SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Objective

Fatigue and pain are important symptoms for patients with rheumatoid arthritis (RA), but their temporal association is unknown. Therefore, the objective of this study was to investigate the longitudinal relationship between fatigue and pain in patients with RA using time-lag models.

Methods

Consecutive RA outpatients (n = 228) were enrolled for this 1-year study. Fatigue was assessed monthly with the Checklist Individual Strength (CIS; range 8–56) and pain was assessed monthly with the bodily pain subscale (inverted, range 0–100) of the Short Form 36. The association between monthly changes in fatigue and pain was analyzed using longitudinal regression (mixed models), using the same months and with a 1-month time lag.

Results

A total of 198 patients were included in the analyses. At baseline, the mean ± SD pain score was 35.23 ± 19.82 and the mean ± SD CIS fatigue score was 31.0 ± 12.4. Severe fatigue at baseline (CIS score ≥35) was present in 42% of the patients. The mean ± SD patient-averaged CIS fatigue score over 1 year was 30.9 ± 6.0 and the mean ± SD patient-averaged pain score over 1 year was 36.4 ± 18.3. The longitudinal regression analysis showed a significant positive relationship between fatigue and pain during the same month (β = 2.04; 95% confidence interval 1.82, 2.27). The models using a time lag showed no significant association between changes in pain and changes in fatigue.

Conclusion

In established RA, pain and fatigue show monthly fluctuations that are synchronous rather than showing a temporal relationship with a time lag; within this timeframe, the results do not indicate that one precedes the other.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Rheumatoid arthritis (RA) is a chronic autoimmune disorder causing inflammation, stiffness, and pain in the joints [1]. In RA, fatigue is a frequently occurring and patient-relevant symptom that often is experienced as debilitating and restricting daily functioning [2]. In cross-sectional studies, it is found that at least 40% of the patients with RA have severe fatigue [3, 4]. Pain and fatigue are the symptoms mentioned by RA patients as the most disturbing symptoms of the disease [5-7].

RA fatigue can be influenced by numerous factors, such as inflammation, pain, disability, and psychosocial factors (mood, beliefs, behavior) [8-10]. Although chronic inflammation may cause fatigue, in RA it has been shown that pain, rather than inflammation, is associated with fatigue severity [4, 7-9, 11-15]. Fatigue and pain often co-occur, and there are previous mainly cross-sectional studies showing that more fatigue is strongly associated with more pain [4, 8, 9, 16]. Few studies have measured fatigue over 1 year in RA [3, 15]. It was shown that severe fatigue was experienced in 50% of the RA patients, both at baseline and at 12-month followup [3]. Also in RA, a higher pain score at baseline predicted worse fatigue 1 year later [15].

The association of pain and fatigue could be synchronous or with a time lag, meaning that a change in pain could be associated with a change in fatigue at the same time (synchronous), or that a change in pain may precede a change in fatigue or a change in fatigue may precede a change in pain.

Temporal associations between pain and fatigue in RA might be day to day (e.g., pain today associated with fatigue tomorrow) or month to month [17-19]. Currently, it is unclear whether a change in pain precedes a change in fatigue or whether a change in fatigue precedes a change in pain. It could be hypothesized both that pain can lead to fatigue because of the energy consumed by prolonged pain suppression and the need to deal with pain, but also that fatigue can lead to pain because of being less able to suppress and deal with pain [20, 21]. Therefore, the question is whether increased levels of pain are followed by an increase in fatigue or whether it is vice versa, i.e., increased levels of fatigue are followed by an increase in pain. It also may be that changes in fatigue and pain tend to fluctuate together and do not show such a temporal association. Therefore, the objective of this study was to investigate the longitudinal relationship between fatigue and pain in patients with RA using time-lag models.

Significance & Innovations

  • Fatigue and pain are important symptoms for patients with rheumatoid arthritis (RA), but their temporal association is unknown. It may be that pain precedes fatigue, but fatigue may also precede pain.
  • This was the first study that investigated the longitudinal relationship between monthly assessed fatigue and pain in RA patients over 1 year, using time-lag models.
  • In established RA, pain and fatigue show monthly fluctuations that are synchronous rather than showing a temporal relationship with a time lag; the results do not indicate that one precedes the other.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Design

This is a prospective cohort study of 1 year with monthly repeated measures of both fatigue and pain in consecutive patients with established RA. Over the course of 2 weeks, daily assessments of pain and fatigue were performed. Approval for this study was obtained from the ethical committee (CMO Arnhem-Nijmegen, The Netherlands).

