Resistance training (RT) is well-established as a primary interventional strategy for increasing muscle strength and mass across populations. Current theory indicates that RT-induced muscular adaptations can be optimized by manipulating exercise variables such as volume, load, and frequency, among others [1
]. In an effort to further enhance adaptations, fitness enthusiasts often seek to employ a variety of advanced training strategies. One of the most popular of these strategies is the drop set method (a.k.a. breakdown sets) [2
]. The drop set method involves performing a set to momentary muscular failure (MMF), then immediately reducing the load (generally by 20% to 25%) and performing as many additional repetitions as possible [3
]. If desired, double or triple drops can be employed to heighten stimulation of working muscle fibers and thus perhaps enhance muscular adaptations [3
The drop set method is largely based on the premise that muscles are not fully fatigued when sets are carried out to MMF, as they are still capable of producing force at lower loads [3
]. Thus, performing additional repetitions at a decreased magnitude of load immediately after reaching muscle failure in a set may elicit heightened fatigue of muscle fibers, potentially leading to a superior anabolic response [4
]. In addition, the combination of a high number of repetitions performed with minimal rest periods induces high levels of metabolic stress [2
], which has been theorized as a potential stimulus for hypertrophic gains [5
Longitudinal research is currently limited as to the effects of drop set training on muscular adaptations of the lower limb. Seminal work by Goto et al. [6
] provided preliminary support for the strategy, showing significantly greater increases in one repetition maximum (RM) leg press and maximal isokinetic knee extension torque for a group performing a single drop set to MMF using 50% one RM compared to a traditional set configuration. Moreover, the drop set group realized a ~2% increase in thigh muscle cross sectional area whereas the traditional group displayed a slight loss in muscle size (~0.5%), although differences in this outcome did not rise to statistical significance (p
< 0.08). While these results are intriguing, several issues must be taken into account when attempting to draw inferences from the data. For one, the sample was small and likely underpowered (17 total subjects in a parallel group design), raising the possibility of a Type II error. In addition, the drop set group performed more sets than the traditional group, which may have confounded hypertrophic adaptations [7
]. Importantly, the drop set group took a 30-s rest period after the fifth set before initiating the drop set. Given that the drop set method customarily is performed with as little rest as possible between drops, the relevance to breakdown training is questionable.
More recently, Angleri et al. [8
] randomized the lower limbs of resistance-trained men so that one leg performed a traditional training routine (three to five sets of six to twelve repetitions with a two-min inter-set rest interval) while the other leg performed the same protocol with up to two drop sets at a 20% reduction in load. After 12 weeks, participants achieved significant increases in measures of muscle strength and hypertrophy with no significant differences observed between conditions. It should be noted that hypertrophy was only assessed in the vastus lateralis (VL) at a single point along the length of the muscle (corresponding to 50% femur length).
Research indicates that exercise-induced quadriceps femoris hypertrophy can manifest in a non-uniform fashion, with different magnitudes of growth observed in the proximal, middle, and distal regions of the musculature [9
]. Thus, it is not clear whether differences may have existed in other regions of the VL, or in other muscles of the quadriceps. Given that drop-set training involves performing additional repetitions after muscle failure, it is conceivable that the strategy could stimulate aspects of muscles containing a higher proportion of fibers with predominantly oxidative characteristics, and consequently, from a theoretical point of view, this may contribute to greater hypertrophy in these sites. The purpose of the present study was to compare the effects of lower-limb drop set training versus traditional training on muscle strength and regional hypertrophy of the quadriceps in a sample of physically active young men. We hypothesized that the drop-set method would promote greater increases in strength and size of the quadriceps femoris compared to traditional training.
4. Resistance Training Protocol
During the eight-week RT intervention, participants trained under direct supervision of personal trainers. The routine consisted of the leg extension exercise, with the number of sets gradually increasing each week. Participants performed a standardized five-min warm-up prior to each training session that consisted of various track-and-field drills (skips, high steps, lateral crossovers), 12–15 bodyweight squats and two sets on the leg extension machine for 8–10 repetitions with 50% of their estimated one RM. The program followed a linear periodization model that readjusted training loads based on each participant’s progression rate. The starting leg was alternated from one session to the next in counterbalanced fashion so that neither condition obtained a performance advantage over time. Once participants completed all the repetitions on one leg, they immediately performed the alternate condition on the contralateral leg, then rested for 120 s before performing the ensuing set. Participants were coached to perform concentric and eccentric actions with a cadence of 1:2 s.
The DS protocol involved performing sets with a ~five RM load carried out to MMF. Immediately thereafter, the load was reduced by 20% and participants continued to perform additional repetitions until reaching MMF, at which point the load was reduced by 10–15% and the set was concluded once the participant reached MMF at that load. To account for strength progression over time, we set a target repetition range of three to seven repetitions in which the participant would have to reach MMF with a given load. Loads were continually adjusted to maintain this target repetition range. Alternatively, the TRAD protocol involved performing sets with a ~15 RM load with repetitions carried out consecutively until the participant reached MMF. To account for strength progression over time, we targeted a range of 13 to 17 repetitions with loads continually adjusted from set to set so as to maintain this repetition range.
