SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgments
  9. REFERENCES
  10. Supporting Information

Objective

Although patients with rheumatoid arthritis (RA) experience fatigue, little is known about its causes and consequences, and a fully developed theoretical model explaining the experience of fatigue in RA is lacking. Our goal was to systematically review studies in RA that examined factors related to fatigue to gain more insight into its possible causes and consequences.

Methods

Medline, Web of Science, Scopus, and PsycINFO were searched for relevant studies. All studies with RA samples about the relationship between fatigue and other variables that defined dependent and independent variables and used multivariate statistical methods were preliminarily included. After reviewing 129 full texts, we identified 25 studies on possible causes of fatigue and 17 studies on possible consequences of fatigue.

Results

The studies found possible causes of fatigue in illness-related aspects, physical functioning, cognitive/emotional functioning, and social aspects. Additionally, being a woman was related to higher levels of fatigue. Inflammatory activity showed an unclear relationship with fatigue in RA. Possible consequences of fatigue were also found among illness-related aspects, physical functioning, cognitive/emotional functioning, and social aspects. The strongest evidence for a relationship between fatigue and other variables was found regarding pain, physical functioning, and depression.

Conclusion

This review summarizes the current knowledge in the field in order to inform future research on causes and consequences of fatigue in RA. However, the results are based on cross-sectional and longitudinal studies with different designs and different fatigue scales. For a better identification of causal associations between fatigue in RA and related factors, longitudinal prospective designs with adequate fatigue measurements are suggested.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgments
  9. REFERENCES
  10. Supporting Information

Fatigue is commonly reported by patients with rheumatoid arthritis (RA) ([1, 2]). Qualitative research has shown that patients experience fatigue as a multidimensional, persistent symptom with far-reaching consequences ([3-6]). A generally accepted definition of fatigue in RA does not exist ([7]), and little is known about its etiology ([8]). The number of studies including fatigue as an outcome measure has rapidly increased over the previous years. However, a theoretical framework explaining the experience of fatigue in RA is lacking ([9]), and the phenomenon of fatigue is not yet understood in detail.

Hewlett et al ([10]) proposed a hypothetical model for fatigue in RA, suggesting interactions between different factors. The first factor, “RA,” includes disease processes. The second factor, “cognitive, behavioural,” contains thoughts, feelings, and behaviors. The third factor, “personal,” is about personal issues in the life of a patient. The model vividly reflects the dynamic and bidirectional relationships between fatigue and physical, psychological, and environmental factors. However, since the authors did not provide evidence for the hypothesized relationships, their model remains on a heuristic level.

An overview about fatigue in rheumatic diseases ([11]) also underlines the multifactorial nature of fatigue, but did not review scientific evidence for bidirectional relationships between fatigue and other variables in RA. Therefore, our aim was to systematically review studies on RA about factors associated with fatigue that indicated potential causes of fatigue. We also reviewed studies that indicated that fatigue has a potential impact on other variables.

Providing an overview about causes and consequences of fatigue is challenging because several, possibly parallel, interdependent and circular processes may account for fatigue in RA ([10]). However, it is not possible to study such processes as a whole. Although the approach is suboptimal, researchers have to define separate concepts and variables and have to choose relevant time periods. Even if the operationalization of concepts and sequences is undertaken thoroughly, it remains a simplified reflection of reality and, consequently, retains an artificial character ([12]). As a result, we face a “causation dilemma.”

In fact, the results of most studies on fatigue in RA did not answer questions related to bidirectional causation and did not use adequate designs to investigate the mutual influence of fatigue with other variables. A majority of studies merely reported bivariate correlations. Furthermore, many studies were cross-sectional. Most studies either statistically predicted fatigue with other factors in a regression model or predicted other outcomes with fatigue. This methodology provides insight into associations and potential causal relationships, but no evidence for a causal relationship ([12]). Studies with a longitudinal design do give some insight into causality provided that adequate controls are conducted, such as controls for the baseline levels of the predicted outcome. Actually, scientific methodology would require researchers to manipulate the processes that shall be unraveled, but due to ethical and practical issues, such manipulation is problematic with fatigue in patients with RA.

Consequently, for this review, we included studies that examined the relationships at least multivariately and that assumed a directional association between fatigue and the other constructs under consideration. Thereby, we reported the directionality according to the authors' intentions of the original studies, since their analyses are based on these assumptions. We also registered whether a study was cross-sectional or longitudinal, and whether analyses controlled for baseline levels of the statistically predicted outcome. The following research questions were formulated and addressed based on the reviewed empirical studies: 1) What is reported about possible causes of fatigue in RA? and 2) What is reported about possible consequences of fatigue in RA?

We used this division for the presentation of the results. However, in the interpretation of the results, we have accounted for the fact that cross-sectional studies are not able to ascertain causal relationships.

Results can inform further research by guiding the formulation of new research questions, encouraging researchers to examine certain relationships in more detail, and discovering how disease-specific processes contribute to fatigue in RA. Basically, a better understanding of the phenomenon of fatigue in RA is essential to be able to compare it with fatigue in other conditions and for the development of interventions and treatments for this symptom.

Box 1. Significance & Innovations

  • This systematic review summarizes the available scientific evidence regarding possible causes and consequences of fatigue in rheumatoid arthritis (RA).
  • In addition to an overview of potential predictors of fatigue in RA, we also discuss nonsignificant study results and potential consequences of fatigue.
  • The reviewed studies showed conflicting evidence for a relationship between fatigue and most of the examined variables, especially regarding inflammatory markers. The most convincing indications for a relationship with fatigue were found for pain, disability, and depression.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgments
  9. REFERENCES
  10. Supporting Information

Search strategy and study selection

A systematic literature search was executed in Medline, Web of Science, Scopus, and PsycINFO. Main search terms were “fatigue / tiredness” in combination with “rheum* / arthritis / musculoskeletal / joint disease” and “model* / theor* / framework / predict* / etiology / pathophysiology / factor*”. Whenever possible, proximity searches were used to ensure that the search terms, for example, “fatigue” and “model,” were mentioned in one sentence of the abstract. The detailed search strategies are included in Supplementary Appendix A (available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.21949/abstract).

The search was conducted in May 2011. All hits were saved in EndNote and duplicates were removed. After that, 1,923 articles were present in the database. First, one researcher (SN) read all titles and abstracts and retrieved the potentially relevant articles. Since our aim was to summarize information about possible causes (statistical predictors) and consequences (fatigue as a statistical predictor of another variable) of fatigue, we only included studies that defined dependent and independent variables and used multivariate statistical methods. We excluded studies that merely provided correlations between variables, case studies, qualitative studies, studies on effects of (medical) interventions, conference papers, letters, and non-English articles. If an abstract was not available or did not give enough information, the study was preliminarily included so that the full text could be screened. This procedure resulted in a preliminary set of 129 full-text articles.

These articles were read and summarized by 3 researchers (SN, CB, ET) and discussed as a team. Thereby, consensus about the essential information of each article was obtained and agreement about the categorization of an article into 1 of the 2 research questions was reached (question 1: statistical predictors of fatigue; question 2: fatigue as a statistical predictor). On closer examination of the 129 abstracts/full-text articles, 92 studies did not fulfill our inclusion criteria and were excluded. The main reason for exclusion was that the study's sample did not include patients with RA or no data were provided for this group separately. Moreover, some studies did not conduct relevant analyses with fatigue as a variable, and for some abstracts, no original research article existed. Complementary to the electronic search, the reference lists of all 129 full texts were searched for additional potentially relevant studies. The selection procedure is shown in Figure 1.

image

Figure 1. Flow chart of the selection of relevant studies from Medline, Web of Science, Scopus, and PsycINFO, and the reference lists of the selected studies. RA = rheumatoid arthritis.

Download figure to PowerPoint

Data analysis

The selected studies were summarized in 2 tables, according to our research questions. We identified 25 studies about statistical predictors of fatigue (Table 1) and 17 studies where fatigue was the statistical predictor of another variable (Table 2).

