To investigate cardiorespiratory and perceptual responses to exercise in patients with chronic fatigue syndrome (CFS), accounting for comorbid fibromyalgia (FM) and controlling for aerobic fitness.
To investigate cardiorespiratory and perceptual responses to exercise in patients with chronic fatigue syndrome (CFS), accounting for comorbid fibromyalgia (FM) and controlling for aerobic fitness.
Twenty-nine patients with CFS only, 23 patients with CFS plus FM, and 32 controls completed an incremental bicycle test to exhaustion. Cardiorespiratory and perceptual responses were measured. Results were determined for the entire sample and for 18 subjects from each group matched for peak oxygen consumption.
In the overall sample, there were no significant differences in cardiorespiratory parameters between the CFS only group and the controls. However, the CFS plus FM group exhibited lower ventilation, lower end-tidal CO2, and higher ventilatory equivalent of carbon dioxide compared with controls, and slower increases in heart rate compared with both patients with CFS only and controls. Peak oxygen consumption, ventilation, and workload were lower in the CFS plus FM group. Subjects in both the CFS only group and the CFS plus FM group rated exercise as more effortful than did controls. Patients with CFS plus FM rated exercise as significantly more painful than did patients with CFS only or controls. In the subgroups matched for aerobic fitness, there were no significant differences among the groups for any measured cardiorespiratory response, but perceptual differences in the CFS plus FM group remained.
With matching for aerobic fitness, cardiorespiratory responses to exercise in patients with CFS only and CFS plus FM are not different from those in sedentary healthy subjects. While CFS patients with comorbid FM perceive exercise as more effortful and painful than do controls, those with CFS alone do not. These results suggest that aerobic fitness and a concurrent diagnosis of FM are likely explanations for currently conflicting data and challenge ideas implicating metabolic disease in the pathogenesis of CFS.
Chronic fatigue syndrome (CFS) is a medically unexplained illness characterized by persistent or relapsing fatigue lasting at least 6 consecutive months that substantially reduces physical activity and is accompanied by 4 or more of the following symptoms: impaired memory or concentration, sore throat, tender lymph nodes, muscle and/or multiple-joint pain, headaches, unrefreshing sleep, and postexertional malaise (1). Proposed etiologies of this illness include cardiovascular, respiratory, and metabolic abnormalities, as well as altered effort sense (2, 3). Dysfunction of these systems could result in the premature development of fatigue and may help to explain the postexertional malaise and exacerbation of symptoms commonly reported by CFS patients. Exercise has been used as a stressor to determine whether stimulation of the cardiorespiratory system could provide insight into the mechanisms that underlie symptoms of fatigue in CFS.
Results of studies to determine the metabolic, cardiorespiratory, and perceptual responses to exercise in CFS have been equivocal (4–15). Explanations for the discordant results include nonstandard testing procedures including different exercise protocols for CFS patients and controls, problematic exercise testing criteria for peak effort, use of physically active controls, and nonstandardized assessment of perceived exertion. Aside from methodologic issues, major limitations have included failure to control for aerobic fitness or to account for common comorbid illnesses, particularly fibromyalgia (FM). A critical issue for research in CFS and FM is whether “abnormal” exercise responses observed in previous studies are truly a phenomenon of the illness or are an epiphenomenon of extreme sedentary behaviors. Controlling for aerobic fitness would allow for a clear test of this issue and determine whether “chronic fatigue syndrome” or physical deconditioning is the likely explanation for previous findings. Moreover, accounting for comorbid illnesses such as FM may help to determine whether illness heterogeneity is an important contributing factor in physiologic and perceptual responses to exercise.
FM is a chronic disorder characterized by diffuse musculoskeletal pain, sleep disturbance, fatigue, stiffness, and presence of multiple tender points. Importantly, FM is the most common comorbid condition in CFS patients, with 43–70% fulfilling the American College of Rheumatology (ACR) diagnostic criteria for FM (16–19). Previous exercise studies have not routinely assessed whether subjects with CFS had comorbid FM and have not evaluated its effects. This is important because patients diagnosed as having both CFS and FM report higher levels of physical disability and pain compared with patients with CFS only (17). It is plausible that the pain associated with FM involves different pathophysiologic mechanisms than the fatigue associated with CFS. Alternatively, painful muscle symptoms could affect a patient's willingness to exert peak effort during exercise. Therefore, the presence of comorbid FM could have specific, albeit presently unknown, effects on both physiologic and perceptual responses to exercise.
The present study was undertaken to compare cardiorespiratory and perceptual responses during a maximal exercise test in patients with CFS alone (CFS only), CFS patients with coexisting FM, and sedentary healthy controls. Based on previous observations that CFS patients did not differ in either peak oxygen consumption or perceived exertion compared with sedentary healthy controls (5) and that CFS patients with comorbid FM reported greater impairment of physical functional status and more distress than patients with CFS only (17), we hypothesized that patients with CFS only would not differ from controls in cardiorespiratory and perceptual responses to exercise when expressed relative to peak exercise capacity, but that patients with CFS plus FM would demonstrate decreased cardiorespiratory capacity and elevated perception of pain and effort compared with sedentary healthy subjects. In addition to controlling statistically for differences in aerobic fitness, we analyzed data from 18 subjects in each group matched (within 1 ml/kg/min−1) for peak oxygen consumption, thereby controlling for differences in both aerobic capacity and exercise time.
