Strength Training

Strength Training

Strength training is the planned application of mechanical tension to muscle, tendon, and bone. Maximal strength, hypertrophy, power, and muscular endurance share tissues and often share exercises. They do not share the same dose, the same test, or the same evidence. Most of the commercial remainder — “heavy is always necessary,” “you must train to failure,” “the hormone pulse after the session decides the month,” “EMG is hypertrophy,” “a strong grip proves that lifting saves lives” — overstates a real finding, spends a surrogate as an outcome, or treats an observational association as a mechanism.

Compiled by South Beach Longevity · 20 August 2026 Copyright 2026 Series SBL-41 / SP-STRENGTH-TRAINING · Register A scientific article Sources peer-reviewed human trials, consensus statements, meta-analyses, and labelled mechanistic or observational work · verified NCBI records Constraint This document describes published research. It is not medical advice. No human use, dose, route, schedule, diagnosis, or individualized training prescription is recommended anywhere in this document.

How to read this document Every finding is labelled, in the sentence that reports it, by the kind of study that produced it. A 1RM is not a vertical jump. Muscle thickness over weeks is not force on day one. Surface EMG in a single set is not hypertrophy. A circulating hormone pulse after a session is not a month of adaptation. A grip-strength mortality cohort is not a training trial. Where two results conflict, both are given. Loads, set counts, and durations appear only as reported experimental parameters, always with the population attached. Nothing here is a recommendation. Findings are graded in place as established, strongly supported, emerging, plausible, or speculative. Two further labels mark careful absences rather than verdicts: not established, where the evidence is too thin to place a claim on the ladder at all — untested or insufficient, an absence of proof rather than disproof; and not supported, where the weight of evidence leans against a claim but stops short of a formal refutation.

01 What strength training is, and four things it is not

Strength training is the systematic use of external resistance — free weights, machines, body mass, bands, or another opposing force — to change the capacity of the neuromuscular system to produce force. Kraemer and Ratamess, in a narrative prescription review, treated progression and exercise selection as the working grammar of that change (Kraemer and Ratamess, 2004). The American College of Sports Medicine’s progression-model position stands are the field’s current maps, not its constitution: they grade how load, volume, frequency, and exercise choice have been used in healthy adults (American College of Sports Medicine, 2002; American College of Sports Medicine, 2009). Those papers are consensus syntheses. They are strongly supported as descriptions of what the College then judged the human record to say. They are not licences to treat every number in them as a prescription, and they are not used that way here.

Four things follow immediately.

First, strength training is not one outcome. Maximal strength, hypertrophy, power, and muscular endurance can move together and can be driven apart. Campos and colleagues randomised 32 untrained healthy men to three repetition bands at the same exercises and showed that the high-load, low-repetition regimen favoured maximal strength, the intermediate band favoured hypertrophy, and the high-repetition band favoured local endurance (Campos et al., 2002). That is a small, male, previously untrained, laboratory allocation. It is strongly supported as evidence that the four outcomes are not interchangeable. It is not a menu.

Second, it is not a session of muscle activity. Surface electromyography records a voltage. Vigotsky and colleagues, in a methods review, warned that amplitude is not a straightforward index of muscle force, activation, or hypertrophy, and that greater EMG does not imply greater adaptation (Vigotsky et al., 2018). An acute hormone pulse is the same class of error. West and Phillips reported that exercise-induced hormone profiles did not explain strength or hypertrophic gains in their human sample (West and Phillips, 2012). Morton, Phillips and colleagues later showed, in a randomised comparison of high- versus low-load training in resistance-trained men, that neither load nor the systemic hormone response determined the hypertrophy or strength gain (Morton et al., 2016). Those two human papers are strongly supported as a rejection of the post-session hormone story. They are not a claim that endocrine status is irrelevant to a life.

Third, it is not a mortality drug. Grip strength predicts death and disability in large cohorts (Rantanen et al., 1999; Newman et al., 2006; Leong et al., 2015). Muscle-strengthening activity, as a self-reported behaviour, is associated with lower risk in pooled observational analyses (Stamatakis et al., 2018; Saeidifard et al., 2019; Momma et al., 2022). Association is not a trial of a program. Part Five returns to that gap and does not close it.

Fourth, it is not medical advice and it is not a plan. This document describes published research. It recommends no load, set, frequency, exercise, or schedule for any person.

Sibling articles in this series take sports nutrition, protein, and aging diets as related problems. This title is the force-production cut of the same material. Dietary-protein evidence is not spent here as if it were a loading scheme. Sarcopenia imaging is not spent as if it were a 1RM.

02 Four outcomes that share a barbell

Maximal strength is the greatest force or torque a person can produce in a defined task, classically a one-repetition maximum (1RM) or a dynamometer peak. Grgic, Lazinica, Schoenfeld and Pedisic, in a systematic review of test–retest reliability, found the 1RM assessment reliable across a range of exercises when familiarisation and standardised procedures were used (Grgic et al., 2020). Reliability is not validity for every question. A squat 1RM is not a jump, a clean, or a gait speed.

Hypertrophy is an increase in muscle size — fibre cross-section, anatomical cross-section, thickness, or volume — measured over weeks to months. Seynnes, de Boer and Narici, using imaging in seven young men during a 35-day high-intensity program, reported early increases in vastus lateralis anatomical cross-sectional area and architecture (Seynnes, de Boer, and Narici, 2007). That is a small, young, male, imaging series. It is strongly supported as evidence that size can change on a timescale of weeks. Acute muscle protein synthesis is not this outcome. Sports nutrition already made that point. It is restated because strength-training marketing spends a biopsy as if it were a circumference.