Table 1. Statistical predictors of fatigue (n = 25)*
Author, year (ref.)Sample (country, population, sex)Study designMain statistical methodMeasure of fatigueResults
Statistically significant predictors of fatigueStatistically nonsignificant predictors
  1. RA = rheumatoid arthritis; MAF = Multidimensional Assessment of Fatigue; VAS = visual analog scale; ESR = erythrocyte sedimentation rate; NRS = numerical rating scale; BMI = body mass index; LPS = lipopolysaccharide; IL-6 = interleukin-6; CRP = C-reactive protein; FM = fibromyalgia; PIEs = positive interpersonal events; NIEs = negative interpersonal events; CIS = Checklist Individual Strength; RF = rheumatoid factor; DAS28 = Disease Activity Score in 28 joints; Hgb = hemoglobin; AD = affective disorder; SF-36 = Short Form 36; HAQ = Health Assessment Questionnaire; MTX = methotrexate; TJC = tender joint count; SJC = swollen joint count; QUEST-RA = Quantitative Patient Questionnaires in Standard Monitoring of Patients with Rheumatoid Arthritis; OA = osteoarthritis; TIRA = Early Intervention in RA; PCA = principal component analysis; DMARD = disease-modifying antirheumatic drug; anti-CCP = antibodies to cyclic citrullinated peptides; OLS = ordinary least squares; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; ANOVA = analysis of variance; PANAS-X = 10-item Positive and Negative Affect Schedule; DI = disability index; GH = general health; SW28 = 28 joint count for swelling; TE28 = 28 joint count for tenderness; EMA = ecological momentary assessment.

  2. a

    Wording of the VAS not reported.

Cross-sectional studies      
Belza et al, 1993 ([14])
  • US
  • RA: n = 133
  • 75% women
Cross-sectionalHierarchical multiple regression analysisMAFBeing female, pain, quality of sleep, comorbid conditions, activity level, functional status, disease durationEducation, age, depression, social support, helplessness
Bergman et al, 2009 ([27])
  • US
  • RA: n = 1,577
  • Sex ratio not reported
  • Other samples also included but not relevant for this review
Cross-sectionalHierarchical regression analysis0–10 fatigue VAS: How much of a problem has fatigue or tiredness been for you in the past week? (anchors: fatigue is no problem, fatigue is a major problem)Patient globalInflammatory activity, joint counts, and ESR
Davis et al, 2008 (15)
  • US
  • RA: n = 58
  • 60% women
Cross-sectionalMultilevel modeling Hierarchical linear regression analysis101-point fatigue NRS (0 = no fatigue, 100 = fatigue as bad as it can be)
  • When pain was included in the initial step of the regression models, no other covariate (age, BMI, etc.) was significantly related to fatigue
  • LPS-stimulated IL-6 level predicted fatigue over and above the contribution of pain
Plasma levels of both CRP and IL-6, age, sex, ethnicity, BMI, sleep disturbance, alcohol use, steroid medication use, pain
Dhir et al, 2009 (26)
  • North India
  • RA: n = 200 (30 with FM, 170 without FM)
  • Controls: n = 200
  • Female:male ratio: 5.7:1 RA, 4.7:1 controls
Cross-sectionalLinear regression analysis10-cm fatigue VAS, with cm markings for the preceding weekaFor RA, pain and fatigue were independently predicted by number of tender points and disease activityNot reported
Finan et al, 2010 ([34])
  • US
  • RA: n = 231
  • 70% women
  • Cross-sectional, longitudinal
  • 30 daily diaries
Multilevel modeling (multilevel regression analysis)Fatigue VAS: What number between 0 and 100 best describes your average level of fatigue today? (0 = no fatigue, 100 = fatigue as bad as it can be)
  • Daily increase in PIE results in decreased fatigue
  • Daily increase in NIE results in increased fatigue
  • Interaction of PIEs/NIEs significant effect on fatigue
  • Fatigue higher on days with few PIEs or many NIEs, reduction of fatigue only with combination of many PIEs with few NIEs (absence of NIEs not enough to decrease fatigue), both PIEs and NIEs related to more fatigue the next day
Neuroticism, extraversion were not moderator
Van Hoogmoed et al, 2010 ([16])
  • The Netherlands
  • RA: n = 228
  • 63% women
Cross-sectionalBackward stepwise multiple regression analysisCISAge, RF, pain severity, bodily pain, physical functioning, role functioning, depressive mood, self-efficacy on fatigue, coping, catastrophizing, sleep disturbancesSex, joint count, optimism, self-esteem, social functioning, social support, physical activity, DAS28, ESR, CRP, Hgb, etc.
Huyser et al, 1998 ([17])
  • US
  • RA: n = 73
  • 45% women
Cross-sectionalPrincipal component factor analysis, stepwise multiple regression analysisPiper Fatigue Self-Report Scale
  • Pain, depressive symptoms, and female sex
  • All other variables subsequently, one at a time, added to the model; the following contributed significantly to an increase in the explained variance of the model: longer symptom duration, less perceived adequacy of social support, and less disease activity
Many potential predictors included in analysis, e.g., anxiety, self-efficacy, coping, sleep, life stress, functional ability
Jump et al, 2004 ([33])
  • US
  • RA with a history of AD: n = 48, RA without: n = 74
  • 93% women
Cross-sectionalHierarchical multiple regression analysisMAF
  • Fatigue related to a history of AD
  • This relationship was not significantly influenced by neuroticism, but mediated by self-efficacy
Control for age, education, illness duration
Pollard et al, 2006 ([18])
  • UK
  • Sample 1

    • RA: n = 228

    • 80% women
  • Sample 2

    • RA: n = 274

    • 75% women
Cross-sectionalSimple linear regression followed by stepwise multiple linear regression analysis100-mm fatigue VASa SF-36 vitality subscale
  • Sample 1: pain, HAQ, depression, MTX, erosions predicted fatigue (VAS)
  • Sample 2: pain, mental health, patient global predicted fatigue (VAS); pain and mental health predicted energy and vitality (SF-36)
DAS28, TJC, SJC, physician global, ESR, CRP, age, sex, disease duration, and other variables (e.g., regarding medication)
Riemsma et al, 1998 ([19])
  • The Netherlands
  • RA: n = 229
  • 61% women
Cross-sectionalStepwise multiple regression analysesTiredness VAS: How tired were you on average during the past week due to your arthritis? (0 = not tired at all, 100 = very tired)Pain; self-efficacy toward coping with RA symptoms as pain, disability, and depression; self-efficacy expectations toward the mobilization of help; problematic social supportAge, sex, education, duration, laboratory, social support
Sokka et al, 2009 ([37])
  • Multinational cohort (QUEST-RA), 25 countries
  • RA: n = 6,004
  • 79% women
Cross-sectionalGroup comparisons, effect sizesFatigue VASaFatigue predicted by sex: women had higher scores (poorer status) than men in all core data set measures
Stebbings et al, 2010 ([24])
  • New Zealand
  • RA: n = 103
  • 71% women
  • OA: n = 103
  • 58% women
Cross-sectionalMultivariate linear regression analysesMAFAnxiety, depressionAge, sex, disability, DAS, erosive damage and pain, CRP, sleep
Thyberg et al, 2009 ([20])
  • Sweden Early
  • RA (<1 year at inclusion): n = 276
  • 69% women
  • Cross-sectional
  • Longitudinal data from Swedish TIRA project: measures at inclusion and 3, 6, and 12 m later, then once per year to 8 years of followup
  • Fatigue data only available at 12-, 24-, and 36-m followup, so only these time points were used for this study
PCA, multiple linear regression analysis100-mm fatigue VAS (0 = no fatigue, 100 = worst possible fatigue) referring to “these last 7 days”
  • At 12 m and 36 m:

    • Women more fatigue than men

    • When comparing fatigue at 12 m with 24 m and 24 m with 36 m, no differences in either men or women
    • Underlying components (same for men and women) that turned out to be related to fatigue:

      • “Physical disability”: disease activity (DAS28), activity limitation (HAQ), pain

