Forty-seven participants with a physician diagnosis of FM (44 female, mean ± SD: 51.3 ± 12.3y; range: 24-75y; BMI: 30.2 ± 6.9) and 47 control (CON) participants (44 female, mean ± SD: 52.5 ± 14.7y; range: 20-74y; BMI: 27.7 ± 5.6) completed this study. Exclusion criteria included known orthopedic, cardiopulmonary, neurological, or unstable medical conditions that would preclude performance of fatiguing exercise or experimental techniques. All participants were screened with the Physical Activity Readiness Questionnaire (ACSM 2018) to verify safety prior to engaging in physical activity. Informed consent was acquired before study initiation and the protocol was approved by the Marquette University Institutional Review Board (HR-3035) according to principles of the Declaration of Helsinki.

Participants completed three sessions, one familiarization that included baseline assessments and two randomized experimental sessions separated by approximately one-week (Figure 1). During the familiarization session, participants were familiarized to pressure algometry and the custom-made pressure pain device twice, first to a remote site and subsequently to the site used during experimental sessions to mimic the study protocol. The FM participants completed the modified 2010 American College of Rheumatology Modified Diagnostic Criteria for Fibromyalgia (ACR) (Wolfe et al. 2011; Wolfe et al. 2010b) and Revised Fibromyalgia Impact Questionnaire (FIQR) (Burckhardt et al. 1991). All participants performed maximal voluntary isometric contractions (MVICs) of the right elbow flexors and voluntary activation of the biceps brachii with the twitch interpolation technique (Gandevia 2001). All participants were familiarized to performing isometric and concentric contractions with the right elbow flexors. Participants were issued an ActiGraph activity monitor (ActiGraph wGT3X-BT, Pensacola, FL) to quantify physical activity for a seven-day period. During the first experimental session, dual-energy x-ray absorptiometry (DXA) (GE Lunar iDXA, Madison, WI) was completed and each participant completed the Physical Activity Assessment Tool (PAAT) (Meriwether et al. 2006). During both randomized experimental sessions perceived fatigue was measured using the PROMIS Short Form v1.0 – Fatigue 7a (PROMIS Fatigue) (Ameringer et al. 2016; Lai et al. 2011). Experimental pain perception, right elbow flexor MVIC, and left handgrip MVIC were measured before and after submaximal intermittent isometric and concentric contractions performed with the right elbow flexors. Self-reported pain and rating of perceived exertion were measured every minute during performance of isometric and concentric contractions.

Submaximal intermittent isometric and concentric muscle contractions were matched for intensity (20% of maximal voluntary isometric contraction), duration (10-minutes), and duty-cycle (2-s contraction: 1-s relaxation) Each participant matched a target force line indicating 20% of MVIC on a computer monitor placed 1 m from the participant. Verbal encouragement was provided throughout both exercise tasks.

Pressure pain thresholds (PPTs) were measured before and after exercise with a pressure algometer (Somedic Algometer Type II, Hörby, Sweden) locally at the exercising limb (right biceps) and systemically in the right quadriceps. Two trials separated by a 10-second intertrial interval were performed at each site with a 1-cm² rubber tip using a ramp protocol increasing applied pressure at a rate of 50 kPa/sec. The average of the two assessments at each site were used for data analysis. PPTs were recorded with participants seated upright in a Biodex System 3 Pro (Biodex Medical Systems, Inc., Shirley, New York) with the upper extremity supported in 40° shoulder flexion and 15° elbow flexion, bilateral feet were rested on a footrest with hips and knees at 90° flexion. The right biceps site was marked at 2/3 distance between the anterior border of the acromion to the superior border of the cubital fossa and the right quadriceps site was marked midway between the anterior superior iliac spine and the mid-patella.

