Could Objective Tests Be Used to Measure Fatigue in Patients with Advanced Cancer?

Abstract

Context

Assessment of cancer-related fatigue is currently based on patient-reported outcomes. We asked whether objective assessments such as muscle strength and nutritional markers can be used as surrogate measures of cancer-related fatigue.

Objective

We examined the association among three fatigue scales, muscle strength, and nutritional markers in patients with advanced cancer.

Methods

In this prospective study, we enrolled hospitalized cancer patients who had been seen in palliative care consultation at MD Anderson Cancer Center. We assessed fatigue with use of three fatigue scales—the Brief Fatigue Inventory (BFI), the Edmonton Symptom Assessment System (ESAS), and the European Organisation for Research and Treatment of Cancer Quality of Life Core Questionnaire 30 (EORTC QLQ-C30)—and determined their association with objective assessments, including handgrip strength, maximal inspiratory pressure (MIP), lean body mass, phase angle, and albumin. Spearman’s correlation test was used to assess associations.

Results

Among 222 patients, the mean age was 55 years; 59% were female. The median overall survival was 106 days. The total BFI score had weak association with handgrip strength (ρ = −0.18, P = 0.007) and no association with the remaining objective measures. ESAS fatigue and EORTC fatigue showed similar findings. Total BFI had moderate to strong association with ESAS (ρ = 0.54, P < 0.0001) and EORTC (ρ = 0.60, P < 0.0001) fatigue.

Conclusion

Our study showed that subjective assessment of fatigue based on patient-reported outcomes correlates only weakly with muscle strength and nutritional markers; thus, patient-reported outcomes remain the gold standard for fatigue assessment.

Introduction

Fatigue, which is present in almost all aspects of a cancer patient’s life, is reported in more than 90% of patients actively receiving chemotherapy and/or radiotherapy and in approximately one-third of patients after completion of therapy.^{1, 2} Cancer-related fatigue (CRF) is one of the most distressing symptoms for many patients and is considered a dose-limiting toxicity for some treatments.^{1, 3–5}

Proper management of fatigue requires routine documentation of this symptom by using patient-reported outcomes, which represent the gold standard for fatigue assessment. Validated instruments for CRF can be divided into unidimensional or multidimensional scales.^6–12 Some commonly used unidimensional scales include the Edmonton Symptom Assessment System (ESAS) and the European Organisation for Research and Treatment of Cancer Quality of Life Core Questionnaire 30 (EORTC QLQ-C30).^{13, 14} Multidimensional scales offer a more comprehensive assessment by capturing the physical, emotional, social, and functional aspects of fatigue.^{15, 16} However, they are more cumbersome to complete, and thus their use is most often limited to the research setting. One of the most commonly used (unidimensional) questionnaires for CRF is the Brief Fatigue Inventory (BFI). This scale consists of 9 items assessing the intensity of fatigue as well as the impact of fatigue on daily functions. This scale, which has been validated extensively in multiple patient populations,^17–21 has been shown to be a fast yet reliable measurement, with 98% of the study patients being able to complete it.⁸

CRF is multidimensional in nature (Figure 1). Because functional and nutritional compromise are often associated with CRF in the advanced cancer setting, we hypothesized that decreased muscle function and nutritional status may be particularly useful as surrogate markers of CRF.^{22, 23} To date, only a handful of studies have examined this issue, and none in the palliative care setting.^23–30 A better understanding of the association between the subjective assessment of fatigue and objective functional measures would potentially allow us to assess and monitor fatigue more objectively, devise treatments targeted at improving muscle function and nutrition, and identify predictors of treatment response.

Figure 1.

Cancer-related fatigue is a multifactorial symptom with various contributors, such as progressive cancer, cancer treatments, pro-inflammatory cytokines, direct neurologic dysfunction by cytotoxic treatments, sleep disruption, hypothalamic-pituitary-adrenal axis disturbances, deconditioning, nutritional compromise and mood and emotional disorders. Many of these contributors may be associated with each other. We hypothesized that subjective assessment of fatigue correlates with objective measures of patients’ muscle and nutritional status.

