Research

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/282534834
Dose–Response Relationships of Resistance Training in Healthy Old Adults:

A Systematic Review and Meta-Analysis
Article in Sports Medicine · September 2015

DOI: 10.1007/s40279-015-0385-9
CITATIONS READS
65 454
3 authors, including:
Ron Borde Urs Granacher

Universität Potsdam Universität Potsdam
7 PUBLICATIONS 116 CITATIONS 208 PUBLICATIONS 2,392 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Metastability in Sport Biomechanics View project
Effects of modality mappings within working memory on postural control, associated neural correlates, and training-induced
modulation of dual-task performance in old age View project
All content following this page was uploaded by Urs Granacher on 06 October 2015.
The user has requested enhancement of the downloaded file.

Sports Med
DOI 10.1007/s40279-015-0385-9
SYSTEMATIC REVIEW
Dose–Response Relationships of Resistance Training in Healthy

Old Adults: A Systematic Review and Meta-Analysis
Ron Borde1 • Tibor Hortobágyi2,3 • Urs Granacher1
Ó The Author(s) 2015. This article is published with open access at Springerlink.com
Abstract Study Eligibility Criteria The initial search identified 506

Background Resistance training (RT) is an intervention studies, with a final yield of 25 studies. Only RCTs that
frequently used to improve muscle strength and morphol- examined the effects of RT in adults with a mean age of 65
ogy in old age. However, evidence-based, dose–response and older were included. The 25 studies quantified at least
relationships regarding specific RT variables (e.g., training one measure of muscle strength or morphology and suffi-
period, frequency, intensity, volume) are unclear in healthy ciently described training variables (e.g., training period,
old adults. frequency, volume, intensity).
Objectives The aims of this systematic review and meta- Study Appraisal and Synthesis Methods We quantified
analysis were to determine the general effects of RT on the overall effects of RT on measures of muscle strength
measures of muscle strength and morphology and to pro- and morphology by computing weighted between-subject
vide dose–response relationships of RT variables through standardized mean differences (SMDbs) between interven-
an analysis of randomized controlled trials (RCTs) that tion and control groups. We analyzed the data for the main
could improve muscle strength and morphology in healthy outcomes of one-repetition maximum (1RM), maximum
old adults. voluntary contraction under isometric conditions (MVC),
Data Sources A computerized, systematic literature and muscle morphology (i.e., cross-sectional area or vol-
search was performed in the electronic databases PubMed, ume or thickness of muscles) and assessed the method-
Web of Science, and The Cochrane Library from January ological study quality by Physiotherapy Evidence Database
1984 up to June 2015 to identify all RCTs related to RT in (PEDro) scale. Heterogeneity between studies was assessed
healthy old adults. using I2 and v2 statistics. A random effects meta-regression
was calculated to explain the influence of key training
variables on the effectiveness of RT in terms of muscle
This article is part of the Topical Collection on Exercise to improve strength and morphology. For meta-regression, training
mobility in healthy aging.
variables were divided into the following subcategories:
& Ron Borde volume, intensity, and rest. In addition to meta-regression,
[email protected] dose–response relationships were calculated independently
Urs Granacher for single training variables (e.g., training frequency).
[email protected] Results RT improved muscle strength substantially (mean
1
SMDbs = 1.57; 25 studies), but had small effects on
Division of Training and Movement Sciences, Research
Focus Cognition Sciences, University of Potsdam, Am Neuen
measures of muscle morphology (mean SMDbs = 0.42;
Palais 10, Building 12, 14469 Potsdam, Germany nine studies). Specifically, RT produced large effects in
2 both 1RM of upper (mean SMDbs = 1.61; 11 studies) and
Center for Human Movement Sciences, University Medical
Center Groningen, University of Groningen, Groningen, lower (mean SMDbs = 1.76; 19 studies) extremities and a
The Netherlands medium effect in MVC of lower (mean SMDbs = 0.76;
3
Faculty of Health and Life Sciences, Northumbria University, four studies) extremities. Results of the meta-regression
Newcastle Upon Tyne, UK revealed that the variables ‘‘training period’’ (p = 0.04)
123
R. Borde et al.
and ‘‘intensity’’ (p \ 0.01) as well as ‘‘total time under important role in improving muscle strength and mor-
tension’’ (p \ 0.01) had significant effects on muscle phology and should be implemented in exercise training
strength, with the largest effect sizes for the longest programs targeting healthy old adults. Still, further
training periods (mean SMDbs = 2.34; 50–53 weeks), research is needed to reveal optimal dose–response rela-
intensities of 70–79 % of the 1RM (mean SMDbs = 1.89), tionships following RT in healthy as well as mobility
and total time under tension of 6.0 s (mean limited and/or frail old adults.
SMDbs = 3.61). A tendency towards significance was
found for rest in between sets (p = 0.06), with 60 s
showing the largest effect on muscle strength (mean
Key Points
SMDbs = 4.68; two studies). We also determined the