Patients

A total of 230 RA patients attending the outpatient clinic of the Radboud University Nijmegen Medical Centre were approached between June 2006 and October 2007 to participate in this study. Patients were informed in writing and orally by their rheumatologist or a nurse specialist and invited to participate in the study.

Inclusion criteria consisted of a diagnosis of RA according to the 1987 American College of Rheumatology classification criteria [22], being age 18–75 years, and being able to read and write in the Dutch language. Study participation was allowed with comorbidities such as secondary Sjögren's syndrome, regulated thyroid disease (values of free T4 of minimally 8 pmoles/liter and thyroid-stimulating hormone, maximum 1.0 units/liter), regulated diabetes mellitus (normalized glucose values between 2.5 and 3.7 mmoles/liter in urine and between 4.0 and 5.6 mmoles/liter in blood and glycosylated hemoglobin <8.0%), mild nonrestrictive chronic obstructive pulmonary disease, and successfully treated not metastasized basal cell carcinoma or squamous cell carcinoma in the skin in medical history. Patients were excluded from study participation if they had a second rheumatic disease (except for secondary Sjögren's syndrome), a history of malignancies or other comorbidities associated with chronic fatigue, a current diagnosis of depression, or current psychological or psychiatric treatment.

Data collection

Patient characteristics (sex and age), disease characteristics (disease duration and rheumatoid factor), and medication use were collected at inclusion (baseline) by research nurses. A blood sample was taken (for erythrocyte sedimentation rate, C-reactive protein level, and hemoglobin level) and disease activity was assessed by the rheumatologist or a research nurse using the Disease Activity Score in 28 joints (DAS28). Restrictions of daily functioning were assessed using the Health Assessment Questionnaire (HAQ) disability index [23].

Fatigue and pain were self-assessed every month for 12 consecutive months. Fatigue severity was measured using the fatigue severity subscale of the 20-item Checklist Individual Strength (CIS20), which also contains subscales of physical activity, motivation, and concentration [24]. The CIS fatigue consists of 8 items on fatigue symptom severity regarding the last 2 weeks and all items are scored on a 7-point Likert scale (range 8–56). Higher scores on the CIS fatigue indicate a higher level of fatigue experienced; a score of <27 is considered normal and a score of ≥35 indicates severe fatigue [24]. The CIS20 has proven to be a reliable and valid instrument in numerous conditions and was also used in RA [24, 25].

Pain was assessed with the bodily pain subscale of the Short Form 36 health survey (SF-36), which asks about pain experienced in the last 2 weeks [26]. The SF-36 bodily pain subscale consists of 2 items, one regarding pain level and one asking about the impact of pain on daily life. Final scores range from 0–100, with higher scores indicating less pain. For the purpose of the analyses, the SF-36 bodily pain scoring was inverted so that higher scores indicate more pain, whereas higher CIS scores indicated more fatigue. The SF-36 was adapted to cover a retrospective timeframe of 2 weeks, in order to make the interval similar to the interval of the CIS fatigue. At baseline and 12 months, pain was also assessed using a numerical rating scale.

During 2 weeks in the first month, the patients completed daily self-assessments of pain and fatigue using a self-observation list, with Likert scales asking about today [27].

The Beck Depression Inventory for primary care (BDI-PC) was used to classify patients for depression. The BDI-PC is a 7-item self-report instrument (range 0–21); a total score of ≥4 is suggestive of depression [28].

Statistical analyses

To assess whether fatigue changed over time on the group level, the course of fatigue over 1 year was analyzed graphically and using a longitudinal regression model (mixed model) correcting for repeated measures within patients, with CIS fatigue as the dependent variable and time as the independent variable. Next, 1-month changes in individual fatigue and individual pain scores were calculated over consecutive months. A scatter plot was made showing the individual monthly changes in pain and fatigue. Next, Pearson's correlations were used to analyze the correlations between monthly changes in pain and monthly changes in fatigue, at the same month and with a 1-month time lag. Finally, 3 longitudinal regression models (mixed models) were used to analyze the relationships between change in pain level and change in fatigue level over time, corrected for repeated measurements within the same patient. It was analyzed whether a change in pain was associated with a change in fatigue over the same month (Figure 1, model 1), or with a change in fatigue 1 month later (Figure 1, model 2), or with a change in fatigue 1 month earlier (Figure 1, model 3).

image

Figure 1. Schematic representation of the models used to analyze the longitudinal association between fatigue and pain.