Power is force times velocity, or work divided by time. Cormie, McGuigan and Newton, in a two-part narrative synthesis, separated the biological basis of maximal neuromuscular power from the training considerations that try to raise it (Cormie, McGuigan, and Newton, 2011a; Cormie, McGuigan, and Newton, 2011b). Suchomel, Nimphius and Stone treated muscular strength as a foundation for athletic power, not as a synonym for it (Suchomel, Nimphius, and Stone, 2016). A heavy 1RM can rise while a jump does not. A jump can rise while a 1RM does not. Those are different tests.

Muscular endurance is the ability to sustain force or to repeat submaximal contractions. Campos and colleagues’ high-repetition arm is the cleanest within-study demonstration that local endurance can be preferentially trained (Campos et al., 2002). It is not “cardio,” and it is not VO₂max.

The four outcomes share motor units, connective tissue, and the need for progressive overload (Kraemer and Ratamess, 2004; American College of Sports Medicine, 2009). They diverge on load, velocity, proximity to failure, and the test used to declare success. Collapsing them into “getting stronger” is how a hypertrophy meta-analysis is sold as a power program and a grip-strength cohort is sold as a reason to deadlift.

03 A short history: DeLorme to the position stands

The modern clinic of progressive resistance is a polio clinic. DeLorme and Watkins described the technic of progressive resistance exercise as a method for restoring strength after injury and disease, built on a small number of sets with a repetition maximum that was raised as the patient improved (DeLorme and Watkins, 1948). DeLorme had already reported the quadriceps response to progressive-resistance exercise in poliomyelitis (DeLorme, 1948). Those papers are established as the English-language origin of the set-and-repetition grammar still used in gyms. They are not evidence about healthy recreational lifters, and they are not evidence about hypertrophy as a cosmetic endpoint.

What followed was a half-century of load-band arguments, military and athletic adoption, and, eventually, college position stands. The ACSM 2002 and 2009 progression models are the documents that tried to turn that argument into a table: novice versus intermediate versus advanced; strength versus hypertrophy versus power versus endurance; recommended ranges offered as evidence-graded summaries (American College of Sports Medicine, 2002; American College of Sports Medicine, 2009). Kraemer and Ratamess restated progression and prescription as a fundamentals review in the same period (Kraemer and Ratamess, 2004). These are syntheses. The primary allocations sit underneath them. The grade is strongly supported as a map of the early-2000s human record. It is weaker wherever later equated-volume trials have shown that a load band the 2009 stand treated as obligatory is not (Mitchell et al., 2012; Morton et al., 2016; Schoenfeld et al., 2015). That is not a reason to discard the stands. It is a reason not to treat them as frozen law.

The still older size-principle physiology is not a training history, but it is why the training history worked at all. Henneman, Somjen and Carpenter showed that motoneurone size ordered recruitment (Henneman, Somjen, and Carpenter, 1965). Lüscher, Ruenzel and Henneman later restated how size determines susceptibility to discharge (Lüscher, Ruenzel, and Henneman, 1979). Sale’s 1988 review is the paper that brought that physiology into the resistance-training literature as “neural adaptation” (Sale, 1988). Part Two takes that up as mechanism, not as nostalgia.

04 How force is produced

Force in a voluntary contraction is the product of how many motor units are recruited, how fast they fire, and how much force each fibre can produce in its current architectural and contractile state. Heckman and Enoka’s motor-unit review is the modern map of that statement (Heckman and Enoka, 2012). Rate coding — the frequency of action potentials — is how a recruited unit grades force after it has been brought in; Enoka’s later review treated rate coding as a primary control of muscle force (Enoka, 2017). Training adaptations in motor-unit behaviour, including recruitment and firing-rate changes, were reviewed by Duchateau, Semmler and Enoka as a human-recording literature, not as a gym slogan (Duchateau, Semmler, and Enoka, 2006).

The fibre cannot be asked to do more than its length, pennation, and moment-arm allow. Fukunaga, Ichinose, Ito, Kawakami and Fukashiro measured fascicle length and pennation in contracting human muscle in vivo (Fukunaga et al., 1997a). Fukunaga, Kawakami, Kuno, Funato and Fukashiro related architecture to function in the same experimental tradition (Fukunaga et al., 1997b). Those ultrasound studies are established as evidence that architecture is a force variable, not a caption. They are not a claim that a particular exercise “shapes” a muscle in a marketed way.

Tendon stiffness changes the time and the amplitude with which fibre force reaches the skeleton. Bone strain, over months, is the mechanostat’s input (Frost, 1987). None of that is visible in a single set’s EMG. The voltage is real. The month is a different measurement.

05 Motor units, the size principle, and rate coding

Henneman’s size principle is established in spinal motoneurones: smaller cells are recruited first; larger, higher-force units are recruited as the drive increases (Henneman, Somjen, and Carpenter, 1965; Lüscher, Ruenzel, and Henneman, 1979). The training implication that is usually stolen from this is false in one direction and true in another. False: that light loads cannot recruit high-threshold units. A high-threshold unit can be recruited by a light load taken near failure, which is why Mitchell and colleagues could produce hypertrophy at ~30% 1RM in young men when the sets were continued to volitional failure (Mitchell et al., 2012). True: that a heavy load recruits those units earlier in the set, which is why strength at high loads is more specific to high-load practice (Campos et al., 2002; Schoenfeld et al., 2015). The size principle does not pick a percentage. It picks an order.

