Content uploaded by Zac Robinson
Author content
All content in this area was uploaded by Zac Robinson on Nov 12, 2023
Content may be subject to copyright.
Part of the Society for Transparency,
Openness and Replication in
Kinesiology (STORK)
Preprint
not peer reviewed
All authors have read and approved this version of the
manuscript. This article was last modified on November, 2023.
The Effect of
Resistance Training
Proximity to Failure on
Muscular Adaptations
and Longitudinal
Fatigue in Trained Men
Supplementary materials:
https://osf.io/58hy3/
For correspondence:
zrobinson2019@fau.edu
Zac P. Robinson1, Christian T. Macarilla1, Matthew C. Juber1, Rebecca M. Cerminaro1, Brian Benitez, 1
Joshua C. Pelland1, Jacob F. Remmert1, Thomas A. John1, Seth R. Hinson1, Shawn Dinh1, Ethan Elkins1,
Laura C. Canteri1, Caitlyn M. Meehan1, Eric R. Helms1,2, Michael C. Zourdos1
1 Department of Exercise Science and Health Promotion, Florida Atlantic University
2 Sport Performance Research Institute New Zealand (SPRINZ), Auckland University
of Technology, Auckland, New Zealand
Please cite as: Robinson ZP, Macarilla CT, Juber, MC, Cerminaro RM, Benitez B, Pelland JC, Remmert JF, John
TA, Hinson SR, Dihn S, Elkins E, Canteri LC, Meehan CM, Helms ER, Zourdos MC, (2023). The Effect of
Resistance Training Proximity to Failure on Muscular Adaptations and Longitudinal Fatigue in Trained Men.
1
ABSTRACT
Purpose:
This study examined the effect of proximity to failure on hypertrophy, strength, and
fatigue. We hypothesized strength gains would be superior in non-failure groups compared to
those that include sets to momentary failure, while hypertrophy would be similar in all groups.
Methods:
38 men were randomized into four groups (4–6 rating of perceived exertion-RPE per
set, 7–9 RPE per set, 7–9+ RPE [last set taken to momentary failure], and 10 RPE per set) and
completed an eight-week program. Back squat and bench press strength, muscle thickness,
subjective fatigue, muscle soreness, and biomarkers (creatine kinase-CK and lactate
dehydrogenase-LDH) were assessed.
Results:
Bench Press strength gains were comparable
between the 4–6 RPE (9.05 kg [95% CI: 6.36, 11.76]) and 7–9 RPE (9.72 kg [95% CI: 7.03, 12.42])
groups, while outcomes in the 7–9+ (5.07 kg [95% CI: 2.05, 8.1]) and 10 RPE (0.71 kg [95% CI: -
4.51, 5.54]) groups were slightly inferior. Squat strength gains were comparable between 4–6
RPE (13.79 kg [95% CI: 7.54, 19.92]) and 7–9 RPE (18.05 kg [95% CI: 12.28, 23.99]) groups, but
data for 7–9+ RPE and 10 RPE are difficult to interpret due to poor feasibility of the protocols.
For muscle hypertrophy, our data do not provide strong conclusions as to the effects of
proximity to failure due to the large variability observed. The indices of fatigue were largely
comparable between groups, without strong evidence of the repeated bout effect.
Conclusion:
These data suggest strength outcomes are comparable when taking sets to either a self-
reported 4–6 RPE or 7–9 RPE, while training that includes sets to momentary failure may result
in slightly inferior outcomes (i.e., 7–9+ and 10 RPE). However, the influence of proximity to
failure on hypertrophy remains unclear and our data did not reveal clear differences between
groups in any measure of fatigue.
Key Words:
RESISTANCE TRAINING, PROXIMITY TO FAILURE, MUSCULAR ADAPTATIONS,
LONGITUDINAL FATIGUE
Statements and Declarations
No funds, grants, or other support was received for this study.
Michael Zourdos, Eric Helms, Zac Robinson, Josh Pelland, and Jacob Remmert are all coaches
and writers in the fitness industry. All other authors declare that they have no conflicts of
interest relevant to the content of this review.
2
1 INTRODUCTION
The necessity of performing resistance training sets to failure to maximize muscle hypertrophy
and strength remains unclear. Recently, meta-analyses from Grgic et al. [1] and Vieira et al. [2]
found that increases in strength and muscle hypertrophy were not significantly different (p =
0.237–0.860, SMD = 0.01–0.59) between training to or not to failure when volume is equated.
However, these meta-analyses examine proximity to failure in binary (i.e., training to or not to
failure) rather than investigating dose-response relationship (i.e., how does training to a 1
repetition in reserve [RIR] influence strength and hypertrophy outcomes compared to 2 RIR,
and 3 RIR, etc.); thus, the minimum proximity to failure needed to maximize adaptations is
unknown. Additionally, training to failure has been shown to elongate acute recovery time
courses [3]; possibly compromising training performance. Thus, in addition to examining the
dose-response relationship between RIR with and training outcomes, more information is
needed regarding how a given proximity to failure influences recovery over time.
Recent studies [4–6] that have examined multiple submaximal proximities to failure reported
that non-failure training may lead to superior strength gains due to better maintenance of
barbell velocity. For example, Pareja-Blanco et al. [4] reported trivial to small effect sizes (d =
0.10–0.35) for Smith machine squat strength improvement over eight weeks in favor of trained
men who did not train to failure (i.e., 0, 10, and 20% velocity losses) versus a group that trained
with a high-velocity loss (i.e., 40% velocity loss) which likely resulted in most sets reaching
failure [7]. Importantly, the proximity to failure (i.e., repetitions in reserve-RIR) on each set was
not directly quantified; however, the 0, 10, and 20% velocity loss groups had faster average
concentric velocities (ACV) across all sets (0.74 ± 0.11, 0.71 ± 0.07, 0.64 ± 0.06 m·s-1,
respectively) compared to the 40% velocity loss group (0.58 ± 0.06 m·s-1). However, a limitation
of this study was that total repetition volume was not equated, with a range of 181–637% more
repetitions performed in the high-velocity loss group (i.e., 40% velocity loss). Thus, it is unclear
which factor (i.e., proximity to failure or repetition volume) mediate the differential strength
gains.
Regarding hypertrophy, most recent data have reported either comparable muscle growth
between conditions training to and not to failure [8–10], with a minority of studies favoring
training to [11,12], or not to failure [13]. However, despite the mixed findings of most studies,
3
similar to strength, proximity to failure has not been directly quantified making it difficult to
infer the RIR required to maximize hypertrophy. For example, using a within-participants
design, Santanielo et al. 2020 [10] reported no significant difference in vastus lateralis cross-
sectional area between limbs training to or not to failure over ten weeks on the unilateral leg
press and knee extension at 75–80% of 1RM. Further, Santanielo et al. [10] reported that the
participants performed 12.0 ± 2.1 repetitions per set when training to failure versus 10.4 ± 2.8
repetitions per set in the limb not training to failure; however, RIR was not directly quantified.
Carroll et al. [13] also reported effect sizes in favor of not training to failure for type I (g = 0.27)
and type II (g = 0.99) fiber hypertrophy of the vastus lateralis, but did not directly quantify the
number of RIR after each set. However, the prescription of the group not training to failure
suggests a wide range of per set RIR (estimated ~1–8 RIR); thus, the minimum proximity to
failure to maximize muscle hypertrophy remains unclear.
Although training to failure may be a viable option, a recent meta-analysis [3] has shown that
recovery time courses are elongated by 24–48 hours and are associated with greater session
rating of perceived exertion (sRPE) [8]. Consequently, if training a muscle group 2-3 times per
week, weekly volume load could be negatively affected. However, all studies included in the
Vieira et al. meta-analysis examined recovery time course following a single training session.
Therefore, it is unknown if the repeated bout effect (RBE) would manifest following chronic
exposure to failure training and the recovery differences would be inconsequential over time. If
the longer recovery time courses are not mitigated over time, then training to failure may lack
long-term feasibility [14].
Therefore, the purpose of this study was to compare muscle hypertrophy and strength
outcomes among trained men using four different volume-equated resistance training
protocols with a different number of self-reported RIR per set (i.e., 4-6 RIR, 1-3 RIR, 0-3 RIR, 0
RIR). A secondary aim was to assess the change in acute fatigue (i.e., indirect markers of muscle
damage, barbell velocity, and subjective fatigue scales) over the course of the eight-week
protocol. We hypothesized larger strength gains in both submaximal training groups compared
to the groups that included some sets to failure, but similar hypertrophy between groups.
2 METHODS
4
2.1 Participants. Thirty-eight males between the ages of 18–40 were recruited. For inclusion,
participants needed to have at least 2 years of resistance training experience as determined by
a physical activity questionnaire. Participants were required to have a 1RM squat of at least
1.25 times body mass, and 1RM bench press of at least 1 times body mass. Further,
participants with contraindications to exercise (i.e., heart disease, hypertension, diabetes, etc.)
as determined via a health history questionnaire were excluded. Prior to participation, all
participants provided written consent and the University’s Institutional Review Board approved
this investigation. Six participants dropped out of the study (n = 3 in 7–9+ RPE and n = 3 in 10
RPE) due to training related discomfort and/or injury. At the investigators’ discretion, the 10
RPE group was terminated due to safety concerns. Data of participants that completed the 10
RPE protocols was analyzed, but the small sample size warrants extreme caution. Similarly, the
lower body training protocol of the 7–9+ RPE group was altered mid-intervention (i.e.,
participants trained with the 4–6 RPE protocol for over half of the intervention) at the
investigator’s discretion. Therefore, the 7–9+ group was not analyzed for any lower body or
systemic outcomes; however all raw data is provided in supplementary file 0. Additionally, one
participant in the 4–6 RPE group was removed from lower body and systemic outcomes due to
training related discomfort that required significant modification to the squat training. Another
participant in the 4–6 RPE group was uncomfortable with blood draws and thus was removed
for systemic outcomes. Participant characteristics for all 4 groups can be seen in Table 1.
