2.1. Study design and ethics
A randomized, controlled crossover trial with two exercise conditions was performed. At least 72 hours after a familiarization session, each participant visited the laboratory on 5 consecutive days. On day 1, eccentric and concentric exercise (randomized order and arm allocation) of the elbow flexors was performed using a dumbbell. Outcome measures (see below) were assessed immediately before (base), immediately after (post 0) and every 24 hours up to 96- hours post-exercise (post 24, 48, 72, and 96). The study was approved by the Edith Cowan University Human Research Ethics Committee, and adhered to the Guidelines of Good Clinical Practice as well as the Declaration of Helsinki. All participants provided a written informed consent prior to the study participation.
2.2. Participants
Eleven (seven male and four female) healthy active students (24.3 ± 1.9 y; 167 ± 6 cm; 67.2 ± 8.1 kg; 293 ± 145 min sporting activity per week) volunteered to participate in the present study. The Edinburgh Handedness Inventory – Short Form [23] showed that all participants were right-handed. None of the participants had performed regular upper arm resistance training in the preceding six months. All individuals were healthy, not reporting any orthopedic, cardiovascular, neurological, endocrine or psychiatric diseases. Intake of anti-inflammatory or analgesic medicine was prohibited during or on the days before the experiment.
2.3. Determination of one-repetition maximum (1-RM)
Immediately before the exercise session (day 2), the one repetition maximum (1-RM) was measured using dumbbell curls (one arm for eccentric exercise, and the other arm for concentric exercise) according to the randomization (e.g., left arm eccentric, right arm concentric). This was done because a first bout of eccentric exercise could protect against muscle damage in subsequent bouts of exercise [16, 24]. After a warm-up with a light (50 % of the estimated 1-RM) dumbbell (10 elbow flexion and extension movements), the participants performed either one isolated eccentric (lowering the dumbbell) or concentric (lifting the dumbbell) contraction with a load of approximately 80 % of the estimated 1-RM. For the 1-RM assessment, each participant sat on a preacher arm-curl bench to stabilize their shoulder. For the eccentric 1-RM, the participant was instructed to lower the dumbbell from 90° elbow flexion to full extension in three seconds, and for the concentric 1-RM, the participant lifted the dumbbell from full extension to 90° flexion in three seconds. After each attempt, a one-minute rest was applied, and the procedure was repeated increasing the weight based on the perceived effort of the participant until a failed attempt occurred. All participants required less than five attempts for each arm to determine the 1-RM.
2.4. Exercise protocols
The participants performed 6 sets of 10 repetitions of eccentric contractions for one arm and concentric contractions of the elbow flexors for the other arm. Repetition duration was 3 seconds and the inter-set interval was 2 minutes. The exercises were performed with a dumbbell, using 80% of the individual 1-RM [25]. After each repetition, the investigator assisted in returning the dumbbell to its starting position to ensure eccentric-only or concentric-only contractions were performed with the load [26]. If necessary, the investigator carefully supported the participant to complete the protocol. After each set, the participants were asked to rate the effort of the exercise using the rating of perceived exertion (CR-10) scale ranging from 1 anchoring “extremely easy,” to 10, “extremely hard” [27].
2.5. Outcome measures
The following measurements were taken from the exercised arm before (base), immediately after (excluding muscle pain measures) and 24, 48, 72 and 96 hours after exercise. To avoid possible influences of circadian rhythm, examinations were performed at approximately the same time of the day for each participant.
2.5.1. Maximal voluntary isometric contraction (MVIC) torque
MVIC torque was measured using an isokinetic dynamometer (Biodex system 3, Biodex medical, Shirley, NY, USA). Data were recorded and stored via the LabChart Pro software (v8.1.5, ADInstruments) to a laptop computer. Each participant was positioned in supine on a therapy bed having the elbow joint aligned with the dynamometer’s axis of rotation (Fig. 1). The elbow joint of the examined arm was positioned at 90° flexion, and both shoulders were stabilized on the therapy bed. Following a brief warm-up, each participant was instructed to flex the elbow joint as fast and strong as possible against the stationary lever arm for 3 seconds. Verbal encouragement was provided to elicit maximal effort for each attempt. Two attempts with a one-minute rest were performed. The average peak force, which could be maintained for 1 second, was considered as the MVIC torque and the mean of both measurements was used for further analysis [28, 29].
2.5.2. Pressure pain threshold (PPT)
PPT of the biceps brachii of the exercised arm was measured using an electronic algometer (Somedic AB, Sollentuna, Sweden). Each participant lay on a therapy bed in a relaxed supine position with the forearm being slightly elevated by a soft pad. The probe of the algometer (diameter of 1 cm) was placed on the biceps brachii muscle belly (9 cm above the elbow crease), and compressive force was gradually increased at a rate of approximately 50 kPa/s until the participant reported the first sensation of pain. The value (kPa) corresponding to the force applied at this moment was recorded. After a 10-second rest interval, the procedure was repeated twice, and the average of the two measures was used for further analysis.
2.5.3. Palpation pain
Muscle pain upon palpation of the elbow flexors was assessed by a 100-mm visual analogue scale (VAS), anchoring 0 being no pain and 100 being worst possible pain [30, 31]. Each participant was in the same position as that of PPT, and the measurement was taken from the same place as that of PPT (9 cm above the elbow crease). The investigator palpated the muscle of each participant with two fingers in longitudinal orientation [32]. The participants were asked to indicate the level of perceived pain by marking it on the VAS. Pressure was always applied by the same investigator and kept as constant as possible (approximately 400 kPa) between days and participants [28]. The measurement site was marked on the skin with a semi-permanent marker and renewed on a daily basis.