Rate coding is how force is graded once a unit is active (Enoka, 2017). Training can change firing rates and the behaviour of the pool (Duchateau, Semmler, and Enoka, 2006). Those are human motor-unit recording findings, sparse and technically difficult. They are strongly supported as a class. They are not a licence to infer a firing-rate change from a feeling of effort.

06 Neural adaptations

Moritani and deVries, using surface EMG and limb circumference across an eight-week program in humans, argued that early strength gain was more neural than hypertrophic, with hypertrophy contributing later (Moritani and deVries, 1979). That paper is the type specimen of the early-neural story. It is strongly supported as a time-course argument in that sample and method. Circumference is a crude size measure; surface EMG is the method Vigotsky later warned against over-reading (Vigotsky et al., 2018). The direction has survived better methods. Sale’s narrative review gathered the reflex, cross-education, and skill evidence and named the class “neural adaptation to resistance training” (Sale, 1988). Aagaard, Simonsen, Andersen, Magnusson and Dyhre-Poulsen then showed, in a human evoked-potential study, that 14 weeks of heavy resistance training increased V-wave and H-reflex amplitudes during maximal contraction, a sign of increased descending drive and motoneurone excitability (Aagaard et al., 2002). That is a mechanistic human experiment, not a program comparison. Grade: strongly supported for a neural contribution to maximal force. It is not evidence that hypertrophy is optional if the goal is size.

Detraining and short-term disuse cut the other way. Hvid, Aagaard, Ørtenblad, Kjaer and Suetta later recorded plasticity in central neural drive with rapid loss and restoration of force after short-term disuse and recovery (Hvid et al., 2018). Hortobágyi, Houmard, Stevenson, Fraser, Johns and Israel measured performance loss after inactivity in power athletes (Hortobágyi et al., 1993). Neural adaptation is reversible. That is a finding, not a reason to train every day.

07 Muscle architecture and hypertrophy

Seynnes, de Boer and Narici’s 35-day imaging series is the early-hypertrophy type specimen: anatomical cross-section and fascicle architecture of vastus lateralis changed while the program was still in its first month (Seynnes, de Boer, and Narici, 2007). Fukunaga’s in-vivo architecture work is why those changes matter for force: pennation and fascicle length alter the fibre’s contribution to joint torque (Fukunaga et al., 1997a; Fukunaga et al., 1997b).

Hypertrophy as a training outcome is a dose problem. Schoenfeld, Ogborn and Krieger’s meta-analysis reported a graded relationship between weekly resistance-training volume and hypertrophic gain, with diminishing returns rather than a single magic set count (Schoenfeld, Ogborn, and Krieger, 2017). Krieger’s earlier meta-regression found multiple sets superior to single sets for strength, with a smaller advantage as set number rose (Krieger, 2009). Wernbom, Augustsson and Thomeé mapped frequency, intensity, volume and mode against whole-muscle hypertrophy in a narrative synthesis of the human record then available (Wernbom, Augustsson, and Thomeé, 2007). These are syntheses. They inherit the bias, the untrained samples, and the measurement heterogeneity of their inputs. Grade: strongly supported that volume is a hypertrophic driver in the ranges those papers modelled. They are not a weekly set target for a reader.

Load is not the same driver. Mitchell and colleagues assigned young men to 30% or 80% 1RM, taken to failure, and found similar hypertrophic gains (Mitchell et al., 2012). Morton and colleagues, in trained men, found the same pattern and added that systemic hormones did not determine the gain (Morton et al., 2016). Schoenfeld, Peterson, Ogborn, Contreras and Sonmez compared low- versus high-load training in well-trained men and reported that hypertrophy could be similar while dynamic strength favoured the heavier load (Schoenfeld et al., 2015). Lasevicius and colleagues, equating volume load across intensities, reported a load threshold below which hypertrophy was inferior even with more repetitions (Lasevicius et al., 2018). That last paper is the counterweight. Conflict is left as conflict: very light loads taken to failure can hypertrophy young and trained men in some allocations; equated-volume work suggests a floor. The 2009 ACSM stand’s heavier hypertrophy band is not cancelled. It is no longer the only human method.

08 Tendon and connective tissue

Tendon adapts more slowly than muscle and is not a decoration on a fibre. Bohm, Mersmann and Arampatzis systematically reviewed human tendon adaptation to mechanical loading and concluded that tendon stiffness and modulus increase with the right loading character, with high-intensity loading better supported than low-intensity (Bohm, Mersmann, and Arampatzis, 2015). Magnusson’s group has spent a career on the same tissue: tendon properties in relation to muscular activity (Reeves, Maganaris, and Narici sit adjacent; Magnusson’s 2003 and later papers in the reviewed record treat training, ageing, collagen homeostasis, and heavy-slow loading as a clinical and physiological problem) (Magnusson et al., 2003; Magnusson and Kjaer, 2019; Kjaer and Magnusson, 2015). These are human imaging, biopsy, and clinical-comparison papers. Grade: strongly supported that tendon is a trainable tissue. Emerging as to which exact loading scheme is best for which tendon, because the tendinopathy trials and the healthy-adaptation trials are not the same experiment.