Download figure to PowerPoint

Age, sex, disease duration, rheumatoid factor positivity, HAQ score, and BDI-PC score were considered as possible confounders. The monthly pain and fatigue absolute scores at the beginning of the monthly differences in fatigue were entered into the mixed model as covariates.

We also tested whether the associations differed by sex and age (effect modification). The assumptions of the linear mixed-model analysis were checked by testing the linear relationship between the difference in CIS fatigue and the predicted values and graphically using a scatter plot of the predicted values versus the residuals. Fatigue was the dependent variable in all 3 models to facilitate comparison of regression coefficients between the 3 models (Figure 1).

These analyses were repeated using the 2-week data set with daily changes in pain and fatigue. Data analysis was performed using the SAS system, version 9.20.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Patient characteristics

Two patients were excluded after the first measurement because of newly acquired sleep apnea and a malignant lung tumor. A total of 228 patients were included and 198 patients who filled in at least 10 monthly CIS fatigue and pain questionnaires were included in the analyses. The 30 patients who were not included in the analysis were not different in baseline characteristics from the 198 patients included in the analysis (data not shown). The included patients were mostly middle aged and the majority were women and rheumatoid factor positive (Table 1). Most patients had established disease, low levels of disease activity (DAS28 <3.2), and low levels of disability (HAQ score). Clinical depression (BDI-PC score ≥4) seldom occurred. At baseline, the mean pain score or pain severity was moderate and the mean level of fatigue was higher than normal (CIS fatigue score <27), but lower than severe (CIS fatigue score ≥35) [24]. Severe fatigue at baseline (CIS fatigue score ≥35) was experienced in 40% of the patients. There were no large differences between baseline and followup in any of the variables.

Table 1. Patient characteristics at baseline and 12 months (n = 198)a
 Baseline12 months
  1. a

    Values are the median (25th, 75th percentiles) unless otherwise indicated. DAS28 = Disease Activity Score in 28 joints; VAS = visual analog scale; SJC28 = 28 swollen joint count; TJC28 = 28 tender joint count; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; NRS = numerical rating scale; SF-36 = 36-item Short Form health survey; CIS = Checklist Individual Strength; HAQ = Health Assessment Questionnaire; DI = disability index; BDI-PC = Beck Depression Inventory for primary care.

  2. b

    CRP levels <5 mg/liter are scored as 0.

  3. c

    Scored on an inverted scale (where a higher score indicates less pain).

Age, mean ± SD years56.7 ± 10.6 
Sex, no. (%) women126 (64) 
Disease duration, years10 (6, 17) 
Rheumatoid factor positivity, no. (%)153 (77) 
DAS28, mean ± SD3.16 ± 1.243.09 ± 1.21
VAS general health30 (15, 46)25 (15, 50)
SJC283 (1, 6)4 (1, 6)
TJC282 (0, 4)1 (0, 4)
ESR, mm/hour8 (4, 18)10 (4, 18)
CRP level, mg/literb0 (0, 8)0 (0, 8)
Pain: NRS pain severity (range 0–10), mean ± SD4.29 ± 2.463.86 ± 2.55
Pain: SF-36 bodily pain (range 0–10), mean ± SDc35.23 ± 19.8234.52 ± 21.34
CIS fatigue (range 8–56), mean ± SD31.0 ± 12.430.0 ± 12.5
CIS fatigue, no. (%)  
Normal fatigue (<27)78 (39.4)87 (43.9)
Heightened fatigue (27–34)41 (20.7)35 (17.7)
Severe fatigue (≥35)79 (39.9)71 (35.9)
HAQ DI score0.63 (0.13, 1.13)0.63 (0.20, 1.13)
Hemoglobin level, mean ± SD mmoles/liter8.21 ± 0.728.15 ± 0.78
BDI-PC score ≥4, no. (%)10 (5)13 (6.6)

Medication

At baseline, 82 (41.4%) of 198 patients received disease-modifying antirheumatic drug (DMARD) monotherapy, most often with methotrexate (n = 50), sulfasalazine (n = 17), and azathioprine (n = 5). Twenty-three (12%) of 198 patients received DMARD combination therapy, usually with methotrexate.