      • “Mental aspects”: mental health, sleep disturbance
  • Women: fatigue explained more by physical than by mental aspects at 12 m, but at 24 m and 36 m, more by mental than by physical aspects; age at inclusion was the only significant predictor of fatigue at 36 m
  • Men: fatigue explained more by mental than by physical aspects at 12 m, at 24 m by physical more than by mental aspects, at 36 m still by physical more than by mental, and age at inclusion was a significant predictor
Wolfe et al, 1996 ([1])
  • RA: n = 628
  • 75% women
  • OA: n = 535
  • 79% women
  • FM: n = 325
  • 94% women
Cross-sectionalMultiple stepwise regression analysis0–3 fatigue VAS: How much of a problem has fatigue or tiredness been for you in the past week? (anchors: fatigue is no problem, fatigue is a major problem)Pain, sleep disturbance, depression, tender points, and disability strongest independent predictors of fatigue in RA groupAge, ESR, morning stiffness, sex
Longitudinal studies      
Brekke et al, 2001 ([30])
  • Norway
  • RA: n = 815
  • 79% women
  • Longitudinal
  • Measures at baseline and 2 years later
Bivariate and multiple linear regression analysis100-mm fatigue VAS (0 = best score)aFatigue after 2 years was predicted by baseline fatigue, higher baseline self-efficacy for pain, and higher baseline self-efficacy for other symptoms (e.g., depression, fatigue)Age, sex, years of education, disease duration
Contreras-Yáñez et al, 2010 ([23])
  • Mexico
  • Recent-onset RA: n = 112 (36 therapy persistent, 76 nonpersistent)
  • 87% women
Longitudinal medical evaluations every 2, 4, or 6 mMultivariate Cox proportional hazards models (Cox regression analysis)Presence/absence of substantial fatigue (not further defined)
  • Persistence on DMARDs
  • Age at baseline, sex, RF and anti-CCP status, baseline value of fatigue and serial DMARD treatment, education, baseline HAQ, SF-36 physical score, both patient overall disease and pain VAS, DAS28, and followup
Davis et al, 2010 ([21])
  • US
  • RA: n = 228
  • 69% women
Longitudinal 30 daily diariesMultilevel and hierarchical linear regression analysesFatigue NRS: What number between 0 and 100 describes your average level of fatigue today? (0 = no fatigue, 100 = fatigue as bad as it can be)
  • Women more daily fatigue than men
  • Same day fatigue predicted by ≥1 comorbid pain conditions, days with more than average pain level, and more than average number of negative interpersonal events
  • Next day fatigue predicted by comorbid pain condition, same day fatigue, pain, negative events and interaction positive events, and being female
  • Women, positive events [RIGHTWARDS ARROW] less same day fatigue [RIGHTWARDS ARROW] more next day fatigue
  • Relationship of negative events and fatigue is mediated by negative affect
  • Age, education, income, employment status, relationship status, general health, daily pain
  • For men, no relationships of positive events with same day and next day fatigue
Fifield et al, 2001 ([29])
  • RA with and without current or prior AD: n = 415
  • 83% women
10-year longitudinal study, fatigue measured once per year from year 2 through 8 (i.e., 7 years)
  • Growth curve analysis (hierarchical linear models using OLS regression estimates)
  • History of AD: fulfilling DSM-IV criteria for major depression or generalized anxiety disorder at 1 time in the past
Fatigue VAS: On a scale of 0 to 100, with 0 being no fatigue at all and 100 being the most fatigue possible, how much arthritis fatigue did you feel in the past week?
  • Those with AD history had higher levels of fatigue in the first year and remained higher in the entire study compared to those without AD
  • 3% of between-subject variance in initial fatigue due to history of AD (56% of total variance)
  • Effect of AD history on fatigue mediated by current distress
Comorbidity, disease duration (RA)
Mancuso et al, 2006 ([32])
  • US
  • RA: n = 122
  • 84% women
  • Controls: n = 122
  • 91% women
  • Followup ∼14 months later RA: n = 91 Controls: n = 89
1-year longitudinal study, measures at enrollment and after 1 yearBackward stepwise multiple regression analysisFatigue Severity Scale
  • For RA group: more fatigue associated with more anxiety, more disability, less social support, and more social stress
  • Enrollment variables associated with worse fatigue at followup, based on longitudinal multivariate regression analysis, less help at home, more anxiety, more disability
Depressive symptoms, role satisfaction, physical activity, sleep quality at enrollment
Odegård et al, 2008 ([36])
  • Norway
  • RA: n = 550 included during 6 years, n = 216 after 10 years
  • 74% women
Longitudinal, inclusion of 6 cohorts (1 per year) and followup after 10 yearsOne-way ANOVA, analysis of covariance, logistic regression100-mm fatigue VASaSex difference in fatigue as measured 10 years after disease onset; women worse than men 
Parrish et al, 2008 ([35])
  • US
  • RA: n = 89
  • OA: n = 76
  • FM: n = 90
  • 100% women
Longitudinal 30 daily diariesMultilevel modeling
  • Daily fatigue: VAS fatigue: What number between 0 and 100 best describes your average level of fatigue today? (0 = no fatigue, 100 = fatigue as bad as it can be)
  • Daily fatigue affect: fatigue subscale of PANAS-X (4 items)
Negative events predicted more fatigue in all patient groups Different reactions of patient groups on positive events; OA next day fatigue decreases, RA and FM next day fatigue increasesAlways controlled for prior day fatigue
Repping-Wuts et al, 2007 ([28])
  • The Netherlands
  • RA: n = 123
  • 68% women
Longitudinal followup design, with 1-year durationLogistic regression analysis
  • CIS (baseline, 1 year)
  • 2 groups of patients (with and without severe fatigue based on CIS cutoff score)
  • Persistent severe fatigue predicted by disability at baseline (baseline HAQ DI) and average GH (average GH VAS)
  • Level of baseline fatigue related to level of fatigue at followup and remained constant for most patients
ESR, average ESR, SW28, average SW28, TE28, average TE28, GH VAS, HAQ DI, followup
Scharloo et al, 1999 ([31])
  • The Netherlands
  • RA without serious comorbidity: n = 71
  • 75% women
Longitudinal 2 time points of measures (baseline and followup ∼1 year later)Stepwise multiple regression analysis10-cm tiredness VAS over the last week (0 = no tiredness, 10 = very severe tiredness)Baseline scores (tiredness at time 1) explained 40% of variance of tiredness 1 year later, perceived consequences of RA, identity perceptions (belief that experienced symptoms as part of RA), and avoidant coping predictor of fatigue 1 year laterIllness duration, disease severity, other illness perceptions, coping strategies
Stone et al, 1997 ([22])
  • US
  • RA: n = 35
  • 71% women
7 days (EMA measures 7 times per day) EMA for better recalling of current states (how patients feel at the moment)
  • OLS regression
  • Repeated-measures ANOVA
7-point fatigue scale (0 = not at all, 7 = extremely)
  • OLS: higher fatigue variability associated with more pain, average EMA fatigue (severity), joint pain, muscle pain, stiffness on awakening, pain on awakening, swelling on awakening, and poorer sleep quality
  • No sign. Differences in pain or fatigue by day of the week, but strong effects of time of day for both; fatigue moderate in the morning, lowest between 10 AM and noon, steep rise throughout the rest of the day
  • Sleep measure on a daily basis: fatigue predicted by sleep quality
  • OLS: neuroticism, anxiety, depression, average number of hours of sleep each night
  • Sleep measure on a daily basis: fatigue not predicted by number of hours slept
  • More pain on stressful days, but no significant difference in fatigue
Treharne et al, 2008 ([25])
  • UK
  • RA: n = 114
  • 74% women
Longitudinal (2 points of measuring: baseline and followup 1 year later)Hierarchical linear regression analysis
  • 100-mm fatigue VAS (anchors: no fatigue, unbearable fatigue)
  • All variables measured at baseline and fatigue also measured at followup
  • Baseline fatigue explained 13% of the variance in fatigue 1 year later
  • Inflammation (ESR), perception of consequences of RA, and self-efficacy were significant predictors
Age, sex, employment status, DMARD use, pain, impact of disability, sleep disruption, depressed mood, praying/hoping, coping
Table 2. Fatigue as a predictor (n = 17)*
Author, year (ref.)Sample (country, population, sex)Study designMain statistical methodMeasure of fatigueResults
Fatigue predicts …Fatigue does not predict …
  1. RA = rheumatoid arthritis; AS = ankylosing spondylitis; HRQOL = health-related quality of life; MFI = Multidimensional Fatigue Inventory; PCS = physical component summary; MCS = mental component summary; SF-36 = Short Form 36; VAS = visual analog scale; DAS = Disease Activity Score; MS = multiple sclerosis; GHQ = General Health Questionnaire; NHP = Nottingham Health Profile; FM = fibromyalgia; SI = symptom intensity; RPS = regional pain scale; RAES = Rheumatoid Arthritis Evaluation Study; ARCK = Arthritis and Rheumatology Clinic of Kansas; SLE = systemic lupus erythematosus; NRS = numerical rating scale; MAF = Multidimensional Assessment of Fatigue; HAQ = Health Assessment Questionnaire; anti-TNF = anti–tumor necrosis factor.

