Advances in medical imaging have allowed for detailed assessments of musculoskeletal structural properties. Although MRI is often considered the gold standard method for these purposes, constraints relating to MRI access and affordability have led researchers and clinicians to use other imaging methods. Two methods that have gained recent popularity are peripheral quantitative computed tomography (pQCT) and ultrasound imaging. The advantages of these methods include their lower cost and faster analysis speed while providing both direct and indirect quantification of multiple musculoskeletal structural properties.
pQCT is commonly used to quantify bone architecture; however, measures of muscle size, adipose tissue content, and muscle quality can also be obtained. Muscle quality is assessed by quantification of either intra- and intermuscular adipose tissue (IMAT) area or muscle density. As an important distinction, muscle density is primarily influenced by IMAT area, but it is also affected by the compactness of muscle fibers (i.e., muscle fibers occupying a given area) and the amount of protein and other soft tissue structures, including tendons, blood vessels, aponeuroses, and fascia, within muscles (1). On the other hand, ultrasound imaging has been used to measure muscle thickness, muscle fiber pennation angle, and length, as well as the signal intensity of echoes reflected by soft tissue structures within muscle; i.e., echo intensity. Muscle thickness is often considered a valid measure of muscle size (2), whereas echo intensity is thought to reflect the quantity of adipose and possibly fibrotic tissue within (and between) muscles and is often regarded as a useful measure of muscle quality (3). The measurement of muscle size, adiposity, and muscle quality are clinically important as they are strongly related to disease severity and functional capacity in several clinical populations (4,5) and of functional relevance to both healthy (6) and athletic individuals (7). Given the possible significant personal and financial ramifications of inaccurate or unreliable measurements of musculoskeletal structural properties, detailed examinations of the validity and reliability of such technologies are necessary.
One factor that might affect pQCT and ultrasound measurement reliability is the level of muscle blood flow or swelling subsequent to physical activity or resulting from disease processes. Any sustained increase in blood flow will directly increase muscle cross-sectional area (CSA). However, intra- or extracellular fluid shifts, resulting from osmotic or hydrostatic pressure gradients triggered by either the accumulation of metabolic end products (8) or the inflammatory processes (9), will also affect CSA measurements. These effects are highly problematic in both research and clinical contexts because muscle CSA is an important factor influencing physical function, is a primary measure indicating sarcopenia or cachexia, and is readily lost in aging and disease (10).
However, muscle blood flow or swelling might also theoretically affect intermuscular adipose estimates obtained using pQCT (e.g., IMAT and muscle density) or ultrasound (e.g., echo intensity) imaging techniques. With regard to pQCT, iron and other elements in the blood may affect the absorption or scatter of x-rays; i.e., x-ray attenuation, and thus possibly affect IMAT area and muscle density estimates calculated from the attenuation data (11). On the other hand, increased blood flow or swelling can affect ultrasound echo intensity by increasing intramuscular pressure and, thus, muscle fiber (fascicle) angulation; i.e., pennation (12), resulting in a decreased intensity of echoes being reflected to the transducer. Supporting this explanation is the fact that the strength of a reflected ultrasound echo is greatest when the acoustic direction is perpendicular to muscle fibers (13). With that being known, increased muscle blood flow or swelling might be expected to reduce echo intensity by altering muscle fiber pennation and, therefore, result in an underestimation of intermuscular adipose or fibrosis and a misrepresentation of the quality of muscle. Nonetheless, in some inflammatory conditions, such as muscle wasting diseases (14) or after high-intensity exercise (9,15), microdamage to the muscle, resulting in muscle swelling, can increase the reflection of sound waves by generating more structural “mirrors” and, therefore, increase echo intensity. In this case, the higher echo intensity would lead to an overestimation of intermuscular adipose or fibrosis. Of course, in various conditions, the phenomena may occur either simultaneously or sequentially, making their effects on echo intensity difficult to determine. On the basis of these arguments, conditions leading to changes in muscle blood flow or swelling, including those emanating from inflammatory conditions (16), may critically affect both pQCT- and ultrasound-based measurements of muscle size, adiposity, and muscle quality. Despite this, the effect of changes in muscle blood flow and swelling on measurement accuracy has remained relatively unexplored.
One occasion in which increased muscle blood flow and swelling occur is in response to resistance exercise (17,18). Resistance exercise causes immediate swelling, but it may also induce a prolonged increase in muscle swelling associated with inflammatory responses and muscle microstructural damage (9,19–21). This reaction provides a rational model in which to study the effects of muscle blood flow and fluid shift into the muscle, and it is relevant given that pQCT and ultrasound are commonly used in exercise-related studies on muscle function and health.
Therefore, the primary purpose of the study was to investigate the effect of a single bout of resistance exercise on muscle and adipose properties measured using pQCT and ultrasound imaging. A secondary purpose was to determine whether any changes in pQCT and ultrasound estimates of muscle and adipose were correlated after the resistance exercise; if so, this would indicate a possible common underlying cause (i.e., muscle swelling affects the accuracy of both techniques similarly). The final purpose was to determine the precision (i.e., repeatability) error and 95% least significant change (LSC) scores for these properties in the nonexercised contralateral limb to determine the measurement error, and whether the changes in these properties in the exercised limb were greater than the LSC. It was hypothesized that pQCT estimates of CSA and muscle area would be increased 24, 48, and 72 h after the resistance exercise, and these changes would be related to changes in pQCT estimates of IMAT area and muscle density. It was also hypothesized that ultrasound estimates of muscle thickness and echo intensity would be increased after the resistance exercise, and these changes would be related to changes in pQCT estimates of muscle and adipose tissues.