At baseline, 70 patients (35.4%) received a biologic agent and all were receiving tumor necrosis factor–inhibiting agents, either as monotherapy or in combination with a DMARD. Eight patients (4%) stopped a biologic agent during the study and 12 (6%) started a biologic agent during the study. Thereby, 25 (13%) of 198 patients received oral prednisone. Medication use was missing in 23 patients.

Fatigue and pain over time

The monthly CIS fatigue scores are shown for those patients who were severely fatigued at baseline, and for those with heightened and normal fatigue levels [24] (Figure 2A). The mean CIS fatigue score for the total group dropped a little within the 1-year period (Table 1); in the 3 subgroups, the mean fatigue scores remained stable over time. According to the linear mixed-model analyses, on average there was no change in the level of fatigue over time (P = 0.80) or in the level of pain over time (P = 0.10) (Figures 2A and B). Figure 2B shows the course of both the monthly CIS fatigue and pain scores of the total sample over 1 year.

image

Figure 2. A, Course of fatigue in patients with no, mild, and severe fatigue. Mean Checklist Individual Strength (CIS) fatigue scores over 12 months of the total group (n = 198; ——), patients who were severely fatigued at baseline (n = 96; – – –), and patients who were not severely fatigued at baseline (n = 102; - - -) are shown. B, The course of fatigue and pain over 1 year.

Download figure to PowerPoint

Monthly changes in pain and fatigue: simple correlations

The “naive” association between all individual changes in pain and changes in fatigue, without correction for repeated measurements, is shown as a scatter plot in Figure 3. Each dot indicates a single time point with a monthly change in pain and a monthly change in fatigue, and each patient contributes up to 12 dots. From the graph it appears that there are fluctuations in pain and fatigue and that a change in fatigue is positively associated with a change in pain. On average, there was no change in pain or fatigue; the mean ± SD monthly difference in pain was −0.022 ± 16.81 (P = 0.95) and the mean ± SD monthly difference in fatigue was −0.14 ± 8.12 (P = 0.41).

image

Figure 3. Scatterplot indicating the individual monthly changes in pain and changes in fatigue.

Download figure to PowerPoint

It can be seen in Table 2 that changes in pain and changes in fatigue in the same month have a positive correlation (r = 0.42), i.e., an increase or decrease in fatigue goes along with a change (increase or decrease) in pain in the same direction. It can also be seen that a change in pain in 1 month is negatively associated with a change in pain the next month (r = −0.41), i.e., an increase in pain is associated with a decrease in pain the next month. The same is found for fatigue (r = −0.39). The same kind of correlations are found for changes in fatigue preceding changes in pain (r = −0.21) and changes in pain preceding changes in fatigue (r = −0.15).

Table 2. Association between fatigue and pain with Pearson's correlation coefficientsa
 Δpain, same monthΔfatigue, 1 month earlierΔpain, 1 month earlier
r95% CIr95% CIr95% CI
  1. a

    Naive correlations between monthly changes in fatigue and pain scores over 12 months. P < 0.0001 for all. 95% CI = 95% confidence interval.

Δfatigue, same month0.420.38, 0.45−0.39−0.43, −0.35−0.15−0.19, −0.11
Δpain, same month−0.21−0.25, −0.17−0.41−0.44, −0.37

Monthly changes in pain and fatigue: longitudinal regression

For all regression models, adjustments were made for age, sex, and HAQ score, while monthly pain and fatigue scores were added as covariates. There were no large differences between the crude and the adjusted models, and only the results of the adjusted models are shown (Table 3).

Table 3. Association between fatigue and pain over time with linear mixed-model analysisa
 β (95% CI)P
  1. a

    A linear mixed-model analysis was performed, adjusted for age and sex, Health Assessment Questionnaire score, and the monthly pain and fatigue absolute scores at the beginning of the monthly differences in fatigue. 95% CI = 95% confidence interval.

