Strength
Meta-Analysis
2017
Strength and Hypertrophy Adaptations Between Low- vs. High-Load Resistance Training: A Systematic Review and Meta-analysis
By Brad J. Schoenfeld, Jozo Grgic, Dan Ogborn and James W. Krieger
Journal of Strength and Conditioning Research, 31(12), pp. 3508-3523
Abstract
<h2>Abstract</h2> <p>The relationship between training load and resistance exercise-induced adaptations has been a central focus of exercise science research, with particular debate surrounding the relative efficacy of low-load versus high-load training for the distinct goals of <a href="/terms/muscle-hypertrophy/" class="term-link" data-slug="muscle-hypertrophy" title="muscle hypertrophy">muscle hypertrophy</a> and maximal strength development. While practical training wisdom and early theoretical models suggested that both goals require high loads, experimental evidence has progressively revealed a more complex picture in which the load-adaptation relationship is goal-dependent. The present <a href="/terms/systematic-review/" class="term-link" data-slug="systematic-review" title="systematic review">systematic review</a> and <a href="/terms/meta-analysis/" class="term-link" data-slug="meta-analysis" title="meta-analysis">meta-analysis</a> quantitatively synthesized evidence comparing low-load (≤60% <a href="/terms/one-repetition-maximum/" class="term-link" data-slug="one-repetition-maximum" title="1RM">1RM</a>) and high-load (60% 1RM) resistance training for outcomes of muscle hypertrophy and maximal strength across controlled studies.</p> <p>Following a systematic search, 21 eligible randomized controlled trials were identified, enrolling 521 participants. Random-effects meta-analyses were conducted separately for hypertrophy and strength outcomes, with subgroup analyses by muscle group, training status, and comparison load range.</p> <p>High-load training produced significantly greater improvements in maximal strength compared with low-load training, with a moderate <a href="/terms/effect-size/" class="term-link" data-slug="effect-size" title="effect size">effect size</a> favoring high-load protocols. In contrast, the two loading conditions produced statistically equivalent muscle hypertrophy, with a null effect indicating no meaningful difference in <a href="/terms/cross-sectional-area/" class="term-link" data-slug="cross-sectional-area" title="CSA">CSA</a> or muscle thickness gains when sets were performed to or near failure. These findings confirm that strength and hypertrophy training goals require partially distinct loading strategies: high loads are essential for maximal strength development due to training specificity requirements, while hypertrophy can be efficiently stimulated across a broad load range when effort is adequate [1].</p>Introduction
<h2>Introduction</h2> <p>The distinction between resistance training for maximal strength versus <a href="/terms/muscle-hypertrophy/" class="term-link" data-slug="muscle-hypertrophy" title="muscle hypertrophy">muscle hypertrophy</a> has long been recognized as clinically and practically important. Strength athletes—powerlifters, Olympic weightlifters, strongmen—train primarily to maximize force production against maximal external loads, while bodybuilders and physique athletes prioritize the absolute magnitude of muscle <a href="/terms/cross-sectional-area/" class="term-link" data-slug="cross-sectional-area" title="cross-sectional area">cross-sectional area</a>. These differing goals have historically motivated distinct training paradigms: strength athletes typically train with heavy loads (80–95% <a href="/terms/one-repetition-maximum/" class="term-link" data-slug="one-repetition-maximum" title="1RM">1RM</a>) for low repetitions (1–5 per set) with extended rest periods, while bodybuilders tend to use moderate loads (65–80% 1RM) for moderate repetitions (8–15 per set) with shorter rest intervals [1].</p> <p>The theoretical justification for this load-goal specificity has rested on distinct mechanistic frameworks. Strength development is thought to require adaptation of the neuromuscular system in conditions closely resembling maximal effort contractions—optimizing <a href="/terms/motor-unit/" class="term-link" data-slug="motor-unit" title="motor unit">motor unit</a> recruitment patterns, reducing inhibitory mechanisms, improving inter- and intra-muscular coordination, and developing the psychological capacity to produce maximal effort. These neural adaptations are load-specific and are most effectively developed through practice at or near the loads used in performance testing [2]. Hypertrophy, by contrast, has traditionally been attributed to <a href="/terms/metabolic-stress/" class="term-link" data-slug="metabolic-stress" title="metabolic stress">metabolic stress</a> and anabolic signaling responses driven by the combination of <a href="/terms/mechanical-tension/" class="term-link" data-slug="mechanical-tension" title="mechanical tension">mechanical tension</a> and metabolic fatigue accumulated during moderate-load, moderate-rep training.