Seventeen male participants volunteered to participate in the study (age, 24 ± 4 yr; height, 180.0 ± 6.3 cm; body mass, 81.8 ± 16.9 kg). All participants had previous resistance training experience to ensure they were familiar with exercising to muscular failure. However, they were unaccustomed to performing biceps curl training with the preacher bench. The participants were free from upper body musculoskeletal injury in the last 3 yr and were not taking any anti-inflammatory or anabolic medications. Participants were advised not to undertake any upper body resistance training for 3 d before the first session and throughout the study period. They were also asked to maintain their normal diet and to maintain hydration. All participants provided written informed consent, and the study was approved by the University Human Research Ethics Committee (no. 16826).
Each participant attended the laboratory on four consecutive days. In the first session (PRE), height and body mass were measured (Seca 763; Seca, Birmingham, UK) before single transverse scans of both upper arms were obtained using pQCT and ultrasound imaging. The assessment of body mass was used as an indirect marker of hydration status throughout the experiment. The participant then performed a bout of unilateral dumbbell biceps curls on an adjustable preacher bench (Sorinex, Columbia, SC). Measurements were repeated 24 h (POST-24), 48 h (POST-48), and 72 h (POST-72) after the bout (±2 h) without any further exercise being performed.
Muscle and adipose properties were measured from a single slice of both upper arms immediately before and 24, 48, and 72 h after a bout of resistance exercise using pQCT (XCT-3000; Stratec Medizintechnik, Pforzheim, Germany). The participant lay supine on a table, which was placed next to the pQCT gantry, and a single arm was placed in an abducted position through the gantry. A customized mount was secured to the pQCT enabling an acrylic support, which had a layer of bubble wrap surrounding it, to be placed through the gantry’s aperture to support the participant’s arm (22). The participant was secured to the support using a Velcro strap while they gently held a tennis ball in their hand. The participant was asked to maintain this position for ~5–10 min before the scans commenced to help limit the effects of posture-induced fluid shifting (23). A 30-mm proximally moving scout view scan was performed slightly distal to the radial head of the radius so that a reference line could be positioned at the proximal endplate of the radial head. The x-ray source then moved proximally from this reference line to 33% of the humeral segment length to undertake the single slice CT scan. The humeral segment length was estimated as 0.186 × standing height (measured at PRE) and used in all other sessions (24). Immediately before the scans, the participant was advised to remain motionless. The scanning time for both the scout view and CT scans was ~7 min. The slice thickness was 2.4 mm, the tube voltage was 61 kV, and the voxel size and the scanner speed were set at 0.4 mm and 20 mm·s−1, respectively. Quality assurance scans were performed using the manufacturer’s standard phantom each morning before testing and with the cone phantom every 30 d.
All images were analyzed with an installed macro into the Stratec software that used edge detection and threshold techniques to differentiate the tissues (25). Image filters were applied to allow contour algorithms to detect the edges of different tissues more precisely (26). To calculate CSA and muscle area, skin, subcutaneous fat, and bone areas were subtracted from both; however, muscle “CSA” included IMAT area whereas muscle “area” excluded IMAT area. Before activating the macro, a region of interest (ROI) was created around the cross-sectional image of the upper arm, which excluded the acrylic support that supported the participant’s arm. The bubble wrap around the acrylic support provided sufficient space for this to be done accurately.
After the image was analyzed, the participant was instructed to maintain their position as the researcher restarted the system. This procedure reactivated the laser, which was kept at the spot as the previously completed scan over the mid–upper arm. Using a pen, a line was marked on the participant’s skin directly over the area of the laser to indicate the point of placement of the ultrasound transducer in the upcoming tests. The system was then aborted before another scan was activated. The participant was advised to maintain the integrity of the marked line over the 4-d study period, and the researcher checked the scanning position in the subsequent sessions to make sure the same region had been scanned. This process was then repeated for the contralateral upper arm. To quantify the soft and hard tissue properties and the movement of the arm in image capture, the database file was opened in a customized Excel spreadsheet. All analyzed images had movement artifacts below 25% of the cortical bone area (27).
Ultrasound images were obtained immediately after pQCT scanning using two-dimensional B-mode ultrasonography (Aloka SSD-α10; Aloka Co., Ltd., Tokyo, Japan) with a linear array transducer (45 mm, Model UST-5412, 13 MHz). All system settings were kept constant throughout the study, including having the dynamic image compression switched off. Images were obtained with the participant lying supine with their arm in a relaxed and supported abducted position; i.e., closely resembling the position used in the pQCT scanning. Scanning commenced after the participant had rested in the supine position for 10 min to help control for posture-induced fluid shifting (23). Before positioning the ultrasound transducer over the muscle, water-soluble gel was placed on the skin directly over the line marked during the pQCT testing. Care was taken to minimize the pressure applied to the skin during measurements. The transducer was placed perpendicular to the long axis of the humerus to align the apex of the humerus and superficial biceps brachii fascia in the middle of the image. A minimum of 12 transverse images were captured of each upper arm for post hoc analysis; however, three images of each were selected for analysis based on several considerations: the distance of the skin interface to the top of the image, the echogenic quality of the sides of the image due to contact of the sides of the probe on the skin, and the alignment of the humerus and superficial biceps brachii fascia in the middle of the image. A single representative image of each upper arm from the first testing session was printed and placed alongside the computer monitor so that anatomical landmarks could be recreated in subsequent testing sessions to ensure that the probe position and orientation were as similar as possible.