Model 1: Δfatigue = Δpain in same month2.00 (1.77, 2.21)< 0.0001
Model 2: Δfatigue = Δpain 1 month earlier0.12 (−0.12, 0.36)0.32
Model 3: Δfatigue = Δpain 1 month later−0.02 (−0.25, 0.20)0.83

The results of the first model, representing the association between change in pain and change in fatigue during the same month, showed a significant relationship (P < 0.0001). The positive beta coefficient (β = 2.00) indicated that more pain was associated with more fatigue during the same month. The second model, analyzing the longitudinal association between change in pain level in the preceding month and change in fatigue 1 month later, indicated that a change in fatigue level was not related (β = 0.12, P = 0.32) to a change in pain level 1 month earlier. This indicates that an increase in pain level was not associated with an increase in fatigue level 1 month later. Model 3, analyzing the longitudinal association between change in fatigue level in the preceding month and change in pain level 1 month later, also indicated that a change in fatigue level is not related (β = −0.02, P = 0.83) to a change in pain level 1 month later. This indicates that an increase in fatigue level is not associated with an increase in pain level 1 month later. When similarly analyzing the longitudinal association between change in daily pain and fatigue scores, the same results were found (data not shown). The associations between change in pain score and change in fatigue score were not significantly different for men and women, nor did they vary with age.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

This longitudinal study is the first study to examine the course of fatigue and pain in patients with RA in a period of 1 year with monthly measurements of fatigue and pain. The aim was to investigate whether changes in pain precede changes in fatigue, or vice versa, or whether pain and fatigue fluctuate together in time. According to the results of this study, pain and fatigue showed monthly fluctuations that were synchronous rather than showing a temporal relationship with a time lag. Within the timeframe of 1 year and monthly assessments, as well as daily assessments in 2 weeks, the results do not indicate that one precedes the other. There also was no indication that the results would differ for age and sex. The results showed that pain and fatigue scores were quite stable over 1 year. However, within the patients there was considerable monthly fluctuation in pain scores and in fatigue scores that was synchronous rather than showing a temporal relationship with a time lag of 1 month. This was shown by the naive correlations of change that, however, were not controlled for confounders and the absolute scores at the baseline of each change. One of the consequences of this correction using the linear mixed models is that the time-lagged effects of the naive correlations “disappeared.” Using the linear mixed-model analysis adjusted for age and sex, HAQ score, and monthly pain and fatigue scores at the beginning of the monthly differences in fatigue, the strongest association was found in the correlations and the regression model, reflecting synchronous changes in pain and fatigue. According to the regression coefficient, a 1-point increase in pain score (range 0–100) was associated with a 2-point increase in fatigue score (range 8–56). Until now, the minimally important difference for the CIS fatigue score is not formally known, but a comparison with the health transition question of the SF-36 using our own data suggested that a minimally important difference may be +5 for improvement and −3 for worsening (data not shown). In a trial of cognitive–behavioral therapy for RA, the mean decrease in CIS fatigue score was 5 in the intervention group and 2 in the control group [29]. Therefore, the size of the relationship between pain and fatigue appears to be relevant.

The models using a time lag showed no significant association between changes in pain and changes in fatigue. With this timeframe of 1 month, it cannot be said that one precedes or “causes” the other. To inform about probable causality, it would be informative if a change in pain precedes a change in fatigue (cause precedes effect) or vice versa. The time scale of such a temporal relationship between pain and fatigue is unclear and might be day to day or month to month [17-19]. Since we were primarily interested in the course of fatigue over 1 year, we chose to assess fatigue and pain every month. The question, however, is whether monthly changes probably are too long. In a period of 2 weeks shortly after baseline, patients filled in a diary with daily pain and fatigue scores. Therefore, we were also able to analyze a temporal relationship between pain and fatigue on a day-to-day basis. In the end, the results of the monthly analyses and the daily analyses were the same, i.e., a strong association of pain and fatigue at the same time point, with no recognizable time lag. We identified several studies that looked into the relationship between pain and fatigue using daily measurements in RA and in fibromyalgia (FM) [17, 30-32]. It was found that patients with RA showed much variability in pain and fatigue levels within days, whereas there were no differences in pain and fatigue levels between days [17]. It was suggested that the pattern of pain and fatigue was not explained by mood cycles [17]. However, in another study of RA patients and consecutive daily fatigue assessments, it was found that days with more frequent positive events were related to lower levels of same-day fatigue and higher levels of next-day fatigue in women, but not for men [18]. In a study of patients with RA, it appeared that diurnal fluctuations in fatigue were independent of the circadian rhythm of cortisol or inflammatory activity, but rather reflect temporal changes as a consequence of sleep, rest, and physical activity throughout the day [17, 30]. In FM, it appeared that there was a diurnal relationship between pain and fatigue that probably was mediated by stress or sleep quality [31, 32]. It may be worthwhile to evaluate sex differences; for example, it has been found that especially younger women with multiple daily roles seemed to be vulnerable to the negative impact of RA fatigue [33]. However, in our study, no difference between men and women was found. To our knowledge, no previous studies have analyzed the longitudinal relationship between fatigue and pain in RA patients. The advantage of longitudinal analysis is that the individual development of both fatigue and pain in time can be investigated. There is one previous study, not in RA, in which the temporal relationship between pain and fatigue among primary care patients presenting with main symptoms of fatigue was analyzed [34]. In this observational cohort study, pain and fatigue were measured at 1, 4, 8, and 12 months after baseline. The longitudinal associations were analyzed with 3 different models that were similar to those models used in the current study. The results indicated that changes in pain and fatigue are directly related in time, rather than showing a time lag in their relationship. This means that the findings of Nijrolder et al in the general care population are similar to the results of the current study [34].