Characteristic!
4-6 RPE (n=10)!
7-9 RPE (n=10)!
7-9+ RPE (n=9)!
10 RPE (n=3)!
Age (years)!
22.50 ± 3.21!
23.30 ± 3.09!
21.33 ± 2.78!
21.67 ± 2.08!
Height (cm)!
177.51 ± 5.75!
174.95 ± 6.14!
173.19 ± 6.23!
168.03 ± 3.57!
Pre Body Mass (kg)!
81.33 ± 12.04!
82.05 ± 12.82!
78.24 ± 7.46!
78.68 ± 14.37!
Post Body Mass (kg)!
82.61 ± 12.19!
83.15 ± 13.14!
79.33 ± 9.05!
77.98 ± 15.49!
Δ Body Mass (kg)!
1.28 ± 1.64!
1.10 ± 0.69!
1.09 ± 2.87!
-0.70 ± 2.08!
Pre Sum of Skinfolds (mm)!
32.20 ± 8.98!
31.40 ± 11.16!
29.94 ± 9.52!
31.50 ± 12.49!
Post Sum of Skinfolds (mm)!
35.32 ± 7.18!
33.70 ± 13.51!
33.44 ± 9.77!
35.50 ± 14.00!
Δ Sum of Skinfolds (mm)!
3.12 ± 3.86!
2.30 ± 3.70!
3.50 ± 2.59!
4.00 ± 5.29!
Pre Estimated Body Fat (%)!
15.42 ± 3.41!
15.24 ± 4.60!
14.35 ± 3.18!
14.90 ± 5.34!
Post Estimated Body Fat (%)!
16.42 ± 3.02!
15.94 ± 5.31!
15.46 ± 3.36!
16.15 ± 5.94!
Δ Estimated Body Fat (%)!
1.00 ± 1.25!
0.70 ± 1.16!
1.11 ± 0.82!
1.25 ± 1.68!
Table 1
: Participant descriptive data. Data are mean ± standard deviation
5
2.2 Experimental Design. The purpose of this study was to compare hypertrophy, strength gains,
and markers of both subjective and objective fatigue over eight weeks between four training
groups. Participants were counterbalanced by relative strength into four training groups in
which each set’s proximity to failure was controlled by an RIR-based RPE rating: 4–6 RPE, 7–9
RPE, 7–9+ RPE, 10 RPE. This intended groups to train at an average of 4–6 RIR, 1–3 RIR, 1–3 RIR
with the last set of each session taken to failure on each exercise, and all sets taken to failure,
respectively. All groups trained the squat and bench press three times per week on non-
consecutive days (i.e., Monday, Wednesday, and Friday). Participants reported to the laboratory
a total of 25 days over eight consecutive weeks. Pre- and post-study testing for 1RM squat and
bench press, MT of the quadriceps and chest, and anthropometrics took place 48–72 hours
prior to and after weeks 1 and 8; respectively. In week 1, following baseline testing, participants
performed a group-specific introductory microcycle. Weeks 2–7 served as the main training
program followed by a group-specific taper during week 8 to prepare for post-study testing. To
examine temporal recovery and the occurrence of the RBE, blood was collected, and muscle
soreness was assessed immediately before and after training on days 1 and 2 (i.e., Monday
and Wednesday) of weeks 1, 2, and 7. Readiness, motivation to train and acute global fatigue
were assessed before and after each training session using the perceived recovery status
(PRS), motivation to train (MTT), and session RPE (sRPE) scales; respectively. Additionally,
average concentric velocity (ACV m·s-1) was assessed on every repetition of every training
session to verify different proximities to failure between conditions.
To control for pre- and post-exercise nutrient timing, participants ingested a Branched Chain
Amino Acid (BCAA) (BCAA, Core Nutritionals, LLC, Arlington, Virginia, United States of America,
22203) supplement containing 3.5g of leucine, 1.75g of Isoleucine, 1.75g of Valine (Ratio of
2:1:1), and 2.5g of glutamine 30 minutes prior to each testing and training session. Immediately
following each training session 30g of whey protein (Core Pro, Core Nutritionals, LLC, Arlington,
Virginia, United States of America, 22203), containing 3.5g of leucine, was ingested by each
subject. These nutrient portions were selected as 3.5g of leucine is the threshold to maximally
stimulate the process of muscle protein synthesis (MPS) [15]. Stimulating MPS prior to and
following high intensity resistance training significantly augments performance [16]. Both BCAA
and whey protein were consumed in powdered form with 10 oz. of water. Participants were
also instructed to discontinue all other supplementation for the duration of the study.
6
2.3 Training Program. All groups trained three times per week on non-consecutive days and the
number of sets and repetitions were the same with only a set’s proximity to failure (4–6 RPE, 7–
9 RPE, 7–9+ RPE, or 10 RPE) differing between groups, as seen in supplementary file 1. Further,
rest periods of 3–5 minutes were administered between all sets for all groups. During week 1
all groups performed group-specific introductory training with reduced volume and RPE
targets. Investigators selected all loads during week 1 using subject-reported RIR-based RPE of
previous sets in addition to barbell velocity and apparent set difficulty. The investigators
explained their load selection decisions to participants to further familiarize them with protocol
expectations since participants selected their own training loads for the rest of the training
program. The main training program occurred between weeks 2–7. For weeks 2 and 3, an
undulating periodized repetition pattern of 10, 8, and 6 was followed on session 1, 2, and 3
(e.g., Monday, Wednesday, Friday), respectively. In weeks 4 and 5, the undulation pattern was 9,
7, and 5 repetitions. In weeks 6 and 7, the undulation pattern decreased, again, to 8, 6, and 4
repetitions. The final training week was a taper microcycle where each participant performed
sets with the average load lifted throughout the training program within that session (e.g., 7
weeks of session 1 loads averaged). 2 sets of 4 repetitions and 2 sets of 2 repetitions were
performed in session 1 and session 2, respectively. At the completion of each set, participants
were asked to report an RIR-based RPE value. The above protocol was performed on the back
squat and bench press exercises. Additionally, to aid in participant recruitment and train more
muscle groups, all participants performed assistance exercises (barbell overhead press, barbell
row, barbell lying triceps extension, barbell curl, and dumbbell lateral raise) to an 8 RPE and
with the same exact prescription. All repetitions were performed with a controlled eccentric
tempo and habitual concentric intended velocity. Failure was defined as the inability to
perform a repetition through its full range of motion despite maximal effort to do so or the
point at which a subject did not feel comfortable attempting another repetition. However, a
distinction was made in record keeping between sets that reached momentary failure as
defined by Steele et al. [17] (i.e., 11 RPE) and those that were terminated volitionally (i.e., 10
RPE).
2.4 Training Load Instruction and Adjustments. The following script was read to each participant
while being shown the RIR-based RPE scale before each testing and training session:
“Please view this scale to remind you of how RPE is scored. Today, working sets should fall within the
RPE range of insert RPE range assigned for the week. Use your knowledge of your prior performances
7
and how the warm-up sets felt to select a load you believe will fall within the assigned RPE range. The
goal is to maintain your loads in a subsequent fashion, therefore, if the load you select falls above or
below the target RPE range, an increase or decrease in load will occur on the next set. If you fall
within the target RPE range, you have the freedom to increase or decrease load as you see fit so long
as you believe this modified load will still fall within the target RPE range. Avoid being overly
conservative or aggressive in your load selection and expect your RPE to rise with each set as fatigue
accumulates.”
In terms of the specific intra-session set-to-set load adjustments, if the 4–6 and 7–9 RPE
groups either under- or over-shot the desired RPE range, load was increased or decreased by
2% for every 0.5 RPE value from the middle of the range in accordance with Helms et al. 2018
[18]. For all sets except the final set of each session in the 7–9+ RPE group, if the subject-
reported RIR-based RPE was under or over the desired RPE range of 7–9, load was increased
or decreased by 2% for every 0.5 RPE value from the range for the subsequent set in
accordance with Helms et al. [18]. Prior to the final set of each session of weeks 2–7,
participants were informed that the goal of the final set was to select a load that would lead to
failure after reaching the session’s repetition target (e.g., failing repetition 11 if the repetition
target is 10). If the second-to-final set was in the RPE 7–9 range, the participant selected the
load for the final set based on these instructions. If the RIR-based RPE for the previous set was
under or over the desired RPE range of 7–9, load was maintained or increased by 2% for every
0.5 RPE value from the desired 10 RPE on the upcoming set.
For the 10 RPE group, load adjustments were made based upon repetitions performed
compared to the prescribed number of repetitions. Pilot testing was conducted on five
individuals prior to data collection to determine the appropriate set-to-set loading changes
with the goal that following a set to failure the participant would be able to complete the
prescribed number of repetitions during the next set to failure. Pilot testing revealed that a 2%
reduction in load should be made from set 1 to 2 to account for fatigue from training to failure
no matter if the target repetition number was met or not. However, this 2% fatigue adjustment
was not applied in sets thereafter. Further, pilot testing revealed that a 1% loading change
(increase or decrease) should be applied from set-to-set for the difference in repetitions
performed and the prescribed number of repetitions. If the target repetitions were completed
on the first set, load was changed by 2%. If target repetitions were completed from sets 2–4,
then the load was kept the same.