The study sample included 3 groups of subjects, ranging in age from 18 to 55 years: patients with a diagnosis of CFS alone (n = 29; 68% female), patients with a diagnosis of CFS and FM (n = 23; 69% female), and healthy controls (n = 32; 53% female). Prior to enrollment in the study, all subjects underwent a thorough medical evaluation by a nurse practitioner trained in the diagnosis of CFS and FM. All CFS patients met the 1994 Centers for Disease Control and Prevention (CDC) case definition of CFS (1). Patients with comorbid FM were diagnosed based on the ACR 1990 criteria (19). Sedentary healthy control subjects were recruited from a control subject pool at the New Jersey CFS Cooperative Research Center (CRC). “Sedentary” was defined as working in an occupation that did not require moderate to intense physical labor and not participating in physical exercise for more than 1 session per week. In order to quantify physical activity and verify the sedentary status of the control group, participants filled out the Godin Leisure-Time Exercise Questionnaire (20). The Godin Questionnaire is a recall measure of usual physical activity level. It asks for the individual to report the average number of times that he or she participates in certain activities for more than 15 minutes over a usual 1-week period. The exercise categories include strenuous (running, jogging, soccer, etc.), moderate (fast walking, tennis, easy swimming, etc.), and mild (yoga, fishing, easy walking, etc.). Scores are obtained by multiplying the number of times reported for each activity level by an estimated metabolic equivalent level for that category (strenuous = 9, moderate = 5, mild = 3). These products are summed for the total activity score.
Exclusion criteria for CFS patients were based on the clinical criteria listed in the CDC case definition (1), including the following: 1) any active medical condition that might explain the patient's chronic fatigue, 2) any unresolved previously diagnosed medical condition that might explain the patient's chronic fatigue, 3) any current or past diagnosis of major depression with psychotic or melancholic features, 4) alcohol or other substance abuse currently or within 2 years prior to the onset of the chronic fatigue, and 5) a body mass index of ≥45 kg/m2. Sedentary healthy controls were excluded if they had a history of cardiovascular, respiratory, neurologic, or major psychiatric disorders or were taking any medications other than oral contraceptives.
In order to further reduce heterogeneity, we restricted our analyses to only those subjects who could complete a maximal exercise test based on standardized criteria for volitional effort. Additionally, the groups were matched on age (±3 years), height (±4 cm), weight (±5 kg), and years of education (±2 years).
Participants were recruited from a large patient pool available through the CRC. Participants volunteered through advertisements for the CRC or were contacted by CRC personnel. All participants were given the opportunity to volunteer for a number of research protocols that were supported by the CRC. A total of 92 participants volunteered for the present study. Of the 92, 5 were excluded from participation prior to completion of the exercise test. All 5 had chronic fatigue, and the reasons for exclusion were as follows: 1) patient (male, age 27 years) did not meet case definition criteria for CFS (determined during history and physical examination); 2) patient (male, age 29 years) had not eaten that day and was feeling light-headed; 3) patient (female, age 28 years) was found to have asthma; and 4) 2 patients (both female, ages 29 years and 42 years) were not comfortable with the experimental preparation and devices. A total of 87 participants completed the exercise test, of whom 84 (96.6%) met standard criteria for peak effort. Of the 3 who did not meet criteria, 2 were female patients with CFS plus FM (ages 53 years and 44 years) and 1 was a male control subject (age 46 years). Thus, the vast majority of volunteers were capable of exerting peak effort during cycling exercise.
The testing was performed under controlled environmental conditions (20–24°C and 40–60% relative humidity), between 11:00 AM and 2:00 PM. Subjects reported to the Human Performance Laboratory located at the Veterans Affairs Medical Center (VAMC) in East Orange, New Jersey, having abstained from smoking for 2 hours, from ingesting caffeine for 4 hours, and from exercising for 24 hours prior to testing. After providing informed consent (approved by human subjects committees of both the VAMC and the University of Medicine and Dentistry of New Jersey), subjects completed the Profile of Mood States (POMS) scale (21). This scale assesses feelings of fatigue and vigor for the week preceding the testing day. They also completed a Short-Form McGill Pain Questionnaire (MPQ) (22) to determine resting muscle pain prior to testing.
Prior to exercise testing, subjects were instrumented for monitoring of heart rate and metabolic responses to exercise. Modified lead II electrocardiography was used to monitor heart rate throughout the test. Expired air was collected breath by breath during exercise, using a 2-way non-rebreathing valve (Hans-Rudolph, Kansas City, MO), and was analyzed with the Max-1 metabolic measurement system (Physio-Dyne Instruments, Quoge, NY) for assessment of O2 consumption (VO2), CO2 production (VCO2), respiratory rate, and ventilation (VE). From the directly collected measures we derived measures of ventilatory equivalents of carbon dioxide (VE/VCO2) and oxygen (VE/VO2) for statistical analysis. The system was calibrated prior to each test, using standard gases with known concentrations and with a calibrated 3-liter syringe. End-tidal CO2 was measured breath by breath using an infrared expired gas analyzer (POET II; Criticare Systems, Waukesha, WI). The ventilatory threshold (VT) was determined using the V-slope method as described by Sue et al (23).