A custom-made pressure pain device (Romus, Inc, Milwaukee WI) was used to measure pain perception with isometric and concentric muscle contractions (Hoeger Bement et al. 2011; Hoeger Bement et al. 2009). A weighted Lucite edge was placed on the left index finger for 2-minutes, the equivalent of 1-kg mass was applied to controls while a 0.75-kg mass was applied to participants with FM. A lesser mass was used with FM participants to facilitate completion of the two-minute trial despite their increased pain sensitivity (Hoeger Bement et al. 2011). Participants reported “pain” when they first perceived pain (i.e., pain threshold) and the time in seconds was recorded. Pain ratings were reported every 20-seconds using a 0 to 10 numerical pain rating scale. Summation of pain during the constant noxious stimulus was calculated as the difference from the last pain rating (120 sec) to the first (0 sec).

During both exercise tasks, participants were asked to rate exercising arm pain and perceived exertion upon initiation, every minute during, and at completion of the 10-minute exercise task. Perceived pain was measured with a 0 to 10 numerical rating scale (NRS) with anchors of 0 = no pain; 5 = moderate pain; and 10 = worst pain (Farrar et al. 2001; Turk et al. 1993). Perceived exertion was measured with the modified Borg Rating of Perceived Exertion (RPE) scale with anchors 0 = “nothing at all” and 10 = “very very strong” (Borg 1998).

Self-reported arm and whole-body pain was rated before and after the 10-minute exercise bout with a 0-10 cm visual analogue scale (VAS) with anchors of no pain and worst pain (Turk et al. 1993). Arm and whole-body pain were assessed in reference to performing gross limb and whole-body movement such as mimicking reaching forward a picking up a cup and walking and squatting to pick something up from the floor respectively.

Maximal voluntary isometric contractions (MVIC) were performed by each participant with the right elbow flexors while seated in a Biodex System 3 PRO. The participant’s forearm was placed in a modified forearm orthosis attached to the dynamometer. Each participant was seated with their hips and knees flexed to 90° flexion, right shoulder in 40° flexion, right elbow in 90° flexion, and a neutral forearm position. The setup and positioning for maximal voluntary isometric testing was maintained throughout sessions. Participants were verbally encouraged to contract as strong as they could and build as much force for 3-5 seconds with visual feedback of torque production on a computer monitor placed 1m from the participant. Torque recordings from the dynamometer were recorded online and digitized using a Power 1401 analog-to-digital converter and Spike 2 software [Cambridge Electronics Design (CED), Cambridge, UK] at 500 samples per second. Participants were familiarized to MVICs during the familiarization session and two MVICs were performed prior to isometric and concentric muscle contractions to identify a load equivalent of 20% of MVIC. MVICs were performed after isometric and concentric muscle contractions to measure the local exercise-induced decrease in muscle force production (i.e., fatigue of the exercising muscle). Post-exercise MVICs were conducted approximately five minutes following exercise to facilitate completion of experimental and clinical pain assessments.

Voluntary activation was assessed using the interpolated stimulus technique (Gandevia 2001) to evaluate the participants’ ability to maximally activate their biceps muscle and generate torque during performance of MVICs. This technique involved superimposing an evoked contraction with electrical stimulation over the biceps brachii during performance of the MVICs. The biceps brachii was stimulated with a paired stimulus (i.e. twitch stimulation) by a constant-current stimulator (Digitimer, DS7AH, Welwyn Garden City, UK) with surface electrodes [30 x 22 mm] (Ambu Neuroline 715, Columbia, MD). Intensity of twitch stimulations were increased by 50mA until there was a plateau in force and no further increase in evoked force with two consecutive stimulations. The stimulation intensity was increased a further 20% to provide supramaximal stimulations during the torque plateau of each MVIC (superimposed twitch) and then while the muscle was relaxed ~3-seconds after cessation of the MVIC (potentiated twitch). To determine voluntary activation using the twitch interpolation technique (Gandevia 2001), the increase in elbow flexion torque evoked by a stimulation during MVIC (superimposed twitch) was expressed as a fraction of the torque amplitude during stimulation while the biceps muscle was relaxed (potentiated twitch) and quantified as a percentage using the following formula: [(1 – superimposed twitch/potentiated twitch) × 100]. The potentiated twitch was further analyzed for peak amplitude and half-relaxation time/rate (Gandevia 2001).