8
For example, if a participant in the 10 RPE group chose to squat 100kg for their first of 3 sets
with a target of 10 repetitions and performed 9 repetitions on the first set, then the load was
reduced by 3% (2% automatic reduction for training to failure on set 1 and a 1% reduction
performing one less repetition than prescribed) to 97kg. If this same individual then performed
12 repetitions on set 2, the load was increased (+2% for 2 repetitions over the prescribed
number) to 99kg for set 3.
2.5 Anthropometric Assessments. Total body mass (kg) was assessed by a calibrated digital scale
(Mettler-Toledo, Columbus, Ohio, USA) and subject’s height (cm) was measured via a wall-
mounted stadiometer (SECA, Hamburg, Germany). Body-fat percentage was estimated using
the average of two skinfold thickness measurements acquired from three sites (chest,
abdomen, anterior thigh) which were then summed. If any measurement was >2mm different
than the previous measure, a third thickness was taken. The Jackson and Pollock equation [19]
was used to estimate body-fat percentage, and the same investigator took all measurements.
2.6 Back Squat and Bench Press Technique. Both the back squat and bench press were
performed in accordance with International Powerlifting Federation standards [20]. Specifically,
for the squat, participants stood straight with the hips and knees locked, and the barbell
placed across the upper back/shoulders. Upon the investigator’s command of “squat”
participants descended by bending the knees until the hip joint was below the top of the knee.
Then participants returned to the starting position upon their own volition. Participants waited
until a rack command was issued to re-rack the barbell. During the bench press, participants
laid supine on a weight bench, maintaining five points of contact (head, butt, and shoulders in
contact with the bench, both feet flat on the floor throughout the movement). Participants
removed the barbell from the rack and held it with arms extended in a stable position.
Investigators issued a start command upon which participants lowered the barbell until it
contacted the chest and then pressed upwards until the arms were once again fully extended.
No pause was required during the bench press. Participants waited until a rack command was
issued to re-rack the barbell.
2.7 One-Repetition Maximum (1RM) Testing. All 1RM testing was performed in accordance with
previously validated procedures [21]. Specifically, all participants completed a 5-minute
dynamic warm-up followed by a squat-specific warm-up consisting of as many repetitions as
9
desired with an empty barbell. Next, participants performed 5 repetitions with 20% of their
estimated 1RM, followed by 50% for 3 repetitions, 70% for 2 repetitions, and 80% of 1RM for 1
repetition. Following the 80% of estimated 1RM warm-up, participants were given 3–5 minutes
of rest before a final warm-up at a load determined by the investigators (between 85–90% of
estimated 1RM). Following the final warm-up, participants took 5–7 minutes of rest while the
investigators determined the load for the first 1RM attempt. Load was increased on each
subsequent attempt until a 1RM was reached and 5–7 minutes of rest was given between each
attempt. On every warm-up and 1RM attempt, RIR-based RPE and ACV were collected to aid in
attempt selection. Following 1RM testing on the back squat, 10 minutes of rest was given and
then an identical protocol was followed for the bench press. A 1RM was accepted as valid if
one of 3 conditions are met: (a) participant reported a “10” on the RPE scale and the
investigators determined an additional attempt with increased load would be unsuccessful, (b)
participant reported a “9.5” RPE and then proceeded to fail the subsequent attempt with a load
increase of 2.5 kg or less, and (c) participant reported an RPE of 9 and failed the subsequent
attempt with a load increase of 5 kg or less. Finally, Eleiko barbells and lifting discs (Chicago, Ill.,
USA) calibrated to the nearest 0.25 kg were used for all 1RM testing. At post-study back squat
testing for the 7–9+ RPE group, the investigators had already determined the group would be
excluded from the lower body and systemic outcomes due to the change in protocol mid-
intervention to decrease risk of injury. However, to ensure all groups performed the bench
press 1RM in a similar state of fatigue, a back squat 1RM was still performed by the 7–9+
group, albeit with conservative load selection. Specifically, weight jumps were limited to 5 kg or
more; thus, the test was concluded if the investigators believed the next attempt would
potentially result in a missed attempt within those constraints.
2.8 Velocity Assessment. The Open Barbell System Version 3 (OBS3): (Squats & Science, New
York, N.Y., USA) was used to assess ACV during every repetition of every testing and training
session. During testing, ACV was used to aid in 1RM attempt selection. During training
sessions, ACV was collected to be compared across sets between groups. In other words, a
difference between groups in mean ACV of each training session verifies that per set proximity
to failure was different between groups. The OBS has a velocity sensor and a display unit. The
OBS was set on the floor to the right side of the participant and attached to the barbell using a
Velcro strap, via a cord, just inside of the “sleeve”. The OBS has been previously validated for
ACV against a gold-standard 3D motion capture system [22]. Data from the OBS was
10
transmitted via Bluetooth to an Apple iPad (Cupertino, California, USA) and data was gathered
and stored in the Open Barbell phone application for later assessment.
2.9 Ultrasonography Assessment. Pectoralis major and vastus lateralis muscle thickness were
assessed via ultrasonography (Bodymetrix Pro System, Intelemetrix Inc., Livermore, Calif., USA)
prior to 1RM pre- and post-study testing on the right side of the body. This method of testing
was previously used to assess the growth response to resistance training [23] and has been
validated with magnetic resonance imaging [24]. Scans were performed with the participant in
the supine position. Sites were scanned from the lateral border of the vastus lateralis to the
medial border of the vastus lateralis with the transducer perpendicular to the skin. Sites were
scanned twice and an average of the two scans was recorded. The site for the chest was
designated as half the distance between the nipple and the anterior axillary line. Vastus
lateralis scans were performed in the supine position. Sites were marked and measured at 50
and 70%, respectively, of the distance from the greater trochanter to the lateral epicondyle of
the femur [25,26]. For each scan, an average of the muscle thickness values was quantified.
Due to the COVID-19 pandemic, 12 participant’s pre- and post-test scans were performed by a
different investigator (4–6 RPE [n = 5], 7–9 RPE [n = 3], 7–9+ RPE [n = 1], 10 RPE [n = 3]). Given
the unique circumstance, inter-rater reliability was not able to be determined. For each
participant, the same investigator took the scan at pre and post study testing, with the
exception of 4 total scans (4–6 RPE [n = 1], 7–9+ RPE [n = 3]). All scans were analyzed by the
same investigator prior to analysis.
2.10 Perceived Recovery Status Scale (PRS). The PRS scale was completed by each participant
prior to each training session. It is a 0–10 scale, which asks participants to rate subjective
recovery [27].
2.11 Motivation to Train Scale. Prior to each training session participants completed a 1–10
Likert scale assessing their “motivation to train” on that specific day [28]. This scale has the
following anchors: 1 – Not Motivated at All, 5 – Somewhat Motivated, and 10 – Highly
Motivated.
2.12 Session Rating of Perceived Exertion (sRPE) Scale. The sRPE scale was completed by
participants immediately following each training session to gauge the difficulty and fatigue of
11
the entire training session [29]. This scale is a 0–10 scale with “0” indicating the participants
were at “rest” meaning they used no effort and a score of “10” indicating “maximal effort”.
2.13 Blood Collection and Analysis. Blood was collected via the antecubital vein using serum
venipuncture techniques and serum separating tubes. Once collected, samples were set at
room temperature for 20–30 minutes for clotting and then were centrifuged at 1,600 x G for
10 minutes to obtain serum. Both biomarkers (creatine kinase – CK and lactate dehydrogenase
– LDH) were measured in duplicate using the Epoch microplate spectrophotometer (BioTek
Instruments, Winooski, VT, USA) through commercially available colorimetric assay kits (cat. no.
K726, Lactate Dehydrogenase Activity Colorimetric Assay Kit, and cat. no. K777, Creatine Kinase
Activity Colorimetric Assay Kit; BioVision, Milpitas, Calif., USA). Blood collection occurred
immediately before and after training on day 1 (i.e., Monday) of weeks 1, 2, and 7 and
immediately before training on day 2 (i.e., Wednesday) of weeks 1, 2, and 7. The CK and LDH
analysis was used to assess the temporal muscle damage response in each group.
2.14 Delayed Onset Muscle Soreness. Pressure-pain threshold was used to assess delayed onset
muscle soreness (DOMS) and was defined as the minimal amount of pressure needed to
induce pain [30,31]; thus, a decrease in pressure-pain threshold indicates an increase in
DOMS. Participants were tested in a relaxed standing position using the probe of an algometer
(Pain Diagnostic & Treatment Inc.; Great Neck, NY, USA) with a 0.9 cm diameter stimulation
area. Palpations occurred at the midline of the vastus lateralis at the midpoint between the
iliac crest and the superior border of the patella and into the midline of the biceps femoris at
the distal 40% point between the articulate interline of the knee and the head of the femur. For
the upper body, algometer palpations occurred on the pectoralis major just medial to the
anterior point of the axillary line. For all palpations, force started at 0 kPa and gradually
increased at a constant rate of 50 to 60 kPa·s-1 until the participant indicated the presence of
pain. All pressure-pain threshold assessments were tested on each subject’s non-dominant
side and participants were instructed to say “now” the instant pain was felt rather than
pressure and this point was recorded. Assessments were completed three consecutive times
with a 30-second interval between measurements. The assessment of DOMS occurred
immediately before training in weeks 1, 2, and 7. Pressure-pain thresholds were used to assess
the temporal muscle soreness response in each group.