In this prospective study, we examined the association between fatigue scales and objective assessments that can broadly evaluate muscle function and nutritional status, including handgrip strength, maximal inspiratory pressure (MIP), phase angle (PA), lean body mass, and albumin in patients with advanced cancer. The BFI, EORTC, and ESAS, with 9, 3, and 1 questions, respectively, are commonly used scales to assess fatigue. We were interested in determining their level of correlation and whether these scales are interchangeable despite differences in participant burden. We hypothesized that subjective assessment of fatigue correlates with objective measures of patients’ muscle and nutritional status.

Methods

Study Population

We herein describe our preplanned secondary analysis of a prospective study of the prognostic utility of functional measures in patients with advanced cancer.³¹ Patients were enrolled in the study if they met the following inclusion criteria: age ≥18 years; diagnosis of advanced cancer; hospitalized, receiving parenteral hydration; and seen by the inpatient palliative care mobile team for consultation at The University of Texas MD Anderson Cancer Center. Patients with delirium; those fitted with a defibrillator or cardiac pacemaker; those unable to use a handheld dynamometer due to a neuromuscular disorder, joint disease, or arm pain; or those with a local infection/wound preventing the use of bioelectric impedance analysis pads were excluded. The Institutional Review Board at The University of Texas MD Anderson Cancer Center approved the current study.

Data Collection

We prospectively collected baseline patient demographics at the time of admission. We assessed fatigue by using the BFI, EORTC QLQ-C30 fatigue, and ESAS fatigue scales. We also assessed muscle function with use of handgrip strength and MIP, as well as nutritional status with use of PA, lean body mass, and dosed serum albumin, as documented previously.³¹ All study assessments required less than 30 minutes to complete.

BFI

The BFI is a fatigue-specific tool used to assess the severity and impact of fatigue on daily functioning in patients with cancer or treatment-related fatigue in the past 24 hours. It consists of three items associated with the level of fatigue (at its usual level, at its worst level, and at the moment of application) and six items focused on how fatigue interferes with the daily aspects of the patient’s life during the previous 24 hours, including general activity, mood, walking ability, normal work, relationships with other people, and enjoyment of life. It uses numeric scales ranging from 0 to 10, with 0 representing “no fatigue” and 10 representing “the worst fatigue you could ever imagine”.⁸ The severity score is calculated by averaging the values given in response to the three questions associated with level of fatigue. The interference score is the average of the values given in response to the six questions associated with how fatigue interferences with daily life. A global score is calculated using the average of all nine questions. Patients who could answer at least one-half (five) of the questions were included. This scale has high internal consistency (96%) to support reliability and has been validated in various languages.^17–21 Severity of fatigue was classified as mild, moderate, or severe based on scores of 1–3, 4–6, and 7–10, respectively.⁸

ESAS

The ESAS measures the average intensity of 10 common symptoms in the past 24 hours, including fatigue (as well as pain, nausea, depression, anxiety, drowsiness, shortness of breath, appetite, sleep, and well-being), with a numeric scale ranging from 0 (indicating none) to 10 (indicating worst).¹³ The ESAS was validated in patients with cancer and in various languages.³²

EORTC QLQ-C30

EORTC QLQ-C30 is a quality-of-life questionnaire for cancer that consists of 30 items. This scale has been validated in patients with advanced solid tumors, leukemia, and lymphoma.³³ Its fatigue subscore has also been validated.³⁴ This study focuses on EORTC fatigue, which is assessed by three questions (Did you need rest? Have you felt weak? Were you tired?) to determine how the patient felt during the past week.¹⁴ The average was transformed into a 0–100 scale, with a higher score corresponding to a higher level of fatigue.¹⁴

Handgrip Strength

Handgrip strength uses a Jamar dynamometer.³⁵ Patients are asked to sit comfortably with their shoulder adducted and forearm neutrally rotated, elbow flexed to 90°, and forearm and wrist in a neutral position. Patients then perform maximal isometric contraction, repeated every 30s. The highest value of three tests is used for the analysis, adjusted for age and sex. Handgrip strength varies with age and sex, normally ranging between 30 and 50 kg.³⁶

Maximal Inspiratory Pressure

MIP is used widely to measure inspiratory muscle strength.³⁷ Measurements were collected with use of an NS 120-TRR 120cm H2O NIF Meter (Instrumentation Industries Inc, Bethel Park, PA) according to American Thoracic Society guidelines.³⁸ We asked patients to breathe tidally for a few breaths and then to exhale maximally before inhaling maximally, maintaining the pressure level for at least 2 seconds. Five consecutive efforts were recorded, with a 1-minute pause between each effort. We used the average of the top 3 measurements that varied by <20% for analysis.³⁸ MIP varies with age and sex, normally ranging between 50 and 100 cm water (H₂O).³⁹