Left handgrip MVIC was assessed before and after isometric and concentric muscle contractions with a handgrip dynamometer (JAMAR Hydraulic Hand Dynamometer, Sammons Preston, Bolingbrook, IL). Participants had their left upper extremity unsupported and positioned with their left shoulder in neutral, elbow flexed to 90°, forearm and wrist in neutral, and elbow tucked against the body. Participants were verbally encouraged to squeeze as strong as possible for 3-5 seconds. Maximal contractions were performed prior to and upon completion of each exercise task to measure systemic fatigue following the exercise task.

Participants wore activity monitors (ActiGraph wGT3X-BT, Pensacola, FL) on the non-dominant wrist for seven days to quantify the percent of time spent in sedentary, light activity, and moderate-to-vigorous activity. Placement on the wrist has shown to increase wear compliance (Freedson & John, 2013; Martin & Hakim, 2011; Troiano et al., 2014; van Hees et al., 2011) and correlate with physical activity measured at the hip (Kamada et al., 2016; Scott et al., 2017). Daily logs tracking sleep time, physical activity, and any removal time were completed by participants. Data were downloaded and analyzed using ActiLife software (ActiLife 6.13.1, Pensacola, FL). Non-wear and sleep time were identified and removed using the Troiano algorithm (Choi et al. 2012) and daily logs to calculate amount of physical activity time. Data from four validated days (wear time of at least 10 hours) were used for analysis as four days has been shown to be representative of a week’s average of physical activity (Migueles et al. 2017). Percent of time spent in sedentary, light, and moderate-to-vigorous activity were estimated using cut points based on Freedson’s algorithm and the worn-on wrist option was selected to mathematically depress counts (Freedson 1998; Rosenberger 2013). Self-reported physical activity was measured with the Physical Activity Assessment Tool (PAAT) (Meriwether et al. 2006) and time spent in moderate to vigorous intensity activity was reported.

Body composition was quantified using a GE Lunar iDXA (GE, Madison, WI). Participants were instructed to refrain from eating and drinking 1-2 hours prior to scanning. Scans were analyzed using Encore software (GE, Madison, WI) and measures of total lean mass (kg), total fat mass (kg), right arm lean mass (kg), and right arm fat mass (kg) were obtained. Because exercise was performed with the right elbow flexors, right arm fat mass and lean mass were used as an indicator of regional body composition.

All data were analyzed using IBM SPSS (Version 26, IBM, Armonk, NY, USA). Normality and linearity were evaluated with the Shapiro-Wilk test and visual inspection via Q-Q plots. Data are reported as mean ± SD in text and tables and displayed as mean ± SEM in figures. Differences between means were tested with paired-samples and independent samples t-test. Repeated measures analysis of variance was used to compare the following variables across session and time with between subject factor of group: elbow MVIC, handgrip MVIC, experimental pain threshold, pain and RPE during exercise, and the change in arm and whole-body pain after exercise. Pressure pain thresholds were analyzed with a repeated measures analysis of variance with variables of session, time, site, and between subject factor of group. Pain summation was analyzed with a repeated measures analysis of variance with variables of session, time, summation, and between subject factor of group. Post hoc tests were applied where appropriate. Pearson product moment correlations were calculated to determine associations between variables. A P-value ≤ 0.02 was used for statistical significance.