12
Scores were recorded in kilograms per centimeters squared and converted to kilopascals (98.1
𝑘𝑃𝑎 =!"
#$!
) and the mean score of the three trials were used for analysis. To maintain reliability
between assessments, each DOMS assessment site was marked by a semi-permanent pen to
maintain homogeneity in repeated assessments. The outlined protocol for DOMS assessment
is in accordance with previously validated measures [30,31].
3 Statistical Analyses
3.1 Sample Size Justification. Sample size was determined by feasibility [32] and no formal power
analysis was performed. Because the sample size of this study is limited, efforts have been
undertaken to ensure that data is as easy as possible to meta-analytically aggregate in the
future.
3.2 Program Observations. To quantify the training completed by each group, descriptive
statistics (mean ± standard deviation) were provided for I) total sets performed, II) total
repetitions performed, III) mean load (% of pre-study 1RM), IV) total volume load, V) total
relative volume load, VI) mean ACV, VII) mean last repetition ACV, VIII) mean intraset % ACV loss,
and IX) mean RIR-based RPE, for both the bench press and back squat, respectively.
Additionally, the number of accessory sets performed was included.
3.3 Primary Outcomes (Strength and Hypertrophy). To evaluate changes in strength (i.e., bench
press and back squat) and muscle thickness (i.e., pectoralis major, and vastus lateralis),
separate linear regression models fit in an analysis of covariance (ANCOVA) structure were
utilized. For each model, the change from baseline was the dependent variable while group
and the mean-centered pre-study value of the outcome being analyzed were included as fixed
effects. For the model evaluating changes in vastus lateralis muscle thickness a main effect for
region (i.e., 50% or 70% of VL), an interaction between region and group, random intercepts for
participant (to account for multiple observations), and random slopes for region were also
included in the model. For all models, marginal effects were estimated using the marginaleffects
package [33]. Uncertainty intervals (i.e., 90 and 95% confidence intervals) for the marginal
effects of all primary outcomes (i.e., strength and hypertrophy) were created via simulation
based methods, similar to bootstrapping [34]. Specifically, 2000 samples were drawn from a
multivariate normal distribution with a mean and variance equal to that of the original model
13
estimates. Quantiles from the resulting distribution were then used to construct the
confidence intervals and standard error.
Additionally, to investigate the practical implications of our findings, statistical equivalence was
formally evaluated [35]. To do so, the uncertainty intervals of the marginal effects were
compared against thresholds denoting the smallest effect size of interest (SESOI). For strength
outcomes, the SESOI was defined as d = ± 0.25 by referencing the threshold for a small effect
in highly trained samples from Rhea et al. [36] and for strength outcomes in Swinton et al. [37].
To allow for strength outcomes to be presented in raw units, this standardized effect threshold
was divided by the pre-test standard deviation (Squat = ± 7.72kg; Bench = ± 5.03kg). For
hypertrophy outcomes, the SESOI was defined as the typical error of measurement calculated
by dividing the standard deviation of the difference between scans at the same time point by
the square root of two [38,39] (Vastus Lateralis = ± 0.79mm; Pectoralis Major = ± 1.95mm).
Finally, a leave-one-out sensitivity analysis was performed for all primary outcomes by re-
estimating all marginal effects after removing one participant at a time. All new estimates were
then compared with the original to see if any single participant substantially influenced the
magnitude and precision of the effects. Visualizations from these analyses can be seen in
supplementary file 2.
3.4 Secondary Outcomes (Muscle Damage, Soreness, and Subjective Recovery). To explore the
longitudinal effects (i.e., repeated measures) of proximity to failure on indirect markers of
muscle damage (i.e., CK and LDH), perceived soreness (i.e., pectoralis major, quadriceps, and
hamstrings), and ratings of subjective recovery (i.e., sRPE, PRS, and MTT), separate linear mixed
effect models were utilized. For the models examining muscle damage, fixed effects and
interactions thereof were included for group, week, and session. For the model examining
soreness fixed effects and interaction thereof were included for group, session, and muscle.
For the models examining subjective ratings of recovery, fixed effects and interaction thereof
were included for group and session. Random intercepts per participant were introduced in all
models to account for repeated measures. Initially, a maximal random slope structure was
attempted [40], but subsequently reduced until the model did not result in a convergence
error. The final CK model included a random slope for session, the LDH model included a
random slope for week, and all of the other models included a random slope for session. Each
model investigating time as a continuous variable (i.e., soreness and subjective proxies of
14
fatigue) included a first order autoregressive covariance matrix. From each model, marginal
effects and 95% confidence intervals were examined to explore differences between groups
and the occurrence of the RBE for each outcome.
Prior to the extraction of estimates from all models the data were visually examined for
violation of model assumptions using the performance package [41]. Finally, in addressing our
research questions we have opted to avoid dichotomizing our findings and therefore did not
employ traditional null hypothesis significance testing which has been extensively critiqued
[42]. Instead, all outcomes compatible with the data were considered, with the greatest
emphasis placed on the point estimates [43]. All analysis was conducted in the R environment
and language for statistical computing (v 4.3.1; R Core Team, https://www.r-project.org/). All raw
data utilized, model outputs, and visualizations are presented in the supplementary materials.
4 RESULTS
4.1 Program Observations. Descriptions of the training completed by each group can be found
in files 3 and 4 in the supplementary materials. Importantly, negligible differences were
observed in all proxies of training volume (i.e., sets, repetitions, volume load, relative volume
load), while all indices of proximity to failure (i.e., ACV, last repetition ACV, intraset % ACV loss,
and RIR-based RPE) showed meaningful differences between groups. Finally, load (% of pre-
study 1RM) was marginally different in the bench press, but likely meaningfully different in the
back squat between groups.
4.2 Back Squat Strength. Increases in back squat strength were observed in the 4–6 (13.79kg
[95% CI: 7.54, 19.92]), 7–9 (18.05kg [95% CI: 12.28, 23.99]), and 10 RPE (5.45kg [95% CI: -5.49,
16.24]) groups (Figure 1A). However, the 90% confidence intervals of the 10 RPE (90% CI: -3.74,
14.48) but not the 4–6 (90% CI: 8.47, 18.91) and 7–9 RPE (90% CI: 13.31, 22.97) groups suggest
that the strength gains were compatible with values less than the SESOI. Contrasts between
groups favored the non-failure conditions, but the width of the 90% confidence intervals were
compatible with differences less than the SESOI (Figure 1C). The leave-one-out sensitivity
analysis revealed that one participant influenced the magnitude and precision of model
estimates. Specifically, upon removing this subject, the contrast of changes in strength
15
between the 4–6 and 7–9 RPE (-0.78kg [90% CI: -5.06, 3.37]) was considered equivalent at the
stated SESOI.
4.3 Bench Press Strength. Increases in bench press strength were observed in the 4–6 (9.05kg
[95% CI: 6.36, 11.76]), 7–9 (9.72kg [95% CI: 7.03, 12.42]), 7–9+ (5.07kg [95% CI: 2.05, 8.1]), and
10 RPE (0.71kg [95% CI: -4.51, 5.54]) groups (Figure 1B). However, the 90% confidence intervals
of the 7–9+ (90% CI: 2.53, 7.5) and 10 RPE (90% CI: -3.72, 4.8) but not the 4–6 (90% CI: 6.83,
11.31) and 7–9 RPE (90% CI: 7.45, 11.98) groups suggest that the strength gains were
compatible with values less than the SESOI. Contrasts between groups favored the non-failure
conditions, but the width of the 90% confidence intervals were compatible with differences less
than the SESOI (Figure 1D). Moreover, the strength gains observed in the 4–6 and 7–9 RPE
groups were considered equivalent at the stated SESOI (-0.66kg [90% CI: -3.79, 2.62]). The
leave one out sensitivity analysis did not meaningfully change the interpretation of the model
estimates.
16
Figure 1
: Marginal effects for changes in back squat (A) and bench press 1RM strength (B), and contrasts thereof between groups (CD).
Vertical dashed lines represent the smallest effect size of interest (SESOI) defined by converting a standardized mean difference of d ± 0.25
to raw units. Black dots and intervals represent the estimated marginal mean and simulated confidence intervals (90 and 95%). Brighter
portions of the distributions are replications that exceed the SESOI. Finally, individual data are visualized below with solid circles. The
marginal effects are adjusted for the mean centered pretest scores of the dependent variable.
4.4 Vastus Lateralis Hypertrophy. Changes in vastus lateralis muscle thickness were compatible
with values less than the SESOI for the 4–6 (0.07mm [95% CI: -1.25, 1.43]), 7–9 (-0.6mm [95%
CI: -1.89, 0.64]), and 10 RPE (-1.64mm [95% CI: -3.81, 0.67]) groups. Moreover, for all contrasts
between groups the width of the 90% confidence intervals were compatible with differences
less than the SESOI. The leave one out sensitivity analysis did not meaningfully change the
interpretation of the model estimates.