Muscle thickness and echo intensity were determined from the ultrasound images using ImageJ software (28). Muscle thickness was measured as the distance between the uppermost part of the humerus and the most superficial point of the biceps brachii fascia using the straight-line function. Muscle thickness was then calculated as the average result of the three selected images. Echo intensity was measured by creating an ROI, which was saved and then used on the three images selected at each testing session, meaning different ROI of varying sizes were used across the testing sessions. The ROI were created using the polygon function and were positioned ~3 mm from both the humerus and the superficial biceps brachii fascia and ~5 mm from the sides of the images. Echo intensity was then calculated as the average of the mean histogram result.
Resistance exercise bout
A single bout of unilateral elbow flexion resistance exercise was performed after the completion of preintervention measurements. The resistance exercise consisted of five sets of 8–12 repetitions of unilateral free weight bicep curls using a single dumbbell on an adjustable preacher bench (Sorinex). The elbow flexor muscles were selected to be exercised as they have been shown to generate more muscle swelling after high-intensity resistance exercise than other large upper and lower limb muscles (29). The angle of the preacher bench was set at 80°, and the bench height and seat were adjusted to allow the participant to have their feet on the ground, and their axilla and back of the upper arm to be in contact with the bench. Before performing the exercise, the participant’s handgrip strength was tested using a handheld dynamometer (Jamar, Sammons Preston Rolyan, Bolingbrook, IL) with elbow angle of 90° and wrist in a neutral position for 3 s. After a 30-s rest, the participant alternated between arms until each had completed three trials. The arm that generated the highest force (kg) in a single trial was determined to be the force-dominant upper limb and was thus selected as the exercise limb. The participant was also asked before commencing the exercise what dumbbell load they believed they could lift with failure occurring at 8–12 repetitions. After the completion of the first set, the dumbbell load was continually decreased by ~5%–10% in the remaining sets to assist the participant in completing the sets within the proposed repetition range (30). If a participant underestimated the load and performed more than two repetitions than that intended in the first set; e.g., ≥15, the load was not changed in subsequent sets until the participant failed to achieve a minimum of 14 repetitions. The range of repetitions performed within each set are shown in Table 1. The participant started each set by lowering the dumbbell until near full elbow extension, and the set was completed when the participant could not return the dumbbell to full elbow flexion. The nonexercising contralateral limb was positioned behind the participant’s back. The participant completed the sets to failure with 3-s eccentric and 2-s concentric phases and a 2-min passive interset rest. A metronome was set at 60 bpm to provide temporal feedback. The nonexercised contralateral limb served as the control.
Normality of data was confirmed using the Shapiro–Wilk testing and from inspection of histogram plots. A one-way repeated-measures ANOVA (time [PRE, POST-24, POST-48, and POST-72]) was used to assess the participant’s body mass. A two-way repeated-measures ANOVA (condition [exercised vs nonexercised limb] × time [PRE, POST-24, POST-48, and POST-72]) was used to test the effects of exercise and recovery on the dependent variables. Sphericity was assessed using Mauchly’s test, and on occasion when this assumption was violated, the degrees of freedom and the F-statistic were corrected using Greenhouse–Geisser correction when the estimated epsilon (ɛ) was <0.75 or the Huynd–Feldt correction when the estimated ɛ was >0.75. When significant condition–time interactions were observed, separate one-way repeated-measures ANOVA with Holm–Bonferroni sequential correction adjustments (for each condition and time point) were undertaken to correct for possible type I errors (31). Effect size is reported using Hedges g and interpreted as trivial (g < 0.2), small (0.2 ≤ g < 0.5), moderate (0.5 ≤ g < 0.8), and large (g ≥ 0.8) (32). Furthermore, the 95% LSC score (i.e., the smallest percentage change that can be considered statistically significant with 95% confidence interval [CI]) was calculated from the results obtained across the four testing sessions in the nonexercised contralateral limb by multiplying the precision error, i.e., root-mean-square coefficient of variation percentage (CV%RMS), by 2.77 (26). Intraclass correlation coefficients (ICC) were also calculated in the nonexercised limb to establish the reproducibility of the dependent variables. Finally, to determine the relationships between dependent variables for which statistical significance was obtained, Pearson’s product moment correlations (r) with 95% CI using bias-corrected accelerated bootstrapping between change results were computed at PRE to POST-24 (effect of exercise) and POST-24 to POST-72 (effect of recovery). Significant relationships were reported when the 95% CI did not cross 0.00. The alpha level was set at 0.05 for all other analyzes, and data are presented as mean ± SD. Statistical computations were performed using a statistical analysis program (version 24.0; SPSS Inc., Chicago, IL).
Post hoc power analysis
A post hoc power analysis was conducted with the direct condition–time interaction IMAT area partial η2 result (ηp2 = 0.327) from the present study, and the alpha level was set at 0.05 with a total sample size of 34 (both conditions [exercised and nonexercised limbs] had 17 samples). The analysis demonstrated that the achieved power (1-β err prob) was 0.999 (G*Power 220.127.116.11 software).
The participants’ body masses (kg) across the experimental period were 81.8 ± 16.9, 81.5 ± 16.7, 81.3 ± 16.8, and 81.4 ± 16.8 kg at PRE, POST-24, POST-48, and POST-72, respectively. There was no significant effect of time (F3, 48 = 1.81, P = 0.158) for changes in body mass.
Training loads during the resistance exercise bout
The training loads and repetitions completed during the unilateral resistance exercise bout are shown in Table 1.
There was no time effect (F3, 48 = 2.00, P = 0.127), but both condition (F1, 19 = 9.97, P = 0.006) and condition–time interaction (F3, 48 = 3.263, P = 0.029) effects were detected for changes in muscle CSA (mm2). Significant increases in muscle CSA in the exercised limb were observed from PRE to POST-24 (Δ = 251.5 ± 290.3 mm2, P = 0.015, g = 0.27), POST-48 (Δ = 249.7 ± 328.9 mm2, P = 0.024, g = 0.27), and POST-72 (Δ = 226.2 ± 220.9 mm2, P = 0.006, g = 0.26). Nonetheless, the percentage change in muscle CSA compared with PRE did not reach the LSC score (7.8%) for any trial comparison. There were no significant differences in muscle CSA in the nonexercised limb.
There were significant effects of condition (F1, 16 = 33.64, P < 0.001) and time (F3, 48 = 8.32, P < 0.001) and an interaction effect (F1.61, 25.76 = 10.61, P = 0.001) for changes in muscle area (mm2). Significant increases in muscle area in the exercised limb were observed from PRE to POST-24 (Δ = 82.3 ± 55.3 mm2, P < 0.001, g = 0.13), POST-48 (Δ = 128.3 ± 128.2 mm2, P = 0.004, g = 0.19), and POST-72 (Δ = 156.8 ± 137.6 mm2, P < 0.001, g = 0.24) (see Figure, Supplemental Digital Content 1, Sampling distribution for the changes in muscle area from PRE in the exercised limb following the unilateral resistance exercise bout, https://links.lww.com/MSS/B484). The percentage change in muscle area compared with PRE was greater than the LSC score (LSC = 2.0%) at POST-24 (2.2%), POST-48 (3.5%), and POST-72 (4.3%). The percentage change from POST-24 to POST-72 (2.0%) also reached the LSC score. There were no significant differences in muscle area in the nonexercised limb.
There were significant effects of condition (F1, 16 = 43.174, P < 0.001) and time (F2.19, 34.10 = 4.71, P = 0.013), and an interaction effect (F2.13, 34.10 = 7.78, P = 0.001), for changes in IMAT area (mm2). Significant increases in IMAT area in the exercised limb were observed from PRE to POST-24 (Δ = 44.2 ± 34.9 mm2, P < 0.001, g = 0.16), POST-48 (Δ = 67.3 ± 75.4 mm2, P = 0.008, g = 0.23), and POST-72 (Δ = 73.2 ± 77.3 mm2, P = 0.005, g = 0.27) (see Figure, Supplemental Digital Content 2, Sampling distribution for the changes in IMAT area from PRE in the exercised limb following the unilateral resistance exercise bout, https://links.lww.com/MSS/B485). The percentage change in IMAT area compared with PRE was greater than the LSC score (LSC = 4.1%) at POST-24 (4.4%), POST-48 (6.7%), and POST-72 (7.3%). There were no significant differences in IMAT area in the nonexercised limb.
There was no significant effect of condition (F1, 16 = 0.656, P = 0.430), although the time effect for muscle density changes (mg·cm−3) approached significance (F2.27, 36.27 = 3.09, P = 0.052), but there was a significant interaction effect (F3, 48 = 5.309, P = 0.003). Significant decreases in muscle density in the exercised limb were observed from PRE to POST-24 (Δ = −0.54 ± 0.74 mg·cm−3, P = 0.032, g = 0.66), POST-48 (Δ = −0.62 ± 0.74 mg·cm−3, P = 0.018, g = 0.80), and POST-72 (Δ = −0.83 ± 1.07 mg·cm−3, P = 0.030, g = 0.94). The percentage change in muscle density compared with PRE did reach the LSC score (LSC = 1.0%) at POST-72 (−1.0%). There were no significant differences in muscle density in the nonexercised limb.
Subcutaneous adipose tissue area
There were no significant effects of condition (F1, 16 = 0.98, P = 0.336) or time (F3, 48 = 0.40, P = 0.751) or an interaction effect (F3, 48 = 1.91, P = 0.106) for changes in the subcutaneous adipose tissue area (mm2). Within the exercised limb, the percentage change in subcutaneous adipose tissue area compared with PRE did not reach the LSC score (21.0%) for any trial comparison. pQCT results in the exercised and nonexercised limbs are shown in Table 2.
There were significant effects of condition (F1, 16 = 38.46, P < 0.001) and time (F3, 48 = 17.80, P < 0.001) and an interaction effect (F3, 48 = 21.11, P = 0.001) for changes in muscle thickness (mm). Significant increases in muscle thickness in the exercised limb were observed from PRE to POST-24 (Δ = 2.35 ± 1.26 mm, P < 0.001, g = 0.55), POST-48 (Δ = 2.91 ± 1.78 mm, P < 0.001, g = 0.68), and POST-72 (Δ = 2.42 ± 2.05 mm, P < 0.001, g = 0.58). The percentage change in muscle thickness compared with PRE was greater than the LSC score (LSC = 1.2%) at POST-24 (6.3%), POST-48 (8.0%), and POST-72 (6.6%). The percentage change from POST-24 to POST-48 (1.6%) was also greater than the LSC score. There were no significant differences in muscle thickness in the nonexercised limb.
There was no significant effect of condition (F1, 16 = 2.90, P = 0.108), but both time (F3, 48 = 11.78, P < 0.001) and condition–time interaction (F1.76, 28.14 = 4.07, P = 0.032) effects were detected for changes in echo intensity (8-bit gray scale: au). Significant increases in echo intensity in the exercised limb were observed from PRE to POST-72 (Δ = 9.39 ± 12.60 au, P = 0.030, g = 0.94), POST-24 to POST-48 (Δ = 6.79 ± 8.87 au, P = 0.030, g = 0.71), and POST-72 (Δ = 10.00 ± 11.24 au, P = 0.012, g = 0.96). The percentage change in echo intensity compared with PRE was greater than the LSC score (LSC = 5.7%) at POST-48 (6.4%) and POST-72 (9.7%). The percentage change from POST-24 to POST-48 (7.1%) and POST-72 (10.5%) was also greater than the LSC score. There were no significant differences in echo intensity in the nonexercised limb. Ultrasound results in the exercised and nonexercised limbs are shown in Table 3. Figure 1 highlights the changes in pQCT and ultrasound images from before to 24, 48, and 72 h after resistance exercise in one participant. pQCT and ultrasound precision error, LSC, and ICC results determined in the nonexercised limb are shown in Table 4.
Relationships between Changes in Dependent Variables
Effect of exercise (PRE to POST-24)
Significant correlations were observed between changes in muscle CSA and IMAT area (r = 0.53, 95% CI = 0.31–0.74), muscle CSA and muscle density (r = −0.88, 95% CI = −0.98 to −0.35), muscle area and echo intensity (r = 0.41, 95% CI = 0.07–0.68), and IMAT area and muscle density (r = −0.45, 95% CI = −0.69 to −0.06) when changes were assessed from before to 24 h postexercise.
Effect of recovery (POST-24 to POST-72)
Significant correlations were observed between changes in muscle CSA and muscle thickness (r = 0.61, 95% CI = 0.18–0.96), muscle area and echo intensity (r = 0.59, 95% CI = 0.05–0.83), muscle area and IMAT area (r = 0.72, 95% CI = 0.26–0.92), IMAT area and echo intensity (r = 0.83, 95% CI = 0.56–0.93), IMAT area and muscle thickness (r = 0.69, 95% CI = 0.24–0.91), IMAT area and muscle density (r = −0.65, 95% CI = −0.91 to −0.19), muscle density and echo intensity (r = −0.69, 95% CI = −0.92 to −0.01), and echo intensity and muscle thickness (r = 0.80, 95% CI = 0.42–0.94) when changes were assessed from 24 to 72 h postexercise. pQCT and ultrasound correlation results in the exercised limb are shown in Table 5.
The present results demonstrate that pQCT-derived CSA and muscle area (Table 2) as well as ultrasound-derived muscle thickness (Table 3) can be increased significantly up to at least 72 h after a single bout of unaccustomed resistance exercise. Although these findings have important implications, perhaps even more clinically important is that estimates of pQCT-derived IMAT area and muscle density as well as ultrasound-derived echo intensity were also affected by an acute bout of resistance exercise. During the first 24 h (PRE to POST-24), the change in muscle area was significantly correlated with the change in echo intensity. In addition, the changes in muscle CSA and IMAT area were significantly correlated with the change in muscle thickness, and the changes in muscle area, IMAT area, and muscle density were significantly correlated with the change in echo intensity during the recovery period (POST-24 to POST-72) (Table 5). These associations suggest that a link exists between these results, providing evidence that swelling subsequent to physical activity affects the reliability of pQCT- and ultrasound-based muscle and adipose estimates.
Although the current results are consistent with previous findings showing an increase in both muscle thickness and echo intensity for 72 h after a bout of resistance exercise (15), they are inconsistent with at least one study reporting that muscle CSA returned to preexercise values and was not statistically different to the control condition, at 75 min after high-intensity resistance exercise (33). Immediately after resistance exercise, plasma volume decreases, indicating that fluid has shifted from the vascular space into active muscles in response to osmotic or hydrostatic pressure changes occurring locally within the active muscle (18). This fluid movement in addition to the increased blood flow likely caused the immediate increase in muscle CSA. However, plasma volume is restored within 60 min after resistance exercise and is closely mirrored by the return of muscle CSA (8,18). Therefore, immediate pressure-driven fluid shifts and vasodilatory effects might be expected to cause muscle swelling and affect estimates of muscle for only a short period after resistance exercise. Thus, reports of prolonged swelling and inflated estimates of muscle size across several days must be caused by different mechanisms.
Other researchers have reported that inflammation can affect muscle CSA over a more extended period after unaccustomed resistance exercise. After an initial inflammatory response that supports the delivery of leukocytes to the damaged area (34), a secondary inflammatory response may be developed because of the generation of reactive oxygen species (35) and the activation of calpain-mediated proteases (36), which magnify the existing damage. Consequently, increased muscle blood flow is needed to cope with the microstructural remodeling resulting from intracellular Ca2+ accumulation caused by the loss of sarcoplasmic reticulum permeability control, damage to the sarcolemma membrane, opening of stretch-responsive channels, or changes in triad and t-tubule arrangement enabling Ca2+ entry via voltage-sensitive channels (37–41). Although ultrasound echo intensity is typically used to determine the extent of IMAT content or fibrosis in muscle tissue; i.e., muscle quality, it is proposed that the delayed increase in echo intensity observed after resistance exercise in the current study is caused by microstructural damage to the muscle due to processes activated by intracellular Ca2+ accumulation. Muscle damage is likely to increase echo intensity by providing more reflective structural “mirrors” for the transmitted sound waves to be projected back to the transducer. However, as no direct measurement of muscle damage was undertaken, the suggestion that muscle damage was generated in the present study can only be speculated. In saying that, the generation of muscle swelling after intense exercise may offer a partial explanation for the prolonged increases in CSA in the current and other studies. The findings that inflammation contributes to the acute resistance training–induced increases in muscle CSA (9,21) and that residual muscle swelling has been observed in muscle fibers removed by muscle biopsy 48 h after repetitive submaximal eccentric exercise (42) support the suggestion that exercise-induced muscle swelling can affect CSA for a prolonged period after resistance exercise. An important additional consideration is that inflammatory processes are also prominent in many clinical conditions, so errors in CSA and muscle area are also likely to be common in studies examining those populations.