Descriptive statistics for the study sample are listed in Table 1. Control (CON) and fibromyalgia (FM) groups were matched for age (CON: 52.5 ± 14.7; FM: 51.3 ± 12.3), sex (CON: 44 Female (F), 3 Male (M); FM: 44 F, 3 M), and BMI (CON: 27.7 ± 5.6; FM: 30.2 ± 6.9) with both groups falling in the overweight to obese category. Both groups had similar total lean mass, total fat mass, right arm lean mass, and right arm fat mass. Both groups self-reported similar amount of moderate-to-vigorous activity minutes per week via the PAAT. Each group spent similar amount of time in sedentary, light, and moderate-to-vigorous activity as measured by the ActiGraph. The FM group had a mean symptom severity score on the 2010 ACR Diagnostic Criteria for Fibromyalgia at time of enrollment was 16.6 ± 5.8. The FM group had a mean total FIQR score of 48.5 ± 19.7. Neither group reported differences in perceived fatigue with the PROMIS Short Form Fatigue 7a across isometric and concentric experimental sessions (p > 0.05) therefore, values from the isometric session were used to evaluate differences between groups and correlations to performance fatigue. The FM group reported greater perceived fatigue on the PROMIS Short Form Fatigue 7a (p < 0.001) compared to controls.

Baseline Voluntary Activation, Muscle Twitch Properties, & Elbow Flexor Strength: Participants with FM had similar activation properties as the control participants that included voluntary activation of the right biceps, baseline twitch amplitude, and half-relaxation time (Table 1). MVIC torque of the right elbow flexors was also similar at the beginning of all sessions and between CON and FM participants (Table 2, Figure 2a).

Right Elbow Flexor Performance Fatigability: Right elbow flexor MVIC torque decreased following exercise (i.e. performance fatigability) (Time (pre- to post-exercise): F(1,92) = 42.47, p < 0.001, η_p² = 0.316) with a mean decline of 7.2% (Figure 2a, Table 2). The performance fatigability differed by group (Time (pre- to post-exercise) x Group: F(1,92) = 5.88, p = 0.017, η_p² = 0.060), with post hoc analysis showing people with FM showing greater declines in MVIC torque (performance fatigability) than CON (CON: −4.0 ± 6.7%, FM: −9.8 ± 13.8%; t(66.44) = 2.56, p = 0.013). The reduction in MVIC torque was similar after the isometric and concentric exercise (Contraction (Iso v Conc) x Time: F(1,92) = 1.53, p = 0.219).

Left Handgrip Strength and Performance Fatigability: Baseline handgrip MVIC was similar across sessions and between CON and FM participants (p > 0.05) (Table 2, Figure 2b). Left handgrip MVIC decreased following exercise (Time (pre- to post-exercise) F(1,92) = 53.04, p < 0.001, η_p² = 0.368) with a mean decline of 6.5 ± 10.2%. The change in left handgrip MVIC was not different between exercise types (Contraction (Iso v Conc) x Time (pre- to post-exercise): F(1,92) = 2.17, p = 0.144) or groups (Group x Time (pre- to post-exercise): F(1,92) = 1.61, p = 0.208).

Perceived arm pain increased during isometric and concentric exercise (Time (during exercise): F(1.71, 157.7) = 154.67, p < 0.001, η_p² = 0.627) and arm pain ratings differed by exercise type (Contraction (Iso v Conc) x Time (during exercise): F(2.75,253.05) = 21.3, p < 0.001, η_p² = 0.188) with post hoc demonstrating greater change in arm pain with concentric (mean change = 5.0 ± 3.6) compared to isometric (mean change = 3.3 ± 3.3) (t(93) = −6.34, p < 0.001) (Table 2, Figure 3a). Arm pain differed by group (Group (CON v FM) X Time (during exercise): F(1.71,157.7) = 17.85, p < 0.001, η_p² = 0.162); post hoc showed a greater increase in arm pain for FM (mean change = 5.8 ± 2.8) compared to CON (mean change = 2.6 ± 2.7) (t(92) = −5.63, p < 0.001). A significant interaction of Contraction (Isometric v Concentric) x Time (during Ex) x Group (CON v FM) (F(2.75,253.05) = 7.10, p < 0.001, η_p² = 0.072) was demonstrated for arm pain during exercise. Post hoc demonstrates the change in arm pain in FM during concentric exercise (6.3 ± 3.3) was greater than the change in arm pain in FM during isometric (5.2 ± 2.9, t(46) = −3.07, p = 0.004), CON during concentric (3.7 ± 3.4, t(92) = 3.84, p < 0.001), and CON during isometric (1.4 ± 2.5, t(85.6) = 8.19, p < 0.001). The change in arm pain in FM during isometric exercise (5.2 ± 2.9) was greater than the change in arm pain in CON during concentric (3.7 ± 3.4, t(92) = 2.35, p = 0.02) and CON during isometric (1.4 ± 2.5, t(92) = 6.77, p < 0.001). The CON reported greater arm pain during concentric (3.7 ± 3.4) compared to isometric (1.4 ± 2.5) (t(46) = −6.03, p < 0.001).