Following instrumentation, subjects were seated on an electronically braked cycle ergometer (Sensormedics, Loma Linda, CA) with the seat and handlebars adjusted for optimal performance, and allowed a few minutes to habituate to the cycle and various monitoring devices. The exercise test began with a 3-minute warm-up pedaling at 20W. Subjects were instructed to maintain a pedaling cadence of 60–70 revolutions per minute. Following the warm-up period, work intensity was increased by 5W every 20 seconds until the subject reached volitional exhaustion or a point when he or she could no longer maintain the prescribed pedaling rate. The subjects were verbally encouraged to continue pedaling as long as possible. Peak effort was determined based on meeting at least 2 of the following criteria: 1) respiratory exchange ratio ≥1.1, 2) achievement of 85% of age-predicted maximum heart rate, 3) rating of perceived exertion (RPE) ≥17, and 4) a change in VO2 of <200 ml with an increase in work. Examination of individual exercise tests revealed that the majority of subjects met the first 3 of these criteria (86% of patients with CFS only, 83% of patients with CFS plus FM, 81% of controls), while the fourth criterion was rarely met. Thus, the groups were well matched for both subjective and objective indices of effort.
RPEs were obtained during the last 10 seconds of every minute during the exercise test and every 30 seconds during recovery, using the Borg 6–20 category scale (24). Before exercise, subjects were given standard instructions on the proper use of the Borg 6–20 category scale. Briefly, subjects were instructed 1) that the Borg 6–20 scale would be used to determine the intensity of effort or stress felt in the legs only, during exercise and recovery, 2) that each number represents a category of sensation that is ordered according to its intensity, and 3) that the verbal anchors (“very light,” “light,” “somewhat hard,” “hard,” etc.) should be used to help determine the level of effort at that particular moment. Subjects were also provided cognitive anchors at the high and low ends of the perceptual continuum. Specifically, they were instructed that a 6 on the scale refers to the amount of effort exerted in their legs while they are sitting in a chair (i.e., no effort), while a 20 refers to the most effort imaginable (i.e., carrying a child out of a burning building or finishing a marathon).
Leg muscle pain intensity ratings were obtained during the last 10 seconds of every minute during exercise and every 30 seconds during recovery. Pain intensity was measured with the use of a validated category-ratio pain intensity scale (25) ranging from 0 (no pain at all) to 10 (extremely intense pain, almost unbearable). Subjects were instructed 1) that the 0–10 pain intensity scale would be used to determine the intensity of pain felt in the leg muscles only, during exercise and recovery, 2) that pain was defined as the intensity of hurt that was felt in the leg muscles, 3) that each number represents a category of sensation that is ordered according to its intensity, and 4) that the verbal anchors (“no pain,” “weak pain,” “moderate pain,” “strong pain,” “extremely intense pain,” etc.) should be used to help determine the level of pain intensity at that particular moment.
Statistical analyses were conducted using SPSS for Windows (version 12.0.1; SPSS, Chicago, IL). Subject characteristics, variables at the VT, and peak exercise variables in experimental and control groups were compared using analysis of variance (ANOVA). Tukey's post hoc analyses were used to further determine group differences. Differences in cardiorespiratory features (VO2, VE, respiratory rate, VE/VO2, VE/VCO2, heart rate, end-tidal CO2) and perception (pain and RPE) during exercise were analyzed using ANOVA with repeated measures for exercise intensity.
To control statistically for the effect of fitness differences on perception and cardiorespiratory responses, dependent variables were expressed and analyzed relative to each individual's peak oxygen consumption, by repeated-measures ANOVA. Complete pain and RPE data were available for workloads associated with 30–100% of peak oxygen consumption, and complete cardiorespiratory data for peak VO2 of 40–100%. Therefore, the repeated-measures analyses were conducted using these respective intensities. Tukey's post hoc analyses were used to further determine significant main effects. Greenhouse-Geisser adjustments for degrees of freedom were used when a significant difference (P < 0.05) was shown with Mauchly's test of sphericity. Linear regressions of log-transformed data were used to further characterize significant interaction terms by providing curve estimates of the growth functions of the physiologic and perceptual variables.
In order to more definitively determine the effect of aerobic fitness on cardiorespiratory and perceptual responses to exercise, we performed the same set of analyses described above on a subset of participants matched (within 1 ml/kg/min−1) for peak oxygen consumption. Eta-square values are reported to demonstrate the magnitude of the differences between the whole group and the fitness-matched subgroup samples.
Demographic and baseline cardiorespiratory and questionnaire data on both the entire group and the subgroup matched for fitness are presented in Table 1. As intended, there were no significant differences among the groups for age, height, weight, years of education, or baseline heart rate.