Perhaps even more important from a clinical perspective is that pQCT estimates of IMAT area and muscle density were also critically affected by the resistance exercise. That is, the IMAT area and muscle density results do not represent “true” increases in adipose accumulation. A likely explanation for the unexpected decrease in muscle density is the sudden increase in IMAT area. With fat tissue being less dense than other soft tissue structures, and IMAT explaining ~50% of the variance in muscle density estimates (1), it is reasonable to assume that the decreased muscle density resulted directly from the increase in IMAT area. The significant negative relationship between changes in muscle density and IMAT area also reinforces this supposition (Table 5). However, an alteration in tissue x-ray attenuation coefficients is likely to explain the increase in IMAT area. pQCT systems emit x-rays that are then scattered, absorbed, or transmitted through tissues with different attenuations. Differences in x-ray attenuation between tissues enable them to be delineated. Typically, tissues that are thicker, denser, and have a higher atomic number absorb more x-ray energy and, thus, show a higher attenuation coefficient. Elemental iron, of which ~75% of whole-body stores are found in the blood, is a potent x-ray absorber. It is known that muscle blood flow and volume are increased in the days after even submaximal eccentric muscle exercise (17). If a similar response was generated in the current study, the resulting microvascular perfusion in the exercised muscles may have affected x-ray transmission and affected the estimation of IMAT area and, hence, muscle density (11). As changes in muscle blood flow were not directly measured, future research is required to explicitly test this hypothesis.
Collectively, the current findings have important clinical implications as these measurements aid in the diagnosis of metabolic dysfunction, including cardiovascular diseases and diabetes (43,44). pQCT testing should be avoided in the days after intense or unaccustomed exercise because muscle swelling, likely resulting from muscle blood flow and inflammation-dependent fluid shifts, may critically affect muscle size, IMAT, and density (i.e., quality) estimates. Also, if inflammatory responses to exercise cause errors in measurements, then errors may also exist in measurements in patients with muscular, or systemic, inflammation. This is an important consideration as, to a large extent, pQCT is used to monitor the musculoskeletal health of individuals with these exact conditions. The current data provide some evidence for the future direct assessment of the relationship between muscle blood flow and errors in pQCT-derived CSA, IMAT, and density estimates, which might allow for the development of correction equations.
Some key points need to be discussed regarding the implications of the results of the current study. First, the magnitude of the observed effects are possibly specific to the elbow flexor muscles as lower body muscles tend to perform more activities of daily living, which may attenuate exercise-induced muscle swelling (29). Future research is needed to confirm this theory as lower body sites are commonly used in clinical settings to assess pQCT- and ultrasound-based muscle and adipose properties. Second, the results may not necessarily apply to longitudinal studies as the performance of repeated bouts of exercise over time is likely to attenuate muscle damage and, hence, muscle swelling (20,21,45). Further research is needed to determine whether adipose estimates continue to be affected after repeated bouts of resistance exercise and, if so, whether the reduction in error is related to the minimization of muscle damage. Third, the discrepancy between the magnitudes of the changes in pQCT- and ultrasound-based muscle size estimates is likely attributable to differences in the muscles included in images. The pQCT generates two-dimensional cross-sectional images, and thus both elbow flexor and extensor muscles are included when scanning the upper arm. In the present study, changes in the trained elbow flexor muscles are therefore assessed within the same images as the untrained elbow extensor muscles. Because of the relatively low resolution of pQCT scans, it is difficult to select the elbow flexor muscles alone for analysis. Thus, a lesser percentage increase in muscle size is observed using pQCT as opposed to ultrasound scanning, where the exercised elbow flexor muscles can be assessed in isolation. Hence, direct comparisons between the percentage changes in pQCT- and ultrasound-based muscle size estimates cannot be made.