Rating of perceived exertion increased during exercise (Time (during exercise): F(2.63,241.82) = 327.66, p < 0.001, η_p² = 0.781) and differed by exercise type (Contraction (Iso v Conc) x Time (during exercise): F(3.25, 22.56) = 7.88, p < 0.001, η_p² = 0.079) with post hoc demonstrating greater change in RPE with concentric (mean change = 6.5 ± 5.5) compared to isometric (mean change = 5.5 ± 2.8) (t(93) = −3.63, p < 0.001) (Table 2, Figure 3b). RPE was greater for FM than CON (Group X Time (during Ex); F(2.63, 241.82) = 5.89, p = 0.001, η_p² = 0.060 with post hoc showing a greater increase in RPE for FM (mean change = 6.7 ± 2.5) compared to CON (mean change = 5.2 ± 2.0) (t(92) = −3.15, p = 0.002).

Self-reported arm pain relative to movement was assessed immediately before and after the exercise protocol increased (Time: F(1, 92) = 88.49, p < 0.001, η_p² = 0.490), which was greater for participants with FM (Group (CON v FM) x Time (pre- to post-exercise): F(1, 92) = 33.38, p < 0.001, η_p² = 0.266) (Table 2). Post-hoc analysis reveals that participants with FM reported a greater increase in arm pain following exercise (CON: 0.7 ± 1.3, FM: 2.9 ± 2.3; t(71.74) = −5.78, p < 0.001). Exercise type also led to differences in arm pain following exercise (Contraction (Iso v Conc) x Time (pre- to post-exercise): F(1,92) = 11.31, p = 0.001, η_p² = 0.110) with greater increases following concentric contractions than the isometric exercise (Iso: 1.4 ± 2.0, Conc: 2.2 ± 2.9; t(93) = −3.35, p = 0.001).

Self-reported whole-body pain relative to movement increased following exercise (Time (pre- to post-exercise): F(1, 92) = 10.53, p = 0.001, η_p² = 0.103) however, post-hoc analysis showed no increase (p> 0.05) (Table 2). The change in whole body pain was similar across exercise types (Time (pre- to post-exercise) x Contraction (Iso v Conc): F(1,92) = .467, p = 0.496) and groups (Time (pre- to post-exercise) x Group: F(1,92) = 5.03, p = 0.027).

Control participants reported higher baseline PPTs compared with FM participants at the biceps and quadriceps across all three sessions (p<0.01) (Figure 4a & 4b). Baseline PPTs were similar across sessions for both groups (p>0.05). Higher PPTs were assessed at the quadriceps than the biceps (Site (bicep v quad): F(1,92) = 164.82, p < 0.001, η_p² = 0.642). Following exercise, the change in PPTs differed by site (Site (bicep v quad) x Time (pre- to post-exercise: F(1,92) = 6.41, p = 0.013, η_p² = 0.065) with post hoc analysis showing PPT at the biceps increased following exercise (Pre: 205.5 ± 100.3, Post: 219.0 ± 109.3, p = 0.004) and remained unchanged at the quadriceps (Pre: 310.4 ± 139.5, Post: 306.5 ± 146.1, p = 0.370). The change in PPTs were not different between groups (Group x Time (pre- to post-exercise): F(1,92) = 1.92, p = 0.169) or exercise type (Contraction (Iso v Conc) x Time: F(1,92) = 0.007, p = 0.933). The change in PPTs (absolute and relative) at the biceps and quadriceps were not correlated with any of the self-report pain assessments (change in arm and whole-body pain, or pain during exercise).