|Overall study population||Fitness-matched subgroup|
|CFS only (n = 29)||CFS plus FM (n = 23)||Controls (n = 32)||CFS only (n = 18)||CFS plus FM (n = 18)||Controls (n = 18)|
|Age, years||39.8 ± 9||40.9 ± 8||37.0 ± 12||41.4 ± 10||41.2 ± 8||40.8 ± 12|
|Height, cm||170.0 ± 10||166.3 ± 9||168.9 ± 12||168.5 ± 10||166.4 ± 10||163.9 ± 9|
|Weight, kg||71.7 ± 13||69.4 ± 18||74.7 ± 19||69.8 ± 12.5||70.8 ± 18.5||70.9 ± 22.6|
|Education, years||16.3 ± 2.7||16.0 ± 2.5||16.7 ± 2.8||16.4 ± 2.7||16.0 ± 2.5||17.4 ± 2.9|
|Heart rate, beats/minute−1||72 ± 10||76 ± 13||72 ± 10||73 ± 10||75 ± 12||74 ± 9|
|Physical activity score†|
|Strenuous||0.4 ± 1.1‡||0.4 ± 1.2‡||1.5 ± 1.9||0.47 ± 1||0.33 ± 1||0.61 ± 1|
|Moderate||1.1 ± 1.9‡||1.2 ± 2.1‡||3.1 ± 3.3||1.2 ± 2‡||0.8 ± 2‡||3.6 ± 4|
|Minimal||1.8 ± 1.8‡||3.9 ± 2.6||2.7 ± 2.5||2.2 ± 2||3.3 ± 2||3.4 ± 3|
|Total||14.9 ± 18.0‡||22.3 ± 24.8‡||37.5 ± 22.4||16.6 ± 21||17.7 ± 21||33.6 ± 21|
|Pain score§||2.1 ± 4‡||5.2 ± 4¶||0.4 ± 1||2.0 ± 3.8||4.8 ± 3.0‡||0.2 ± 0.4|
|Fatigue||16.1 ± 7‡||19.4 ± 4‡||2.9 ± 3||16.4 ± 7‡||19.6 ± 5‡||3.4 ± 4|
|Depression||11.2 ± 12‡||6.2 ± 6||3.2 ± 5||12.0 ± 13‡||7.1 ± 7||4.0 ± 6|
|Tension/anxiety||10.5 ± 9‡||7.6 ± 5‡||4.8 ± 4||10.0 ± 8‡||8.4 ± 6‡||4.9 ± 5|
|Anger/hostility||5.9 ± 7||4.1 ± 5||3.3 ± 4||6.0 ± 7||4.4 ± 5||4.0 ± 4|
|Vigor||9.2 ± 7‡||8.3 ± 5‡||22.8 ± 6||9.9 ± 8‡||8.4 ± 5‡||22.6 ± 6|
|Confusion||11.1 ± 6‡||10.0 ± 4‡||3.3 ± 3||11.9 ± 7‡||9.8 ± 4‡||3.2 ± 3|
|Total||45.7 ± 39‡||38.0 ± 17‡||−5.1 ± 18||46.4 ± 40‡||39.4 ± 18‡||−3.1 ± 21|
Although our control subjects would be considered sedentary by normal physical activity standards (26), there was a significant group main effect of total physical activity (P = 0.001). Tukey's post hoc analyses indicated that the control group was significantly more active than both the CFS only group (P = 0.001) and the CFS plus FM group (P = 0.04), while the CFS only and CFS plus FM groups were not significantly different from one another. The range of scores for self-reported physical activity was large (CFS only group 0–64, CFS plus FM group 0–92, control group 9–84). Examination of the subscales of the Godin questionnaire revealed that controls engaged in strenuous and moderate activities significantly more than either the CFS only or the CFS plus FM patients (for strenuous activity, P = 0.02 and P = 0.03, respectively; for moderate activity, P = 0.01 and P = 0.03, respectively) and in minimal effort activities significantly more than the CFS only group (P = 0.006).
Group differences in total physical activity were reduced but not eliminated when the data were examined in the groups matched for fitness (P = 0.03), with controls engaging in moderate activities significantly more frequently than the CFS only group (P = 0.03) or the CFS plus FM group (P = 0.007).
MPQ data on the entire sample indicated a significant main effect for group (P < 0.001), with both the CFS only group (P = 0.05) and the CFS plus FM group (P < 0.001) reporting significantly more resting pain than controls and the CFS plus FM group reporting more pain than the CFS only group (P = 0.001). POMS data indicated significant main effects for fatigue (P < 0.001), vigor (P < 0.001), depression (P = 0.007), tension/anxiety (P = 0.003), and confusion (P < 0.001), but not for anger/hostility. Healthy control subjects reported different levels of each measured symptom than the CFS only group, and their responses for each variable, with the exception of self-reported depression and anger/hostility, were significantly different from those of the CFS plus FM group. There were no significant differences between the CFS only and CFS plus FM groups.
In the matched sample, only the CFS plus FM group reported greater resting muscle pain than controls (P = 0.01). The results on the POMS followed the same pattern as described for the whole group analyses.
Heart rate responses during exercise are depicted in Figure 1. Repeated-measures ANOVA indicated a significant main effect for exercise intensity (P < 0.001) and a significant group-by–exercise intensity interaction (P = 0.03, η2 = 0.08), but no group main effect. Linear regression of the log-transformed heart rate and relative oxygen consumption data revealed a lower slope for the CFS plus FM group (0.46) compared with the CFS only group (0.54) and the control group (0.54), indicating a slower increase in heart rate in the CFS plus FM group compared with the other groups.
Group differences in heart rate were eliminated (P = 0.16, η2 = 0.05) when the data were examined in the subgroups matched for aerobic fitness.
Ventilation responses during exercise are shown in Figure 1. There were significant main effects for exercise intensity (P < 0.001) and group (P = 0.02, η2 = 0.14), as well as a significant group-by–exercise intensity interaction (P = 0.02, η2 = 0.12). Post hoc analyses indicated that the CFS plus FM group had significantly lower VE throughout exercise compared with the control group (P = 0.02), but not with the CFS only group. There were no significant differences between the CFS only and control groups. Linear regression of the log-transformed ventilation and relative oxygen consumption data revealed no differences in the slopes among the groups (CFS plus FM 1.20, CFS 1.25, control 1.28).