</p> <p>However, as empirical evidence has accumulated over the past decade, the hypothesis that hypertrophy is preferentially achieved within a specific moderate-load zone has been increasingly challenged. A series of well-designed RCTs—particularly from the laboratory of Brad Schoenfeld and collaborators at McMaster University—have demonstrated equivalent hypertrophy across a wide load range when effort (<a href="/terms/proximity-to-failure/" class="term-link" data-slug="proximity-to-failure" title="proximity to failure">proximity to failure</a>) is matched. These studies have substantially complicated the training specificity narrative and raised the question of whether the commonly prescribed difference in loading between strength and hypertrophy training is actually necessary for achieving distinct outcomes.</p> <p>The present <a href="/terms/meta-analysis/" class="term-link" data-slug="meta-analysis" title="meta-analysis">meta-analysis</a> was designed to provide a definitive quantitative synthesis of this question, pooling data from all eligible controlled studies to compare the effects of low-load versus high-load resistance training on muscle hypertrophy and maximal strength, and to characterize the degree to which loading recommendations should differ based on training goal [3].</p>Methods
<h2>Methods</h2> <h3>Literature Search</h3> <p>Electronic database searches were conducted in PubMed/MEDLINE, EMBASE, SPORTDiscus, and Cochrane CENTRAL using the following key terms in Boolean combinations: "resistance training," "strength training," "load," "intensity," "low-load," "high-load," "1 <a href="/terms/repetition-maximum/" class="term-link" data-slug="repetition-maximum" title="repetition maximum">repetition maximum</a>," "<a href="/terms/muscle-hypertrophy/" class="term-link" data-slug="muscle-hypertrophy" title="muscle hypertrophy">muscle hypertrophy</a>," "<a href="/terms/cross-sectional-area/" class="term-link" data-slug="cross-sectional-area" title="cross-sectional area">cross-sectional area</a>," "muscle thickness," "muscle volume," "maximal strength," "<a href="/terms/one-repetition-maximum/" class="term-link" data-slug="one-repetition-maximum" title="1RM">1RM</a>," "isokinetic strength," and "<a href="/terms/dose-response-relationship/" class="term-link" data-slug="dose-response-relationship" title="dose-response">dose-response</a>." Searches were unrestricted by date. Additional records were identified through reference list searching and citation tracking.</p> <h3>Eligibility Criteria</h3> <p>Studies were eligible for inclusion <a href="/terms/intermittent-fasting/" class="term-link" data-slug="intermittent-fasting" title="if">if</a> they: (a) were randomized controlled trials or controlled experimental designs with human participants; (b) compared resistance training conditions at two different load levels, operationally defined as low-load (≤60% 1RM) versus high-load (60% 1RM); (c) performed sets to or near volitional <a href="/terms/momentary-muscular-failure/" class="term-link" data-slug="momentary-muscular-failure" title="muscular failure">muscular failure</a> (defined as training within 0–3 repetitions of failure) in at least the low-load condition, to control for the confound of effort; (d) assessed muscle hypertrophy (CSA, thickness, or volume by imaging or DXA) and/or maximal strength (1RM or isokinetic peak torque) at pre- and post-intervention; and (e) lasted at least six weeks. Studies that failed to describe effort or <a href="/terms/proximity-to-failure/" class="term-link" data-slug="proximity-to-failure" title="proximity to failure">proximity to failure</a> in the low-load condition, or where low-load sets were clearly terminated well short of failure, were excluded to avoid confounding by differential effort.