2.5.4. Fascia thickness
A high-resolution ultrasound (US) device (Aloka ProSound F75, Hitachi Healthcare, Tokyo, Japan) with an 8.0 x 1.5 cm linear array transducer was used (frequency range of 7.5 MHz, display depth of 3.0 cm, dynamic range of 60 dB and image gain of 50) to measure fascia thickness. Each participant was positioned in supine with the examined arm being attached to the lever arm of the isokinetic dynamometer at 35° flexion (Fig. 1). The ultrasound probe was positioned longitudinally on the biceps muscle belly (with the center of the probe 9 cm above the elbow crease). The investigator ensured to stabilize the probe on the skin with minimal pressure in order to avoid compressing the tissue. Three static images were taken when the deep fascia was clearly recognizable as a hyperechoic region overlying the muscle. To ensure identical positioning during follow-up measurements, the US-transducer location was marked on the skin with a semi- permanent marker and renewed every day.
Fascia thickness was calculated using ImageJ (Image J 1.52 k software, USA). Within each US image, five regions of interest (ROIs) at equidistant points were selected for the thickness measurement (Fig. 2a). The average of the five ROIs in each of the three images was chosen to determine the fascia thickness.
2.5.5. Fascia and muscle displacement
High-resolution ultrasound imaging was also used to evaluate fascia and muscle displacement during passive movements of the elbow joint of the exercised arm. Each participant was positioned as explained for the MVIC torque measure. The linear array ultrasound transducer was positioned longitudinally on the belly of the biceps brachii with the center of the probe at 9 cm above the elbow crease, and fixed with elastic bandages and tape, ensuring the muscle was not compressed. The isokinetic dynamometer passively moved the elbow joint three times between 90° flexion and full extension at an angular velocity of 5°/s (33). Previous studies have shown that reflexive muscle contraction does not occur at this velocity [34, 35]. Participants were instructed to remain completely relaxed, avoiding any voluntary muscle activity. In order to familiarize the participants with the device and measurement conditions, a warm-up of three flexion-extension cycles was performed prior to the actual measurements [36, 37]. Video recordings, depicting fascia and muscle tissue, were made via ultrasound at 10 Hz during the passive movement. Additionally, elbow joint position [°], relative to the neutral zero position, was recorded from the dynamometer signal to the laptop computer. Ultrasound videos were cut into extension and flexion parts of the three repetitions using the synchronized joint position. If a participant could not reach full extension in the follow-up days due to increased muscle stiffness, all videos of this participant were cut to the lowest achieved extension angle.
The maximal horizontal displacement of biceps brachii fascia and muscle was quantified using a frame-by-frame cross correlation algorithm proposed by Dilley [38]. To determine tissue displacement, rectangular ROIs were selected in the ultrasound videos. The used software [38] computes the correlation coefficient between the pixel gray levels of the consecutive frames. The pixel shift with the highest correlation coefficient represents the tissue displacement between the two successive frames. Three equidistant ROIs were defined in the fascia starting 2.5 cm from the myotendinous junction. Additionally, three equidistant ROIs were defined in the muscle on the same level. Horizontal tissue displacement was calculated by calculating the mean of three ROIs and three repetitions for the corresponding tissue (Fig. 2b and 2c).
2.6. Statistical analyses
All data were tested for normal distribution using a Shapiro Wilk test and inference statistics were applied as appropriate. Baseline values were compared between conditions (eccentric vs. concentric) using a paired t-test or a Wilcoxon test. To examine the possible influence of exercise order (eccentric or concentric arm first), baseline values of participants starting with the respective conditions were compared by means of independent t tests or Mann-Whitney-U tests. As an additional check, we used point biserial correlation or contingency tables with Cramer-V to detect potential association of pre-post differences and treatment order.
Prior to the main analysis comparing between eccentric and concentric exercise for changes in the dependent variables over time, sphericity was checked by a Mauchly’s test and the normal distribution of the residuals was examined by a Shapiro-Wilk test. If requirements were met, we performed 2-factorial ANOVAs (condition * time). Resulting effect sizes (η2) were interpreted as small (0.01), medium (0.06) or large (0.14) according to Cohen (1988) [39]. In case of main effects for time, Bonferroni-Holm adjusted post hoc tests (paired t-test) were performed. Effect size interpretation followed the recommendations of Cohen (1988), distinguishing small (d = 0.2 to 0.5), medium (0.5 to 0.8) or large (0.8 or higher) effects [39]. In case of violations of the assumptions for parametric testing, Kubinger’s bifactorial rank variance analyses with Bonferroni-Holm corrected post-hoc Wilcoxon tests were computed. We also investigated possible correlations between fascia thickness, fascia displacement, muscle displacement and palpation pain by a Pearson or Spearman correlation analysis. According to Cohen (1980), correlation coefficients (CC) were interpreted a small (r = 0.1 to 0.3), medium (0.3 to 0.5) or large (0.5 and higher) [40].
All analyzes were performed with SPSS 26 (SPSS Inc., Chicago, Illinois, USA) and BiAs Statistics (version 11.10, Goethe University, Frankfurt am Main, Germany). The level of significance was set to α = 0.05.