A strength of our study is the monthly measurement of pain and fatigue during 1 year. By measuring pain and fatigue with relatively short time intervals, we consider that we had a reasonable precise measurement of the courses of pain and fatigue to analyze their longitudinal association. The recall period of fatigue and pain was 2 weeks because a recall period of 4 weeks is considered relatively long for patients to remember their fatigue and pain levels. The bodily pain subscale of the SF-36 was modified to assess pain experienced in the last 2 weeks instead of the last 4 weeks to compare it with the CIS20, in which fatigue severity was assessed for the last 2 weeks. Another strength is few loss of data; 198 of the 228 patients filled in at least 10 monthly CIS fatigue and pain questionnaires. It was hypothesized that patients with a well-controlled comorbidity, such as regulated diabetes mellitus or regulated thyroid disease, would not experience extra fatigue from this underlying condition. Patients with multiple rheumatic diseases were not included; secondary Sjögren's syndrome was allowed and regarded as an extraarticular manifestation of RA.

A limitation of this study concerns the observational nature leading to a risk of confounding. However, in the adjusted model, we corrected for the most important confounders. No other confounders were identified. The patients were consecutively included at regular visits, not at indication of having high pain levels or high fatigue severity. Therefore, large changes in fatigue and pain after baseline were not found; on the other hand, the sample is representative for the RA population regarding levels of pain and fatigue. Another important consideration is the different scaling for measuring pain and fatigue. The SF-36 bodily pain subscale was used to assess pain instead of a single pain visual analog scale because the subscale is composed of 2 items, which may be more valid than 1 item. However, these 2 items have a limited number of response options, which may make it difficult to detect smaller changes in pain. One of the pain items considers the impact of pain on daily life, but it may not confound a relationship of pain with fatigue because the CIS fatigue subscale is about symptom severity, and impact on daily life is assessed using other CIS subscales.

Seeing the important meaning of pain and fatigue for patients with RA, reductions of pain and fatigue are recommendable goals for RA management [10]. Further research is recommended in factors that can cause or perpetuate fatigue in RA. Although pain could probably “cause” fatigue or fatigue could probably “cause” pain, probably both pain and fatigue are driven by common factors. Common factors maybe psychological factors, which could be investigated by a multidimensional model of fatigue using structural equation modeling.

In summary, the results of this 12-month longitudinal study of the monthly temporal relationship between pain and fatigue show that in established RA, pain and fatigue have a synchronous association with a fluctuating pattern around an individual mean, rather than showing temporal associations. There is no indication that one precedes or causes the other, regardless of the time scale being days or months. The clinical implication is that both manifestations should be treated because it cannot be expected that an improvement in one is followed by an improvement in the other.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

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. Ms van Dartel 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. Van Dartel, Repping-Wuts, Bleijenberg, van Riel, Fransen.

Acquisition of data. Van Dartel, van Hoogmoed.

Analysis and interpretation of data. Van Dartel, Repping-Wuts, van Hoogmoed, Bleijenberg, van Riel, Fransen.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES