Fatigue and plasma cytokine concentrations at rest and during exercise in patients with sarcoidosis
Dr. Baydur designed the study, recruited subjects, monitored some of the exercise testing, centrifuged some of the plasma samples for subsequent assay, and wrote the paper. Dr. Alavy recruited study subjects, scheduled some of the pulmonary function testing and and monitored some of the exercise tests, and centrifuged some of the plasma samples for subsequent assay. Dr. Nawathe recruited study subjects, scheduled some of the pulmonary function and and monitored some of the exercise tests, and centrifuged some of the plasma samples for subsequent assay. Dr. Liu performed the high-sensitivity cytokine assays and statistical analyses to detect differences amongst cohorts and correlation analyses. Dr. Louie provided advice and guidance for the cytokine assays and provided the kits used to perform the assays. Dr. Sharma conceived of the idea for the study – namely to attempt to seek a relationship between fatigue in patients with sarcoidosis and plasma cytokine levels.
The study was approved by our institutional review board, and all subjects read and signed an informed consent document.
Conflict of interest
The authors have stated explicitly that there are no conflicts of interest in connection with this article
Ahmet Baydur, MD, Division of Pulmonary and Critical Care Medicine, Keck School of Medicine, University of Southern California, 2020 Zonal Avenue, IRD 723
Los Angeles, CA 90033, USA.
Tel: +323 226 7923
Fax: +323 226 2738
Background: Patients with sarcoidosis exhibit exercise intolerance-related fatigue and increased levels of circulating proinflammatory cytokines at rest. Exercise may result in increased plasma cytokine levels (PCLs) in healthy adults, but such a relationship has not been studied in sarcoidosis patients.
Objectives: To assess relationship of fatigue in sarcoidosis with PCLs at rest and with cardiopulmonary exercise testing (CPET).
Methods: We assessed lung function, CPET data, multidimensional fatigue inventory, plasma tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) concentrations before, immediately after, and 4–6 h following CPET in 22 sarcoidosis patients (13 receiving immunomodulatory drugs) and 22 controls.
Results: Patients exhibited greater fatigue, reduced cardiorespiratory function, higher Medical Research Council (MRC) scores and higher plasma TNF-α concentrations than controls at all times. Plasma IL-1β levels did not differ between cohorts. Patients exhibited a 28% increase (statistically not significant) in TNF-α level immediately post exercise. Plasma IL-β concentrations did not change among cohorts. Treated patients exhibited higher MRC and physical fatigue scores and lower breathing reserve, but no differences in cardiorespiratory function or PCLs compared to untreated patients. In treated patients, pre-exercise plasma IL-1β correlated with physical fatigue, reduced motivation and total fatigue; TNF-α levels only correlated with general fatigue score.
Conclusion: Treated sarcoidosis patients exhibit a relation between physical fatigue, reduced motivation and total fatigue and pre-exercise plasma IL-1β concentrations. Acute exercise does not increase PCLs. Whether the reduced MRC score and physical fatigue in treated patients is related to the therapy or to the underlying inflammatory process is difficult to determine.
Please cite this paper as: Baydur A, Alavy B, Nawathe A, Liu S, Louie S and Sharma OP. Fatigue and plasma cytokine concentrations at rest and during exercise in patients with sarcoidosis. Clin Respir J 2011; 5: 156–164.
Fatigue in sarcoidosis is frequent and incompletely understood. Clinically, it correlates poorly with traditional objective measurements of sarcoidosis including chest radiography and lung function assessment. The presence of fatigue and its effects on quality of life (QOL) can be gauged by using instruments like the World Health Organization Quality of Life-100 (WHOQOL-100) questionnaire (1), the Fatigue Assessment Scale (2) and the Multidimensional Fatigue Inventory (MFI) (3). Tissue cytokine profiles demonstrate an increase in CD4 Th1-mediated inflammatory cytokines such as tumor necrosis factor-α (TNF-α) (4) and interleukin-1β (IL-1β) (5) where both can induce granuloma formation. Furthermore, case reports of sarcoidosis patients treated with anti-TNF-α agents show reduction in fatigue and improvement in QOL (6). Clinical trials with infliximab [using the short form 36 health survey questionnaire (SF-36) instrument](7) and etanercept (using the St George's respiratory questionnaire, SGRQ) (8), however, have not been able to demonstrate improvement in QOL over time or vs placebo.
Fatigue and general weakness may be the reason why patients with sarcoidosis frequently experience exercise intolerance (9), a symptom also noted in other chronic inflammatory states, such as chronic fatigue syndrome (CFS) (10, 11). Some investigators have hypothesized that cytokines may mediate certain of the symptoms and immunologic disturbances associated with inflammatory conditions (10, 11). Exercise is known to increase plasma TNF-α and IL-1β cytokine levels in healthy subjects (12), yet there is little information on exercise modulation of previously activated inflammatory cells in chronic inflammatory disorders (10, 11, 13) and no information with respect to sarcoidosis. Since most studies of cytokines have been conducted in resting patients and physical exertion is known to exacerbate fatigue, the present study was designed to test the hypothesis that acute exercise would provoke serum cytokine abnormalities of potential pathogenic importance in sarcoidosis. We thus (i) compared plasma cytokine concentrations before and immediately after cardiopulmonary exercise testing in sarcoidosis patients and control subjects, and (ii) sought relations between plasma cytokine levels and fatigue as assessed by a multifactorial inventory instrument (3).
Materials and methods
Twenty-two patients (18 women) who had not smoked within the previous 5 years were recruited from the outpatient clinic (Table 1). The diagnosis of sarcoidosis was based on clinical and radiographic features, and was confirmed by previously obtained biopsy-proven noncaseating epithelioid granulomas. The Scadding radiographic criteria (14) were used to assess the extent of thoracic involvement with sarcoidosis. Patients with cardiac, neurologic, hepatic or arthritic involvement with sarcoidosis, hemoglobin below 10 mg/dL and hypercalcemia were excluded. Patients were compared to a control group of 22 nonsmoking subjects (16 women) free of cardiorespiratory illness. Subjects were recruited between August 2002 and February 2007.
Table 1. Anthropometric and lung function data in patients with sarcoidosis and control subjects
|Age (years)||46.4 ± 10.8||46.4 ± 9.9||0.9626|
|BMI (kg/m2)||29.6 ± 7.7||27.5 ± 6.2||0.4114|
|Chest radiograph stage 0/I/II/III/IV (n)||4/5/6/6/1||–||–|
|MRC scale||2.9 ± 1.5||1.3 ± 0.6||0.0006|
|FVC (% pred.)||83.7 ± 17.5||100.7 ± 12.3||0.0013|
|FEV1 (% pred.)||88.6 ± 22.2||109.2 ± 14.9||0.0009|
|FEV1/FVC (%)||80.3 ± 9.4||83.7 ± 5.3||0.3433|
|IC (% pred.)||88.6 ± 31.4||114.5 ± 20.6||0.0048|
|TLC (% pred.)||87.9 ± 15.2||101.0 ± 11.1||0.0038|
|RV (% pred.)||95.5 ± 35.1||100.2 ± 22.2||0.1936|
|FRC (% pred.)||83.3 ± 20.5||90.2 ± 17.9||0.1289|
|MVV (% pred.)||77.4 ± 22.4||112.5 ± 21.5||<0.0001|
|DLCO (% pred.)||80.8 ± 24.5||100.1 ± 16.2||0.0185|
|D/VA (% pred.)||108.0 ± 20.5||105.7 ± 14.0||0.6706|
Pulmonary function testing
Pulmonary function studies were obtained as part of routine evaluation and with the subjects in the seated position. At least three forced vital capacity (FVC) maneuvers were obtained using a 10-liter water displacement spirometer (Collins Medical, Inc., Braintree, MA, USA). Their predicted values were obtained from Schoenberg et al.(15). Maximum voluntary ventilation was measured directly and its reference values were derived from Grimby and Soderholm (16). Lung volumes were measured by the helium dilution technique (Collins), and their reference values were from Crapo et al.(17). The transfer factor of the lung for carbon monoxide was measured by the single breath method (DLCO) (Collins) (18). Reference values for DLCO were obtained from Cotes(18). Dyspnea was evaluated by means of the British Medical Research Council (MRC) Scale (19).
Cardiopulmonary exercise testing
Each subject underwent symptom-limited, incremental cardiopulmonary exercise testing on a treadmill using a steady state modified Bruce protocol (20). The protocol targeted a test duration of 11–15 min. All variables were recorded for 2 min at rest, throughout the test and for the first 5 min of recovery. Peripheral O2 hemoglobin saturation was monitored by pulse oximetry. Heart rates and rhythms were monitored by a 12-lead electrocardiogram. Blood pressure was measured every 2 min with a standard mercury sphygmomanometer. The subjects were encouraged to exercise until they perceived they could no longer continue. Exercise end points for every patient were leg fatigue and/or dyspnea.
Gas exchange was studied with each subject breathing through a low resistance valve with a clamped nasal passage. Expired flow, O2 consumption (V'O2) and carbon dioxide output (V'CO2) were continuously measured by a breath-by-breath technique with a Collins equipment. Peak V'O2, V'CO2 and minute ventilation (V'E) were defined as the average value obtained during 20 s of exercise. The anaerobic threshold was determined by the V-slope technique (21), and the result was confirmed by a diagram on which the respiratory equivalent for O2 (V'E/V'O2) and CO2 (V'E/V'CO2) were plotted simultaneously against time. The ventilatory response was calculated as the slope of the relation between V'E and V'CO2 from the beginning of exercise to the anaerobic threshold and was obtained by the V-method. The O2 pulse was obtained as the ratio of peak V'O2 to peak heart rate. Breathing reserve was computed as 1- (V'Emax/MVV).
To measure the level of various aspects of fatigue, the MFI-20 was used according to Smets et al.(3). We chose this instrument as it was the one fatigue assessment scale that was available at the time of initiation of this study. This questionnaire assesses five dimensions of fatigue: general, physical and mental fatigue, reduced activity and reduced motivation. General fatigue refers to fatigue expressed in terms like ‘I feel tired’ and ‘I feel rested’. Physical fatigue refers to physical sensations related to the feelings of tiredness. Reduction in activities and lack of motivation to start any activity are covered by the scales ‘reduced activities’ and ‘reduced motivation’, respectively. Cognitive symptoms such as having difficulties concentrating are included in the scale for ‘mental fatigue’. Each scale contains four items for which the person had to indicate on a 7-point scale to what extent the particular statement applies to him or her (3). An equal number of items is worded in a positive and in a negative direction to counteract response tendencies. Validity of the MFI-20 was demonstrated by results of univariate testing showing significant differences between groups for all the scales. All scales discriminated between cohorts. In addition, the scales of the MFI-20 showed good internal consistency in all samples in both cohorts.
Plasma cytokine determinations
We chose TNF-α and IL-1β for measurement as others have suggested they are the most important cytokines present in sarcoidosis (4, 5) at the time of planning of this study (2000). Immediate pre- and post-exercise venous samples were drawn in heparin-containing test tubes just prior to the beginning and end of exercise, and 4–6 h after the end of exercise testing. The blood samples were allowed to settle in ice and centrifuged at 2000 rpm at for 10 min. The plasma was separated from the cells and flash frozen at −70°C until cytokine levels were determined in undiluted form. Plasma IL-1β was measured by using Quantikine HS IL-1β immunoassay ELISA kits (R&D, Minneapolis, MN, USA). Plasma TNF-α was quantified by using Quantikine HS TNF-α immunoassay ELISA kits (R&D, Minneapolis, MN, USA). Procedures of cytokine measurements strictly follow manufacturer's suggested protocols. According to the protocol, the TNF-α concentration range for undiluted heparinized plasma is from nondetectable to 1.411 pg/mL, and for IL-1β, from nondetectable to 0.218 pg/mL. Furthermore, the standard curve in each assay had a good linear correlation and coefficacy, and the value for healthy volunteers was within the recommended range.
For variables, descriptive statistics [mean, standard deviation (SD) or median (interquartile range), as appropriate] were generated separately for the sarcoidosis and control groups. With respect to each of the study variables, the Mann–Whitney test was used to assess the significance of differences in central tendency in variables of respiratory function and cytokine concentrations between the two cohorts because of the skewed nature of many of the data distributions (22). Comparisons seeking significant difference in group means among multiple disease stages, as well as between patients receiving immunomodulatory drugs, those that were not and control subjects were performed by analysis of variance using the Kruskal–Wallis test because of non-Gaussian data distributions. Multiple linear regression tests were performed to assess associations between adjusted independent MFI-20 component scores and selected physiologic measurements and plasma cytokine levels, while F-tests were performed to assess the significance of difference between zero and the fitted regression slopes. Because treatment with immunomodulatory medications (such as corticosteroids) may lead to fatigue and weakness, linear regression was performed separately in patients receiving such drugs and those not receiving medication. All significance testings were two-tailed and were conducted using an alpha level of 0.05. Statistical analyses were performed by using PRISM 4.0 (Prism Software Co., Irvine, CA, USA).
Anthropometric, lung and exercise functional data
Using the Scadding radiographic criteria (11), stages 0, I (nonparenchymal), II, III, IV (parenchymal) were exhibited in 4, 5, 6, 6 and 1 patients, respectively. Pulmonary parenchymal involvement by sarcoidosis was confirmed in 18 of 22 patients by transbronchial biopsy. Four patients (stage 0) were diagnosed by skin biopsy, three of whom exhibited pulmonary parenchymal or hilar nodal involvement by imaging studies. Thirteen patients (59%) were receiving oral immunomodulatory drugs, singly or in combination [7 on prednisone (maximum dose 20 mg daily), 6 on methotrexate (15 mg weekly) and 3 on hydroxychloroquine] at the time of the study. On average, both groups were mildly overweight. The MRC scale in the sarcoidosis patients was more than twice that in the control subjects. Patients with sarcoidosis exhibited a number of significantly reduced functional values (Table 1). Absolute values of FVC, inspiratory capacity, total lung capacity and single breath diffusion capacity were, respectively, 25%, 27%, 22% and 23% less than the corresponding values in control subjects (all statistically significant differences).
As compared with control subjects free of cardiorespiratory illness, functional and peak exercise capacity was lower in the sarcoidosis patients (Table 2). Peak oxygen consumption (V'O2max, as mL/kg/min) in sarcoidosis patients was 23% less than in control subjects (P = 0.0042). The O2 pulse at peak exercise was normal in both cohorts, indicating adequate cardiovascular or pulmonary vascular reserve. In the sarcoidosis cohort, percent predicted values of minute ventilation, ventilatory equivalents for oxygen and carbon dioxide and physiologic dead space at peak exercise were, respectively, 17%, 29%, 19% and 26% greater (all statistically significant) than in control subjects, indicating a ventilatory impairment to exercise. Breathing reserve was 12% less in the sarcoidosis patients. When separated by radiographic stage, patients with radiographically evident parenchymal involvement (stages II–IV, n = 13) exhibited mean forced expiratory volume in 1 sec (FEV1), FVC and DLCO values that were 24%, 17% and 15% less than the corresponding values in the group with no parenchymal involvement (stages 0–I, n = 9) (all P < 0.01), but there were no significant differences in the MRC scores, breathing reserve or V'O2max between the two groups. Mean duration of exercise for the patients was 10.5 min. No patients experienced chest pain, dizziness, serious cardiac arrhythmias or electrocardiographic findings typical of ischemia during the exercise testing.
Table 2. Peak exercise data in patients with sarcoidosis and control subjects
|V'O2 (ml/kg/min)||19.7 ± 6.1||25.7 ± 5.3||0.0042|
|V'O2 (% pred.)||63.6 ± 21.7||82.0 ± 20.5||0.0448|
|HR (% pred.)||85.6 ± 14.3||94.59 ± 8.4||0.0318|
|O2 pulse (% pred.)||109.4 ± 38.6||112.8 ± 31.1||0.1769|
|V'E (% pred.)||117.6 ± 49.5||98.1 ± 63.7||0.0378|
|Breathing reserve (%)†||29.2 ± 19.5||40.6 ± 14.9||0.001|
|V'E/V'O2 (% pred.)||36.5 ± 8.5||26.1 ± 9.8||0.0005|
|V'E/V'CO2 (% pred.)||36.5 ± 5.4||29.4 ± 10.5||0.0124|
|VD/VT (% pred.)||138.0 ± 37.4||102.1 ± 31.2||0.0143|
Table 3 shows selected physiologic and MFI-20 symptom rating variables for sarcoidosis patients receiving oral immunomodulatory drugs and those who were not. As can be seen, there were no significant differences in spirometric and peak exercise variables. The physical fatigue and MRC scale were, however, significantly greater in patients taking immunosuppressive medications.
Table 3. Comparison of selected variables between patients receiving oral immunomodulatory drugs and those who are not
|Age (years)||44.8 ± 6.7||48.8 ± 15.1||0.55|
|BMI||28.3 ± 6.9||31.4 ± 8.8||0.39|
|Chest radiograph stage 0/I/II/III/IV (n)||3/2/5/2/1||1/3/1/4/0||–|
|MRC scale||3.3 ± 1.7||2.3 ± 1.0||0.015|
|FVC (% pred.)||84.8 ± 20.1||82.2 ± 14.2||0.67|
|TLC (% pred.)||90.1 ± 18.7||85.0 ± 8.9||0.30|
|DLCO (% pred.)||79.7 ± 27.1||82.3 ± 21.9||0.94|
|V'O2max (% pred.)||63.5 ± 22.8||63.7 ± 21.2||0.97|
|O2-pulse (% pred.)||105.0 ± 36.3||115.7 ± 43.1||0.62|
|V'Emax (% pred.)||117.8 ± 54.3||117.3 ± 44.1||1.0|
|Breathing reserve (%)||25.3 ± 15.7||33.3 ± 24.4||0.10|
|General fatigue||16.8 ± 3.4||13.7 ± 6.1||0.30|
|Physical fatigue||15.9 ± 4.0||10.2 ± 4.9||0.015|
|Mental fatigue||10.7 ± 4.2||8.1 ± 3.7||0.29|
|TNF-α (pg/mL)†||1.86 (1.11–11.46)||2.10 (1.26–2.89)||0.28|
|IL-1β (pg/mL)†||0.17 (0.13–0.19)||0.16 (0.13–0.20)||0.85|
MFI-20 questionnaire scores
All patients and control subjects completed the MFI without omitting items. All of the mean categorical scores for fatigue, activity and motivation in the sarcoidosis patients were 1.6–2.5 times greater than corresponding scores for the control subjects (all statistically significant) (Table 4). The total fatigue score for the sarcoidosis patients was twice as high as for the controls.
Table 4. Scores for the multidimensional fatigue inventory (MIF-20) in patients with sarcoidosis and control subjects
|General fatigue||15.5 ± 4.9||7.3 ± 3.2||<0.0001|
|Physical fatigue||13.6 ± 5.1||6.9 ± 3.2||<0.0001|
|Reduced activity||13.2 ± 5.6||5.4 ± 2.2||<0.0001|
|Reduced motivation||11.0 ± 5.0||5.5 ± 2.0||0.0004|
|Mental fatigue||9.6 ± 4.1||6.2 ± 2.9||0.0071|
|Total||62.9 ± 21.9||31.2 ± 9.3||<0.0001|
Plasma cytokine levels
Plasma concentrations of TNF-α and IL-β before, immediately after and 4–6 h after exercise are shown in Table 5. Plasma TNF-α concentrations were, respectively, 42%, 80% and 62% greater than in the control subjects, at the times sampled (all statistically significant differences). Plasma IL-1β levels did not differ significantly between patients and control subjects. There were no significant differences in plasma IL-β concentrations among the three collection times in either cohort. A 28% increase in plasma TNF-α concentration was noted immediately post exercise, but this change was not statistically significant. Pre-exercise TNF-α and IL-1β concentrations did not differ between patients who were receiving immunomodulatory medications and those who did not (Table 3).
Table 5. Plasma levels of cytokines immediately before, immediately after and 6 h after exercise in patients with sarcoidosis and control subjects
|Sarcoidosis|| || || || |
| TNF-α (pg/mL)||1.72 (1.01–9.75)||2.21 (0.87–12.02)||1.96 (1.11–11.46)||0.66|
| IL-1β (pg/mL)||0.17 (0.13–0.20)||0.16 (0.12–6.74)||0.16 (0.11–2.81)||0.97|
|Control subjects|| || || || |
| TNF-α (pg/mL)||1.21 (0.11–3.63)||1.23 (0.10–3.73)||1.21 (0.10–4.16)||0.96|
| IL-1β (pg/mL)||0.16 (0.10–3.78)||0.15 (0.11–7.90)||0.14 (0.10–3.55)||0.30|
|P value**|| || || || |
| TNF-α||0.0008||0.008||0.0088|| |
| IL-1β||0.9855||0.4461||0.0626|| |
We found that mean pre-exercise plasma TNF-α concentration in the patients with stages II–IV (n = 13) was more than twice the value exhibited by those with stages 0–I (n = 9) (3.6 ± (SD) 3.2 pg/mL and 1.7 ± 0.3 pg/mL, respectively, P < 0.01). We found no significant differences, however, in mean pre-exercise plasma IL-β concentrations between the two groups (0.17 ± 0.03 pg/mL and 0.16 ± 0.21 pg/mL, respectively).
Relation between MFI-20 and physiologic and pre-exercise plasma cytokine data
Table 6 shows correlation factors and corresponding P values for components of the MFI-20 and selected physiologic data and pre-exercise plasma cytokine concentrations in patients receiving immunomodulatory drugs. As can be seen, plasma IL-1β correlated with 3 of 5 MFI-20 fatigue categories resulting in a significant correlation with the total fatigue score. Only general fatigue correlated with pre-exercise plasma TNF-α level. There were no correlations between selected physiologic variables or pre-exercise plasma cytokine levels with MFI-20 fatigue scores in patients not receiving drugs. In control subjects, only V'O2max correlated with general fatigue, and FVC correlated with the mental fatigue score. There was no difference in the general fatigue scores between patients with stages 0 and I, and those with stages II–IV (15.7 ± 5.4 and 15.4 ± 4.7, respectively).
Table 6. Correlations of MFI-20 components scores with selected physiologic and pre-exercise plasma cytokine measurements in patients with sarcoidosis receiving immunomodulatory medications
|FVC (% pr)||0.06/0.44||0.02/0.63||0.01/0.73||0.03/0.59||0.01/0.73||0.02/0.68|
|MVV (% pr)||0.30/0.07||0.09/0.34||0.004/0.84||0.14/0.23||0.04/0.53||0.11/0.29|
To our knowledge, this is the first study assessing plasma proinflammatory cytokine concentrations during rest and immediately following acute exercise in patients with sarcoidosis. The new findings in this study are (1) the association between physical fatigue, reduced motivation and total fatigue scores and pre-exercise plasma IL-1β concentrations, and between general fatigue and baseline plasma TNF-α concentrations in individual sarcoidosis patients receiving immunomodulatory drugs; (2) no significant increase in plasma concentrations of TNF-α and IL-β immediately following exercise. We did find a trend toward an increase in plasma TNF-α concentrations immediately following exercise in the sarcoidosis patients as a whole.
Our findings of reduced QOL and increased fatigue scores in the sarcoidosis cohort are similar to those of Cox et al.(23) who used the SF-36 and SGRQ instruments, and to those of Spruit et al.(24) who used the SF-36. In addition, patients in the group treated with immunomodulatory medications exhibited higher MRC and physical fatigue scores than patients not receiving such medications, yet cardiorespiratory function variables, distribution of radiographic staging or in baseline plasma concentrations of IL-1β did not differ from the untreated patients. Does this mean that the changes noted in the treated group are due to the drugs or to the disease? Cox et al.(23) who used a Sarcoidosis Health Questionnaire to assess QOL in their cross-sectional study, similarly found that treated patients exhibited more symptoms than those who took no medications, and lung function did not differ between their two groups. It is difficult to determine in our study, as in that of Cox et al.(23), as to whether the reduced MRC score and physical fatigue is related to the therapy or the underlying inflammatory process.
Intense exercise is known to cause an increase in the plasma levels of many pro- and anti-inflammatory cytokines in healthy adults (12, 25, 26). Although TNF-α and IL-1β have traditionally been understood to be the main inducer of cytokine production in acute phase reactions, the majority of studies have shown that the circulating concentration of these cytokines is either unchanged following exercise in healthy adult subjects, or exhibit relatively small, delayed increments (27–33). For example, Peterson et al.(11), using ELISAs manufactured by R&D Systems (sensitivities not described), found that TNF-α and IL-1β plasma concentrations were undetectable in adults with CFS before and after undergoing mild exercise. In our study, the absence of a significant increase in TNF-α and IL-1β plasma concentrations immediately after exercise in the sarcoidosis patients and control subjects was likely due to the less intense and shorter duration of exercise used in contrast to other investigations employing strenuous exercise (27–33). Alternatively, acute exercise resulted in no further increase in cytokine concentrations because of suppression by anti-inflammatory mechanisms such as anti-inflammatory cytokines or by an increase in TNF-α and IL-1β receptor antagonists (33). Our findings are similar to those of investigations of adults with CFS (10, 11) and suggest there is no consistent relationship between cytokine production by inflammatory cells and exercise. Spruit et al.(24) reported that quadriceps muscle force in sarcoidosis patients was inversely related to fatigue but not to the circulating levels of several proinflammatory cytokines, suggesting that skeletal muscle weakness rather than the circulating cytokines are responsible for the complaints of fatigue in these patients.
Our values for plasma TNF-α concentrations were lower than those reported by Boots et al.(34). One explanation for these findings is the different sensitivities between the ELISA kits used in the two studies. In the study of Boots et al.(34), the plasma TNF-α level of healthy controls is 5 ± 0.3 pg/mL. They used the PeliKine Compact human TNF-α ELISA kit (CLB/Sanquin, Amsterdam the Netherlands). According to the manufacturer's protocol, their kit has the lowest detectable TNF-α concentration of 1–3 pg/mL, and the value for normal serum is below 10 pg/mL. The ELISA kits we used were high-sensitivity assays, with much lower ranges of detection for TNF-α and IL-1β. Our plasma TNF-α-values are comparable to those reported by Cartier et al.(35) and Ostrowski et al.(27) who also used the high-sensitivity ELISA kits manufactured by R&D. Gender differences may also have contributed to the differences in plasma TNF-α and IL-1β concentrations. In contrast to the study of Boots et al.(34), most of our subjects were female (8/31 vs 34/44, respectively). Cartier et al.(35) reported that, compared with men, premenopausal women (body mass index 26.6 vs. ours 26.0, respectively) had lower mean resting plasma TNF-α concentrations (1.50 pg/mL compared with 1.71 pg/mL; P < 0.001). In a smaller study, Imahara et al.(36) reported that lipopolysaccharide-induced TNF-α and IL-1β production in venous blood samples from women was less than in men.
Limitations of our study include its cross-sectional design, which precluded the measurement of changes in fatigue scores and cytokine levels over time as well as the characterization of the cumulative effect of immunomodulatory medications on the MFI-20. Because of the controversial effects of corticosteroids on long-term outcomes of sarcoidosis (23), such as progression to pulmonary fibrosis, our finding that drug use was associated with worse physical fatigue scores, but not with general or mental fatigue scores, bears further examination in future studies. Therapy may be identifying more ill patients; however, it may affect not only fatigue but also cytokine levels. The relationship between cytokines and treatment is therefore complex, and the relatively limited number of patients makes it difficult to reach a definite conclusion. Serial data before and after therapy may help separate the effects of disease vs therapy on both fatigue and cytokine levels.
In conclusion, physical fatigue, reduced motivation and total fatigue scores are related to resting IL-1β concentrations in patients with sarcoidosis receiving immunomodulatory therapy, a finding not observed in untreated patients. Patients with a greater perception of dyspnea and physical fatigue are more likely to be receiving drugs, but their baseline plasma proinflammatory cytokine concentrations and cardiorespiratory function do not differ from those not receiving medications, making it difficult to separate the effects of disease vs therapy on fatigue scores and cytokine levels. Acute exercise is not associated with increases in cytokine levels, possibly because of the less strenuous nature of the exercise level or because of the effects of anti-inflammatory factors suppressing an increase in plasma cytokine concentrations. Future exercise-related studies assessing serial cytokine and fatigue data before and after therapy may be able to separate the effects of disease vs therapy.