To multidimensionally assess fatigue in rheumatoid arthritis (RA) and to evaluate the impact of fatigue on health-related quality of life (HRQOL).
To multidimensionally assess fatigue in rheumatoid arthritis (RA) and to evaluate the impact of fatigue on health-related quality of life (HRQOL).
The study was conducted in 1999 among 490 RA patients with varying disease duration. Fatigue was measured with the Multidimensional Fatigue Inventory (MFI-20) and HRQOL with a validated Dutch version of the RAND 36-Item Health Survey. We evaluated the impact of fatigue on HRQOL by multiple linear regression analyses taking into account RA-related pain and depressive symptoms.
Different aspects of fatigue selectively explained different dimensions of HRQOL. The MFI-20 was entered last to the linear regression models, resulting in an additional increase of explained variance of 1% (mental health) to 14% (vitality).
The multidimensional portrayal of RA-related fatigue can be used to develop intervention strategies targeted to specific aspects of fatigue. Fatigue, supplementary to RA-related pain and depressive symptoms, appears to be a feasible and treatable target in the clinical management of RA to increase HRQOL.
Rheumatoid arthritis (RA) is a systemic, chronic inflammatory disease with a prevalence of about 1% in western countries. Women are affected ∼3 times more often than men (1). RA primarily affects joints, which leads to pain, deformities, joint destruction, and disability, but it also produces such extraarticular symptoms as fatigue. Because of its chronic, painful, and disabling character, RA tends to have a profound impact on health-related quality of life (HRQOL). The multidimensional approach of HRQOL has shown that this impact is not restricted to physical aspects only, but also influences the other domains of HRQOL, e.g., psychological and social functioning (2–4). Recently, fatigue was identified as the consequence of RA that distinguished best, i.e., better than pain, between RA patients who are doing well and those who are doing less well with respect to HRQOL (4). This finding emphasizes the need for elucidating the symptom “fatigue” in RA to make it a more feasible target for clinical management.
Fatigue is a frequent and debilitating problem for patients with RA (5, 6). Because of differences in definition and measures, varying prevalence rates have been reported. Some studies have found prevalence rates of fatigue in adults with RA of 80% and more (7–9), whereas Wolfe et al described clinically important levels of fatigue in 42% of RA patients (10). Several studies have been conducted on predictors and correlates of RA-related fatigue. Tack was one of the first to investigate self-reported fatigue in RA. She found that fatigue, pain, and depression were significantly and positively correlated (5). Crosby reported that RA patients most frequently mentioned RA disease activity, disturbed sleep, and increased physical effort as factors that contribute to fatigue (11). Belza et al suggested that fatigue has predominantly disease- and sex-related components, including comorbid conditions, disease duration, pain rating, functional status, sleep quality, and female sex (8). Wolfe et al confirmed that pain, sleep disturbance, and depression were the major predictors of fatigue in RA, osteoarthritis, and fibromyalgia (10). Huyser et al concluded that RA-related fatigue appeared to be strongly associated with psychosocial variables, apart from disease activity per se, and suggested that pain, depressive symptoms, and fatigue form an interrelated matrix of RA symptoms that are relatively unrelated to RA disease activity (12). Depression is considered an important potential sequel of RA (13, 14). However, diagnosing depression among patients with RA remains complicated because of the overlap of symptoms of depression and RA (e.g., fatigue, insomnia, lack of appetite) (15–18).
In the literature, RA patients described fatigue as an “overall sense of tiredness and heaviness that was associated with a desire to sleep,” and as “that kind of fatigue which one never recuperates from” (6). Belza states that with the high prevalence of fatigue in rheumatic diseases, clinicians need to better assess and manage fatigue relative to exercise prescription and performance (19). Wolfe et al states that fatigue assessment adds much to the understanding and management of patients and diseases (10). To cover a complete description of the fatigue experience of patients, a multidimensional approach to fatigue has been advocated (20). Although the multidimensional concept of fatigue is widely accepted and despite the fact that fatigue is an important symptom in RA, relatively little attention has been paid to the multidimensional nature of RA-related fatigue and its consequences on HRQOL.