4−6 RPE
7−9 RPE
10 RPE
Condition
−10 0 10 20 30 40 50
∆ Back Squat 1RM (kg)
A
4−6 RPE
7−9 RPE
7−9+ RPE
10 RPE
−10 −5 0 5 10 15
∆ Bench Press 1RM (kg)
B
(4−6 RPE) − (10 RPE)
(4−6 RPE) − (7−9 RPE)
(7−9 RPE) − (10 RPE)
Contrast
−20 0 20
Difference in ∆ Back Squat 1RM (kg)
C
(4−6 RPE) − (10 RPE)
(4−6 RPE) − (7−9 RPE)
(4−6 RPE) − (7−9+ RPE)
(7−9 RPE) − (10 RPE)
(7−9 RPE) − (7−9+ RPE)
(7−9+ RPE) − (10 RPE)
−10 0 10 20
Difference in ∆ Bench Press 1RM (kg)
D
1RM Strength Outcomes
17
4.5 Pectoralis Major Hypertrophy. Changes in pectoralis major muscle thickness were
compatible with values less than the SESOI for the 4–6 (1.83mm [95% CI: -0.96, 4.66]), 7–9
(0.52mm [95% CI: -2.28, 3.27]), 7–9+ (2.99mm [95% CI: -0.25, 6.24]), and 10 RPE (-5.54mm [95%
CI: -10.88, -0.63]) groups. The contrasts between the 4–6 and 10 RPE groups (7.37mm [90% CI:
2.41, 12.41]) and the 7–9+ and 10 RPE (8.53mm [90% CI: 3.49, 13.63]) groups favored the
conditions that included non-failure training. The 90% confidence intervals of the remaining
contrasts were compatible with differences less than the SESOI. The leave-one-out sensitivity
analysis revealed that two participants influenced the magnitude and precision of model
estimates. Specifically, upon removing one subject, contrast between the 4–6 and 10 RPE
(4.98mm [90% CI: -0.69, 10.61]), and 7–9+ and 10 RPE (5.67mm [90% CI: -0.26, 11.69]) groups
became compatible with differences less than the SESOI. Additionally, upon removing another
subject, the contrast between the 7–9 and 10 RPE (7.39mm [90% CI: 2.7, 12.21]) groups no
longer was compatible with differences less than the SESOI and favored the non-failure
condition.
18
Figure 2
: Marginal effects for changes in vastus lateralis (A) and pectoralis major muscle thickness (B), and contrasts thereof between
groups (CD). Vertical dashed lines represent the smallest effect size of interest (SESOI) defined by the typical error of measurement. Black
dots and intervals represent the estimated marginal mean and simulated confidence intervals (90 and 95%). Brighter portions of the
distributions are replications that exceed the SESOI. Finally, individual data are visualized below with solid circles. The marginal effects are
adjusted for the mean centered pretest scores of the dependent variable for both outcomes and region (i.e., 50% or 70%) for the vastus
lateralis outcomes
4.6 Indirect Muscle Damage. Averaged across week, there were increases in CK immediately post
exercise in the 4–6 (43.15U·L-1 [95% CI: 31.86, 54.44]), 7–9 (60.57U·L-1 [95% CI: 51.12, 70.01]),
and 10 RPE (49.14U·L-1 [95% CI: 31.89, 66.38]) groups. CK then returned closer to baseline 48
hours post exercise in 4–6 (-51.68U·L-1 [95% CI: -64.99, -38.37]), 7–9 (-62.64U·L-1 [95% CI: -
73.77, -51.5]), and 10 RPE (-55.66U·L-1 [95% CI: -75.99, -35.33]) groups. Averaged across
session, CK decreased from week 1 to 2 (-1.29U·L-1 [95% CI: -11.9, 9.32]) and from week 2 to 7
(-6.47U·L-1 [95% CI: -17.07, 4.14]) in the 4–6 RPE group, but increased in the 7–9 RPE group
4−6 RPE
7−9 RPE
10 RPE
Condition
−6−4−2 0 2 4 6
∆ Vastus Lateralis MT (mm)
A
4−6 RPE
7−9 RPE
7−9+ RPE
10 RPE
−15 −10 −5 0 5 10
∆ Pectoralis Major MT (mm)
B
(4−6 RPE) − (10 RPE)
(4−6 RPE) − (7−9 RPE)
(7−9 RPE) − (10 RPE)
Contrast
−4−2 0 2 4 6 8
Difference in ∆ Vastus Lateralis MT (mm)
C
(4−6 RPE) − (10 RPE)
(4−6 RPE) − (7−9 RPE)
(4−6 RPE) − (7−9+ RPE)
(7−9 RPE) − (10 RPE)
(7−9 RPE) − (7−9+ RPE)
(7−9+ RPE) − (10 RPE)
−10 −5 0 5 10 15 20
Difference in ∆ Pectoralis Major MT (mm)
D
Muscle Hypertrophy Outcomes
19
(Week 1 to 2: 10.1U·L-1 [95% CI: 1.22, 18.98]; Week 2 to 7: (2.59U·L-1 [95% CI: -6.29, 11.46])). The
10 RPE group saw an decrease from week 1 to 2 (-1.89U·L-1 [95% CI: -18.09, 14.32]) but a
increase from week 2 to 7 (6.96U·L-1 [95% CI: -9.24, 23.17]). When examining the interaction
contrasts (i.e., pre to post exercise change in CK compared between weeks 1 and 7) between
groups, all were compatible with a null point estimate.
Averaged across week, there were increases in LDH immediately post exercise in the 4–6
(0.29U·L-1 [95% CI: -11.87, 12.46]), 7–9 (23.3U·L-1 [95% CI: 13.13, 33.48]), and 10 RPE (21.08U·L-1
[95% CI: 2.5, 39.67]) groups. LDH then returned closer to baseline 48 hours post exercise in 4–
6 (-11.26U·L-1 [95% CI: -23.42, 0.91]), 7–9 (-20.15U·L-1 [95% CI: -30.33, -9.97]), and 10 RPE (-
30.74U·L-1 [95% CI: -49.32, -12.16]) groups. Averaged across session, LDH decreased from
week 1 to 2 in the 4–6 RPE group (-2.69U·L-1 [95% CI: -19.71, 14.32]), but increased for the 7–9
(9.54U·L-1 [95% CI: -4.69, 23.78]), and 10 RPE (3.58U·L-1 [95% CI: -22.41, 29.58]) groups. From
week 2 to 7, LDH increased in the 4–6 (9.03U·L-1 [95% CI: -7.4, 25.45]), 7–9 (14.39U·L-1 [95% CI:
0.65, 28.13]), and 10 RPE (22.51U·L-1 [95% CI: -2.58, 47.6]) groups . When examining the
interaction contrasts (i.e, pre to post exercise change in LDH compared between weeks 1 and
7) between groups, all were compatible with a null point estimate.
20
Figure 3
: Marginal effects for longitudinal trends in creatine kinase (A) and lactate dehydrogenase (B). Each panel contains three
timepoints pre-exercise, immediately post-exercise, and 48 hours post-exercise. Columns represent the different weeks and rows for the
different groups. Dots and intervals represent the estimated marginal means and confidence intervals (95%). Finally, individual data are
visualized with faded lines.
4.7 Soreness. Averaged across time and muscle, soreness was highest in the 4–6 RPE group
(10.31kPa [95% CI: 8.47, 12.16]), followed by the 7–9 (10.03kPa [95% CI: 8.2, 11.86]), and 10
RPE (9.79kPa [95% CI: 6.44, 13.14]) groups. The 4–6 (0.21kPa [95% CI: 0.09, 0.33]), 7–9 (0.23kPa
[95% CI: 0.12, 0.35]), and 10 RPE (0.24kPa [95% CI: 0.03, 0.46]) groups all exhibited a positive
slope of soreness over the training program. The contrasts of these slopes between groups all
contained a null point estimate.
30
60
90
120
0
30
60
90
120
30
60
90
Creatine Kinase (U·L−1 )
4−6 RPE
7−9 RPE
10 RPE
Week 1
Week 2
Week 7
A
100
150
0
50
100
150
75
100
125
150
175
Lactate Dehydrogenase (U·L−1 )
4−6 RPE
7−9 RPE
10 RPE
Session 1 (Pre) Session 1 (Post) Session 2 (Pre) Session 1 (Pre) Session 1 (Post) Session 2 (Pre) Session 1 (Pre) Session 1 (Post) Session 2 (Pre)
Timepoint
B
Longitudinal Trends in Biomarker Proxied Muscle Damage
21
Figure 4
: Marginal effects for longitudinal trends in perceived muscle soreness. Each panel contains trends over the course of the training
program. Columns represent trends for each of the different groups. Dark lines and intervals represent the estimated marginal means and
confidence intervals (95%). Finally, individual data are visualized with faded lines.
4.8 Subjective Recovery. Averaged across time, sRPE was lowest in the 4–6 RPE group (3.91a.u.
[95% CI: 3.07, 4.74]) followed by the 7–9 (4.44a.u. [95% CI: 3.64, 5.23]), and 10 RPE (5.17a.u.
[95% CI: 3.72, 6.62]) groups. The 4–6 (-0.08a.u. [95% CI: -0.13, -0.03]), 7–9 (-0.08a.u. [95% CI: -
0.13, -0.04]), and 10 RPE (-0.08a.u. [95% CI: -0.15, 0]) groups all exhibited negative slopes of
sRPE over the training program. The contrasts of these slopes between groups all contained a
null point estimate.