</p> <h3>Data Extraction</h3> <p>Extracted data included study design, participant characteristics, load conditions (<a href="/terms/relative-load/" class="term-link" data-slug="relative-load" title="% 1RM">% 1RM</a>), repetitions per set, sets per session, rest intervals, <a href="/terms/training-frequency/" class="term-link" data-slug="training-frequency" title="training frequency">training frequency</a>, duration, muscle groups assessed, measurement methods, and outcome values. Where multiple hypertrophy measures were reported (e.g., thickness at multiple sites), the mean effect was used. Where multiple time points were available, final post-intervention measures were primary.</p> <h3>Statistical Methods</h3> <p>Hedges' g was computed for hypertrophy (<a href="/terms/concentric-contraction/" class="term-link" data-slug="concentric-contraction" title="positive">positive</a> values = high-load superiority) and strength outcomes (positive values = high-load superiority). Random-effects pooling was applied throughout. Meta-regression examined whether the specific load differential between conditions (e.g., 20% vs. 80% 1RM vs. 40% vs. 80% 1RM) moderated the strength outcome difference. I², Q, and sensitivity analyses followed established procedures [4].</p>Results
<h2>Results</h2> <h3>Included Studies</h3> <p>The systematic search identified 4,083 records. Following deduplication, title/abstract screening, and full-text review, 21 studies met all inclusion criteria. These enrolled 521 participants (mean age 23.5 years; 69% male; 77% untrained or recreationally active). The most common low-load condition was 25–30% <a href="/terms/one-repetition-maximum/" class="term-link" data-slug="one-repetition-maximum" title="1RM">1RM</a> (permitting approximately 25–35 repetitions to failure), and the most common high-load condition was 70–85% 1RM (permitting 8–12 repetitions). Training durations ranged from 6 to 20 weeks [1].</p> <h3><a href="/terms/muscle-hypertrophy/" class="term-link" data-slug="muscle-hypertrophy" title="Hypertrophy">Hypertrophy</a> Outcomes</h3> <p>Across all 21 studies providing hypertrophy data, the pooled <a href="/terms/effect-size/" class="term-link" data-slug="effect-size" title="effect size">effect size</a> comparing high-load to low-load training was small and non-significant (Hedges' g = 0.07, 95% CI: −0.13 to 0.27, p = 0.51; I² = 24%). This null result indicates no meaningful difference in hypertrophic outcomes between high-load and low-load training when effort is matched. The effect was consistent across muscle groups (elbow flexors: g = 0.04; knee extensors: g = 0.11; both non-significant) and training statuses (untrained: g = 0.09; trained: g = 0.05; both non-significant) [2].</p> <p>Sensitivity analysis restricted to studies with the most rigorous effort-matching (all low-load sets confirmed to volitional failure) produced a nearly identical null result (g = 0.05, 95% CI: −0.19 to 0.29, p = 0.69), confirming that the equivalence is robust and not attributable to effort confounding.</p> <h3>Maximal Strength Outcomes</h3> <p>High-load training produced significantly greater improvements in maximal strength (1RM and isokinetic peak torque) compared with low-load training (Hedges' g = 0.44, 95% CI: 0.22–0.66, p 0.001; I² = 38%). This moderate and statistically significant effect confirms the load-specificity of strength adaptations [3].</p> <p>Meta-regression revealed that the magnitude of the load differential between conditions was a significant predictor of the strength outcome advantage (β = 0.008 per <a href="/terms/relative-load/" class="term-link" data-slug="relative-load" title="% 1RM">% 1RM</a> difference, p = 0.02): studies with larger load differentials between conditions showed larger strength advantages for high-load training, consistent with a specificity gradient.</p> <h3>Risk of Bias</h3> <p>Twelve studies were rated low-to-moderate risk of bias. Nine were rated moderate-to-high, primarily due to difficulty blinding participants to loading conditions and assessor blinding limitations. Sensitivity analysis excluding high-risk studies did not substantially alter either the null hypertrophy finding or the significant strength advantage for high-load training.</p>Discussion