The final purpose of the current study was to determine the precision error and LSC score of pQCT- and ultrasound-based muscle and adipose tissue estimates in the nonexercised contralateral limb, and whether the changes in these estimates in the exercised limb were greater than the LSC (i.e., measurement error). The measurement errors obtained in the nonexercised limb demonstrate that pQCT-measured adipose estimates and other estimates that include adipose tissue; e.g., muscle CSA, are less accurate than estimates of lean muscle (Table 4). The measurement of adiposity is an important diagnostic tool, so it is vital that researchers and clinicians are aware of measurement imprecisions to ensure that decisions relating to a person’s health or disease status are reflective of “real” outcomes rather than being influenced by analytical or biological variations. Interestingly, the precision errors in pQCT muscle area, muscle density, and IMAT area measurements in young male adults in the present study (CV%RMS; 0.7, 0.3, and 1.5, respectively) were lower than the reported errors of the same variables in postmenopausal women using the same analysis method (25). The upper arm was examined in the current study, whereas Frank-Wilson et al. (2015) studied the lower leg (CV%RMS; 2.6, 0.7, and 3.3) and forearm (CV%RMS; 5.3, 1.4, and 7.0). The possibility exists, therefore, that the reliability of pQCT-measured muscle and the adipose estimates are specific to the muscle group tested. On the other hand, the percentage changes in muscle area, IMAT area, and muscle density in the exercised limb (Table 2) were greater than the LSC scores obtained from the nonexercised limb (Table 4), signifying that the changes were statistically larger than the measurement error. These statistically meaningful changes in the exercised limb determined by assessing the errors in the nonexercised limb corroborate the finding that the changes were significantly different when inferences were made by analyzing their variance (i.e., ANOVA). However, this outcome was not observed for changes in muscle CSA. The percentage change in muscle CSA in the exercised limb did not reach or surpass the LSC, although a statistical increase was observed using ANOVA. This conflicting result suggests that the measurement error calculated by testing the nonexercised contralateral limb over four consecutive days may provide a more accurate indication of the reliability of the pQCT, and in doing so provides greater confidence that changes reaching or surpassing these values are “real” and not influenced by extraneous factors. To the best of the authors’ knowledge, the present study is the first to examine precision error, LSC, and ICC results of pQCT- and ultrasound-based muscle and adipose estimates using an upper arm model.
In conclusion, a single bout of resistance exercise affected the accuracy of pQCT-derived muscle CSA and muscle area measurements as well as ultrasound-derived muscle thickness for at least 72 h. Of significant clinical importance is that pQCT-derived IMAT area and muscle density as well as ultrasound-derived echo intensity were similarly affected by the exercise, suggesting that exercise-induced muscle swelling, likely caused by muscle blood flow and inflammation-dependent fluid shifts in muscle, affects the reliability of pQCT- and ultrasound-based muscle and adipose tissue estimates. The correlational findings support the theory that there is a link between these results, indicating that muscle swelling can affect both pQCT- and ultrasound-based estimates of muscle size, adiposity, and muscle quality. These estimate errors indicate that researchers and clinicians who take these measurements as indicators of metabolic disease or muscle function may draw incorrect conclusions relating to health or disease status if the measurements were taken in the days after strenuous physical activity or when testing individuals with (muscle) inflammatory conditions. Theoretically, these estimate errors may persist beyond 72 h as they did not return to baseline before the end of the study period.
The authors acknowledge all the participants of this study. They also acknowledge Jethro Nagle for his assistance conducting the experiments and Daniel J. Schiferl for his expert advice on collection and analysis of pQCT data. Lastly, they acknowledge the staff and researchers of the Exercise Medicine Research Institute at Edith Cowan University for their help and support.
No funds were received to complete this work. The authors declare that they have no conflict of interest. All procedures performed herein were in accordance with the ethical standards of the institutional research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.
The present study does not constitute endorsement by the American College of Sports Medicine. The results of this study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.
1. Wong AK, Beattie KA, Min KK, et al. Peripheral quantitative computed tomography-derived muscle density and peripheral magnetic resonance imaging-derived muscle adiposity: precision and associations with fragility fractures in women. J Musculoskelet Neuronal Interact
2. Abe T, DeHoyos DV, Pollock ML, Garzarella L. Time course for strength and muscle thickness changes following upper and lower body resistance training in men and women. Eur J Appl Physiol
3. Reimers K, Reimers CD, Wagner S, Paetzke I, Pongratz DE. Skeletal muscle sonography: a correlative study of echogenicity and morphology. J Ultrasound Med
4. Marden FA, Connolly AM, Siegel MJ, Rubin DA. Compositional analysis of muscle in boys with Duchenne muscular dystrophy using MR imaging. Skeletal Radiol
5. Kramer HR, Fontaine KR, Bathon JM, Giles JT. Muscle density in rheumatoid arthritis: associations with disease features and functional outcomes. Arthritis Rheum
6. Visser M, Kritchevsky SB, Goodpaster BH, et al. Leg muscle mass and composition in relation to lower extremity performance in men and women aged 70 to 79: the health, aging and body composition study. J Am Geriatr Soc