Averaged across time, PRS was lowest in the 10 RPE group (5.93a.u. [95% CI: 4.65, 7.2])
followed by the 7–9 (6.76a.u. [95% CI: 6.06, 7.45]), and 4–6 RPE (6.81a.u. [95% CI: 6.07, 7.55])
groups. The 4–6 (-0.01a.u. [95% CI: -0.06, 0.04]), and 7–9 (-0.03a.u. [95% CI: -0.08, 0.01]) groups
all exhibited negative slopes of PRS over the training program. However, the 10 RPE group
(0.12a.u. [95% CI: 0.03, 0.21]) exhibited a positive slope of PRS over the training program. The
contrasts between the 4–6 and 10 RPE (-0.13a.u. [95% CI: -0.23, -0.03]), and 7–9 and 10 RPE (-
0.15a.u. [95% CI: -0.26, -0.05]) groups all did not contain a null point estimate.
Averaged across time, MTT was lowest in the 10 RPE group (7.05a.u. [95% CI: 5.72, 8.38])
followed by the 4–6 (7.54a.u. [95% CI: 6.77, 8.31]), and 7–9 RPE (7.58a.u. [95% CI: 6.85, 8.31])
groups. The 4–6 (-0.06a.u. [95% CI: -0.1, -0.01]), and 7–9 RPE (-0.06a.u. [95% CI: -0.1, -0.02])
groups all exhibited negative slopes of MTT over the training program. However, the 10 RPE
4−6 RPE
7−9 RPE
10 RPE
0 5 10 15 20 0 5 10 15 20 0 5 10 15 20
0
10
20
30
Timepoint
Pain Pressure Threshold (kPa·s−1 )
Longitudinal Trends in Muscle Soreness
22
group (0a.u. [95% CI: -0.08, 0.08]) exhibited a positive slope of MTT over the training program.
The contrasts of these slopes between groups all contained a null point estimate.
Figure 5
: Marginal effects for longitudinal trends in subjective markers of fatigue. Each panel contains trends over the course of the
training program. Columns represent trends for each of the different groups while each row represents a different scale (i.e., PRS, MTT,
sRPE). Dark lines and intervals represent the estimated marginal means and confidence intervals (95%). Finally, individual data are
visualized with faded lines.
0.0
2.5
5.0
7.5
10.0
PRS (a.u.)
4−6 RPE
7−9 RPE
10 RPE
0.0
2.5
5.0
7.5
10.0
MTT (a.u.)
0.0
2.5
5.0
7.5
10.0
sRPE (a.u.)
0 5 10 15 20 0 5 10 15 20 0 5 10 15 20
Session
Longitudinal Trends in Subjective Proxies of Fatigue
23
5 DISCUSSION
To our knowledge, this is the first longitudinal study to quantify proximity to failure via RIR-
based RPE, corroborate these ratings with objective barbell velocity (ACV), and assess
longitudinal fatigue with varying resistance training proximities to failure. As hypothesized,
strength outcomes were comparable between the 4–6 RPE and 7–9 RPE groups, while both
groups saw marginally greater changes in strength compared to the 7–9+ and 10 RPE groups.
Our hypothesis that hypertrophy would be similar between groups was partially supported as
the changes in muscle thickness were not meaningfully different nor practically similar at all
sites measured, with the width of the uncertainty intervals suggesting inconclusive findings. All
secondary analyses of both objective and subjective markers of longitudinal fatigue resulted in
negligible differences between groups and did not reveal strong evidence in favor of the RBE.
Overall, these data suggest that when both repetition and set volume are equated, muscle
strength outcomes are likely similar when taking sets to either 4–6 RIR or 1–3 RIR in trained
men, while both approaches seem to be slightly more effective than training that includes sets
performed to momentary failure (i.e., 7–9+ and 10 RPE groups). However, muscle hypertrophy
outcomes remain inconclusive and markers of fatigue (objective and subjective) were
comparable between groups.
As noted in supplementary files 3 and 4, the 7–9 RPE group trained at a higher percentage of
1RM (Bench Press: 83.44 ± 2.88%; Back Squat: 82.41 ± 4.40%) than the 4–6 RPE group (Bench
Press: 79.50 ± 3.85%; Back Squat: 73.18 ± 3.95%). Given higher loads seem to be
advantageous for strength development [8], it may be that the lower intraset velocity loss
accumulated in the 4–6 RPE group (Bench Press: 30.80 ± 4.38%; Back Squat: 14.36 ± 4.58%)
compared to the 7–9 RPE group (Bench Press: 41.63 ± 4.54%; Back Squat: 24.81 ± 5.88%)
contributed to similar 1RM improvements [4]. Moreover, meta-analytic data suggest, training
that includes sets performed to failure does not seem to further enhance strength outcomes
[1,2], which also reflects our data with less strength gains observed in the 7–9+ and 10 RPE
groups.
To explain these outcomes, it is possible that high loads (i.e., % of 1RM) and low intraset fatigue
(i.e., velocity loss) have independent influences on strength outcomes, leading to the
equivalent results in our study. Indeed, meta-analyses by Refalo et al. [44] and Jukic et al. [45]
24
suggest greater strength gains are expected with high loads and low velocity loss thresholds,
respectively. Moreover, training to momentary failure results in high levels of intraset fatigue,
and thus, greater decrements in performance would be expected [46,47]. Practically, greater
reductions in performance would necessitate decreased loading to remain within a given
repetition target, potentially reducing the potency for strength gain. Therefore, it seems that
various program design variables may contribute to maximal strength gains and in the present
study the 4–6 RPE and 7–9 RPE groups exhibited a favorable design in one of the
aforementioned variables, while the groups that included sets to momentary failure (i.e., 7–9+
and 10 RPE) tended to observe inferior outcomes.
The findings for proximity to failure and muscle hypertrophy were inconclusive, as each group
failed to experience hypertrophy greater than the SESOI defined by measurement error.
Moreover, nearly all contrasts between conditions were compatible with the SESOI. The
absence of a clear disadvantage, along with the lack of feasibility/safety we observed with
regular momentary failure, may indicate that non-failure training is preferable for multi-joint
lower-body exercises in trained individuals. Multiple studies [8,13,48] have observed training
far from failure (i.e, >4 RIR) to be sufficient to maximize hypertrophy when using moderate to
heavy loads (>60% of 1RM), as was done in the present study. Additionally, it seems that
hypertrophy can be maximized far from failure in trained participants [13,18], as was done in
the present study. Further, exercise selection may also be crucial to consider when
determining the optimal per-set proximity to failure. Both Carroll et al. [13] and Helms et al.
[18] used multi-joint exercise and observed a group training with ~4+ RIR achieved greater or
similar hypertrophy, respectively, compared to a group training closer to failure (~0–3 RIR).
Importantly, given we elected momentary failure (i.e., failing a repetition despite maximal effort
to do so), our results (and lack of feasibility of the RT program) may not extrapolate to other,
less strict definitions of failure [49].
A meaningful limitation of our study is the fact that multiple investigators performed
ultrasound scans. Due to the COVID-19 pandemic, this limitation could not be avoided nor
could inter-rater reliability be established. Considering the large amount of variability in our
data, these results should be viewed as exploratory. While our study agrees with previous data
that muscle hypertrophy does not seem to be meaningfully affected by proximity to failure
when moderate to heavy loads are used in trained individuals using multi-joint exercises,
extreme caution is warranted due to the aforementioned limitations.
25
As intended, the 4–6 RPE (Bench Press: 5.06 ± 0.28 RPE; Back Squat: 5.05 ± 0.79 RPE), 7–9 RPE
(Bench Press: 7.51 ± 0.35 RPE; Back Squat: 7.34 ± 0.39 RPE), 7–9+ RPE (Bench Press: 8.39 ±
0.62 RPE; Back Squat: 6.04 ± 0.53 RPE), and 10 RPE (Bench Press: 10.07 ± 0.15 RPE; Back
Squat: 10.06 ± 0.13 RPE) groups trained at different self-reported RPE in both the bench press
and back squat; however, previous research has shown that trained lifters over-estimated RPE
(i.e., under-estimated RIR) by ~5 and ~3 repetitions, when asked to estimate when they
believed they had reached a 5 and 7 RPE during a set to failure at 70% of 1RM on the back
squat [50]. A recent meta-analysis also confirmed that, on average, participants tend to under-
estimate RIR [51]. Thus, these self-reported RPE values should be interpreted with caution.
To further inform per set proximity to failure, our lab established RIR/ACV relationships in the
free weight bench press and back squat on the group-level [52]. In the current study, the
average last repetition ACV of the 4–6 RPE group on the back squat was 0.55 ± 0.06 m.s-1,
which corresponds to approximately an 8 RIR (0.55 ± 0.02 m.s-1) based on the data from our
lab. This comparison may indicate that the 4–6 RPE group (target 5 RIR) may have actually
trained with more RIR than intended, on average, for each set for the back squat. The average
last repetition ACV’s of the 7–9, 7–9+, and 10 RPE groups on the back squat (7–9 RPE = 0.43 ±
0.05 m.s-1; 7–9+ RPE = 0.44 ± 0.08 m.s-1; 10 RPE = 0.31 ± 0.06 m.s-1) were close to the velocities
associated with 3 RIR (0.42 ± 0.01 m.s-1) and 0 RIR (0.35 ± 0.02 m.s-1) from our lab, potentially
verifying the desired proximity to failure. Finally, the average last repetition ACV of the 4–6 RPE
(0.34 ± 0.04 m.s-1), 7–9 RPE (0.25 ± 0.05 m.s-1), 7–9+ RPE (0.24 ± 0.03 m.s-1) and 10 RPE (0.19 ±
0.04 m.s-1) groups were closest to RIR values within the desired range from our lab on the
bench press (6 RIR: 0.35 ± 0.02 m.s-1; 3 RIR: 0.26 ± 0.02 m.s-1; 2 RIR: 0.23 ± 0.02 m.s-1; 0 RIR:
0.17 ± 0.02 m.s-1).