The aim of our study was to elucidate fatigue in RA and to evaluate the impact of fatigue on HRQOL taking into account 2 other important potential sequelae of RA: RA-related pain and depressive symptoms. We assessed fatigue multidimensionally, giving us the opportunity to identify which aspects of fatigue are related to different aspects of HRQOL. These insights could be used to improve the current fatigue treatment strategies by identifying those aspects of fatigue needing targeted interventions.
We conducted the present study as part of a longitudinal survey, starting in 1997, on health and health outcomes among patients with RA registered at an outpatient center for rheumatology and rehabilitation in Amsterdam or at 1 of its affiliated outpatient clinics. Inclusion criteria were having rheumatoid arthritis according to the 1987 revised American College of Rheumatology (formerly American Rheumatism Association) criteria for RA (21), being >16 years of age, and having sufficient command of the Dutch language. Eligible patients (n = 1,200) were asked to participate and to sign a letter of informed consent. The response rate in 1997 was 74%. In the present study we used the third wave of the study (1999), consisting of 841 eligible participants.
Data were collected through a self-administered mailed questionnaire. Respondents to the questionnaire were invited for a clinical examination in which the 28-joint count (22) and erythrocyte sedimentation rate (ESR) were assessed. In addition, information on disease duration was retrieved from the patients' medical records.
A total of 683 patients returned the questionnaire (81% of the eligible participants); 31 of these 683 patients had missing data with respect to fatigue, RA-related pain, or depressive symptoms. For 490 of the remaining 652 patients, complete clinical data (i.e., ESR and 28-joint count) were available.
HRQOL was assessed with a validated Dutch version of the RAND 36-Item Health Survey (RAND-36) (23). The RAND-36 is a validated, self-administered, internationally used questionnaire measuring health status with respect to 8 dimensions: physical functioning, social functioning, role limitations caused by physical health problems, role limitations caused by emotional health problems, pain, mental health, vitality, and general health perception. Additionally, 1 single item assesses change in perceived health during the last 12 months. Scores in the range of 0–100 are computed for each dimension, with lower scores indicating poorer health status.
For a global assessment of fatigue severity, we used a 100-mm visual analog scale (VAS) ranging from 0 (no fatigue) to 100 (fatigue as bad as it could be). More detailed information was obtained with the Multidimensional Fatigue Inventory (MFI-20) (20). The MFI-20 covers 5 dimensions of fatigue: general fatigue, physical fatigue, reduced activity, mental fatigue, and reduced motivation. Scores within the 5 dimensions range from 4 to 20, with higher scores indicating higher levels of fatigue. The psychometric properties of the MFI-20 have been tested in various populations and the results, by and large, support the validity of the MFI-20 (20, 24). The MFI-20 has frequently been used in oncology research (25–30). It has also been applied in research on chronic conditions, e.g., Parkinson's disease (31), chronic obstructive pulmonary disease (32, 33), liver disease (34), and rheumatic diseases (35, 36).
Mental health was assessed with a Dutch version of the Center for Epidemiologic Studies Depression Scale (CES-D) (37). The CES-D is an internationally used 20-item scale designed to measure depressive symptoms in the general population. Scores range from 0 to 60, with higher scores indicating more depressive symptoms. We additionally computed an adjusted scale of 13 items with a score range of 0–39, taking into account possible criteria contamination due to RA-related items (15, 17, 38) or overlap in symptomatology with the MFI-20. To determine the adjusted CES-D score, we deleted the following items: “I could not get going,” “My sleep was restless,” “I felt hopeful about the future,” “I felt that everything I did was an effort,” “I was bothered by things that usually don't bother me,” “I had trouble keeping my mind on what I was doing,” and “I did not feel like eating; my appetite was poor.”