Muscle can outrun its tendon. That mismatch is a plausible injury story and a thin causal demonstration in healthy training studies. It is named so that “get strong fast” is not allowed to pretend the connective tissue signed the same contract.

09 Bone

Frost’s mechanostat is a theoretical account: bone mass is regulated by mechanical strain, with muscle force a principal source of that strain (Frost, 1987). It is plausible as a framework and not a human loading trial. The ACSM position stand on physical activity and bone health, with Kohrt as lead, treated resistance exercise and impact as osteogenic candidates, with site-specificity and a thinner interventional record than the muscle literature (Kohrt et al., 2004). Grade: strongly supported that bone is load-sensitive and that resistance and impact are among the loads the stand judged useful. Emerging as to the dose that changes bone mineral density in a particular site in a particular decade of life. High-intensity resistance and impact protocols in older adults with low bone mass exist in the wider literature; this review’s verified ACSM bone stand is the conservative citation. A protocol is not recommended here.

10 Specificity and progressive overload

Specificity is the least glamorous law in the file and the one the commercial literature violates first. Campos and colleagues’ three-regimen trial is the within-study demonstration: the test you improve is the test you practised, with some spillover (Campos et al., 2002). Sale’s neural review is why: skill, recruitment, and the task are part of the adaptation (Sale, 1988). Suchomel’s athletic-strength review is the sport version of the same sentence (Suchomel, Nimphius, and Stone, 2016).

Progressive overload is DeLorme’s technic restated for people who were not recovering from polio: the work must increase as the tissue allows (DeLorme and Watkins, 1948; Kraemer and Ratamess, 2004). The ACSM stands made progression their title (American College of Sports Medicine, 2002; American College of Sports Medicine, 2009). Overload is not a licence to add load every session. Bickel, Cross and Bamman showed that the dose required to retain a resistance-training adaptation in young and older adults was smaller than the dose that produced it (Bickel, Cross, and Bamman, 2011). Acquisition and maintenance are different experiments.

11 Intensity and load

“Heavy weights are always necessary” is the first required challenge. Partially accepted.

Necessary for what. For maximal strength at a 1RM or a high-load test, heavy practice is strongly supported. Schoenfeld and colleagues’ well-trained comparison and Campos and colleagues’ three-band trial both favoured heavier loading for dynamic maximal strength (Schoenfeld et al., 2015; Campos et al., 2002). The ACSM 2009 stand said the same as a synthesis (American College of Sports Medicine, 2009). For hypertrophy, heavy is sufficient and not unique. Mitchell 2012, Morton 2016, and Schoenfeld 2015 are the human allocations that broke the monopoly (Mitchell et al., 2012; Morton et al., 2016; Schoenfeld et al., 2015). Lasevicius 2018 is the paper that put a floor back under the claim: equated volume load at very low intensities underperformed (Lasevicius et al., 2018). For power, load is a velocity problem, not a 1RM-percentage slogan (Cormie, McGuigan, and Newton, 2011b). For local endurance, the light, high-repetition arm in Campos is the evidence, not a light day in a strength block (Campos et al., 2002).

West and Phillips, and then Morton and colleagues, closed the hormone-as-dose version of the heavy-day argument (West and Phillips, 2012; Morton et al., 2016). A hard session can raise circulating anabolic hormones. That pulse did not determine the month.

12 Volume, sets, and frequency

Krieger’s meta-regression is the multiple-set paper most often stolen: more than one set beat one set for strength, with diminishing returns (Krieger, 2009). Schoenfeld, Ogborn and Krieger’s weekly-volume meta-analysis is the hypertrophy twin (Schoenfeld, Ogborn, and Krieger, 2017). Wernbom, Augustsson and Thomeé had already warned that the human hypertrophy record was a scatter of frequencies, intensities and modes, not a single curve (Wernbom, Augustsson, and Thomeé, 2007). Grade: strongly supported that weekly volume is a hypertrophic driver in the modelled range; strongly supported that set number is not linear.

Frequency is the second required challenge. Partially accepted as “higher can work,” rejected as “higher is superior once volume is equated.” Grgic, Schoenfeld and colleagues’ hypertrophy frequency meta-analysis found an advantage for higher frequency that largely tracked greater weekly volume (Grgic, Schoenfeld, Davies, et al., 2016; Grgic and Schoenfeld, 2018; Schoenfeld, Grgic and Krieger, 2019). Grgic and colleagues’ strength-frequency systematic review found small or negligible differences when weekly volume was comparable (Grgic et al., 2018). High-frequency training is a volume-delivery method. It is not a law.

13 Proximity to failure

“Failure training is required” is the third required challenge. Partially accepted for low-load hypertrophy; rejected as a general requirement.

Lasevicius and colleagues reported that muscle failure promoted greater hypertrophy in low-load but not in high-load training in their allocation (Lasevicius et al., 2022). Helms, Refalo and colleagues’ later proximity-to-failure syntheses and experiments treated repetitions in reserve as a continuous variable: stopping short of failure can produce similar hypertrophy to momentary failure when the set is already hard, with possible advantages for fatigue management (Refalo, Helms, et al., 2022; Robinson, Helms, et al., 2023; Refalo et al., 2023; Robinson et al., 2024; Robinson, Pelland, et al., 2024). Hermann, Mohan, Enes and colleagues’ 2025 single-set-to-failure paper is a recent human test of the other extreme — one hard set — and is emerging as a challenge to high-volume necessity, not a replacement for the volume metas (Hermann et al., 2025). Conflict is left as conflict. Failure is a tool that recruits high-threshold units when the load is light. It is not a moral.