Differences in VE were eliminated (group P = 0.46, η2 = 0.03; interaction P = 0.4, η2 = 0.04) when the data were examined in the subgroups matched for aerobic fitness.
VE/VO2 and VE/VCO2 responses during exercise are depicted in Figure 2. For VE/VO2, there was a significant main effect for exercise intensity (P < 0.001), but no significant main effect for group or for the interaction. For VE/VCO2, there were significant main effects for exercise intensity (P < 0.001) and group (P = 0.04, η2 = 0.08), but no group-by–exercise intensity interaction. There were no significant group differences in the pairwise comparisons.
There were no significant differences among the groups for either VE/VO2 (P = 0.7, η2=0.01) or VE/VCO2 (P = 0.5, η2 = 0.03) when the data were analyzed in the subgroups matched for aerobic fitness.
For end-tidal CO2, significant main effects for exercise intensity (P < 0.001) and group (P = 0.017, η2 = 0.11), but not for the interaction, were observed (Figure 2). Post hoc analyses indicated that end-tidal CO2 throughout exercise was significantly reduced in the CFS plus FM group compared with the control group (P = 0.016), but not the CFS only group. There were no significant differences between the CFS only group and the control group.
Differences in end-tidal CO2 were eliminated (P = 0.08, η2 = 0.10) when the data were analyzed in the fitness-matched subgroups.
Pain intensity ratings during exercise are shown in Figure 3. There were significant main effects for exercise intensity (P < 0.001) and group (P = 0.005, η2 = 0.12), but no interaction. Post hoc analyses indicated that the CFS plus FM group rated exercise as significantly more painful than either the CFS only group (P = 0.02) or the control group (P = 0.007). There were no significant differences between the CFS only and control groups.
The pattern of results was unchanged (P = 0.03, η2 = 0.13) when differences in fitness and exercise time were accounted for in the matched subgroup analysis, except that findings in the CFS plus FM group were no longer significantly different from those in the CFS only group.
Figure 3 shows the RPE data in each group. There were significant main effects for exercise intensity (P < 0.001) and group (P < 0.001, η2 = 0.21), as well as a significant group-by–exercise intensity interaction (P = 0.04, η2 = 0.05). Post hoc analyses indicated that both the CFS plus FM group (P < 0.001) and the CFS only group (P = 0.01) rated exercise as significantly more effortful than did controls. However, there were no significant differences between the CFS plus FM and CFS only groups. Linear regression of the log-transformed data on perceived exertion revealed similar slopes among the groups (CFS plus FM 0.56, CFS only 0.57, control 0.62).
When differences in fitness were accounted for, there was a significant main effect for group (P = 0.001, η2 = 0.25). Post hoc analyses indicated that the RPEs of the CFS only group were not significantly different from those of the controls, while the CFS plus FM group continued to rate exercise as more effortful than the control group (P < 0.001), but with no significant difference from the CFS only group (P = 0.06).
Exercise responses at the VT are shown in Table 2. Each group reached the VT at a similar relative intensity, as demonstrated by the percentage of peak VO2. There were no differences among groups for respiratory rate, respiratory exchange ratio, or heart rate. Significant group main effects occurred for VO2 (P = 0.001), VCO2 (P = 0.001), VE (P = 0.02), VE/VO2 (P = 0.02), VE/VCO2 (P = 0.009), watts (workload) (P = 0.001), pain (P = 0.03), and perceived exertion (P = 0.02). Compared with controls, both the CFS only and the CFS plus FM groups had significantly lower VO2 (CFS P = 0.009, CFS plus FM P = 0.001), VCO2 (CFS P = 0.008, CFS plus FM P = 0.003), and watts (CFS P = 0.03, CFS plus FM P = 0.002). In addition, the CFS plus FM group exhibited significantly lower VE (P = 0.04) as well as higher VE/VO2 (P = 0.02), VE/VCO2 (P = 0.009), and RPE (P = 0.02) compared with controls and significantly higher pain levels compared with the CFS only group (P = 0.03) and the control group (P = 0.001).
|Overall study population||Fitness-matched subgroup|
|CFS only||CFS plus FM||Controls||CFS only||CFS plus FM||Controls|
|% peak VO2||58 ± 0.08||59 ± 0.06||57 ± 0.06||58 ± 0.08||59 ± 0.05||58 ± 0.06|
|VO2, ml||1,001 ± 217†||929 ± 196†||1,257 ± 432||914 ± 132||968 ± 192||962 ± 239|
|VCO2, ml||909 ± 234†||867 ± 203†||1,173 ± 427||822 ± 154||908 ± 202||882 ± 203|
|RER||0.9 ± 0.08||0.9 ± 0.08||0.9 ± 0.08||0.9 ± 0.08||0.9 ± 0.08||0.9 ± 0.08|
|VE, liters/minute−1||29 ± 6||28 ± 6†||34 ± 9||27 ± 4||30 ± 5||27 ± 5|
|RR, breaths/minute−1||24 ± 5||23 ± 5||25 ± 4||24 ± 6||23 ± 4||25 ± 4|
|VE/VO2||29 ± 4||31 ± 3†||28 ± 3||29 ± 4||31 ± 3||29 ± 4|
|VE/VCO2||32 ± 4||33 ± 3†||30 ± 4||33 ± 4||34 ± 3||32 ± 4|
|Heart rate, beats/minute−1||109 ± 17||107 ± 14||112 ± 17||104 ± 16||107 ± 12||106 ± 13|
|Watts||70 ± 19†||62 ± 17†||90 ± 37||65 ± 12||62 ± 19||62 ± 14|
|RPE, 6–20 scale||11 ± 2||13 ± 2†||10 ± 2||11 ± 2||13 ± 2†||10 ± 2|
|Pain, 0–10 scale||1.6 ± 2||3.2 ± 2‡||1.1 ± 1||1.4 ± 2||2.5 ± 2†||0.9 ± 1|
In analyses of the subgroups matched for fitness, all differences observed at the VT were eliminated with the exception of pain and RPE; the CFS plus FM group rated exercise as significantly more painful (P = 0.03) and more effortful (P = 0.001) than did the controls.