independent effects of the remaining training variables on
Meta-regression of data from 25 studies revealed that
muscle strength. The following independently computed
a resistance training (RT) program with the goal to
training variables are most effective in improving measures
increase healthy old adults’ muscle strength is
of muscle strength: a training frequency of two sessions per
characterized by a training period of 50–53 weeks, a
week (mean SMDbs = 2.13), a training volume of two to
training intensity of 70–79 % of the one-repetition
three sets per exercise (mean SMDbs = 2.99), seven to nine
maximum (1RM), a time under tension of 6 s per
repetitions per set (mean SMDbs = 1.98), and a rest of
repetition, and a rest in between sets of 60 s.
4.0 s between repetitions (SMDbs = 3.72). With regard to
Selecting a training frequency of two sessions per
measures of muscle morphology, the small number of
week, a training volume of two to three sets per
identified studies allowed us to calculate meta-regression
exercise, seven to nine repetitions per set, and a rest
for the subcategory training volume only. No single
of 4.0 s between repetitions could also improve
training volume variable significantly predicted RT effects
efficacy of training.
on measures of muscle morphology. Additional training
variables were independently computed to detect the lar- The meta-regression revealed that none of the
gest effect for the single training variable. A training period examined training variables of volume (e.g., period,
of 50–53 weeks, a training frequency of three sessions per frequency, number of sets, number of repetitions)
week, a training volume of two to three sets per exercise, predicted the effects of RT on measures of muscle
seven to nine repetitions per set, a training intensity from morphology. Yet, RT to improve muscle
51 to 69 % of the 1RM, a total time under tension of 6.0 s, morphology seems to be effective using the
a rest of 120 s between sets, and a rest of 2.5 s between following independently computed training
repetitions turned out to be most effective. variables: a training period of 50–53 weeks, a
Limitations The current results must be interpreted with training frequency of three sessions per week, a
caution because of the poor overall methodological study training volume of two to three sets per exercise,
quality (mean PEDro score 4.6 points) and the considerable seven to nine repetitions per set, a training intensity
large heterogeneity (I2 = 80 %, v2 = 163.1, df = 32, from 51 to 69 % of the 1RM, a total time under
p \ 0.01) for muscle strength. In terms of muscle mor- tension of 6.0 s, a rest of 120 s between sets, and a
phology, our search identified nine studies only, which is 2.5-s rest between repetitions.
why we consider our findings preliminary. While we were This meta-analysis provides preliminary data for
able to determine a dose–response relationship based on therapists, practitioners, and clinicians regarding
specific individual training variables with respect to muscle relevant RT variables and their dose–response
strength and morphology, it was not possible to ascertain relationships to improve muscle strength and
any potential interactions between these variables. We morphology in healthy old adults.
recognize the limitation that the results may not represent
one general dose–response relationship.
Conclusions This systematic literature review and meta-
analysis confirmed the effectiveness of RT on specific
measures of upper and lower extremity muscle strength and 1 Introduction
muscle morphology in healthy old adults. In addition, we
were able to extract dose–response relationships for key With the onset of the sixth decade in life, degenerative
training variables (i.e., volume, intensity, rest), informing processes affect the neuromuscular system in terms of
clinicians and practitioners to design effective RTs for losses in muscle strength (dynapenia) and muscle mass
muscle strength and morphology. Training period, inten- (sarcopenia) [1–3]. Neural (e.g., numerical loss of alpha
sity, time under tension, and rest in between sets play an motoneurons) and morphological factors (e.g., reduced
123
Resistance Training in Old Age
number and size of particularly type-II muscle fibers) as of non-randomized controlled trials (RCTs) [12, 13] }, the
well as their interaction are responsible for age-related number of included training variables (e.g., traditional
declines in muscle strength and mass [4]. There is evidence variables such as training period, frequency, volume,
that muscular weakness is highly associated with impaired intensity only) [14–16], and by focusing only on direct
mobility and an increased risk for falls [5]. Moreover, comparisons of intervention groups (e.g., high- vs. low-
lower extremity muscle weakness was identified as the intensity) [14], it seems imperative and timely to quantify
dominant intrinsic fall-risk factor with a five-fold increase the dose–response relationships through a systematic
in risk of falling [5]. Although the age-related decline in review and meta-analysis. To the best of our knowledge, a
muscle strength is associated with the loss in muscle size meta-analysis that only includes RCTs and is based on a
(r = 0.66–0.83, p \ 0.001) [6], longitudinal studies found comparison between an intervention group and a physically