Phase Angle and Lean Body Mass

PA is a novel marker of nutritional and functional status.⁴⁰ PA and lean body mass were assessed by using the Quantum IV bioelectrical impedance analysis system (RJL Systems, Clinton Township, MI). The electrodes were placed over the middle of the dorsal surface of the right hand between the distal prominence of the radius and the ulnar styloid, and over the right foot between the medial and the lateral malleoli at the ankle. A small single frequency (50 hertz) was applied, alternating low-voltage electrical current, to detect the voltage drop. PA has previously been shown to have high reliability and predictive validity.^{41, 42} In healthy individuals, PA generally ranges between 5 and 7.⁴³ In addition to PA, bioelectrical impedance analysis provided data regarding body composition (i.e., lean body mass). This bedside test takes fewer than 5 minutes to complete.

Albumin

Albumin is a known prognostic marker associated with nutritional and inflammatory status.⁴⁴ Serum albumin was routinely collected in our hospitalized patients.

Statistical Analysis

Baseline characteristics were summarized with use of descriptive statistics such as mean, standard deviation (SD), median, interquartile range, frequency, and percentage. Spearman’s rank correlation test was used to assess correlations between fatigue scales and objective measurements.

Statistical analyses were conducted with use of Statistical Analysis System software (SAS version 9.3, SAS Institute, Cary, NC). A P value of <0.05 was considered statistically significant. Correlation coefficient values of 0.0 to 0.20, 0.21 to 0.40, 0.41 to 0.60, 0.61 to 0.80, and greater than 0.80 were classified as weak, fair, moderate, substantial, and almost perfect, respectively.⁴⁵

Results

Patient Characteristics

This study included 222 hospitalized patients. The mean age was 55 years (range, 22–79 years). The median overall survival was 106 days (95% confidence interval [CI], 71–128 days). The patient characteristics are detailed in Table 1.

Table 1.

Patient Characteristics and Results

Characteristic	No.
Age, average (range) (years)	54.5 (22 – 79)
Female sex	130 (59.1%)
Ethnicity
White	146 (66.4%)
Black	43 (19.5%)
Hispanic	29 (13.2%)
Others (e.g., Asian)	2 (0.9%)
Education
High school or lower	116 (52.8%)
College	74 (33.6%)
Advanced	30 (13.6%)
Cancer
Breast	28 (12.7%)
Gastrointestinal	73 (33.2%)
Genitourinary	19 (8.6%)
Gynecological	24 (10.9%)
Head and neck	10 (4.5%)
Hematological	12 (5.5%)
Others	18 (8.2%)
Respiratory	36 (16.4%)
Karnofsky performance status (KPS_50), average (SD)	0.5 (0.5)
Karnofsky performance status (PaP_KPS), average (SD)	54.8 (12.4)
ESAS, median (IQR)
Pain	5 (4, 8)
Fatigue	6 (4, 8)
Nausea	1.5 (0, 5)
Depression	1 (0, 4)
Anxiety	3 (0, 5)
Drowsiness	5 (2, 7)
Appetite	5 (3, 8)
Well-being	5 (3, 6)
Dyspnea	2 (0, 5)
Sleep	5 (3, 7)
EORTC QLQ-C30, average (SD)
Global health status	36.6 (24.1)
Physical	57.2 (27.4)
Role	33.9 (31.8)
Emotional	61.3 (29.4)
Cognitive	62.6 (30.2)
Social	40.6 (33.4)
Fatigue	65.2 (24.9)
Nausea	34 (32.5)
Pain	73.3 (27)
Dyspnea	33.3 (37)
Insomnia	55.2 (35)
Appetite	57.7 (36.8)
Constipation	45.3 (37.8)
Diarrhea	23.4 (31.7)
Financial	43.7 (39.3)
BFI, average (SD), median (IQR)
Severity	5.6 (2.3), 5.7 (4, 7.3)
Interference	4.2 (2.5), 4.4 (2.4, 6.1)
Global	4.7 (2.2), 4.9 (3.1, 6.5)
Handgrip strength, average (SD), median (IQR)	24 (10.7), 21.7 (16.5, 30)
MIP, average (SD), median (IQR)	45.6 (22.6), 40 (28.7, 60)
Lean body mass, median (IQR), kg	51.7 (44.2–61.5)
Serum albumin, average (SD), g/dL	3.3 (0.7)
Unadjusted phase angle, median (IQR),°	4.4 (3.5, 5.3)
Standardized phase angle, median (IQR)	−2.1 (−3.2, −0.89)