Table 3 shows peak responses to exercise in the various study groups. There were no significant differences in peak exercise responses for pain, RPE, heart rate, respiratory exchange ratio, or breathing frequency. Significant group main effects were found for peak VO2 (P = 0.002), watts (P = 0.001), total exercise time (P = 0.001), and VE (P = 0.01). Post hoc analyses showed a significant difference between the CFS plus FM group and controls for each of these parameters. The range of aerobic fitness values in the groups was large (CFS only 17–40 ml/kg/min−1, CFS plus FM 14–33 ml/kg/min−1, controls 18–49 ml/kg/min−1). However, there were no significant differences between the CFS only and the CFS plus FM groups.
|Overall study population||Fitness-matched subgroup|
|CFS only||CFS plus FM||Controls||CFS only||CFS plus FM||Controls|
|VO2, ml/kg/minute−1||25.7 ± 6||23.4 ± 4†||29.7 ± 8||24.0 ± 4||24.2 ± 4||24.3 ± 4|
|Heart rate, beats/minute−1||169 ± 14||163 ± 20||173 ± 16||169 ± 14||169 ± 10||166 ± 16|
|RER||1.2 ± 0.1||1.2 ± 0.1||1.2 ± 0.1||1.2 ± 0.1||1.2 ± 0.1||1.2 ± 0.1|
|VE, liters/minute−1||77 ± 20||68 ± 15†||88 ± 30||72 ± 12||72 ± 14||68 ± 13|
|RR, breaths/minute−1||47 ± 10||41 ± 8||46 ± 10||47 ± 11||41 ± 8||43 ± 9|
|Watts||150 ± 39||129 ± 26†||177 ± 60||138 ± 22||133 ± 23||135 ± 31|
|Exercise time, minutes||9.5 ± 2.6||8.1 ± 1.7†||11.2 ± 3.9||8.7 ± 1.4||8.4 ± 1.5||8.4 ± 1.9|
|RPE, 6–20 scale||18.1 ± 1.3||18.5 ± 1.4||18.0 ± 1.5||18.0 ± 1||18.4 ± 1||18.0 ± 2|
|Pain, 0–10 scale||4.8 ± 2.8||6.7 ± 3.6||5.5 ± 3.2||4.6 ± 3||6.4 ± 4||4.5 ± 3|
All differences in peak exercise responses, including peak oxygen consumption and peak exercise time, were eliminated when the data were analyzed in the fitness-matched subgroups.
In the present study, we sought to extend and improve on previous investigations of high-intensity exercise among patients with CFS. Earlier studies have yielded equivocal results for various reasons, including use of poorly matched control groups and questionable exercise testing procedures. Some investigations have utilized tests lasting twice as long as standard protocols designed to identify a metabolic maximum, others have used different exercise protocols for CFS patients and control participants, and others have failed to adhere to accepted criteria for maximal effort. However, even carefully controlled studies with appropriate exercise tests and objective indicators of maximal effort have not produced consistent results.
We believe the results of the present investigation provide 2 primary explanations for inconsistencies in findings with regard to exercise responses among CFS patients: 1) many of the patients included in previous investigations also had FM, and 2) previous studies did not account adequately for differences in aerobic fitness. The present investigation is the first study of short-term exercise in CFS patients in which a concomitant diagnosis of FM was controlled for and data were examined in groups closely matched for aerobic fitness and exercise performance. Several of our findings would not have been revealed had we failed to account for the presence of comorbid FM. Indeed, for virtually all of the primary dependent variables, the CFS only group did not differ from a group of sedentary healthy controls. Further, the results of the study would have been misleading had differences in aerobic fitness not been accounted for. After fitness was controlled for, we found that CFS patients with comorbid FM perceived exercise as more painful and effortful, but did not differ in their cardiorespiratory responses. Thus, concurrent diagnoses and differing levels of disease burden combined with the use of control subjects who had greater aerobic capacity may have confounded the results of many earlier investigations.
Several studies have depicted the exercise capacities of CFS patients as being reduced compared with healthy controls (7, 8, 10, 15), whereas others have shown “low normal” VO2max values (4, 5, 11, 13). Severe deconditioning may result in abnormal exercise responses and may partly explain why some patients report postexercise exacerbation of symptoms. However, this hypothesis has not been systematically tested. Further, investigators in those studies have concluded that cardiovascular, ventilatory, or metabolic abnormalities revealed through exercise testing may explain the CFS patients' exercise intolerance and provide insight into pathophysiologic mechanisms of the disease. Specifically, depressed maximal and submaximal heart rates, lower respiration during exercise, slower oxygen saturation recovery in muscle following exercise, and general autonomic differences in CFS patients compared with healthy controls have been reported (10, 27–30).