14 Rest, tempo, and range of motion

Rest intervals are a strength-versus-hypertrophy split in the ACSM stands: longer rest for high-load strength, shorter rest historically offered for hypertrophy (American College of Sports Medicine, 2009). Later human work has not made short rest a hypertrophic requirement; it has made rest a fatigue-and-volume-delivery problem. This harvest did not recover a single definitive rest-interval mega-trial that closes the question. Grade: strongly supported that very short rest can reduce acute load and therefore volume; emerging as an independent hypertrophic variable once volume is caught up.

Tempo and time-under-tension are the same class. The ACSM stands mention controlled tempos; the equated-volume human record does not make a slow eccentric into a unique hypertrophic drug. Grade: plausible as a way to keep tension on a fibre; not established as superior to a moderate tempo at the same proximity to failure.

Range of motion is the fourth required challenge. Partially accepted. Newmire and Willoughby’s 2018 systematic review found that full ROM generally produced similar or greater hypertrophy than partial ROM, with exercise-specific exceptions (Newmire and Willoughby, 2018). Schoenfeld and Grgic’s 2020 narrative update kept that direction (Schoenfeld and Grgic, 2020). Pedrosa, Lima, Schoenfeld and colleagues then reported that partial ROM, particularly at long muscle lengths, can elicit favourable hypertrophic adaptations (Pedrosa et al., 2022). Pedrosa and colleagues’ 2023 randomised comparison found that training in the initial (lengthened) ROM promoted greater muscle adaptations than the final ROM (Pedrosa et al., 2023). Full ROM is not cancelled. The monopoly of “always full” is. The honest statement is lengthened-position tension, not a religious complete lockout.

15 Exercise selection: free weights, machines, one limb or two

“Free weights beat machines” is the fifth required challenge. Rejected as a hypertrophy law; partially accepted as a skill-and-specificity claim.

Schwanbeck, Chilibeck and Binsted compared free-weight squat to Smith-machine squat with surface EMG in 6 men and 6 women and reported greater vastus and gastrocnemius EMG in the free-weight condition (Schwanbeck, Chilibeck, and Binsted, 2009). That is an acute EMG study in twelve people. Vigotsky’s methods paper is why it cannot be spent as hypertrophy or strength (Vigotsky et al., 2018). The ACSM stands treat both free weights and machines as legitimate tools, with free weights favoured when the goal is coordination in a standing task (American College of Sports Medicine, 2009). Specificity, not moral superiority, is the remaining claim. A Smith-machine 1RM is not a free-squat 1RM. A vastus lateralis thickness change does not have to care.

Unilateral versus bilateral training is a laterality-and-stability problem, not a hypertrophy monopoly. The ACSM stands allow both (American College of Sports Medicine, 2009). Open versus closed kinetic chain is an old rehabilitation argument; Escamilla-type kinetic-chain language is useful in the clinic and is not a general-population loading law. This harvest did not recover a decisive healthy-adult RCT that makes chain type a primary hypertrophic variable. Grade: plausible as task specificity; not established as a size or 1RM law.

16 Periodization

“Periodization is superior” is the sixth required challenge. Partially accepted.

Kiely’s narrative essay asked whether twenty-first-century periodization was evidence-led or tradition-driven and answered: mostly tradition, with a thinner experimental core than the textbooks imply (Kiely, 2012). Williams, Tolusso, Fedewa and Esco’s meta-analysis compared periodized and non-periodized resistance training on maximal strength and found an advantage for periodization that was small and sensitive to study quality (Williams et al., 2017). A later volume-equated comparison of periodization models on strength and hypertrophy did not make a single model obligatory (Moesgaard and colleagues, 2022). Grade: strongly supported that some form of planned variation can beat a static program for strength in some metas. Rejected as proof that a named model — linear, undulating, block — is the adaptation. Variation and progression can be written without a Soviet diagram. Issurin’s block-periodization writing sits in the wider literature; this review’s verified Kiely and Williams papers are the ones that constrain the claim.

17 Autoregulation, RPE/RIR, and velocity

Zourdos and colleagues validated a resistance-training-specific RPE scale anchored to repetitions in reserve (Zourdos et al., 2016). Helms, Cronin, Storey and Zourdos then described the application of that RIR-based RPE scale for resistance training (Helms et al., 2016). Later work from the same groups tested RPE versus percentage-1RM loading and RPE as volume autoregulation inside periodized programs (Helms et al., 2018a; Helms et al., 2018b). Grade: strongly supported that trained lifters can estimate proximity to failure with usable accuracy, especially later in a set. Emerging as a claim that RPE-loaded programs beat percentage programs when volume and proximity are otherwise matched.

Velocity-based training is the instrumented version of the same idea. Pareja-Blanco, Rodríguez-Rosell, Sánchez-Medina and colleagues showed that the velocity loss allowed in a set changes the adaptation: larger losses look more like hypertrophy-and-fatigue; smaller losses look more like strength-and-power with less fatigue (Pareja-Blanco et al., 2016). Banyard, Tufano, Weakley and colleagues reported that velocity-based methods changed jump, sprint and change-of-direction more than they changed maximal strength in their comparison (Banyard et al., 2021). Behm, Konrad, Nakamura and colleagues’ 2025 narrative review restated contraction-velocity best practice (Behm et al., 2025). Grade: strongly supported that velocity loss is a real acute and chronic variable. Emerging as a claim that a velocity device is required rather than useful.