7. Blazevich AJ, Gill ND, Bronks R, Newton RU. Training-specific muscle architecture adaptation after 5-wk training in athletes. Med Sci Sports Exerc
8. Collins MA, Hill DW, Cureton KJ, DeMello JJ. Plasma volume change during heavy-resistance weight lifting. Eur J Appl Physiol Occup Physiol
9. Damas F, Phillips SM, Lixandrão ME, et al. Early resistance training-induced increases in muscle cross-sectional area are concomitant with edema-induced muscle swelling
. Eur J Appl Physiol
10. Visser M, Goodpaster BH, Kritchevsky SB, et al. Muscle mass, muscle strength, and muscle fat infiltration as predictors of incident mobility limitations in well-functioning older persons. J Gerontol A Biol Sci Med Sci
11. Rittweger J, Möller K, Bareille M-P, Felsenberg D, Zange J. Muscle x-ray attenuation is not decreased during experimental bed rest. Muscle Nerve
12. Vieira A, Blazevich A, Souza N, et al. Acute changes in muscle thickness and pennation angle in response to work-matched concentric and eccentric isokinetic exercise. Appl Physiol Nutr Metab
13. Madaras EI, Perez J, Sobel BE, Mottley JG, Miller JG. Anisotropy of the ultrasonic backscatter of myocardial tissue: II. Measurements in vivo. J Acoust Soc Am
14. Heckmatt JZ, Leeman S, Dubowitz V. Ultrasound imaging in the diagnosis of muscle disease. J Pediatr
15. Radaelli R, Bottaro M, Wilhelm EN, Wagner DR, Pinto RS. Time course of strength and echo intensity recovery after resistance exercise in women. J Strength Cond Res
16. Pober JS, Sessa WC. Inflammation and the blood microvascular system. Cold Spring Harb Perspect Biol
17. Selkow NM, Herman DC, Liu Z, Hertel J, Hart JM, Saliba SA. Blood flow after exercise-induced muscle damage. J Athl Train
18. Ploutz-Snyder LL, Convertino VA, Dudley GA. Resistance exercise-induced fluid shifts: change in active muscle size and plasma volume. Am J Physiol
19. Nosaka K, Clarkson PM. Changes in indicators of inflammation after eccentric exercise of the elbow flexors. Med Sci Sports Exerc
20. Damas F, Phillips SM, Libardi CA, et al. Resistance training-induced changes in integrated myofibrillar protein synthesis are related to hypertrophy only after attenuation of muscle damage. J Physiol
21. Damas F, Libardi CA, Ugrinowitsch C. The development of skeletal muscle hypertrophy through resistance training: the role of muscle damage and muscle protein synthesis. Eur J Appl Physiol
22. Weatherholt AM, Avin KG, Hurd AL, et al. Peripheral quantitative computed tomography predicts humeral diaphysis torsional mechanical properties with good short-term precision. J Clin Densitom
23. Hagan RD, Diaz FJ, Horvath SM. Plasma volume changes with movement to supine and standing positions. J Appl Physiol
24. Sievänen H, Koskue V, Rauhio A, Kannus P, Heinonen A, Vuori I. Peripheral quantitative computed tomography in human long bones: evaluation of in vitro and in vivo precision. J Bone Miner Res
25. Schiferl DJ. Fat vs. Lean—Analysis Manual 7.4.2 ed; 2017. Spring Branch, TX. [cited 20 Jun 2017]. Available from: http://www.bonediagnostic.com
26. Frank-Wilson AW, Johnston JD, Olszynski WP, Kontulainen SA. Measurement of muscle and fat in postmenopausal women: precision of previously reported pQCT imaging methods. Bone
27. Blew RM, Lee VR, Farr JN, Schiferl DJ, Going SB. Standardizing evaluation of pQCT image quality in the presence of subject movement: qualitative versus quantitative assessment. Calcif Tissue Int
28. Schneider CA, Rasband WS, Eliceiri KW. NIH image to ImageJ: 25 years of image analysis. Nat Methods
29. Chen TC, Lin K-Y, Chen H-L, Lin M-J, Nosaka K. Comparison in eccentric exercise-induced muscle damage among four limb muscles. Eur J Appl Physiol
30. Medeiros HS Jr, Mello RS, Amorim MZ, Koch AJ, Machado M. Planned intensity reduction to maintain repetitions within recommended hypertrophy range. Int J Sports Physiol Perform
31. Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat
32. Cohen J. Statistical power analysis for the behavioural sciences. Hillside. NJ: Lawrence Earlbaum Associates 1988. 567 p.
33. Freitas ED, Poole C, Miller R, et al. Time course change in muscle swelling
: high-intensity vs. blood flow restriction exercise. Int J Sports Med
34. McCully KK, Faulkner JA. Injury to skeletal muscle fibers of mice following lengthening contractions. J Appl Physiol (1985)
35. Hellsten Y, Frandsen U, Orthenblad N, Sjødin B, Richter EA. Xanthine oxidase in human skeletal muscle following eccentric exercise: a role in inflammation. J Physiol
36. Belcastro AN, Shewchuk LD, Raj DA. Exercise-induced muscle injury: a calpain hypothesis. Mol Cell Biochem
37. Lynch GS, Fary CJ, Williams DA. Quantitative measurement of resting skeletal muscle [Ca2+]i following acute and long-term downhill running exercise in mice. Cell Calcium
38. Lowe DA, Warren GL, Hayes DA, Farmer MA, Armstrong RB. Eccentric contraction-induced injury of mouse soleus muscle: effect of varying [Ca2+]o. J Appl Physiol
39. Armstrong RB, Duan C, Delp MD, Hayes DA, Glenn GM, Allen GD. Elevations in rat soleus muscle [Ca2+] with passive stretch. J Appl Physiol
40. Warren GL, Hayes DA, Lowe DA, Armstrong RB. Mechanical factors in the initiation of eccentric contraction-induced injury in rat soleus muscle. J Physiol
41. Duan C, Delp MD, Hayes DA, Delp PD, Armstrong RB. Rat skeletal muscle mitochondrial [Ca2+] and injury from downhill walking. J Appl Physiol
42. Fridén J, Sfakianos PN, Hargens AR, Akeson WH. Residual muscular swelling after repetitive eccentric contractions. J Orthop Res
43. Yim JE, Heshka S, Albu J, et al. Intermuscular adipose tissue rivals visceral adipose tissue in independent associations with cardiovascular risk. Int J Obes (Lond)
44. Goodpaster BH, Krishnaswami S, Resnick H, et al. Association between regional adipose tissue distribution and both Type 2 Diabetes and impaired glucose tolerance in elderly men and women. Diabetes Care
45. McHugh MP. Recent advances in the understanding of the repeated bout effect: the protective effect against muscle damage from a single bout of eccentric exercise. Scand J Med Sci Sports