Inbar et al (10) reported that patients with CFS exhibited significantly lower heart rates at equal relative exercise intensities compared with sedentary healthy controls and concluded that the observed pattern of slow cardiac acceleration was not characteristic of a deconditioned response, but instead suggested that insufficient cardiac pacing or abnormally low sympathetic drive might explain the low heart rate response and thus the patients' exercise intolerance. Similarly, Montague et al (30) reported normal resting cardiac function in CFS patients, but a slow acceleration of heart rate during exercise, leading to limitations in capacity. Inbar and colleagues (10) also noted that CFS patients had lower VE and end-tidal CO2 at relative and peak exercise intensities, but concluded that these were probably the result, not the cause, of low exercise tolerance. DeBecker et al (7), studying a large cohort of CFS patients (n = 427) and controls (n = 204), reported significantly reduced exercise capacity in the CFS group and speculated that the marked exercise impairment in that group was due to an inability to achieve target heart rates and thus was reflective of impaired cardiac function. However, the use of questionable criteria for maximal exercise and different exercise protocols for patients and controls could have contributed to these observations. Importantly, in none of these studies were differences in aerobic capacity controlled for or the occurrence of comorbid FM determined in the CFS participants.
In the present study, similar conclusions to those discussed above would have been drawn had all subjects who met standard criteria for peak volitional effort been included. Statistically controlling for differences in peak exercise capacity by expressing the data relative to each subject's peak oxygen consumption value allowed for the inclusion of subjects who had low aerobic fitness, but who could complete the maximal exercise protocol based on a priori–set standards. Thus, when the entire sample was considered, patients with CFS only did not differ from healthy controls with respect to virtually any cardiorespiratory variable including maximal or submaximal heart rates, total minute ventilation, or end-tidal CO2 level. However, patients with both CFS and FM exhibited slower accelerating heart rates during exercise compared with healthy controls, indicated by lower heart rates at each percentage of their peak exercise capacity and a lower slope throughout exercise. They also demonstrated lower VE and end-tidal CO2 and higher VE/VCO2 levels at given relative workloads compared with healthy controls. These results occurred throughout the maximal exercise test as well as when the data were examined at the VT, indicating that the differences occurred when the groups were exerting similar metabolic efforts.
One potential explanation for these results is simply that patients with CFS and FM exerted inadequate effort during the exercise test, failing to attain a true “peak.” This explanation is unlikely, however, since all of the subjects met objective criteria for maximal effort. Another potential explanation could relate to the fact that the CFS plus FM group exhibited lower end-tidal CO2 levels during exercise than controls. Thus, chemoreceptor stimulation by partial pressure of CO2 may have been diminished in the CFS plus FM group, leading to lower VE during exercise. However, when we compared the groups after matching for peak oxygen consumption, and thus exercise time, every observed cardiorespiratory difference was eliminated. This finding strongly suggests that results obtained in the entire sample were the result of comparing groups at equivalent relative intensities but different total exercise times.
To illustrate the above point, we compare 2 subjects, 1 from the CFS plus FM group and 1 from the control group. The CFS plus FM participant (female; peak VO2 25 ml/kg/min−1) exercised for a total of 7 minutes and reached 55% of peak oxygen consumption after only 2 minutes of exercise. The control participant (female; peak VO2 39 ml/kg/min−1) exercised for a total of 13 minutes and reached 55% of peak oxygen consumption after 6 minutes of exercise. Corresponding heart rates were 111 beats per minute and 136 beats per minute, respectively. Thus, statistically controlling for differences in aerobic fitness by expressing the data as a function of relative intensity (i.e., 55%) may diminish differences at the group level, but cannot fully account for meaningful differences in total exercise time between less fit and more fit subjects.
The present findings also indicate that, in general, the aerobic capacities of CFS patients do not differ significantly from those of sedentary control subjects, unless there is a comorbid diagnosis of FM. Even then, it was possible to match the majority of the CFS plus FM patients (18 of 23) to both CFS only patients and sedentary controls. Thus, group differences in the entire sample were due to the inclusion of controls who were fit although sedentary (mean ± SD peak VO2 36.6 ± 5 ml/kg/min−1) and patients with CFS only or CFS plus FM who had relatively lower levels of fitness (peak VO2 28.7 ± 8 ml/kg and VO2 20 ± 4 ml/kg/min−1, respectively).
The above results contrast with those obtained for physical activity. Despite our attempts to enroll healthy controls who were matched with the CFS subjects in terms of levels of physical activity, the 2 CFS groups engaged in significantly less self-reported physical activity than the control subjects. However, this difference did not translate into widespread differences in aerobic capacity among the groups. This may indicate a floor effect for the relationship between physical activity and aerobic fitness, since the mean aerobic capacities in the entire sample were at or below the 20th percentile for population averages. In fact, total physical activity was weakly (r = 0.13) correlated with peak oxygen consumption (aerobic fitness) in the total sample. This finding highlights the importance of obtaining an objective measure of aerobic capacity (i.e., peak oxygen consumption) to determine fitness, but does not dismiss the utility of self-reported physical activity to screen for physical activity behaviors.