The degree of RA-related pain was measured with a VAS ranging from 0 (no RA-related pain) to 100 (RA-related pain as bad as it could be).
The sociodemographic variables included in the assessment were age, sex, and marital status. Marital status was rated as single or married/cohabiting.
Comorbidity was assessed by a list, adapted from the Health Interview Survey of the Statistics Netherlands (39), comprising 17 chronic conditions and supplemented with questions on the presence of fractures, Sjögren's syndrome, ulcers of the legs, and thrombosis. In the analyses, comorbidity was used as a continuous variable.
Disease duration was defined as the time between first diagnosis of RA and the date the questionnaire was returned by the patient. Disease activity was assessed with the modified Disease Activity Score including separate 28-joint counts for tenderness and for swelling and without the VAS for general health assessment (40).
All analyses were carried out using SPSS statistical software 10.1 for windows (SPSS, Chicago, IL). Results were considered statistically significant when P values were < 0.05. We restricted analyses to patients with complete clinical data (n = 490), but analyses among 652 patients yielded very similar conclusions (not shown). We quantified correlations between different aspects of fatigue, RA-related pain, depressive symptoms, and HRQOL by means of Spearman's rank correlation coefficients (ρ). The correlation between the adjusted CES-D0–39 score and the original CES-D0–60 score was used to determine whether the original CES-D0–60 score could be used for further analyses.
To determine the impact of fatigue on HRQOL, taking also into account RA-related pain and depressive symptoms, 8 separate linear regression models were built. With the exception of the single item “change in perceived health,” the 8 dimensions of the RAND-36 were used as dependent variables. Fatigue (physical fatigue, reduced activity, mental fatigue, and reduced motivation), RA-related pain, and depressive symptoms were entered as predicting independent variables. To reduce the risk of artifacts arising from (multi)collinearity, we left out general fatigue in the multivariate analyses. Smets et al had previously shown that a 4-factor model of the MFI-20 (i.e., by removing general fatigue) turned out to be equally acceptable as the original 5-factor model (20). All models were adjusted for sociodemographic variables (age, sex, marital status), disease duration, disease activity, and comorbidity. The impact of each variable was expressed with the regression coefficient of the linear regression model, with 95% confidence intervals.
The scales of role limitations (physical and emotional problems) showed a bimodal distribution, meaning that linear regression analysis was not statistically suitable. We dichotomized these 2 scales (according to the median) and additionally performed logistic regression analyses to check whether the statistically less suitable linear regression analyses showed similar, and thus reasonable, results. Generally, the logistic regression analyses led to similar conclusions as the linear regression analyses. For consistency in presenting results, the logistic regression analyses are not shown.
The mean age of the patients was 60.7 years (range 23.4–91.3 years; Table 1). The majority (72.7%, n = 356) were women and 65.2% (n = 319) were cohabiting. The mean disease duration was 10.7 years. Patients with low and high disease activity participated in the study. The majority of the patients (60.0%) reported at least 1 comorbid condition. Mean scores of the RAND-36 were relatively low to moderate on all dimensions, confirming the profound impact of RA on HRQOL (Table 2) (23, 41).