To date, multiple studies [46,47] have observed that training to failure on the squat and bench
press elongates the time course of recovery compared to not training to failure. However,
these studies have only examined temporal recovery over one week; thus, it’s possible that the
RBE mitigates this difference over time. In the present study, we observed largely comparable
CK and LDH responses in all groups, without clear evidence of the RBE in either group. CK and
LDH consistently elevated immediately post exercise and returned closer to baseline 48 hours
later, but this pattern did not change meaningfully over the course of the program. Further,
there were no meaningful differences between groups in measurements of muscle soreness.
26
When averaging across groups and muscles, soreness did tend to increase over the course of
the program, but a meaningful limitation is that many participants demonstrated bruising from
the repeated measurement that could have altered the pressure-pain threshold independent
of the training intervention. Thus, these results should be interpreted cautiously.
Finally, subjective fatigue ratings of sRPE, MTT, and PRS were largely not meaningfully different
between groups, with most having a similar trajectory over time. One explanation for the lack
of convincing group differences in all indices of fatigue may be that the range of proximity to
failure investigated (i.e., a difference of ~2–4 RIR) is considerably smaller than in previous
research. Indeed, previous research has found significantly greater sRPE when comparing
groups training to or not failure that likely differed substantially in the average RIR trained (i.e.,
>4 RIR) [8]; however, the current study is novel in comparing two different submaximal
proximities to failure.
Another limitation of this study is a small sample size (n = 38) with deviation to the original
research plan; thus, our findings should be interpreted cautiously. Data collection was first
halted in the 10 RPE group (n = 3) due to safety precautions, then recruitment for the entire
study was ceased in response to the COVID-19 pandemic. Recruitment then began again for
the newly added 7–9+ RPE group (n = 9), which was constructed to safely investigate a group
that included training to momentary failure, but then had to be modified for the lower body
due to additional safety precautions.
6 CONCLUSION
In summary, these data suggest that muscle strength outcomes are similar when taking sets to
either a self-reported 4–6 RIR or 1–3 RIR in trained men, while training that includes sets to
momentary failure may result in slightly inferior strength development. However, our data do
not provide robust conclusions as to the influence of proximity to failure on muscle
hypertrophy due to the large variability observed. All indices of objective and subjective fatigue
were comparable between groups, without strong evidence of the repeated bout effect. We
urge future research to continue to use RIR-based RPE, a practical tool, while also tracking last
repetition ACV, an objective tool, to report accurate proximities to failure [49].
27
REFERENCES
1. Grgic J, Schoenfeld BJ, Orazem J, Sabol F. Effects of resistance training performed to
repetition failure or non-failure on muscular strength and hypertrophy: A systematic review
and meta-analysis. Journal of Sport and Health Science [Internet]. 2022;11:202–11. Available
from: https://doi.org/10.1016/j.jshs.2021.01.007
2. Vieira AF, Umpierre D, Teodoro JL, Lisboa SC, Baroni BM, Izquierdo M, et al. Effects of
resistance training performed to failure or not to failure on muscle strength, hypertrophy, and
power output: A systematic review with meta-analysis. Journal of Strength and Conditioning
Research [Internet]. 2021;35:1165–75. Available from:
https://doi.org/10.1519/jsc.0000000000003936
3. Vieira JG, Sardeli AV, Dias MR, Filho JE, Campos Y, Sant’Ana L, et al. Effects of resistance
training to muscle failure on acute fatigue: A systematic review and meta-analysis. Sports
Medicine [Internet]. 2021;52:1103–25. Available from: https://doi.org/10.1007/s40279-021-
01602-x
4. Pareja-Blanco F, Alcazar J, Sánchez-Valdepeñas J, Corenjo-Daza PJ, Piqeras-Sanchiz F, Mora-
Vela R, et al. Velocity loss as a critical variable determining the adaptations to strength training.
Medicine & Science in Sports & Exercise [Internet]. 2020;52:1752–62. Available from:
https://doi.org/10.1249/mss.0000000000002295
5. Pareja-Blanco F, Alcazar J, Cornejo-Daza PJ, Sánchez-Valdepeñas J, Rodriguez-Lopez C, Mora
JH, et al. Effects of velocity loss in the bench press exercise on strength gains, neuromuscular
adaptations, and muscle hypertrophy. Scandinavian Journal of Medicine
&$\mathsemicolon$ Science in Sports [Internet]. 2020;30:2154–66. Available from:
https://doi.org/10.1111/sms.13775
6. Rodiles-Guerrero L, Pareja-Blanco F, León-Prados JA. Effect of velocity loss on strength
performance in bench press using a weight stack machine. International Journal of Sports
Medicine [Internet]. 2020;41:921–8. Available from: https://doi.org/10.1055/a-1179-5849
7. Pareja-Blanco F, Rodríguez-Rosell D, Sánchez-Medina L, Sanchis-Moysi J, Dorado C, Mora-
Custodio R, et al. Effects of velocity loss during resistance training on athletic performance,
strength gains and muscle adaptations. Scandinavian Journal of Medicine & Science in Sports
[Internet]. 2016;27:724–35. Available from: https://doi.org/10.1111/sms.12678
8. Lasevicius T, Schoenfeld BJ, Silva-Batista C, Souza Barros T de, Aihara AY, Brendon H, et al.
Muscle failure promotes greater muscle hypertrophy in low-load but not in high-load
28
resistance training. Journal of Strength and Conditioning Research [Internet]. 2019;36:346–51.
Available from: https://doi.org/10.1519/jsc.0000000000003454
9. Lacerda LT, Marra-Lopes RO, Diniz RCR, Lima FV, Rodrigues SA, Martins-Costa HC, et al. Is
performing repetitions to failure less important than volume for muscle hypertrophy and
strength? Journal of Strength and Conditioning Research [Internet]. 2019;34:1237–48. Available
from: https://doi.org/10.1519/jsc.0000000000003438
10. Santanielo N, Nóbrega S, Scarpelli M, Alvarez I, Otoboni G, Pintanel L, et al. Effect of
resistance training to muscle failure vs non-failure on strength, hypertrophy and muscle
architecture in trained individuals. Biology of Sport [Internet]. 2020;37:333–41. Available from:
https://doi.org/10.5114/biolsport.2020.96317
11. Martorelli S, Cadore EL, Izquierdo M, Celes R, Martorelli A, Cleto VA, et al. Strength training
with repetitions to failure does not provide additional strength and muscle hypertrophy gains
in young women. European Journal of Translational Myology [Internet]. 2017;27. Available
from: https://doi.org/10.4081/ejtm.2017.6339
12. Goto K, Ishii N, Kizuka T, Takamatsu K. The impact of metabolic stress on hormonal
responses and muscular adaptations. Medicine & Science in Sports & Exercise [Internet].
2005;37:955–63. Available from: https://journals.lww.com/acsm-
msse/Fulltext/2005/06000/The_Impact_of_Metabolic_Stress_on_Hormonal.9.aspx
13. Carroll K, Bazyler C, Bernards J, Taber C, Stuart C, DeWeese B, et al. Skeletal muscle fiber
adaptations following resistance training using repetition maximums or relative intensity.
Sports [Internet]. 2019;7:169. Available from: https://doi.org/10.3390/sports7070169
14. Haddad M, Stylianides G, Djaoui L, Dellal A, Chamari K. Session-RPE method for training
load monitoring: Validity, ecological usefulness, and influencing factors. Frontiers in
Neuroscience [Internet]. 2017;11. Available from: https://doi.org/10.3389/fnins.2017.00612
15. Stark M, Lukaszuk J, Prawitz A, Salacinski A. Protein timing and its effects on muscular
hypertrophy and strength in individuals engaged in weight-training. Journal of the International
Society of Sports Nutrition [Internet]. 2012;9. Available from: https://doi.org/10.1186/1550-
2783-9-54
16. Tang JE, Moore DR, Kujbida GW, Tarnopolsky MA, Phillips SM. Ingestion of whey hydrolysate,
casein, or soy protein isolate: Effects on mixed muscle protein synthesis at rest and following
resistance exercise in young men. Journal of Applied Physiology [Internet]. 2009;107:987–92.
Available from: https://doi.org/10.1152/japplphysiol.00076.2009
29
17. Steele J, Fisher J, Giessing J, Gentil P. Clarity in reporting terminology and definitions of set
endpoints in resistance training. Muscle & Nerve [Internet]. 2017;56:368–74. Available from:
https://doi.org/10.1002/mus.25557
18. Helms ER, Byrnes RK, Cooke DM, Haischer MH, Carzoli JP, Johnson TK, et al. RPE vs.
Percentage 1RM loading in periodized programs matched for sets and repetitions. Frontiers in
Physiology [Internet]. 2018;9. Available from: https://doi.org/10.3389/fphys.2018.00247
19. Jackson AS, Pollock ML. Generalized equations for predicting body density of men. British
Journal of Nutrition [Internet]. 1978;40:497–504. Available from:
https://doi.org/10.1079/bjn19780152
20. Federation IP. International powerlifting federation technical rules 2019 [Internet]. Available
from: http://www.powerlifting-ipf.com/rules/technical-rules.html.