  2. a

    Wording of the VAS not reported.

Cross-sectional studies      
Chorus et al, 2003 (38)
  • The Netherlands
  • RA: n = 1,056
  • 72% women
  • AS: n = 658
  • 30% women
Cross-sectional
  • Stepwise multiple linear regression analysis
  • To predict: HRQOL
MFI: general fatigue, physical fatigue, reduced motivation, reduced activity, mental fatigue
  • RA and AS: general fatigue worse in women than in men, no differences regarding other fatigue MFI
  • General fatigue predicts physical and mental HRQOL (PCS and MCS scores of SF-36), other fatigue aspects not included in regression model
Covic et al, 2006 (47)
  • Australia
  • RA: n = 134
  • 77% women
  • 54 depressed, 66 nondepressed
Cross-sectionalDiscriminant analysis with depression categories as dependent variables and 12 predictors as independent variables To predict: depression10-cm fatigue VAS (1 = no fatigue, 10 = extreme fatigue)Fatigue predictor of depression after tension, self-esteem, and perceived RA impact
De Croon et al, 2005 ([50])
  • The Netherlands
  • Early RA (working): n = 78
  • >67% women
Cross-sectional
  • Logistic regression analysis
  • To predict: work ability
Checklist Individual Strength total score from 20–140, dichotomized in fatigued and nonfatigued patients by using cutoff point of >76
  • Fatigued employees with RA (n = 17) report lower levels of work ability compared to nonfatigued employees (n = 61) with RA
  • After adjustment for age, pain, and DAS, relationship still significant
Gronning et al, 2010 ([42])
  • Norway
  • RA: n = 310
  • 72% women
Cross-sectional
  • Hierarchical multiple linear regression analysis
  • To predict: HRQOL
VAS tiredness over the last weeksaFatigue predicts decreased physical and mental HRQOL
Katz, 1998 ([45])
  • US
  • RA: n = 446
  • 79% women
Cross-sectional
  • Stepwise multiple linear regression analysis
  • To predict: stressors of RA (e.g., impact of fatigue)
Fatigue item; rate fatigue or tiredness during the past 2 weeks from no fatigue to very mild, mild, moderate, severe, very severe (for analysis, responses were dichotomized; e.g., severe and very severe grouped together)Fatigue severity associated with fatigue impact
Parker White et al, 2009 ([53])
  • US
  • Mothers with RA: n = 68
  • Mothers with MS: n = 103
  • Healthy mothers: n = 91
  • 100% women
Cross-sectional
  • Hierarchical multiple linear regression analysis
  • To predict: caregiving environment (the mother's experience of daily hassles of parenting, discipline style she employed, how she monitored her child's whereabouts)
21-item Modified Fatigue Impact Scale; physical fatigue, cognitive fatigue, psychosocial fatigue
  • Higher levels of fatigue related to both a greater frequency and intensity of parenting daily hassles for mothers with RA
  • Mothers with RA who reported more fatigue also reported more difficulties monitoring their child
RA fatigue did not explain significant variance in predicting laxness and overreactivity
Rupp et al, 2004 (43)
  • The Netherlands
  • RA: n = 490
  • 73% women
Cross-sectional
  • Linear and logistic and regression analyses with 8 dimensions of RAND-36 as dependent variables and fatigue (physical fatigue, reduced activity, mental fatigue, and reduced motivation), RA-related pain, and depressive symptoms as independent variables (general fatigue left out to reduce risk of [multi]collinearity)
  • To predict: HRQOL
  • 100-mm fatigue VAS (0 = no fatigue, 100 = fatigue as bad as it could be)
  • MFI-20 (general fatigue, physical fatigue, reduced activity, mental fatigue, reduced motivation)
  • HRQOL is related to fatigue; different aspects of fatigue explain different dimensions of HRQOL while taking into account pain and depressive symptoms:

    • Physical fatigue: physical functioning, social functioning, role limitations physical, vitality, pain, general health perceptions

    • Reduced activity: social functioning, role limitations emotional, vitality, pain
    • Reduced motivation: vitality
    • Mental fatigue: role limitations physical and emotional, mental health
Smedstad et al, 1996 ([49])
  • Norway
  • RA: n = 238
  • 74% women
  • Controls: n = 116
  • 69% women
Cross-sectional
  • Multiple linear regression analyses with GHQ (subscales for anxiety, depression, social dysfunction, and somatization) as dependent variables and patient status (patient or control) and pain, disability, and fatigue as independent variables
  • To predict: psychological distress
NHP: 3 fatigue itemsDifferences between patients and controls regarding mental distress (sum and all subscales of the GHQ) and all subscales no longer significant when controlling for fatigue, pain, and disability
Wolfe and Michaud, 2008 (39)
  • US
  • RA: n = 9,921
  • 75% women
  • Noninflammatory rheumatic disorder (not FM): n = 2,851
  • FM: n = 2,867
  • Sex ratios not reported
Cross-sectional
  • Logistic regression in univariate and multivariate analysis, Kendall's τ to examine association between combined dryness and RA symptom scales
  • To predict: oral and ocular dryness
  • Fatigue VAS†
  • SI scale: consisting of 1 fatigue VAS and RPS
Fatigue 1 of 4 predictors of dryness in RA (beside SI scale, pain, and sleep scale)
Wolfe and Michaud, 2009 (44)
  • US
  • RA: n = 20,268
  • RA patients from RAES and ARCK data sets: n = 1,075
  • Sex ratio not reported
Cross-sectional
  • Generalized linear model, Cox time-varying regression analyses
  • To predict: RA outcome and treatment success and failure (surrogated by satisfaction with health)
VAS fatigue over the last week, 21 points from 0–10 at 0.5-unit intervalsaPatients without comorbidity who had more fatigue were less satisfied with their health
Wolfe and Michaud, 2009 (48)
  • US
  • RA: n = 22,131
  • 77% women
  • Noninflammatory disease: n = 3,717
  • SLE: n = 1,002
  • FM: n = 2,674
Cross-sectional
  • Fractional polynomial logistic regression
  • To predict: depression
  • Fatigue VAS†
  • SI scale: consisting of 1 fatigue VAS and RPS
  • In RA, fatigue significantly higher for depressed than for nondepressed patients
  • Fatigue and pain important predictors of self-reported depression in RA
Yazici et al, 2004 (40)
  • US
  • Early RA: n = 337
  • 75% women
Cross-sectional and longitudinal design (patients with minimal 2 visits during 3.5 years)
  • Cross-sectional data analysis: ordered logit regression models
  • Longitudinal analysis: random-effects logistic regression
  • To predict: morning stiffness
10-cm fatigue VASaFatigue at baseline was significantly associated with morning stiffness, beside age, disease duration, disability, pain, patient global, and number of symptoms
Longitudinal studies      
Breedveld et al, 2005 ([41])
  • The Netherlands
  • RA: n = 428
  • Sex ratio not reported
  • Longitudinal
  • 54-week measures at 4-week intervals from baseline to week 54
  • Stepwise multiple linear regression and logistic regression analysis
  • To predict: physical disability
10-cm fatigue VASa
  • Results of baseline data (cross-sectional analysis at baseline):

    • Linear regression model: fatigue significantly related to disability

    • Logistic regression model: fatigue related to disability
  • Longitudinal analysis:

    • Multiple linear regression: baseline fatigue associated with disability at week 54