|Sex, no. (%)|
|Age, mean ± SD (range) years||60.7 ± 13.4 (23.4–91.3)|
|Marital status, no. (%)|
|Disease duration, mean ± SD (range) years||10.7 ± 9.2 (1.5–57.8)|
|Disease activity, mean ± SD (range)*||3.5 ± 1.3 (0.46–7.40)|
|Comorbidity, no. (%)|
|No comorbidity||196 (40.0)|
|1 comorbid condition||130 (26.5)|
|2 comorbid conditions||79 (16.1)|
|3–10 comorbid conditions||85 (17.4)|
|Mean ± SD (range)|
|Fatigue by MFI-20 (4–20)†|
|General fatigue (GF)||13.4 ± 4.9 (4–20)|
|Physical fatigue (PF)||12.4 ± 4.6 (4–20)|
|Reduced activity (Rac)||11.1 ± 4.8 (4–20)|
|Reduced motivation (RM)||9.9 ± 4.3 (4–20)|
|Mental fatigue (MF)||8.2 ± 4.2 (4–20)|
|Fatigue by VAS (0–100)||47.9 ± 25.4 (0–100)|
|RA-related pain by VAS (0–100)||38.3 ± 26.7 (0–100)|
|Depressive symptoms (0–39)||6.1 ± 5.6 (0–31.0)|
|Depressive symptoms (0–60)||11.4 ± 8.5 (0–49.0)|
|HRQOL by RAND-36 (0–100)|
|Physical functioning||48.9 ± 26.8 (0–100)|
|Social functioning||72.3 ± 24.3 (0–100)|
|Role limitations (physical)||45.6 ± 42.1 (0–100)|
|Role limitations (emotional)||71.8 ± 41.4 (0–100)|
|Mental health||72.3 ± 17.4 (16–100)|
|Vitality||53.2 ± 20.0 (0–100)|
|Pain||56.9 ± 22.6 (0–100)|
|General health perception||51.8 ± 21.1 (0–100)|
|Changes in health||44.2 ± 23.5 (0–100)|
The multidimensional assessment of fatigue through the MFI-20 resulted in a detailed portrayal of the fatigue experienced by RA patients. RA patients experienced especially general (mean ± SD 13.4 ± 4.9) and physical (12.4 ± 4.6) fatigue, followed by reduced activity (11.1 ± 4.8) and reduced motivation (9.9 ± 4.3), and less mental fatigue (8.2 ± 4.2). Mean scores on the 5 dimensions of the MFI-20 are presented in Table 2, and the distribution and variability of MFI-20 scores among the study population are illustrated in Figure 1.
All correlations between fatigue, RA-related pain, and depressive symptoms were statistically significant, but differed in strength (Table 3). The VAS for fatigue correlated highly with general fatigue (ρ = 0.786) and physical fatigue (ρ = 0.719), but only moderately with the other dimensions of the MFI-20 (ρ = 0.335–0.574). Within the MFI-20, mental fatigue showed the weakest correlation with the other dimensions, indicating that mental fatigue stands apart most from the other dimensions of fatigue (ρ = 0.321–0.407). The correlation between mental fatigue and RA-related pain was also rather weak (ρ = 0.173). General fatigue and physical fatigue, on the other hand, were strongly correlated (ρ = 0.806). The correlation between the adjusted CES-D0–39 score and the original CES-D0–60 score was very strong (ρ = 0.938), which justified use of the original CES-D0–60 score for further analyses.
|General fatigue||Physical fatigue||Reduced activity||Reduced motivation||Mental fatigue||Fatigue (VAS)||RA-related pain (VAS)||CES-D (0–39)||CES-D (0–60)|
|RA-related pain (VAS)||0.468||0.487||0.354||0.278||0.173||0.601†||1.0|
All aspects of HRQOL were significantly correlated with fatigue, RA-related pain, and depressive symptoms, but strength of the correlations varied between and within the different dimensions of the RAND-36 (Table 4). Vitality, for instance, correlated relatively strongly with 4 dimensions of the MFI-20 (except mental fatigue), whereas high correlations (ρ ≤ −0.6) with respect to physical and social functioning were only observed for physical fatigue.