Abbreviations: MIP, maximal inspiratory pressure; BFI, Brief Fatigue Inventory; ESAS, Edmonton Symptom Assessment System; EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer Quality of Life Core Questionnaire 30; SD, standard deviation; IQR, interquartile range.

Fatigue Scores

The mean global BFI score was 4.7 (standard deviation [SD], 2.2). The EORTC QLQ-C30 average fatigue score was 65.2 (SD, 24.9), and the ESAS median fatigue score was 6 (interquartile range [IQR], 4, 8) (Table 1).

Objective Measures

A total of 212 patients (95%) completed handgrip strength and MIP measurements, and 212 completed PA measurement. The main reasons for noncompletion were device malfunction and patient refusal.

The median handgrip strength was 22 kg (IQR, 18, 32 kg), and the MIP was 40 cm H₂O (IQR, 28.7, 60 cm H₂O). The median unadjusted PA was 4.4° (IQR, 3.5°, 5.3°). The median lean body mass was 51.7 kg (IQR, 44.2 kg, 61.5 kg), and the mean albumin was 3.3 g/dL (SD, 0.7 g/dL) (Table 1).

Association Between Objective Measures with Fatigue Scores

As shown in Table 2, total BFI had a weak association with handgrip strength (ρ = −0.182, P = 0.007) and PA (ρ = −0.155, P = 0.026). A trend toward negative association was seen with EORTC QLQ-C30 and ESAS. MIP, albumin, and lean body mass were not significantly correlated with these patient-reported measurements of fatigue.

Table 2.

Correlation Between Fatigue Scores and Objective Measurements

	Global BFI	ESAS fatigue	EORTC fatigue	HGS	MIP	LBM	Albumin	PA
	N Correlation coefficient P value	N Correlation coefficient P value	N Correlation coefficient P value	N Correlation coefficient P value	N Correlation coefficient P value	N Correlation coefficient P value	N Correlation coefficient P value	N Correlation coefficient P value
BFI total	—	219 0.543 <.0001	219 0.602 <.0001	220 −0.182 0.007	210 −0.121 0.079	204 −0.041 0.557	210 −0.095 0.171	205 −0.155 0.026
BFI severity	219 0.802 <.0001	218 0.628 <.0001	218 0.555 <.0001	219 −0.226 0.001	209 −0.162 0.019	203 −0.093 0.185	209 0.108 0.119	204 −0.189 0.007
BFI interference	220 0.953 <.0001	219 0.450 <.0001	219 0.564 <.0001	220 −0.121 0.072	210 −0.086 0.216	204 −0.006 0.936	210 −0.07 0.311	205 −0.125 0.075
ESAS fatigue	219 0.543 <.0001	—	218 0.422 <.0001	219 −0.120 0.076	209 −0.057 0.414	203 −0.002 0.974	209 −0.007 0.919	204 −0.009 0.901
EORTC fatigue	219 0.602 <.0001	218 0.422 <.0001	—	219 −0.134 0.048	209 −0.121 0.082	203 0.029 0.685	209 −0.061 0.382	204 −0.131 0.061

Abbreviations: BFI, Brief Fatigue Inventory; ESAS, Edmonton Symptom Assessment System; EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer Quality of Life Core Questionnaire 30; HGS, handgrip strength; MIP, maximal inspiratory pressure; LBM, lean body mass; PA, phase angle.

Association Among Fatigue Scores

The total BFI score had a moderate correlation with the ESAS fatigue score (ρ = 0.543, P < 0.001) and the EORTC QLQ-C30 fatigue score (ρ = 0.602, P < 0.001). The ESAS fatigue score’s correlation with EORTC QLQ-C30 fatigue score was 0.422 (Table 2).