Davis et al, 2006 ([52])
  • US
  • RA: n = 184
  • 72% women
Longitudinal 30 daily diaries
  • Multilevel modeling, multivariate, multilevel random-effects regression analysis
  • To predict: daily interpersonal events in pain patients
  • NRS fatigue: What number between 0 and 100 describes your average level of arthritis fatigue today? (0 = no fatigue, 100 = fatigue as bad as it can be)
  • Diaries: daily average level of fatigue
  • Also assessed: current day's deviation of fatigue from one's average level
Fatigue is a significant predictor (mean daily fatigue/current day's fatigue) of negative daily events (yes/yes), positive daily events (no/yes), daily relationship stress with friends (no/yes), daily relationship stress with family (yes/no), daily relationship enjoyment with family (no/yes), and enjoyment with spouse (no/yes) on the same dayFatigue not a predictor of daily relationship enjoyment with friends, stress with coworkers, enjoyment with coworkers, stress with spouse
Mancuso et al, 2005 ([51])
  • US
  • RA: n = 122
  • 84% women
  • Healthy controls: n = 122
  • 91% women
Longitudinal measures at baseline and after 1 year
  • Multivariate logistic regression analysis
  • To predict: negative work place events
MAFFor patients with RA, more fatigue was associated with having a negative event in the work place (minor work place events: commonly occurring stressful daily events)No significant association with major work place events
Waltz, 2000 ([54])
  • RA Dutch: n = 136
  • RA German: n = 98
  • 70% women
  • Longitudinal
  • 3 points of measures (baseline and after 12 and 24 m)
  • Hierarchical multiple linear regression analyses
  • To predict: consumption patterns of medical care
  • Composite Index of Fatigue Impairment Symptoms: 11-point NRS and 3 NHP items
  • Sample split up into 4 groups according to level of fatigue
Fatigue contributes to physician consultations and referral to a physical therapist in 2 years following baseline measures as assessed at 12 and 24 m
Wolfe et al, 2004 ([48])
  • US
  • RA: n = 852
  • 80% women
  • RA: n = 12,217
  • Sex ratio not reported
Longitudinal At least 2 consecutive semiannual measures
  • Multivariable ordered logistic regression analysis, general estimating equations
  • To predict: symptom importance in RA
Fatigue VAS: How much of a problem has fatigue or tiredness been for you in the past week? (0 = fatigue is no problem, 10 = fatigue is a major problem)In multivariable-ordered logistic regression with fatigue VAS, HAQ, pain, depression, age, and sex as independent variables, only fatigue VAS predicted the ranked position for fatigue importanceRanking of fatigue not related to use of anti-TNF agents or prednisone

Subsequently, the findings of the studies were summarized according to 5 categories: 1) illness-related aspects, e.g., disease activity, pain, tender joint count, and radiographic damage; 2) physical functioning, e.g., measures of disability, physical functioning, health-related quality of life, and sleep; 3) cognitive/emotional functioning, e.g., depression and anxiety; 4) social/environmental aspects, e.g., work, social support, and life events; and 5) demographic aspects, e.g., sex and age.

We conducted no numerical quality assessment of the studies included in this systematic review because the studies were more diverse than randomized controlled trials, where clear quality assessment criteria such as the Cochrane guidelines ([13]) exist. Explicit standards for quality assessment of studies other than intervention studies are lacking, and a comparable review about the factors associated with fatigue in RA does not yet exist. A standard methodology was not available; therefore, we worked from an explorative viewpoint. Nevertheless, in the description and interpretation of the results, we took into account the different design quality of the included studies (e.g., whether longitudinal data were used). We compared the results from the cross-sectional studies to those of the longitudinal studies and concluded that there were no systematic differences, suggesting that the overall results reported in our study are consistent among the different study designs.

To distinguish between cross-sectional and longitudinal study designs in the tables, we grouped the studies in the tables according to their design. The cross-sectional studies exclusively provided information about associations between variables, but not about causal relationships. By the term “predictor,” we do not intend to indicate causal relationships. “Predictor” refers to the fact that many of the reviewed cross-sectional and longitudinal studies used regression analysis to statistically predict fatigue with other variables, or to statistically predict another variable with fatigue. In the allocation of studies in 1 of the 2 tables, we followed the intended direction of the analyses of the original studies.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgments
  9. REFERENCES
  10. Supporting Information

This results section consists of 2 parts. First, we provide a summary of the results regarding the possible causes of fatigue, and then we describe the results regarding the possible consequences of fatigue.

Possible causes of fatigue in RA

Table 1 shows the reviewed studies that reported results about possible causes of fatigue in RA.

Illness-related aspects

Among illness-related variables, elevated pain most often appeared as a symptom associated with increased levels of fatigue. This relationship was supported by cross-sectional studies ([1, 14-20]) and a longitudinal study ([21]) that controlled for previous fatigue levels. Another longitudinal study found that pain was associated with fatigue variability ([22]). Only 3 studies reported that pain was not significantly related to fatigue ([23-25]). Strikingly, 3 longitudinal studies ([21, 23, 25]) showed conflicting results regarding pain and fatigue severity. Davis et al ([21]) demonstrated a relationship between fatigue and pain, but Contreras-Yáñez et al ([23]) and Treharne et al ([25]) did not. This might be explained by the differences between the studies. Davis et al ([21]) conducted a diary study with 30 consecutive measures of fatigue and pain, whereas Treharne et al ([25]) measured fatigue and pain at baseline and at followup 1 year later and Contreras-Yáñez et al ([23]) measured fatigue and pain every 2, 4, or 6 months. Moreover, these studies included different variables beside pain and fatigue. For example, Treharne et al ([25]) included sleep disruption in their study and the others did not.

Characteristics of inflammatory activity (e.g., erythrocyte sedimentation rate [ESR] and Disease Activity Score in 28 joints [DAS28]) showed an unclear relationship with fatigue in RA. In some studies, these markers were significantly related to fatigue ([15, 17, 20, 25, 26]); in other studies, they did not contribute to the severity of fatigue at all ([1, 18, 23, 24, 27, 28]). Four of the studies that found a significant relationship were cross-sectional ([15, 17, 20, 26]) and 1 study was longitudinal ([25]). Davis et al ([15]) reported that lipopolysaccharide-stimulated interleukin-6 (IL-6) level was still significantly related to fatigue when controlled for pain, but fatigue was not related to plasma levels of both C-reactive protein (CRP) and IL-6. In the study by Dhir et al ([26]), fatigue was significantly associated with DAS28 scores. Huyser et al ([17]) found that less disease activity was related to increased fatigue. Thyberg et al ([20]) showed a significant association with fatigue by a cluster of disease activity, activity limitations, and pain, labeled as physical disability. In the longitudinal study by Treharne et al ([25]), a significant relationship between inflammation (ESR) and fatigue was reported. Van Hoogmoed et al ([16]) found a significant relationship between the presence of rheumatoid factor and fatigue, but inflammatory variables such as the DAS28, ESR, CRP level, and hemoglobin level did not turn out to be related to fatigue in multivariate analyses of this cross-sectional study ([16]).

A longitudinal study found that higher fatigue variability was significantly associated with more stiffness and swelling on awakening ([22]), whereas in a cross-sectional analysis by Wolfe et al ([1]), morning stiffness was not significantly associated with fatigue. Two cross-sectional studies ([1, 26]) reported that the number of tender points was significantly related to fatigue in RA. In contrast, tender and swollen joint counts were not associated with fatigue in a cross-sectional study ([18]) and a longitudinal study ([28]).

Conflicting results were found regarding 2 illness-related aspects: comorbidity and disease duration. A significant association between fatigue and comorbidity was reported in a cross-sectional study ([14]) and a longitudinal study that controlled for previous fatigue levels ([21]). Another longitudinal study did not find this association ([29]). Longer disease/symptom duration turned out to be significantly related to fatigue in 2 cross-sectional studies ([14, 17]). However, 1 cross-sectional study ([18]) and 3 longitudinal studies ([29-31]) did not support the relationship between fatigue and disease duration.

Another longitudinal study reported that worse general health was associated with higher levels of fatigue ([28]). Contreras-Yáñez et al ([23]) did not find this relationship in a longitudinal study. They examined the impact of therapy persistence on RA outcomes, reporting no significant relationship between the persistence of disease-modifying antirheumatic drugs and fatigue.

Physical functioning

Several studies also considered physical functioning and indications of disability. These characteristics were significantly related to fatigue in nearly all cross-sectional studies that included them ([1, 14, 16, 18-20, 27]). Only 1 cross-sectional study ([24]) and 1 longitudinal study ([23]) found no association between fatigue and disability. Two longitudinal studies that controlled for baseline fatigue also reported a significant relationship between fatigue and disability ([28, 32]). In 4 cross-sectional studies, significant associations with level of fatigue were also found for quality of sleep ([14]) or sleep disturbances ([1, 16, 20]). A longitudinal study ([22]) that controlled for previous fatigue levels reported that fatigue variability and severity were significantly related to sleep quality, but not to the number of hours slept. However, some studies found no support for a relationship between sleep and fatigue cross-sectionally ([15, 17, 24]) and longitudinally ([25, 32]).