18 Measuring strength

The 1RM is reliable when familiarised and standardised (Grgic et al., 2020). Prediction equations from submaximal repetitions exist in the wider literature; this review did not lock a single equation as canonical, and no equation is offered as a tool. Velocity at a relative load can estimate a daily 1RM; that is the VBT literature above, not a new test.

What 1RM is not: a power test, a hypertrophy test, a bone test, a mortality test, or a reason to test untrained older adults without the familiarisation the reliability review assumed. Fiatarone’s nonagenarian work used high-intensity efforts in a supervised setting with a different risk calculus than a garage 1RM (Fiatarone et al., 1990). Measurement is a protocol. It is not a personality.

19 Training status

The ACSM stands stratified novice, intermediate and advanced and assigned different progression rates to each (American College of Sports Medicine, 2002; American College of Sports Medicine, 2009). That stratification is strongly supported as a qualitative fact: untrained people gain faster; trained people need more precise overload. It is not established as a set of universal week-counts. Schoenfeld’s 2015 well-trained comparison exists because the untrained literature does not travel unchanged into people who already lift (Schoenfeld et al., 2015). Morton 2016 is the same warning with hormones removed (Morton et al., 2016). A meta-analysis that is 80% untrained men is a fact about untrained men.

20 Sex

“Women need different programs” is the seventh required challenge. Partially accepted as a statement about baseline strength and fatigue; rejected as a statement that hypertrophy cannot occur on similar relative doses.

Roberts, Nuckols and Krieger’s systematic review and meta-analysis of sex differences in resistance training found that relative hypertrophic and strength gains were broadly similar, with men stronger in absolute terms at baseline (Roberts, Nuckols, and Krieger, 2020). Nuzzo’s 2023 narrative review of sex differences in strength, endurance, activation, size, fibre type and hypertrophy is the current map of the same terrain: large absolute-strength differences, smaller relative-training-response differences, and a literature that still under-samples women in convenient trials (Nuzzo, 2023). Grade: strongly supported for similar relative hypertrophy; established for absolute-strength dimorphism; emerging wherever menstrual-cycle periodization or a unique female loading scheme is claimed. This document does not prescribe a cycle-based program. It records that the claim has a thinner trial base than the relative-gain metas.

21 Youth

“Youth resistance training is unsafe or stunts growth” is the eighth required challenge, youth half. Rejected as a growth-and-safety claim when the work is supervised and progressed; accepted as a warning against uncontrolled maximal loading and poor instruction.

Faigenbaum, Kraemer, Blimkie, Jeffreys, Micheli, Nitka and Rowland’s NSCA updated position statement treated youth resistance training as effective and acceptably safe under qualified instruction (Faigenbaum et al., 2009). Lloyd, Faigenbaum and the 2014 international consensus restated that position for a wider set of organisations (Lloyd et al., 2014). Faigenbaum’s later “mythology” essay named the persistent folk claims — growth-plate ruin, lost flexibility, wasted time — and treated them as claims the position statements had already answered (Faigenbaum, 2022). These are position statements and a narrative, not a new RCT. Grade: strongly supported that supervised youth resistance training is not the injury epidemic the folklore describes. They are not a licence for an unsupervised adolescent 1RM contest.

22 Older adults

“Older adults cannot lift heavy” is the eighth required challenge, age half. Rejected as a physiological prohibition; partially accepted as a supervision and comorbidity statement.

Fiatarone, Marks, Ryan, Meredith, Lipsitz and Evans assigned ten institutionalised nonagenarians to a high-intensity strength program and reported large relative strength gains and muscle-area increases (Fiatarone et al., 1990). Fiatarone, O’Neill, Ryan, Clements, Solares, Nelson, Roberts, Kehayias, Lipsitz and Evans then randomised 100 frail nursing-home residents (mean age 87) to resistance training, a nutritional supplement, both, or neither; exercise, not the supplement, increased strength and gait velocity (Fiatarone et al., 1994). Those two papers are established as the demonstration that very old, frail people can respond to high-intensity resistance training in a supervised setting. They are not a community gym protocol.

Peterson, Rhea, Sen and Gordon’s meta-analysis found that resistance exercise increased muscular strength in older adults, with higher intensity associated with larger gains in that synthesis (Peterson et al., 2010). Borde, Hortobágyi and Granacher mapped dose–response relationships in healthy old adults (Borde, Hortobágyi, and Granacher, 2015). Fragala, Cadore, Dorgo, Izquierdo, Kraemer, Peterson and Ryan’s NSCA position statement is the current older-adult map: resistance training is indicated for ageing muscle, strength, function and selected health outcomes, with load and supervision treated as clinical variables (Fragala et al., 2019). Grade: established that older adults adapt; strongly supported that intensity is not forbidden by age. Comorbidity, osteoporosis, and fall risk are why Fragala’s paper is a position statement and not a blog.

23 Athletes and the power problem

Suchomel, Nimphius and Stone treated muscular strength as a contributor to athletic performance, not as the performance (Suchomel, Nimphius, and Stone, 2016). Cormie, McGuigan and Newton’s two-part synthesis is the power textbook: biological basis first, then training considerations, including the load–velocity trade-off (Cormie, McGuigan, and Newton, 2011a; Cormie, McGuigan, and Newton, 2011b). Banyard’s velocity-based comparison is a reminder that the test you want — jump, sprint, change of direction — may move when the 1RM does not, and the reverse (Banyard et al., 2021). Grade: strongly supported that strength is a foundation for many sports. Rejected as a claim that the heaviest squat is the best power program. Specificity survives contact with a barbell.