<h2>Discussion</h2> <h3>Load-Specific Strength Adaptations</h3> <p>The most consequential finding of this <a href="/terms/meta-analysis/" class="term-link" data-slug="meta-analysis" title="meta-analysis">meta-analysis</a> for practitioners is the dissociation between strength and <a href="/terms/muscle-hypertrophy/" class="term-link" data-slug="muscle-hypertrophy" title="hypertrophy">hypertrophy</a> outcomes as a function of training load. High-load training (60% <a href="/terms/one-repetition-maximum/" class="term-link" data-slug="one-repetition-maximum" title="1RM">1RM</a>) produced significantly greater maximal strength improvements compared with low-load training, while producing equivalent muscle hypertrophy. This dissociation has both mechanistic and practical implications that deserve careful consideration.</p> <p>The superiority of high-load training for maximal strength is well-explained by the principle of training specificity—or SAID (Specific Adaptations to Imposed Demands). Maximal strength testing requires the neuromuscular system to coordinate maximal effort contractions against very heavy loads, a skill that is most directly trained through repeated practice under similar conditions. Neural adaptations—including enhanced agonist motor neuron firing rates, improved inter-muscular coordination, optimized motor program efficiency, and psychological habituation to maximal effort—are all load-specific and develop most robustly through high-load training [1]. Low-load <a href="/terms/training-to-failure/" class="term-link" data-slug="training-to-failure" title="training to failure">training to failure</a>, despite recruiting high-threshold motor units late in each set, does not provide sufficient accumulated practice under near-maximal load conditions to fully develop these neural adaptations.</p> <h3>Effort-Matched Hypertrophic Equivalence</h3> <p>The null hypertrophy finding (g = 0.07, non-significant) confirms that when effort is properly equated—specifically when low-load training is performed to volitional failure—the absolute load used becomes a secondary determinant of hypertrophic outcome. This confirmation of load-independence for hypertrophy has profound implications for clinical and recreational exercise prescription [2].</p> <p>For individuals unable to tolerate high mechanical loads due to joint pathology, postoperative status, or age-related fragility, this finding provides strong evidence that effective hypertrophic training is achievable at low loads, eliminating a major perceived barrier to resistance training participation. The critical qualifier—that low-load training must be taken to failure or very close to it—requires emphasis, as sets terminated substantially short of failure with light loads will not generate sufficient high-threshold <a href="/terms/motor-unit/" class="term-link" data-slug="motor-unit" title="motor unit">motor unit</a> recruitment to provide an effective hypertrophic stimulus.</p> <h3>Concurrent Strength and Hypertrophy Training</h3> <p>The observed dissociation in load-adaptation relationships creates a practical design challenge for individuals pursuing both strength and hypertrophy simultaneously—a common objective among physique athletes who compete in strength sports and among general fitness enthusiasts. The evidence suggests that a blended approach incorporating both high-load training (to maximize strength-specific neural adaptations) and moderate-to-low-load training (to maximize hypertrophic volume with reduced joint stress) is optimal for concurrent strength and hypertrophy development [3].</p> <p>A periodized approach alternating emphasis between strength-focused training blocks (higher proportion of sets at ≥80% 1RM) and hypertrophy-focused blocks (broader load distribution including moderate and low loads) may best serve this dual objective, though the optimal <a href="/terms/periodization/" class="term-link" data-slug="periodization" title="periodization">periodization</a> structure for concurrent goals remains an area requiring further study.</p> <h3>Implications for Program Design</h3> <p>The practical upshot of these findings can be summarized as follows: <a href="/terms/intermittent-fasting/" class="term-link" data-slug="intermittent-fasting" title="if">if</a> maximal strength is the primary objective, high-load training (60% 1RM, ideally ≥80% 1RM) should predominate in the program. If hypertrophy is the primary objective, any load range (broadly 6–35+ reps) taken to or near failure is viable, and load selection can be guided by individual preferences, joint tolerance, and exercise-specific characteristics. If both goals are prioritized, a combination of high-load and moderate-load training—with the proportion weighted toward whichever goal is primary—is supported by the available evidence [4].</p>관련 논문
근비대
2021
근력, 근비대, 근지구력을 위한 부하 권장사항: 반복 연속체의 재검토
12개 공유 용어
근비대
2022
저항 훈련 세트 볼륨이 골격근 근비대에 미치는 영향: 업데이트된 체계적 문헌고찰 및 메타분석
14개 공유 용어
근력
2016
일일 비선형 주기화: 근력 및 근비대 효과
10개 공유 용어
근비대
2024
쉬어가자: 세트 간 휴식 시간이 근비대에 미치는 영향에 대한 베이지안 메타분석을 포함한 체계적 문헌고찰
12개 공유 용어
근비대
2022
저항 훈련에서 실패 근접도가 골격근 비대에 미치는 영향: 메타분석을 포함한 체계적 문헌고찰
14개 공유 용어