Cognitive and emotional functioning

Regarding cognitive and emotional functioning, the most often investigated construct in relation to fatigue was depression, operationalized as major depression or depressive mood. In several cross-sectional studies ([1, 16-20, 24]), depression was significantly associated with fatigue. Exceptions were one of the cross-sectional studies ([14]) and 2 of the longitudinal studies ([25, 32]) that controlled for baseline fatigue. A cross-sectional study ([33]) and a longitudinal study ([29]) found associations between fatigue and being diagnosed at least once during the individual's lifetime with an affective disorder (major depression or generalized anxiety disorder). A cross-sectional study ([24]) found a significant relationship between fatigue and anxiety; the same relationship was found by a longitudinal study that controlled for baseline fatigue ([32]). However, the results of another cross-sectional study ([17]) and a longitudinal study ([22]) controlling for previous levels of fatigue did not support this relationship.

A reverse relationship was reported for self-efficacy and fatigue. In 3 cross-sectional studies ([16, 19, 33]) and 2 longitudinal studies ([25, 30]) that controlled for baseline fatigue, significant negative associations were reported. In contrast, a cross-sectional study did not find any support for a relationship between self-efficacy perceptions and fatigue ([17]). Two longitudinal studies that controlled for baseline fatigue ([25, 31]) showed an association between perceived consequences of RA and fatigue. Catastrophizing and avoidant coping were also found to be associated with higher levels of fatigue in a cross-sectional study ([16]) and a longitudinal study ([31]). Relationships between fatigue and reduced role functioning were found in a cross-sectional study ([16]) and between fatigue and current distress ([29]) in a longitudinal study.

Social and environmental aspects

Several cross-sectional ([17, 19, 34]) and longitudinal studies ([21, 32, 35]) that controlled for baseline or previous levels of fatigue pointed to the role of social and environmental aspects in the explanation of fatigue. Negative interpersonal events, for example, were associated with higher levels of fatigue ([21, 34, 35]). Less social support ([32]) or inadequate support (from the perspective of the patient) ([17, 19]) was significantly related to worse fatigue. The same applied to social stress and when there was less help at home ([32]). However, 2 cross-sectional studies found that social support was not related to fatigue ([14, 16]).

Demographic aspects

Age and sex were considered as potential predictors of fatigue in most of the reviewed studies. However, a relationship between fatigue and the age of patients was not demonstrated ([1, 14, 15, 18, 19, 21, 23, 25, 30]) except in 2 cross-sectional studies ([16, 20]) that reported conflicting directions of the relationship. Regarding the sex of the patients, several studies showed associations with fatigue ([14, 17, 20, 21, 36-38]); female sex was significantly related to worse fatigue. However, there were also numerous studies that reported no sex differences ([1, 15, 16, 18, 19, 23, 25, 30]).

Finally, fatigue at baseline was significantly related to fatigue at followup 1 year later, as reported by longitudinal studies that included fatigue at baseline in their prediction model of fatigue at followup ([25, 28, 31]). Moreover, in another longitudinal study, fatigue variability was predicted by fatigue severity ([22]).

Possible consequences of fatigue in RA

Table 2 shows the reviewed studies that reported results about possible consequences of fatigue in RA. In the studies that used fatigue as an independent variable, fatigue turned out to be significantly related to several variables.

Illness-related aspects

Fatigue was significantly related to ocular and oral dryness ([39]) and morning stiffness ([40]) in cross-sectional studies. A possible impact of fatigue on inflammatory processes was not examined in the included studies.

Physical functioning

A longitudinal study found that fatigue statistically predicted physical functioning ([41]) while controlling for baseline physical functioning. Cross-sectional studies reported that fatigue was significantly related to reduced physical health-related quality of life ([38, 42, 43]), and that different aspects of health-related quality of life were associated with different aspects of fatigue ([43]).

Cognitive and emotional functioning

Regarding cognitive/emotional functioning, some cross-sectional studies reported that fatigue was significantly related to reduced mental health-related quality of life ([38, 42, 43]) and satisfaction with health ([44]). A cross-sectional study ([45]) and a longitudinal study ([46]) found that the level of fatigue statistically predicted the importance of fatigue as perceived by patients. Two cross-sectional studies reported that fatigue was significantly associated with depression ([47, 48]) and another study reported that fatigue was associated with global psychological distress ([49]).

Social aspects

The following social aspects were reported to be related to fatigue. A cross-sectional study found that fatigue was significantly related to reduced work ability ([50]). In a longitudinal study ([51]), fatigue was also associated with minor negative work place events (commonly occurring stressful events), but not with major work place events. In another longitudinal study ([52]), fatigue turned out to be related to same day negative daily events, positive daily events, stress with friends, stress with family, enjoyment with family, and enjoyment with the spouse, but not to enjoyment with friends, stress or enjoyment with coworkers, or stress with the spouse ([52]). A cross-sectional study about the impact of fatigue on parenting found that fatigue was related to greater frequency and intensity of daily hassles while parenting as well as having less energy to monitor a child's whereabouts ([53]). Fatigue was not significantly associated with laxness and/or overreactivity in mothers with RA ([53]). Additionally, a longitudinal study found that fatigue was significantly associated with more physician consultations and more frequent referrals to a physical therapist over 2 years ([54]).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgments
  9. REFERENCES
  10. Supporting Information

This systematic review provided an overview of the existing literature on possible causes and consequences of fatigue in RA. Studies found possible causes for fatigue among illness-related aspects (pain), physical functioning (sleep quality, sleep disturbances, global health ratings, and disability), cognitive/emotional functioning (depression, anxiety, and perceptions as self-efficacy), social/environmental aspects (interpersonal events and social support), and female sex. However, none of these variables showed profound and stable relationships with fatigue across the studies. Even if several studies found a significant association with fatigue, at least one study did not show this result. Inconsistent results were especially reported with regard to characteristics of inflammatory activity (e.g., ESR, DAS28, flares).

The following possible consequences of fatigue emerged in the reviewed studies: illness-related aspects (ocular and oral dryness and morning stiffness), physical functioning (physical health-related quality of life), cognitive/emotional functioning (depression, global psychological distress, mental health-related quality of life, and satisfaction with health), and social/environmental aspects (work ability, negative and positive daily events, and parenting). Fatigue also statistically predicted physician consultations and referrals to a physical therapist.

Some of the variables appeared as both possible causes as well as consequences. This means that although some studies reported that the variables statistically predicted fatigue, other studies found that these variables were statistically predicted by fatigue. This applied for variables among physical functioning, cognitive/emotional functioning (health evaluation, depression, and psychological distress), and social/environmental aspects (interpersonal events). This finding clearly points to the previously mentioned causation problem. Based on the reviewed studies, it is not possible to draw profound conclusions about the causal direction of relationships between fatigue and other variables. Many studies were cross-sectional, and some of the longitudinal studies had shortcomings, for example, not controlling for baseline fatigue or using small or selected samples. Moreover, it is unclear how the proposed directions between fatigue and the other variables correspond with the real processes. Multiple circular processes likely underlie fatigue in RA ([10]), and therefore it is not possible to identify single causes of fatigue without considering interactions with other potential pathways. Furthermore, the multiple causes for fatigue in RA may vary between individuals and occur in different combinations and strengths at different times ([10]). However, we compared the results from the cross-sectional studies with the results from the longitudinal studies and found no systematic differences, assuming that the overall results reported in our study are consistent among the different study designs.

Based on this systematic review, we conclude that 3 variables have a high probability to be involved in the complex process of fatigue in RA: pain, disability/physical functioning, and depression/depressive mood. For each of the 3 variables, approximately 10 of the reviewed studies reported significant relationships with fatigue, and these studies included both cross-sectional and longitudinal studies. Regarding pain, 2 longitudinal studies found a significant association with fatigue (one with fatigue severity and the other with fatigue variability), whereas 2 longitudinal studies did not find such a relationship. Physical functioning was significantly associated with fatigue in 2 longitudinal studies that controlled for baseline fatigue. One longitudinal study did not support this association. One longitudinal study found a significant relationship between fatigue and depression, whereas 2 others did not. In conclusion, these 3 variables could be regarded as potential causes for fatigue in RA. However, more evidence was found for fatigue being related to pain and physical functioning than to depression.