Discussion

Our objective was to evaluate the correlation between functional measurements and the level of fatigue assessed by validated scores. We found a weak correlation between BFI (particularly the severity score) and handgrip strength and PA, and a moderate association among the various patient-reported outcomes (BFI, ESAS fatigue, and EORTC QLQ-C30 fatigue). Our findings showed that patient-reported outcomes are still the preferred method to assess fatigue, and that muscle function and nutrition are distinct entities from CRF.

Handgrip strength and MIP are objective measures of skeletal muscle function. Decreased muscle strength is one of the many phenomena associated with chronic diseases, particularly in cancer patients. Sarcopenia is diagnosed by both reduced muscle strength and mass. Patients with both cancer and sarcopenia had significantly higher levels of fatigue than did those without sarcopenia.²³

Only a handful of studies, however, have examined the association between CRF and muscle function alone. Kilgour et al. showed a total BFI score of 4.85 (0–90) in the handgrip strength’s 25th percentile group, compared with 18.8 in the ≤10th percentile group in patients with advanced cancer (P < 0.05).²⁵ The same author discovered a weak association (−0.34, P = 0.018) of BFI with handgrip strength in men, but not in women.²⁶ Other studies, however, failed to demonstrate a significant correlation between handgrip strength and fatigue.^{29, 30}. Stone et al. found that although inpatients undergoing palliative care were malnourished and had impaired voluntary muscle function, the negative correlation of 0.16 with fatigue assessed by the Fatigue Severity Scale (FSS) was not statistically significant (P = 0.08).³⁰ Brown et al. showed that fatigue measured by Functional Assessment of Chronic Illness Therapy (FACIT) also did not correlate with handgrip strength (P = 0.331) in patients with advanced lung cancer.²⁹ Thus, our findings are consistent with the literature that fatigue has minimal, if any, correlation with handgrip strength.

Only one study looked at associations between PA and symptoms in cancer patients. Norman et al. found that patients below the fifth reference PA percentile had a higher fatigue score than did patients above the fifth percentile, using EORTC QLQ-C30 fatigue (fatigue score of 59.4 vs 46.4, P < 0.0001).²⁸ The same study also demonstrated a significant correlation between PA and other objective measures, including handgrip strength and others, which was observed and published in our original report as well.³¹

Poor nutritional status is a known factor associated with fatigue.⁴⁶ However, in our study, albumin and lean body mass, which are usually decreased in malnourished patients, were not independently associated with fatigue. We could find no studies that directly evaluated the association between MIP and fatigue.

The severity score of BFI had the strongest, albeit still weak, correlation with the functional measures. This could be justified by one of its questions assessing fatigue “now” (at the moment of the questionnaire), in a similar moment in which the objective tests were also being conducted. Still unanswered, however, is why the correlation between fatigue and objective measures is not stronger. Although skeletal muscle wasting peripherally may contribute to fatigue by decreasing functionality and physical fitness, CRF is a centrally perceived multifactorial syndrome (Figure 1).^47–53 Considering all aspects, the combination of contributing factors that directly or indirectly affect the central nervous system seems to play a more major role in the development of fatigue than in muscle wasting. Future studies should also examine whether these objective variables are associated with fatigability (i.e., how fast a patient becomes fatigued).^54–56

In this study, we also found that the correlation among the various patient-reported outcomes was moderate, which was consistent with multiple cross-sectional studies.^{19, 21, 29, 30, 57} Potential reasons that the correlation was only moderate and not strong may include differences in time frames analyzed by each scale and different question phrasing and content.

An accurate and quick method of screening for fatigue is extremely important, given the high prevalence of fatigue among cancer patients. Ultimately, patient-reported outcomes are still the gold standard. In particular, the single-item ESAS fatigue score can provide rapid insight into a patient’s symptoms, which is appropriate in a busy clinical setting.

This single-center study had limitations. It enrolled only hospitalized patients, which can significantly influence the obtained results based on the reason for hospitalization. In an outpatient setting, after resolution of the acute problem, the same patients could present with very distinct results. In addition, some correlations were weak but statistically significant; this may be related to either the large sample size or to multiple testing. Thus, these findings should be considered preliminary, and further studies are needed for confirmation.

In summary, our results indicated that the objective measures that we studied had limited association with CRF. Patient-reported outcomes remain the gold standard.