24 Detraining and maintenance

Hortobágyi and colleagues measured what happened when power athletes stopped: performance declined on a timescale of weeks, not years (Hortobágyi et al., 1993). Bickel, Cross and Bamman then asked the maintenance question in young and older adults after a resistance-training period: a reduced frequency could retain much of the adaptation that a higher dose had built, with age affecting how completely the gain was kept (Bickel, Cross, and Bamman, 2011). Grade: strongly supported that the trained state is perishable and that the maintenance dose is smaller than the acquisition dose. “Detraining” is not a moral failure. It is a measured decay.

25 Safety

Aasa, Svartholm, Andersson and Berglund systematically reviewed injuries among weightlifters and powerlifters and found that injury rates were modest relative to many team sports, that the shoulder, back and knee dominated, and that the literature was heterogeneous in definition and surveillance (Aasa et al., 2017). Grade: strongly supported as a description of those two sports’ recorded injuries. It is not a safety certificate for an unsupervised novice, a contested youth 1RM, or a medical population. Fiatarone’s supervised geriatric work and Faigenbaum’s supervised youth statements are the other two poles of the same sentence (Fiatarone et al., 1994; Faigenbaum et al., 2009). Technique, load progression, and comorbidity are why this section is not a waiver.

26 Cardiovascular and metabolic effects

Cornelissen and Smart’s systematic review and meta-analysis of exercise training for blood pressure included resistance-training arms and found reductions that were real and smaller than the most optimistic clinic posters (Cornelissen and Smart, 2013). MacDonald, Johnson, Collins and colleagues then meta-analysed dynamic resistance training as stand-alone antihypertensive lifestyle therapy and reported clinically meaningful reductions in some hypertensive samples (MacDonald et al., 2016). Strasser, Siebert and Schobersberger systematically reviewed resistance training in the metabolic syndrome and found improvements in selected metabolic markers (Strasser, Siebert, and Schobersberger, 2010). Westcott’s “resistance training is medicine” essay is a narrative of that health literature, not a new trial (Westcott, 2012). Maestroni, Read, Bishop, Papadopoulos, Cooper, Jardine and Turner reviewed musculoskeletal-system benefits with a practical-application frame (Maestroni et al., 2020). Grade: strongly supported that resistance training can lower blood pressure and improve selected metabolic markers. Rejected as a claim that it is interchangeable with aerobic training for cardiorespiratory fitness. The WHO 2020 guidelines and the 2018 Physical Activity Guidelines for Americans keep both aerobic and muscle-strengthening activity in the same sentence for a reason (Bull et al., 2020; Piercy et al., 2018).

27 Functional outcomes

Fiatarone 1994 is still the cleanest functional RCT in the very old: gait velocity and strength moved with the training, not with the supplement (Fiatarone et al., 1994). Fragala’s NSCA older-adult statement treats sit-to-stand, gait, and fall-risk-adjacent function as reasons the position exists (Fragala et al., 2019). Peterson’s and Borde’s metas are strength metas first; function is a downstream claim that travels with the strength change more cleanly in frail samples than in healthy recreational adults who were never functionally limited (Peterson et al., 2010; Borde, Hortobágyi, and Granacher, 2015). Grade: established in frail older adults that high-intensity resistance training can change function. Emerging as a claim that a recreational 1RM change will change an already-able adult’s life.

28 Healthspan and mortality associations

This is the tenth required challenge. Partially accepted as a robust observational signal; rejected as proof that a training program caused the deaths that did not happen.

Rantanen, Guralnik, Foley, Masaki, Leveille, Curb and White reported that midlife hand-grip strength predicted old-age disability in a large Hawaiian cohort of men (Rantanen et al., 1999). Newman, Kupelian, Visser, Simonsick, Goodpaster, Kritchevsky, Tylavsky, Rubin and Harris, in the Health ABC study, found that strength, but not muscle mass, was associated with mortality (Newman et al., 2006). Ruiz, Sui, Lobelo, Morrow, Jackson, Sjöström and Blair reported that muscular strength was associated with mortality in men in an aerobically adjusted cohort (Ruiz et al., 2008). Leong, Teo, Rangarajan, Lopez-Jaramillo, Avezum, Orlandini, Seron, Ahmed, Rosengren, Kelishadi and the PURE investigators found that grip strength predicted all-cause and cardiovascular mortality across 17 countries (Leong et al., 2015). García-Hermoso, Cavero-Redondo, Ramírez-Vélez, Ruiz, González-Gross, Martínez-Vizcaíno and colleagues meta-analysed muscular strength as a predictor of all-cause mortality in apparently healthy populations (García-Hermoso et al., 2018). Volaklis, Halle and Meisinger reviewed the same association as a narrative (Volaklis, Halle, and Meisinger, 2015).