For a closer examination of the chronological sequence of the considered symptoms, it would be necessary to conduct prospective, longitudinal studies with representative samples. Such studies should use an adequate measurement instrument for fatigue and apply multivariate analysis techniques with a check for confounders. Not all of the reviewed studies clearly indicated whether they checked for confounders and, if so, for which ones. Moreover, the reviewed studies differed in their precision of describing their fatigue measurement. More than half of the studies merely used a single visual analog scale or numerical rating scale, and information about the wording was not always provided. This measurement does not correspond to the multidimensional character of fatigue as reported by the patients ([55, 56]). Only a few studies used a fatigue questionnaire containing more than one dimension. Those questionnaires that appeared in more than one study were the Multidimensional Assessment of Fatigue (MAF) ([14, 24]), Checklist Individual Strength ([16, 28]), and Multidimensional Fatigue Inventory ([38, 43]). Only one of these instruments, the MAF, is an RA-specific questionnaire. Some items of the other 2 scales could be related to disability or disease activity in RA, and therefore may give a biased estimation of fatigue levels ([55]).

In evaluating our conclusions, the eventuality of positive publication bias also has to be taken into account. Statistically significant results have a higher chance of being published than those showing nonsignificant results ([57]). In our search strategy, we included all relevant studies, irrespective of their results. Nevertheless, it could be the case that studies not finding significant associations between fatigue and other variables were not published and, therefore, could not feasibly appear in any strategic search. Moreover, it is impossible to control whether authors only reported the significant results and not the nonsignificant ones. This study is a systematic review and, therefore, dependent on the available literature. Therefore, we are not able to compensate for possible erroneous judgments in original studies, for example, for the choice of authors to include certain variables and their analytical methodology. Obviously, a systematic review cannot improve the quality of the included studies. However, the great advantage of a systematic review is that it can consider the weaknesses and strengths of the included studies in the interpretation of its results.

Furthermore, we used proximity searches, meaning that the search in the databases was constrained to studies that included a certain combination of search terms within a relatively small area of words in the abstract. Another approach would have been unmanageable because the titles and abstracts to read would have far outnumbered the hits received. To compensate for this limitation, we thoroughly checked the reference lists of the included studies for additional relevant studies.

This study provides an overview of the state of the art on research about fatigue and factors related to fatigue in RA. When comparing our findings with the hypothetical model of Hewlett et al ([10]), obviously only partial support for the proposed relationships exists. The strongest support was found for relationships between fatigue and variables on the RA dimensions of disability, pain, and sleep. However, the proposed relationships between fatigue and the illness-related aspects of cortisol response, inflammation, joint damage, muscle effort and decondition, drugs, and anemia did not match well with the reviewed literature. On the dimension “cognitive, behavioural,” referring to behaviors, emotions, and cognitions, we identified the association between fatigue and depression as most frequent. Regarding the other elements of this dimension (e.g., illness beliefs and stress, activity) and also the “personal” dimension referring to work/caretaking responsibilities, environment, health, and social support, indications for the existence of relationships with fatigue were also found in the literature, but by clearly fewer studies than those reporting about associations between fatigue and pain, disability, and depression.

It should be noted that this review did not test for the causality of considered bivariate associations, but mainly described relationships. We summarize the current knowledge in the field to inform future questions with regard to the challenging task of conducting research into causes and consequences of fatigue in RA and to develop theoretical models of fatigue.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgments
  9. REFERENCES
  10. Supporting Information

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Nikolaus had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Nikolaus, Bode, Taal, van de Laar.

Acquisition of data. Nikolaus, Bode, Taal.

Analysis and interpretation of data. Nikolaus, Bode, Taal.

Acknowledgments

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgments
  9. REFERENCES
  10. Supporting Information