|Physical functioning||Social functioning||Role limitations (physical)||Role limitations (emotional)||Mental health||Vitality||Pain||General health perception||Changes in health|
|RA-related pain (VAS)||−0.535||−0.462||−0.500||−0.337||−0.328||−0.445||−0.754†||−0.359||−0.473|
As seen in Table 5, different aspects of fatigue selectively explained different dimensions of HRQOL while taking into account pain and depressive symptoms. With respect to physical functioning, we found that physical fatigue and RA-related pain had a statistically significant negative impact. Social functioning was negatively influenced by the 2 more physical aspects of fatigue (physical fatigue and reduced activity), RA-related pain, and depressive symptoms. Mental health was negatively associated with mental fatigue and depressive symptoms, but no statistically significant impact of reduced motivation could be detected. With respect to vitality, negative impacts were found for physical fatigue, reduced activity, reduced motivation, and depressive symptoms. Physical fatigue, reduced activity, RA-related pain, and depressive symptoms were related to more pain on the pain scale of the RAND-36. General health perception worsened with physical fatigue and depressive symptoms. Physical fatigue, mental fatigue, depressive symptoms, and RA-related pain were statistically significantly related to role limitations due to physical problems. Role limitations due to emotional problems were statistically significantly associated with depressive symptoms, RA-related pain, reduced activity (not found with logistic regression), and mental fatigue (not found with logistic regression). The amount of explained variance ranged from 36% (role limitations due to emotional problems) to 67% (pain; Table 5). The MFI-20 was entered last in the linear model, resulting in an additional increase of explained variance from 1% (mental health) to 14% (vitality).
|Physical functioning||Social functioning||Role limitations (physical)||Role limitations (emotional)||Mental health||Vitality||Pain||General health perception|
|Linear regression coefficient B (95% CI)|
|Physical fatigue||−2.2 (−2.7, −1.6)†||−1.2 (−1.7, −0.7)†||−2.7 (−3.7, −1.6)†||0.8 (−0.3, 1.9)||0.1 (−0.2, 0.5)||−1.2 (−1.6, −0.8)†||−1.0 (−1.5, −0.6)†||−1.9 (−2.3, −1.4)†|
|Reduced activity||−0.5 (−1.1, 0.1)||−0.9 (−1.4, −0.3)†||−0.7 (−1.7, 0.4)||−1.2 (−2.3, −0.1)†||−0.04 (−0.4, 0.3)||−0.5 (−0.9, −0.1)†||−0.7 (−1.2, −0.3)†||0.2 (−0.3, 0.7)|
|Reduced motivation||0.1 (−0.5, 0.7)||0.1 (−0.5, 0.7)||0.5 (−0.6, 1.6)||−0.03 (−1.2, 1.2)||−0.2 (−0.6, 0.2)||−0.7 (−1.1, −0.3)†||0.4 (−0.1, 0.8)||−0.5 (−1.0, 0.03)|
|Mental fatigue||0.1 (−0.3, 0.6)||−0.2 (−0.7, 0.2)||−1.1 (−2.0, −0.3)†||−1.0 (−1.9, −0.1)†||−0.4 (−0.7, −0.1)†||0.001 (−0.3, 0.3)||0.2 (−0.2, 0.5)||−0.1 (−0.5, 0.3)|
|Depressive symptoms||−0.2 (−0.5, 0.1)||−0.9 (−1.1, −0.6)†||−0.6 (−1.1, −0.1)†||−2.0 (−2.5, −1.4)†||−1.5 (−1.6, −1.3)†||−0.9 (−1.0, −0.7)†||−0.3 (−0.5, −0.1)†||−0.6 (−0.9, −0.4)†|
|RA-related pain||−0.2 (−0.3, −0.1)†||−0.1 (−0.2, −0.1)†||−0.3 (−0.5, −0.2)†||−0.3 (−0.4, −0.1)†||−0.02 (−0.1, 0.02)||−0.04 (−0.1, 0.01)||−0.5 (−0.5, −0.4)†||0.01 (−0.1, 0.1)|
|Explained variance (R2)|
|+ Disease characteristics||0.252||0.126||0.155||0.048||0.071||0.108||0.185||0.069|
|+ RA-related pain||0.426||0.276||0.291||0.157||0.165||0.249||0.585||0.237|
|+ Depressive symptoms||0.467||0.451||0.367||0.337||0.627||0.518||0.622||0.394|
|R2 change MFI-20||0.101||0.073||0.071||0.019||0.010||0.137||0.047||0.115|