Baseline (pre-exercise) pain threshold with the two-minute pressure pain device was similar between groups and consistent across all three sessions (Table 2). There were no changes in pain threshold following performance of isometric and concentric exercise (Time: F(1,87) = 1.37, p = 0.246) and change in pain threshold was not related to exercise type (Time (pre- to post-exercise) x Contraction (Iso v Conc): F(1,87) = 2.37, p = 0.127) or groups (Time (pre- to post-exercise) x Group: F(1,87) = 1.15, p = 0.286) (Figure 5a). Summation of pain from the constant noxious stimulus occurred during all assessments in CON and FM (Summation: (0 to 2-min): F(1.88, 173.10) = 309.53, p < 0.001, η_p² = 0.771) (Table 2 & Figure 5b). Control and FM participants demonstrated similar baseline (pre-exercise) pain summation across sessions (CON: 6.9±3.0, 6.3±3.1, 6.4±3.0; FM: 7.3±2.9, 7.3±2.8, 7.2±3.1). (Figure 5b). Summation of pain decreased following exercise (Summation X Time (pre- to post-exercise): F(3.30, 303.65) = 5.88, p < 0.001, η_p² = 0.060) with post hoc demonstrating a reduction in pain summation from pre-exercise (6.8 ± 2.9) to post-exercise (6.5 ± 2.9) (t(93) = 2.52, p = 0.013). The change in summation did not differ by exercise type (Contraction x Time x Summation: F(3.86, 352.0) = 1.43, p = 0.227) or between groups (Group x Time x Summation: F(3.30, 303.65) = 1.10, p = 0.351). The change in pain summation (absolute and relative) was not correlated with any self-report pain assessments (change in whole-body pain or pain during exercise).

Correlations to performance fatigability for isometric and concentric contractions were combined due to lack of task specificity for elbow flexor and handgrip fatigability. Correlations between baseline clinical pain (ACR, FIQR, whole-body pain VAS) did not correlate with performance fatigability of the right elbow flexors (ACR: r = −0.141, p = 0.345; FIQR: r = 0.009, p = 0.951; whole-body pain VAS: r = 0.041, p = 0.783) or left handgrip (ACR: r = −0.289, p = 0.049; FIQR: r = −0.227, p = 0.124; whole-body pain VAS: r = −0.161, p = 0.280). Baseline perceived fatigue (PROMIS-Fatigue) did not correlate with performance fatigability at the right elbow flexors (r = −0.011, p = 0.942) or left handgrip (r = −0.108, p = 0.468). Self-reported pain during exercise (NPRS) did not correlate with performance fatigability at the elbow flexors (r = −0.088, p = 0.558) or handgrip (r = −0.065, p = 0.664).

Fibromyalgia symptom severity may have fluctuated throughout the three weeks of study participation as prior research demonstrates people with FM experience fluctuating intensity of symptomology (Harris et al. 2005; Wolfe et al. 2011) which may have influenced experimental and clinical assessments. However, our study design included assessment of pain and fatigue before each exercise session and differences were not seen between sessions. Experimental pain was evaluated immediately after each exercise bout, though further research needs to investigate whether longer time durations are needed to capture systemic changes in pressure pain sensitivity following resistance exercise. Post-exercise assessment of elbow flexor and handgrip MVIC were performed after completion of all post-exercise clinical and experimental pain assessments. Post-exercise assessment of elbow flexor and handgrip MVIC were performed after completion of all post-exercise clinical and experimental pain assessments. Therefore, 5 minutes separated the completion of the exercise task and reassessment of MVICs which likely resulted in some recovery of elbow flexor and handgrip strength. The magnitude of exercise-induced reductions in force production may be underestimated in this study. Despite this limitation, both groups experienced exercise-induced decreases in force production. Further research should investigate whether exercise-induced reductions in muscle force persist for longer duration in people with FM.