21. Zourdos MC, Klemp A, Dolan C, Quiles JM, Schau KA, Jo E, et al. Novel resistance training-
specific rating of perceived exertion scale measuring repetitions in reserve. Journal of Strength
and Conditioning Research [Internet]. 2016;30:267–75. Available from:
https://doi.org/10.1519/jsc.0000000000001049
22. Goldsmith JA, Trepeck C, Halle JL, Mendez KM, Klemp A, Cooke DM, et al. Validity of the
open barbell and tendo weightlifting analyzer systems versus the optotrak certus 3D motion-
capture system for barbell velocity. International Journal of Sports Physiology and Performance
[Internet]. 2019;14:540–3. Available from: https://doi.org/10.1123/ijspp.2018-0684
23. Schoenfeld BJ, Ratamess NA, Peterson MD, Contreras B, Sonmez GT, Alvar BA. Effects of
different volume-equated resistance training loading strategies on muscular adaptations in
well-trained men. Journal of Strength and Conditioning Research [Internet]. 2014;28:2909–18.
Available from: https://doi.org/10.1519/jsc.0000000000000480
24. Reeves ND, Maganaris CN, Narici MV. Ultrasonographic assessment of human skeletal
muscle size. European Journal of Applied Physiology [Internet]. 2004;91:116–8. Available from:
https://doi.org/10.1007/s00421-003-0961-9
25. ABE T, BRECHUE WF, FUJITA S, BROWN JB. Gender differences in FFM accumulation and
architectural characteristics of muscle. Medicine & Science in Sports & Exercise [Internet].
1998;30:1066–70. Available from: https://doi.org/10.1097/00005768-199807000-00007
26. Abe T, Kondo M, Kawakami Y, Fukunaga T. Prediction equations for body composition of
japanese adults by b-mode ultrasound. American Journal of Human Biology [Internet].
1994;6:161–70. Available from: https://doi.org/10.1002/ajhb.1310060204
27. Laurent CM, Green JM, Bishop PA, Sjökvist J, Schumacker RE, Richardson MT, et al. A
practical approach to monitoring recovery: Development of a perceived recovery status scale.
30
Journal of Strength and Conditioning Research [Internet]. 2011;25:620–8. Available from:
https://doi.org/10.1519/jsc.0b013e3181c69ec6
28. Colquhoun RJ, Gai CM, Walters J, Brannon AR, Kilpatrick MW, DAgostino DP, et al.
Comparison of powerlifting performance in trained men using traditional and flexible daily
undulating periodization. Journal of Strength and Conditioning Research [Internet].
2017;31:283–91. Available from: https://doi.org/10.1519/jsc.0000000000001500
29. Day ML, McGuigan MR, Brice G, Foster C. Monitoring exercise intensity during resistance
training using the session RPE scale. The Journal of Strength and Conditioning Research
[Internet]. 2004;18:353. Available from: https://doi.org/10.1519/r-13113.1
30. Rocha CS, Lanferdini FJ, Kolberg C, Silva MF, Vaz MA, Partata WA, et al. Interferential therapy
effect on mechanical pain threshold and isometric torque after delayed onset muscle soreness
induction in human hamstrings. Journal of Sports Sciences [Internet]. 2012;30:733–42.
Available from: https://doi.org/10.1080/02640414.2012.672025
31. Fischer AA. Pressure algometry over normal muscles. Standard values, validity and
reproducibility of pressure threshold. Pain [Internet]. 1987;30:115–26. Available from:
https://doi.org/10.1016/0304-3959(87)90089-3
32. Lakens D. Sample size justification. Collabra: Psychology [Internet]. 2022;8. Available from:
https://doi.org/10.1525/collabra.33267
33. Arel-Bundock V. Marginaleffects: Predictions, comparisons, slopes, marginal means, and
hypothesis tests [Internet]. 2023. Available from: https://CRAN.R-
project.org/package=marginaleffects
34. King G, Tomz M, Wittenberg J. Making the most of statistical analyses: Improving
interpretation and presentation. American Journal of Political Science [Internet]. 2000;44:347.
Available from: https://doi.org/10.2307/2669316
35. Lakens D, Scheel AM, Isager PM. Equivalence testing for psychological research: A tutorial.
Advances in Methods and Practices in Psychological Science [Internet]. 2018;1:259–69.
Available from: https://doi.org/10.1177/2515245918770963
36. Rhea MR. Determining the magnitude of treatment effects in strength training research
through the use of the effect size. The Journal of Strength and Conditioning Research
[Internet]. 2004;18:918. Available from: https://doi.org/10.1519/14403.1
37. Swinton PA, Burgess K, Hall A, Greig L, Psyllas J, Aspe R, et al. Interpreting magnitude of
change in strength and conditioning: Effect size selection, threshold values and bayesian
updating. Journal of Sports Sciences [Internet]. 2022;40:2047–54. Available from:
https://doi.org/10.1080/02640414.2022.2128548
31
38. Weir JP. Quantifying test-retest reliability using the intraclass correlation coefficient and the
SEM. The Journal of Strength and Conditioning Research [Internet]. 2005;19:231. Available
from: https://doi.org/10.1519/15184.1
39. Swinton PA, Hemingway BS, Saunders B, Gualano B, Dolan E. A statistical framework to
interpret individual response to intervention: Paving the way for personalized nutrition and
exercise prescription. Frontiers in Nutrition [Internet]. 2018;5. Available from:
https://doi.org/10.3389/fnut.2018.00041
40. Barr DJ, Levy R, Scheepers C, Tily HJ. Random effects structure for confirmatory hypothesis
testing: Keep it maximal. Journal of Memory and Language [Internet]. 2013;68:255–78.
Available from: https://doi.org/10.1016/j.jml.2012.11.001
41. Ludecke D, Ben-Shachar MS, Patil I, Waggoner P, Makowski D. performance: An R package
for assessment, comparison and testing of statistical models. Journal of Open Source Software.
2021;6:3139.
42. Amrhein V, Greenland S, McShane B. Scientists rise up against statistical significance.
Nature [Internet]. 2019;567:305–7. Available from: https://doi.org/10.1038/d41586-019-00857-
9
43. Cumming G. The new statistics. Psychological Science [Internet]. 2013;25:7–29. Available
from: https://doi.org/10.1177/0956797613504966
44. Refalo MC, Hamilton DL, Paval DR, Gallagher IJ, Feros SA, Fyfe JJ. Influence of resistance
training load on measures of skeletal muscle hypertrophy and improvements in maximal
strength and neuromuscular task performance: A systematic review and meta-analysis. Journal
of Sports Sciences [Internet]. 2021;39:1723–45. Available from:
https://doi.org/10.1080/02640414.2021.1898094
45. Jukic I, Castilla AP, Ramos AG, Hooren BV, McGuigan MR, Helms ER. The acute and chronic
effects of implementing velocity loss thresholds during resistance training: A systematic review,
meta-analysis, and critical evaluation of the literature. Sports Medicine [Internet]. 2022;53:177–
214. Available from: https://doi.org/10.1007/s40279-022-01754-4
46. Pareja-Blanco F, Rodríguez-Rosell D, Aagaard P, Sánchez-Medina L, Ribas-Serna J, Mora-
Custodio R, et al. Time course of recovery from resistance exercise with different set
configurations. Journal of Strength and Conditioning Research [Internet]. 2020;34:2867–76.
Available from: https://doi.org/10.1519/jsc.0000000000002756
47. Morán-Navarro R, Pérez CE, Mora-Rodríguez R, Cruz-Sánchez E de la, González-Badillo JJ,
Sánchez-Medina L, et al. Time course of recovery following resistance training leading or not to
32
failure. European Journal of Applied Physiology [Internet]. 2017;117:2387–99. Available from:
https://doi.org/10.1007/s00421-017-3725-7
48. Andersen V, Paulsen G, Stien N, Baarholm M, Seynnes O, Saeterbakken AH. Resistance
training with different velocity loss thresholds induce similar changes in strengh and
hypertrophy. Journal of Strength and Conditioning Research [Internet]. 2021;Publish Ahead of
Print. Available from: https://doi.org/10.1519/jsc.0000000000004067
49. Pelland JC, Robinson ZP, Remmert JF, Cerminaro RM, Benitez B, John TA, et al. Methods for
controlling and reporting resistance training proximity to failure: Current issues and future
directions. Sports Medicine [Internet]. 2022;52:1461–72. Available from:
https://doi.org/10.1007/s40279-022-01667-2
50. Zourdos MC, Goldsmith JA, Helms ER, Trepeck C, Halle JL, Mendez KM, et al. Proximity to
failure and total repetitions performed in a set influences accuracy of intraset repetitions in
reserve-based rating of perceived exertion. Journal of Strength and Conditioning Research
[Internet]. 2021;35:S158–65. Available from: https://doi.org/10.1519/jsc.0000000000002995
51. Halperin I, Malleron T, Har-Nir I, Androulakis-Korakakis P, Wolf M, Fisher J, et al. Accuracy in
predicting repetitions to task failure in resistance exercise: A scoping review and exploratory
meta-analysis. Sports Medicine [Internet]. 2021;52:377–90. Available from:
https://doi.org/10.1007/s40279-021-01559-x
52. Hickmott LM. Relationship between velocity and repetitions in reserve in the back squat,
bench press, and deadlift [Internet] [PhD thesis]. ProQuest Dissertations and Theses. 2020. p.
59. Available from:
https://go.openathens.net/redirector/fau.edu?url=https://www.proquest.com/dissertations-
theses/relationship-between-velocity-repetitions-reserve/docview/2444653562/se-2