Fatigue is a frequently reported symptom in RA. The multidimensional assessment of fatigue by means of the MFI-20 provided a complete and detailed portrayal of fatigue as experienced by RA patients. Compared with previously reported data in a general population sample (28), RA patients experienced high levels of general fatigue, physical fatigue, reduced activity, and reduced motivation, which confirmed that fatigue is a major problem in RA. The mental fatigue component was less important in the profile of RA-related fatigue; its level was not high in our patient group. Nevertheless, mental fatigue predicted HRQOL, indicating that these associations are probably independent of RA. We found relatively weak correlations between mental fatigue and the other dimensions of the MFI-20, which is in concordance with previously reported inter-correlations (25), pointing out that mental fatigue behaves somewhat differently than the other scales. We evaluated the impact of fatigue on HRQOL taking into account pain and depressive symptoms. We found that different aspect of fatigue were associated with different domains of HRQOL, which is in concordance with recently reported results among patients with ankylosing spondylitis (36). The total proportion of explained variance in HRQOL varied between the different domains, but was, by and large, rather high. Fatigue, entered last to the model, added 1–14% to the explained variance of HRQOL.
The clinical relevance of the detected statistically significant associations could be disputed given the relatively small predicted changes in RAND-36 scores, and given the relatively large sample we worked with. The minimal clinically important difference (MCID) for the RAND-36 is typically in the range of 3–5 points (42), but absolute MCID thresholds are suspect and also even lower differences might arguably be clinically important (43). Besides, we should be aware that the changes we found were predicted for just 1 point change of fatigue. A change of 2 points on the MFI-20 would double the improvement of HRQOL. It should be noted that interventions targeted at physical fatigue would be more quickly effective—as indicated by changes of HRQOL of 3–5 points—than interventions targeted at mental fatigue. However, to our best knowledge no information is available on how easily changes on the MFI-20 are obtained. The MFI-20 was developed as a research instrument and, up to now, use only on a group level is recommended; no cut-off scores for clinically significant fatigue have yet been determined (44). Taking these considerations into account, we believe that the statistically significant outcomes we found represent potentially clinically meaningful outcomes.
Our findings suggest that with fatigue intervention strategies, extra improvement in RA patients' HRQOL could be gained. A positive effect of exercise on fatigue in RA has been reported previously (45). Riemsa et al suggested that fatigue may improve by enhancing self efficacy with self-management courses (46). The portrayal of RA-related fatigue, obtained by the MFI-20, and the detection of specific associations between different aspects of fatigue and different domains of HRQOL could be used to develop even more specifically targeted intervention strategies. In this regard, it is important to realize that mental fatigue stands apart from the other aspects of fatigue and should probably be addressed by other types of interventions.
In conclusion, our results support the need for a multidimensional approach of both fatigue and HRQOL. We confirmed that fatigue, pain, and depressive symptoms in RA are interrelated but according to our results, these interrelations may not be as strong as generally assumed. We supported the relevance of fatigue for clinical practice by showing that fatigue has a major unique contribution to the explained variance of HRQOL. The portrayal of RA-related fatigue can be used to develop intervention strategies targeted to specific aspects of fatigue. The finding that different aspects of fatigue selectively explain different dimensions of HRQOL can help by pointing out the dimensions of fatigue most promising for gaining benefit in HRQOL. It seems that the treatment of physical fatigue (e.g., through exercise programs) may be particularly effective. Other aspects of fatigue, and thus other dimensions of HRQOL, may be responsive to behavioral therapy or self-management courses. Fatigue, supplementary to RA-related pain and depressive symptoms, appears to be a feasible and treatable target in the clinical management and treatment of RA.
We would like to thank M. Kammeijer for her large contribution in the data collection.