Behaviour is a different exposure than a dynamometer. Stamatakis, Lee, Bennie, Freeston, Hamer, O’Donovan, Ding, Bauman and Mavros pooled population data on strength-promoting exercise and health (Stamatakis et al., 2018). Saeidifard, Medina-Inojosa, West, Olson, Somers, Bonikowske, Prokop, Vinciguerra and Lopez-Jimenez systematically reviewed and meta-analysed resistance training and mortality (Saeidifard et al., 2019). Momma, Kawakami, Honda and Sawada associated muscle-strengthening activities with lower risk and mortality in major non-communicable diseases (Momma et al., 2022). Giovannucci, Rezende and Lee reported a related observational association with cardiovascular disease, type 2 diabetes and cancer (Giovannucci, Rezende, and Lee, 2021). The WHO 2020 guidelines and the US Physical Activity Guidelines recommend muscle-strengthening activity on at least two days a week as a public-health behaviour, not as a 1RM target (Bull et al., 2020; Piercy et al., 2018).

What these papers cannot do is randomise a nation to a squat program and count the dead. Residual confounding — people who lift also walk, earn, smoke less, and survive long enough to be asked — is the epidemiologist’s first objection. Reverse causation — impending illness lowers strength — is the second. Newman’s dissociation of strength from mass is a warning against treating hypertrophy as the mortality exposure (Newman et al., 2006). Grade: established as an observational association. Speculative as a claim that any particular program added years. The guideline two-days-a-week sentence is a public-health synthesis, not a training prescription in this document.

29 What the four outcomes do not share

Maximal strength wants high-load practice and a high-load test (Campos et al., 2002; Schoenfeld et al., 2015; American College of Sports Medicine, 2009). Hypertrophy wants weekly volume and hard sets; it will accept a range of loads if recruitment is high (Schoenfeld, Ogborn, and Krieger, 2017; Mitchell et al., 2012; Morton et al., 2016; Lasevicius et al., 2018). Power wants velocity and the test you will perform (Cormie, McGuigan, and Newton, 2011b; Weakley et al., 2021). Local endurance wants repeated submaximal work (Campos et al., 2002). Neural drive can raise a 1RM before a fibre has thickened (Moritani and deVries, 1979; Sale, 1988; Aagaard et al., 2002). A tendon and a bone have not signed the fibre’s calendar (Bohm, Mersmann, and Arampatzis, 2015; Frost, 1987; Kohrt et al., 2004). A hormone pulse and an EMG peak have signed nothing (West and Phillips, 2012; Vigotsky et al., 2018; Schwanbeck, Chilibeck, and Binsted, 2009). A grip dynamometer has signed a death certificate in a cohort, not a program (Leong et al., 2015; Newman et al., 2006).

The barbell is shared. The claim is not.

Standing constraint This document describes published research. It is not medical advice. No human use, dose, route, schedule, diagnosis, or individualized training prescription is recommended anywhere in this document. Reported loads, set structures, and frequencies are experimental or consensus parameters, always with the population that produced them.

Table A. Four outcomes distinguished

Quantities are study or consensus parameters, not recommendations.

Table B. Landmark human studies

Table C. Programming variables by outcome

Table D. Adversarial resolutions

Evidence handling

Study type is named in the reporting sentence. Animal and in-vitro results, when mentioned at all, are not phrased as human outcomes. Position stands are secondary maps to primary trials. Meta-analyses inherit the bias of their inputs. Conflict is left as conflict: load-as-optional for hypertrophy versus Lasevicius’s floor; full versus lengthened-partial ROM; periodization’s small meta-analytic edge versus Kiely’s tradition critique; grip strength as a risk marker versus training as a cause of longevity.

Evidence grades used in the prose:

Adversarial resolution

The evidence and argument were tested against six critical perspectives; the objections and where the text answers them are set out below.

Unresolved: a large, equated-volume, multi-year RCT of periodization models in trained women; a current, sport-wide injury surveillance series that uses one injury definition; a randomised test of whether raising grip strength — rather than being the sort of person who has one — changes mortality; a definitive rest-interval mega-trial that isolates rest from weekly volume.

Limitations and sibling scope

Project 06 is a general-science corpus. Its on-topic harvest for this title was sarcopenia- and ageing-heavy, not a resistance-training library; classic programming primaries were largely absent as titled records and were recovered from NCBI. Project 07 journalism was scanned read-only (32,659 articles). The dominant 07 signal was lifestyle and GLP-1 coverage that mentioned lifting. It was not used as evidence. Local Firecrawl retrieved WHO 2020, US Physical Activity Guidelines 2018, NSCA public position pages, and CDC muscle-strengthening basics as institutional discovery; ACSM 2009 was recovered from Europe PMC / NCBI (PMID 19204579), not from a paywalled PDF. This title does not replace the sports-nutrition article. It does not treat therapeutic peptides, doping protocols, or clinical rehabilitation as its subject.

Glossary

1RM. One-repetition maximum — a task-specific strength test, not a tissue. Motor unit. A motoneurone and the fibres it innervates. Size principle. Smaller motoneurones recruited before larger ones as drive increases. Rate coding. Firing frequency as a force grader. Hypertrophy. Increase in muscle size over weeks to months — not an acute tracer. RIR / RPE. Repetitions in reserve and the resistance-specific effort scale anchored to them. Velocity loss. Fractional slowing inside a set; a fatigue and adaptation variable. Mechanostat. Frost’s account of bone mass as strain-regulated. Energy availability. A sports-nutrition construct; not redefined here.

References

Numbered, surname-sorted, generated from NCBI-verified records. In-prose citations are author–year. No identifier in this document was assigned from memory: every PMID used above was confirmed against its NCBI record.