value of 60% (50,70) at week 6 (mean increase, from baseline, of 0.13, p = 0.074 with some improvement from week 3). A higher proportion of patients, 18% at baseline and 23% at 6 weeks, reported being “somewhat” tired in Fig. 2b. The pictogram FP2 “How much does feeling tired prevent you from doing what you want” showed a significant increase from baseline to week 6 (proportion 39% vs 57% mean increase of 0.21, p = 0.001).

The mean FS value at baseline was 4.1 (3.1,5.1) and increased slightly to 4.5 (3.5,5.5) at week 3 and significantly to 5.1 (3.9, 6.3) at week 6 (p = 0.025).

Table 4 compares the treatment groups at baseline, week 3 and week 6. Treatment groups were significantly different for FP1 at baseline (p = 0.02), week 3 (p = 0.04) and week 6 (p = 0.04), although WP+PB and CRT report similar levels of fatigue. Patients receiving WP+PB radiotherapy tended to report higher levels of fatigue for FQ1 but no significant treatment group differences at baseline or week 3 and marginally significant (p = 0.08) at week 6. No significant treatment group differences were observed for the FS at these time points.

Table 4 Treatment group comparisons (see also Table 5)
Full size table

Table 5 shows that if all time points of the FS score were considered in a linear mixed model, there was a significant increase in FS score for the WP+PB group (p = 0.04) across time, but the slight increase in FS score across time for the other groups was not significant.

Table 5 Linear mixed model results for FS
Full size table

The internal reliability of the FQ is good with Cronbach’s alpha equal to 0.83 (all data) and ranging from 0.78 to 0.85 for each week separately. The internal reliability of the FP is slightly lower with Cronbach’s alpha equal to 0.79 (all data) and ranging from 0.72 to 0.84 for each week separately.

Discussion

At baseline, a significant number of patients report some level of fatigue, 39% for the FQ1 and 60% for the FP1. Similar high baseline level of fatigue was reported by Hichok who found that 57% of 372 patients reported some degree of fatigue at the initiation of radiotherapy [22]. When compared to healthy individuals, there were no difference in the mean level of fatigue experienced by cancer patients and the mean level experienced by healthy controls before the start of cancer treatment [23]. Fatigue is therefore common in men referred for a curative course of radiotherapy for localised prostate cancer to our Cancer Centre. The reason for this baseline fatigue is not known and worthy of further study.

Fatigue is a common side effect of radiotherapy and ranked second to worries about the success of therapy, but above all other physical symptoms including pain [24]. As treatment progressed, we observed an increase in the number reporting fatigue to the questionnaire and pictogram. The main change was from “not al all tired/fatigued” to “a little bit tired” for the FQ. There was a significant change in the FQ item “I feel tired/fatigued” with similar changes in some other items during a course of radiation therapy for localised prostate cancer. As treatment progressed the increase in fatigue level is more marked between week 3 and week 6. This is similar to what is reported in the literature [6, 25]. Prospective studies showed that fatigue increases over the course of radiotherapy [23, 26]. It reaches a peak at week 4 [25] and usually improves over the week’s end. The mean fatigue scale scores also increased slightly across time. In this study, the patients were evaluated every Friday, at a time when their fatigue may be expected to be at a peak for that week. This prospective study confirms that fatigue is a common side effect of external beam RT for prostate cancer and increases during treatment from week 3 to week 6.

The questionnaire and pictogram are both measuring fatigue/tiredness. Those two instruments are measuring fatigue/tiredness over different time frames. The questionnaire item measures “I feel tired/fatigued” while the pictogram asks “How tired have you felt during the last week?” The response to the questionnaire, indicate how they feel at the time that the questionnaire is administered. The questionnaire does not define what is meant by “tired/fatigued”. Patients may include different subjective components in their evaluation of fatigue. The pictogram has a more specific time frame and provides a descriptive illustration of the level of fatigue. The more specific question and added information provided by the visual cartoons depicting degrees of fatigue may explain the discrepancy in the percentage reporting fatigue between the questionnaire and pictogram. In addition, the FP1 better differentiates the levels of fatigue for treatment groups compared to the FQ1 where the majority indicated no fatigue. This may favour the pictogram as an assessment tool for screening and monitoring purposes.

The FS showed that those receiving WP+PB reported higher scores compared to CRT or PBRT across time. Patients receiving WP+PB tend to be older, have higher PSA and are more often receiving concomitant hormone therapy. Hormone alone or combined with RT may contribute to fatigue. The volume of tissue irradiated is also larger for those receiving WP+PD compared to CRT or PBRF. The prevalence of fatigue depends on the region treated and the radiotherapy volume [26]. These factors and the administration of hormone may independently contribute to fatigue and explain the observed differences between the treatment groups. Although a difference was noted for the treatment groups, the numbers were too small to detect the contribution of age, stage or hormone treatment independent of treatment group.

Several studies reported that fatigue has an adverse impact on quality of life (QOL) [8, 9, 27]. In our study, the FQ and FP indicated that fatigue has an adverse impact on activity with a significant impact on QOL and limitation of activities by week 6 compared to baseline, although the mean difference in scores was not large.

The FQ asks about specific qualitative dimensions of fatigue and is useful as a research tool. Although fatigue is a multi-dimensional concept, the need to have a short clinically useful screening tool guided the design of the fatigue pictogram. The tool was designed to measure only the two dimensions that patients most often discussed in clinical conversations. The physical feeling of fatigue or tiredness and the inability to do what the individual would usually do on a daily basis were the dimensions patients discussed most often when given the opportunity in open-ended interviews [28]. The initial psychometric evaluation of the fatigue pictogram revealed high correlation between each of the FP questions and, respectively, the physical and reduced activity sub-scales of the multi-dimensional fatigue inventory, while correlations with the other dimensions of the MFI were low [29].

Limitations of this study include the observational nature of the data. Patients are grouped by treatment given and were not randomly assigned to treatment. It would have been inappropriate and unethical to randomly allocate patients to the three radiation techniques to eliminate the confounding effect of factors such as age, stage, PSA level and concomitant hormone treatment. The number of patients assigned to each group limits the power to detect factors contributing to fatigue and analysis of those confounding factors independent of treatment group.

Our results indicate that the pictogram is better at differentiating between treatment groups and is similar to the FQ in evaluating activity level. Because the FP items are more specific and have the pictorial demonstration of fatigue and activity levels, it offers an easy to administer clinical tool to screen patients with baseline fatigue and radiation-induced fatigue for interventional studies in a busy clinic for future studies.

Conclusions

Both instruments showed that a large percentage of patients reported fatigue at baseline. As treatment progressed, fatigue significantly increased, but stayed close to the baseline. The level of fatigue appears higher using the pictogram, which was better than the questionnaire or fatigue scale in differentiating between treatment groups at specific time points. The pictogram could be used in a busy clinic to screen patients for interventional fatigue studies.