The total training volume, determined as the number of sets per session, was gradually increased from week to week with participants performing the same number of sets for both conditions. In the first week of the study, participants performed only one training session consisting of three sets in total; in the second week they performed two training sessions consisting of four sets per session. This gradual increase in volume facilitated adaption to the training stimulus so as to minimize the potential for delayed onset muscle soreness (DOMS) and establish a repeated bout effect whereby DOMS would not interfere with training performance [17
]. Thereafter, participants performed three weekly sessions for the duration of the intervention. The total number of sets peaked in Week 7, with participants performing a total of 15 working sets throughout the week. Volume was tapered in the eighth week to promote recovery and restoration (see Table 1
5. Statistical Analysis
Statistical analyses were carried out using SPSS Statistics software for Windows (version 27.0; IBM Corp.) and Excel 365 (Microsoft Corp.), with data expressed as mean and standard deviation (SD) values. We used a two-way, repeated-measures ANOVA to test for the time (pre and post intervention) x condition (experimental and control) interaction. If no statistically significant interaction was detected, a Bonferroni correction was used to determine the presence of main effects. A paired samples T-test was employed to determine whether differences in total training volume existed between the two conditions. Cohen’s d effect size (ES) was calculated as the mean pre-post change divided by the pooled SD. ESs of 0.00 to 0.19, 0.20 to 0.49, 0.50 to 0.79, and >0.80 were considered to represent trivial, small, moderate and large effects, respectively [18
]. In addition, percent changes were calculated as the mean pre-post change divided by the pre-study mean multiplied by 100. Cumming estimation plots were created for MT data using computer-based software [19
]. The statistical significance level was set a priori at p
From the initial 24 participants who began the study, 8 dropped out prior to completion: 5 dropped out for personal reasons (could not sustain training volume, lost interest because lack of time, wanted to focus on studies), while 3 dropped out due to injuries sustained outside the intervention. Thus, a total of 16 participants were ultimately included in the final analysis.
6.2. Muscle Strength
Mean ± SD, effect size (ES), and percentage of change in tests of maximum strength from pre- to post-intervention are shown in Table 2
For the one RM test, there was a significant main effect for time (p < 0.001) but no main effect for group (p = 0.483). We did not observe a significant group x time interaction (p = 0.378). For peak torque we observed significant main effects for both factors, group (p = 0.016) and time (p < 0.001) without a significant group x time interaction (p = 0.988). For average peak torque there was a significant main effect favoring time (p < 0.001), but not a significant main effect for group (p = 0.058) nor a group x time interaction (p = 0.783).
6.3. Muscle Thickness
Mean ± SD, effect size (ES), and percent change in MT of the RF and VL from pre- to post-intervention are shown in Table 2
. A Cumming estimation plot of individual and mean data are shown in Figure 2
For the RF at 30% muscle length, we observed a significant main effect for time (p < 0.001) as well as a group × time interaction favoring DS (p = 0.001); we did not observe a main effect for group (p = 0.80). For the RF at 50% muscle length, we observed a significant main effect for time (p < 0.001) as well as a group x time interaction favoring DS (p = 0.034); we did not observe a significant main effect for group (p = 0.25). For the RF at 70% muscle length, we observed a significant main effect for time (p = 0.006). We did not observe a significant main effect for group (p = 0.39) nor a group × time interaction (p = 0.70).
For the VL at 30% muscle length, we observed a significant main effect for time (p < 0.001); we did not observe a significant main effect for group (p = 0.14) nor a group × time interaction (p = 0.439). For the VL at 50% muscle length, we did not observe any significant main effects nor any group × time interaction (ES = 0.24). For the VL at 70% muscle length, we did not observe any significant main effects nor any group × time interaction (p = 0.874).
6.4. Total Training Volume
Paired samples t-test revealed no statistically significant differences in total training volume (calculated as sets × reps) between the conditions (p = 0.92). However, volume load (calculated as reps × sets × load) showed statistically significant differences between conditions favoring DS (p > 0.001).
This study aimed to compare the effects of drop-set training with traditional training on measures of muscle strength and hypertrophy. A novel finding was that DS elicited superior hypertrophy of the RF muscle in a non-uniform manner. Alternatively, hypertrophy of the VL was similar between conditions. In addition, DS did not appear to enhance strength-related adaptations compared to TRAD. We discuss the implications of our findings below.
7.1. Muscle Thickness
Assessment of changes in MT showed that DS promoted greater growth in two of the three measurement sites of the RF compared to TRAD; alternatively, changes in MT of the VL were statistically similar between conditions. Differences in RF hypertrophy were most pronounced in the proximal portion (30% of muscle length), where DS elicited a 17.7% increase versus a 3.7% increase for TRAD. The effect size difference equated to 0.87, indicating a large magnitude of effect. DS also produced greater increases at the midpoint of the RF (50% of muscle length) compared to TRAD, although the relative magnitude of differences was smaller (8.3% versus 3.6%, respectively) with a moderate effect size difference (0.58). These results suggest that drop set training may help to enhance regional hypertrophy of the rectus femoris. No hypertrophic differences between conditions were observed for the distal portion of the RF.