The authors would like to thank Dr. Marjolein Drent for supporting the development of the search strategy.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgments
  9. REFERENCES
  10. Supporting Information
  • 1
    Wolfe F, Hawley DJ, Wilson K.The prevalence and meaning of fatigue in rheumatic disease.J Rheumatol1996;23:140717.
  • 2
    Hewlett S, Carr M, Ryan S, Kirwan J, Richards P, Carr A, et al.Outcomes generated by patients with rheumatoid arthritis: how important are they?Musculoskeletal Care2005;3:13142.
  • 3
    Belza Tack B.Fatigue in rheumatoid arthritis: conditions, strategies, and consequences.Arthritis Care Res1990;3:6570.
  • 4
    Hewlett S, Cockshott Z, Byron M, Kitchen K, Tipler S, Pope D, et al.Patients' perceptions of fatigue in rheumatoid arthritis: overwhelming, uncontrollable, ignored.Arthritis Rheum2005;53:697702.
  • 5
    Repping-Wuts H, Uitterhoeve R, van Riel P, van Achterberg T.Fatigue as experienced by patients with rheumatoid arthritis (RA): a qualitative study.Int J Nurs Stud2008;45:9951002.
  • 6
    Nikolaus S, Bode C, Taal E, van de Laar MA.New insights into the experience of fatigue among patients with rheumatoid arthritis: a qualitative study.Ann Rheum Dis2010;69:8957.
  • 7
    Hewlett S, Hehir M, Kirwan JR.Measuring fatigue in rheumatoid arthritis: a systematic review of scales in use.Arthritis Rheum2007;57:42939.
  • 8
    Repping-Wuts H, van Riel P, van Achterberg T.Fatigue in patients with rheumatoid arthritis: what is known and what is needed.Rheumatology (Oxford)2009;48:2079.
  • 9
    Hewlett S, Nicklin J, Treharne GJ. Fatigue in musculoskeletal conditions.2008. URL: http://eprints.uwe.ac.uk/10682/1/ARC_topic_Review_Fatigue08.pdf.
  • 10
    Hewlett S, Chalder T, Choy E, Cramp F, Davis B, Dures E, et al.Fatigue in rheumatoid arthritis: time for a conceptual model [editorial].Rheumatology (Oxford)2011;50:10046.
  • 11
    Stebbings S, Treharne GJ.Fatigue in rheumatic disease: an overview.Int J Clin Rheumatol2010;5:487502.
  • 12
    Rothman KJ, Greenland S, Lash TL.Modern epidemiology.Philadelphia:Wolters Kluwer, Lippincott Williams & Wilkins;2008.
  • 13
    Higgins JP, Green S. Cochrane handbook for systematic reviews of interventions: version 5.1.0.2011. URL: www.cochrane-handbook.org.
  • 14
    Belza BL, Henke CJ, Yelin EH, Epstein WV, Gilliss CL.Correlates of fatigue in older adults with rheumatoid arthritis.Nurs Res1993;42:939.
  • 15
    Davis MC, Zautra AJ, Younger J, Motivala SJ, Attrep J, Irwin MR.Chronic stress and regulation of cellular markers of inflammation in rheumatoid arthritis: implications for fatigue.Brain Behav Immun2008;22:2432.
  • 16
    Van Hoogmoed D, Fransen J, Bleijenberg G, van Riel P.Physical and psychosocial correlates of severe fatigue in rheumatoid arthritis.Rheumatology (Oxford)2010;49:1294302.
  • 17
    Huyser BA, Parker JC, Thoreson R, Smarr KL, Johnson JC, Hoffman R.Predictors of subjective fatigue among individuals with rheumatoid arthritis.Arthritis Rheum1998;41:22307.
  • 18
    Pollard LC, Choy EH, Gonzalez J, Khoshaba B, Scott DL.Fatigue in rheumatoid arthritis reflects pain, not disease activity.Rheumatology (Oxford)2006;45:8859.
  • 19
    Riemsma RP, Rasker JJ, Taal E, Griep EN, Wouters JM, Wiegman O.Fatigue in rheumatoid arthritis: the role of self-efficacy and problematic social support.Br J Rheumatol1998;37:10426.
  • 20
    Thyberg I, Dahlstrom O, Thyberg M.Factors related to fatigue in women and men with early rheumatoid arthritis: the Swedish TIRA study.J Rehabil Med2009;41:90412.
  • 21
    Davis MC, Okun MA, Kruszewski D, Zautra AJ, Tennen H.Sex differences in the relations of positive and negative daily events and fatigue in adults with rheumatoid arthritis.J Pain2010;11:133847.
  • 22
    Stone AA, Broderick JE, Porter LS, Kaell AT.The experience of rheumatoid arthritis pain and fatigue: examining momentary reports and correlates over one week.Arthritis Care Res1997;10:18593.
  • 23
    Contreras-Yanez I, Cabiedes J, Villa AR, Rull-Gabayet M, Pascual-Ramos V.Persistence on therapy is a major determinant of patient-, physician- and laboratory-reported outcomes in recent-onset rheumatoid arthritis patients.Clin Exp Rheumatol2010;28:74851.
  • 24
    Stebbings S, Herbison P, Doyle TC, Treharne GJ, Highton J.A comparison of fatigue correlates in rheumatoid arthritis and osteoarthritis: disparity in associations with disability, anxiety and sleep disturbance.Rheumatology (Oxford)2010;49:3617.
  • 25
    Treharne GJ, Lyons AC, Hale ED, Goodchild CE, Booth DA, Kitas GD.Predictors of fatigue over 1 year among people with rheumatoid arthritis.Psychol Health Med2008;13:494504.
  • 26
    Dhir V, Lawrence A, Aggarwal A, Misra R.Fibromyalgia is common and adversely affects pain and fatigue perception in North Indian patients with rheumatoid arthritis.J Rheumatol2009;36:24438.
  • 27
    Bergman MJ, Shahouri SS, Shaver TS, Anderson JD, Weidensaul DN, Busch RE, et alIs fatigue an inflammatory variable in rheumatoid arthritis (RA)? Analyses of fatigue in RA, osteoarthritis, and fibromyalgia.J Rheumatol2009;36:278894.
  • 28
    Repping-Wuts H, Fransen J, van Achterberg T, Bleijenberg G, van Riel P.Persistent severe fatigue in patients with rheumatoid arthritis.J Clin Nurs2007;16:37783.
  • 29
    Fifield J, McQuillan J, Tennen H, Sheehan J, Reisine S, Hesselbrock V, et al.History of affective disorder and the temporal trajectory of fatigue in rheumatoid arthritis.Ann Behav Med2001;23:3441.
  • 30
    Brekke M, Hjortdahl P, Kvien TK.Self-efficacy and health status in rheumatoid arthritis: a two-year longitudinal observational study.Rheumatology (Oxford)2001;40:38792.
  • 31
    Scharloo M, Kaptein AA, Weinman JA, Hazes JM, Breedveld FC, Rooijmans HG.Predicting functional status in patients with rheumatoid arthritis.J Rheumatol1999;26:168693.
  • 32
    Mancuso CA, Rincon M, Sayles W, Paget SA.Psychosocial variables and fatigue: a longitudinal study comparing individuals with rheumatoid arthritis and healthy controls.J Rheumatol2006;33:1496502.
  • 33
    Jump RL, Fifield J, Tennen H, Reisine S, Giuliano AJ.History of affective disorder and the experience of fatigue in rheumatoid arthritis.Arthritis Rheum2004;51:23945.
  • 34
    Finan PH, Okun MA, Kruszewski D, Davis MC, Zautra AJ, Tennen H.Interplay of concurrent positive and negative interpersonal events in the prediction of daily negative affect and fatigue for rheumatoid arthritis patients.Health Psychol2010;29:42937.
  • 35
    Parrish BP, Zautra AJ, Davis MC.The role of positive and negative interpersonal events on daily fatigue in women with fibromyalgia, rheumatoid arthritis, and osteoarthritis.Health Psychol2008;27:694702.
  • 36
    Odegard S, Kvien TK, Uhlig T.Incidence of clinically important 10-year health status and disease activity levels in population-based cohorts with rheumatoid arthritis.J Rheumatol2008;35:5460.
  • 37
    Sokka T, Toloza S, Cutolo M, Kautiainen H, Makinen H, Gogus F, et al.Women, men, and rheumatoid arthritis: analyses of disease activity, disease characteristics, and treatments in the QUEST-RA Study.Arthritis Res Ther2009;11:R7.
  • 38
    Chorus AM, Miedema HS, Boonen A, van der Linden SJ.Quality of life and work in patients with rheumatoid arthritis and ankylosing spondylitis of working age.Ann Rheum Dis2003;62:117884.
  • 39
    Wolfe F, Michaud K.Prevalence, risk, and risk factors for oral and ocular dryness with particular emphasis on rheumatoid arthritis.J Rheumatol2008;35:102330.
  • 40
    Yazici Y, Pincus T, Kautiainen H, Sokka T.Morning stiffness in patients with early rheumatoid arthritis is associated more strongly with functional disability than with joint swelling and erythrocyte sedimentation rate.J Rheumatol2004;31:17236.
  • 41
    Breedveld FC, Han C, Bala M, van der Heijde D, Baker D, Kavanaugh AF, et al.Association between baseline radiographic damage and improvement in physical function after treatment of patients with rheumatoid arthritis.Ann Rheum Dis2005;64:525.
  • 42
    Gronning K, Rodevand E, Steinsbekk A.Paid work is associated with improved health-related quality of life in patients with rheumatoid arthritis.Clin Rheumatol2010;29:131722.
  • 43
    Rupp I, Boshuizen HC, Jacobi CE, Dinant HJ, van den Bos GA.Impact of fatigue on health-related quality of life in rheumatoid arthritis.Arthritis Rheum2004;51:57885.
  • 44
    Wolfe F, Michaud K.Proposed metrics for the determination of rheumatoid arthritis outcome and treatment success and failure.J Rheumatol2009;36:2733.
  • 45
    Katz PP.The stresses of rheumatoid arthritis: appraisals of perceived impact and coping efficacy.Arthritis Care Res1998;11:922.
  • 46
    Wolfe F, Michaud K, Pincus T.Fatigue, rheumatoid arthritis, and anti-tumor necrosis factor therapy: an investigation in 24,831 patients.J Rheumatol2004;31:211520.
  • 47
    Covic T, Tyson G, Spencer D, Howe G.Depression in rheumatoid arthritis patients: demographic, clinical, and psychological predictors.J Psychosom Res2006;60:46976.
  • 48
    Wolfe F, Michaud K.Predicting depression in rheumatoid arthritis: the signal importance of pain extent and fatigue, and comorbidity.Arthritis Rheum2009;61:66773.
  • 49
    Smedstad LM, Moum T, Vaglum P, Kvien TK.The impact of early rheumatoid arthritis on psychological distress: a comparison between 238 patients with RA and 116 matched controls.Scand J Rheumatol1996;25:37782.
  • 50
    De Croon EM, Sluiter JK, Nijssen TF, Kammeijer M, Dijkmans BA, Lankhorst GJ, et al.Work ability of Dutch employees with rheumatoid arthritis.Scand J Rheumatol2005;34:27783.
  • 51
    Mancuso CA, Rincon M, Sayles W, Paget SA.Longitudinal study of negative workplace events among employed rheumatoid arthritis patients and healthy controls.Arthritis Rheum2005;53:95864.
  • 52
    Davis MC, Affleck G, Zautra AJ, Tennen H.Daily interpersonal events in pain patients: applying action theory to chronic illness.J Clin Psychol2006;62:1097113.
  • 53
    Parker White C, White MB, Fox MA.Maternal fatigue and its relationship to the caregiving environment.Fam Syst Health2009;27:32545.
  • 54
    Waltz M.The disease process and utilization of health services in rheumatoid arthritis: the relative contributions of various markers of disease severity in explaining consumption patterns.Arthritis Care Res2000;13:7488.
  • 55
    Hewlett S, Dures E, Almeida C.Measures of fatigue: Bristol Rheumatoid Arthritis Fatigue Multi-Dimensional Questionnaire (BRAF MDQ), Bristol Rheumatoid Arthritis Fatigue Numerical Rating Scales (BRAF NRS) for Severity, Effect, and Coping, Chalder Fatigue Questionnaire (CFQ), Checklist Individual Strength (CIS20R and CIS8R), Fatigue Severity Scale (FSS), Functional Assessment Chronic Illness Therapy (Fatigue) (FACIT-F), Multi-Dimensional Assessment of Fatigue (MAF), Multi-Dimensional Fatigue Inventory (MFI), Pediatric Quality of Life (PedsQL) Multi-Dimensional Fatigue Scale, Profile of Fatigue (ProF), Short Form 36 vitality subscale (SF-36 VT), and visual analog scales (VAS).Arthritis Care Res (Hoboken)2011;63 Suppl:S26386.
  • 56
    Nikolaus S, Bode C, Taal E, van de Laar MA.Which dimensions of fatigue should be measured in patients with rheumatoid arthritis? A Delphi study.Musculoskeletal Care2012;10:137.
  • 57
    Petticrew M, Roberts H. Exploring heterogeneity and publication bias. In: Petticrew M, Roberts H, editors.Systematic reviews in the social sciences: a practical guide.Malden (MA):Blackwell Publishing;2006. p.21546.

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgments
  9. REFERENCES
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ACR_21949_sm_SupplApp.doc24KSupplementary Data

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.