Several reports have described elevated ratings of effort in CFS, and the authors have speculated that central dysregulation of effort sense may be a contributing factor to both the increased perception of fatigue and the decreased exercise tolerance in these subjects (8, 9, 12, 14, 31, 32). In a recent study, Wallman et al (32) used a submaximal exercise protocol to determine physiologic and perceptual responses in CFS patients and controls matched for physical activity history and found that physiologic responses to the test were similar, but that RPEs were elevated in the CFS patients. They concluded that reduced exercise capacity in CFS may be due to “abnormal” perception of exertion, perhaps due to impairment of the neural mechanisms that regulate effort sense. However, they also found that peak exercise capacity was significantly lower in their CFS patients than in their controls. Therefore, comparisons made at any given absolute workload represented a greater relative workload for CFS patients. This could in part explain the elevated effort ratings. Wallman and colleagues did not determine whether their CFS patients also had comorbid FM.
We previously reported that perceived exertion was elevated in women with CFS when the data were expressed as a function of absolute (exercise time) exercise intensity, but these differences did not remain when the data were expressed as relative (percent peak oxygen consumption) to peak exercise capacity (5). This finding challenged the notion that CFS was an illness of altered effort sense and highlighted the importance of appropriate reference criteria when making group comparisons. In the present study, when we analyzed the entire sample, both the CFS only and the CFS plus FM patients rated exercise as requiring more effort than did healthy controls, even after statistical controlling for differences in peak aerobic fitness. However, when we compared groups matched for aerobic fitness, only the CFS plus FM patients rated cycling exercise as requiring more effort than controls. This demonstrates that controlling for fitness and accounting for comorbid FM are important considerations when assessing exercise perceptions and tolerance in patients with CFS. Thus, with these controls in place and with the use of identical exercise testing procedures, CFS patients do not perceive a short-term session of exercise as requiring more effort than do matched sedentary controls. These results are consistent with our previous findings and extend them to a different mode of exercise (cycling). The findings suggest that reports of elevated RPEs in patients with CFS are likely an epiphenomenon of extremely low aerobic fitness and/or the inclusion of subjects with comorbid FM, and not a phenomenon of central alterations in effort sense.
Another potential, yet unexplored, explanation for reports of elevated RPEs in CFS is that exercise may be more painful and that nociceptive signals from the contracting muscle contribute to an increase in effort sense. Indeed, some have speculated that reduced motivation or increased symptomatology, such as pain, may limit the CFS patient's ability to attain maximal exertion (2). To our knowledge, this is the first study to determine naturally occurring muscle pain during exercise in CFS. Our patients with CFS only did not report greater leg muscle pain during exercise than the healthy controls. This was the case for both the entire sample tested and the subgroup that was matched for aerobic fitness. This finding may appear odd, particularly since muscle and joint pain are part of the symptom list used in diagnosing CFS and are common in most CFS patients (1). However, naturally occurring muscle pain from exercise may represent a different phenomenon from chronic symptoms of muscle and joint pain that occur during rest (25). These results also suggest that CFS patients without comorbid FM have normal nociceptive processing during exercise and that muscle contractions do not result in a hyperalgesic state in these patients.
In contrast to the CFS only group, the CFS plus FM patients reported greater muscle pain throughout exercise compared with the control group; muscle pain during exercise in the CFS plus FM group was also significantly higher than that reported by patients in the CFS only group. While the initial muscle pain ratings were similar among the groups, muscle pain increased at a greater rate in the CFS plus FM patients. These results suggest that normal inhibition of pain signals from the contracting muscle may be intact in patients with CFS but impaired in patients with comorbid FM, and are consistent with data suggesting abnormal central processing and regulation of pain (33) and hyperalgesia resulting from skeletal muscle contractions (34) in FM. The exacerbation of symptoms following exercise reported by CFS patients could be due to the presence of comorbid FM and augmented afferent nociceptive signaling from contracting skeletal muscle.
The fact that patients with CFS plus FM rate exercise as more painful than patients with CFS only has important implications with regard to exercise prescription. Extra care should be taken when prescribing an exercise program for an individual with diagnoses of both CFS and FM, since painful skeletal muscle contractions could lead to peripheral and/or central hyperalgesia with a resulting increase in widespread pain. In theory, pain perception could be used to regulate the mode and intensity of exercise and to individually tailor the exercise program to account for differences in physical function and levels of disability due to pain.
The primary limitation of the present study is the failure to include an FM only group. Testing patients with FM but without CFS would have allowed us to draw more definitive conclusions regarding the impact of FM alone on cardiorespiratory and perceptual responses to exercise. Future research to determine these responses during exercise, as well as in response to exercise training, in both CFS patients and FM patients would be an important next step.
The results of the present investigation highlight the importance of considering aerobic fitness level and the presence of comorbid illness in studies of exercise in CFS and provide insight regarding the functional and aerobic capacity of individuals with single and dual diagnoses. Specifically, we demonstrate that a concomitant diagnosis of FM may greatly impact the psychological responses to exercise in CFS patients, but that neither the diagnosis of CFS nor the diagnosis of FM has a meaningful effect on cardiorespiratory responses to short-term cycling exercise as used in this study. Given the substantial overlap of CFS and FM, equivocal findings of previous research are likely due to a failure to effectively control for differences in aerobic capacity between patients and controls and a failure to account for comorbidity in CFS patients. These results also seriously challenge hypotheses implicating metabolic disease in the pathogenesis of medically unexplained fatigue and pain.