Our findings expand on those of Angleri et al. [8
], who found similar increases in cross sectional area of the VL when training in a traditional fashion versus employing drop sets. Importantly, Angleri et al. [8
] only assessed hypertrophy in the VL at a single-site (midpoint of the muscle) whereas we evaluated both the VL and RF along multiple aspects of the muscle length. Consistent with the results of Angleri et al. [8
], we found no differences in muscle thickness at the midpoint of the VL between conditions; in addition, we did not observe any differences in the proximal and distal regions of the muscle as well. Alternatively, we demonstrated that greater hypertrophic increases occurred in the proximal- and mid-regions of the RF with the use of drop set training; Angleri et al. [8
] did not assess changes in this muscle, precluding comparison between studies.
Although we did not attempt to determine potential mechanistic explanations for our findings, we can hypothesize that the observed differences in MT between the various segments of the RF and VL muscles may be related to the use of a single-joint knee extension exercise. Emerging research indicates that single-joint knee extension RT preferentially activates the RF [20
], which in turn may induce greater growth of this muscle with consistent training [22
]. This may be due the fact that the RF is placed under loaded stretch during leg extension exercise, which has been shown to enhance hypertrophic adaptations [23
]. Our results further this line of research, providing evidence that drop set training may in fact enhance increases in RF muscle mass without having such an effect on the VL. Although speculative, it is possible that the drop set method heightened mechanical stress to the RF, given that the muscle was trained in a lengthened position with a higher volume load. This hypothesis warrants further study.
Notably, we provide additional evidence that RT-induced hypertrophy of the quadriceps femoris can occur in a non-uniform fashion. Our findings indicate that changes in muscle size varied considerably along the length of the muscles studied, as demonstrated in previous research [9
]. Thus, to obtain a true understanding of muscular adaptations pursuant to longitudinal RT designs, researchers should endeavor to measure hypertrophy at proximal, middle, and distal sites when investigating hypertrophic changes of the quadriceps.
7.2. Muscle Strength
Both set configurations elicited similar increases in strength. Large changes were observed for DS and TRAD in the estimated one RM test (34.6% vs. 32.0%, respectively) as well as dynamometry-assessed isokinetic peak torque (21.7% vs. 22.5%, respectively) and average torque (23.6% vs. 22.5%, respectively). No statistical differences were observed between conditions, indicating that DS training does not confer an advantage for improving strength-related measures. Our findings are consistent with those of Angleri et al. [8
], who reported significant increases in the one RM leg press and leg extension after twelve weeks of drop set vs traditional training, with no observed differences between conditions.
It is interesting to note that the initial loads for each set in our study were substantially heavier in DS compared to TRAD (~five RM vs. ~fifteen RM). Given research showing a strength-related benefit to the use of heavier loads [7
], it therefore, might be expected that the DS group would have outperformed TRAD on tests of force capacity. However, this was not the case here. One possible explanation for the finding is that we employed a submaximal strength test to estimate one RM. Based on the principle of specificity, it can be speculated that transfer of strength would be greatest when training as close to maximum loading capacity as possible. Our test could be considered more of a “strength-endurance” assessment as opposed to a pure maximal strength test (i.e., true one RM). Similarly, the employed isokinetic tasks lacked specificity to the training protocols, which involved using a dynamic constant external resistance exercise. Evidence indicates that strength-related changes between loading ranges are mitigated when testing is carried out on a device dissimilar to that used during the exercise intervention [7
], thus providing a viable rationale for the lack of observed difference between conditions.
Our study had several limitations that should be considered when attempting to draw practical inferences. First, results are specific to the leg extension exercise and thus cannot necessarily be extrapolated to compound lower body movements or exercises/muscles for other areas of the body. Second, the results are specific to recreationally trained young men and cannot necessarily be generalized to women, adolescents, the elderly, or those with considerable RT experience. Third, the study had a relatively short duration of eight weeks; although this time-frame has consistently proven sufficient to elicit robust increases in muscle strength and hypertrophy, as was the case in our study, it is unclear whether results might have changed had training been carried out over longer time periods. Fourth, we estimated one RM using submaximal testing methods; although the employed formula has been validated for assessment of maximal dynamic strength [15
], it is susceptible to error and thus may have influenced extrapolation of results on this outcome. Finally, although the within-subject design can be considered a study strength in that it reduces variability and hence increases statistical power, it remains possible that strength adaptations may have been confounded by a cross-education effect [24
]. A number of studies have demonstrated that training one limb results in a strength increase of the contralateral limb, conceivably via modifying motor pathways that innervate the opposite limb [24
]. However, evidence of such an effect occurring is limited to an untrained contralateral limb versus when both limbs perform an RT intervention. No study to date has shown that training both limbs with different methods results in a cross-education effect, raising skepticism that this phenomenon unduly influenced results. Moreover, the body of evidence does not seem to indicate that cross education appreciably affects measures of muscle hypertrophy in an untrained contralateral limb [25
], and the possibility seemingly would be even less likely in the case where both limbs are trained. We recommend replication of the study with a parallel design to rule out potential confounding from cross-education.