a 1.5 to five times greater decline in muscle strength inactive control group is currently missing in the literature.
compared with muscle size [2, 7]. In addition, there was a In contrast to direct comparisons (high- vs. low-intensity
stronger relationship between muscle strength and physical intervention groups), we investigate the effects of RT in
performance or disability compared with the relationship sedentary older adults when starting RT compared with
between muscle strength and mass [3]. physically inactive control groups to mitigate the age-re-
Even though exercise cannot fully prevent aging of the lated loss of muscle strength and morphology. A review of
neuromuscular system, resistance training (RT) has a great existing data concerning so far overlooked variables such
potential to mitigate age-related changes. Over the past as time under tension and rest time would more compre-
25–30 years, numerous studies have examined the effects hensively inform clinicians and practitioners on how to
of RT on measures of muscle strength and morphology in standardize RT. Finally, potential influences of the inclu-
old adults. Frontera and Bigard [8] reviewed RT’s potential ded training variables on the investigated effects of RT on
to improve old adults’ muscle strength and morphology [6]. muscle strength and morphology will be examined using
The review highlighted two studies that examined (a) the meta-regression. Meta-regression will be performed for
impact of aging on muscle strength (i.e., maximal isoki- relevant subcategories of training variables (i.e., volume,
netic knee extensor torque) and muscle size [i.e., cross- intensity, rest). Thus, the purpose of the present systematic
sectional area (CSA) of the knee extensors] in elderly men review and meta-analysis is to determine the general
with a mean age of 65 years, followed over a 12-year effects of RT on measures of muscle strength and mor-
period [7], and (b) the effects of a 12-week RT program phology. Furthermore, the present meta-analysis, using
(three sessions/week) on the same variables of muscle meta-regression, examines how specific training variables
strength and size in a cohort of 60- to 72-year-old men [9]. affect muscle strength and morphology. We constructed
Findings from the 12-year longitudinal study revealed a dose–response relationships for key RT variables [17]
loss in isokinetic knee extensor torque of -24 % and in through the analysis of RCTs that have clearly improved
quadriceps CSA of -16 %. In contrast, 12 weeks of RT at measures of muscle strength and morphology in healthy
80 % of the one-repetition maximum (1RM) resulted in an old adults.
increase in isokinetic torque of 16 % and in knee extensor
CSA of 11 %. Even though different cohorts were inves-
tigated in the two studies, the reported percentage rates are
2 Methods
impressive and may allow a cautious and preliminary
conclusion that biological aging of the neuromuscular
The present meta-analysis follows the recommendations of
system can be mitigated or even reversed to a certain extent
the ‘Preferred Reporting Items for Systematic Reviews and
[8].
Meta-Analyses’ (PRISMA) [18].
Relying on an extensive database comprising individual
experimental studies and reviews, the American College of
Sports Medicine (ACSM) issued what is considered as the 2.1 Search Strategy
gold standard of RT exercise prescription for healthy old
adults [10]. However, a careful examination of this position A systematic literature search was conducted from January
stand suggests that the position stand was based on cate- 1984 to June 2015 in the online databases PubMed, Web of
gory 4 or ‘expert level’ evidence on the evidence pyramid, Science, and The Cochrane Library. The following Medi-
the lowest compared with evidence level 1 provided by cal Subject Headings (MeSH) of the United States National
systematic reviews and meta-analyses [11]. Considering Library of Medicine (NLM) and search terms were inclu-
that the already published meta-analyses are methodolog- ded in our Boolean search syntax: (‘‘resistance training’’
ically limited in terms of study selection criteria {inclusion OR ‘‘strength training’’ OR ‘‘weight training’’ OR ‘‘weight-
123
R. Borde et al.
bearing exercise program’’) AND (old* OR elderly) AND 2.3 Coding of Studies
(sarcopenia OR dynapenia OR ‘‘muscle strength’’ OR
‘‘muscle morphology’’). The search was limited to English The studies were coded for the following variables:
language, human species, age 65? years, full text avail- (a) cohort; (b) age; (c) training variables [i.e., period, fre-
ability, and RCTs. quency, volume (i.e., number of sets per exercise, number
of repetitions per set), intensity, time under tension (total,
2.2 Selection Criteria/Study Eligibility isometric, concentric, eccentric), and rest (rest in between
sets and repetitions)]; (d) strength tests (i.e., 1RM, MVC);
Inclusion criteria were decided by the consensus state- (e) body region (i.e., upper limbs, lower limbs); and
ments of two reviewers (RB, UG). In cases where RB and (f) assessment of muscle morphology (i.e., CSA, muscle
UG did not reach agreement on inclusion of an article, TH volume, muscle thickness). The RT groups were subdi-
was contacted. In accordance with the PICOS approach vided according to the applied training intensity: high-in-
[18], inclusion criteria were selected by (a) population: tensity RT: C70 % 1RM; moderate-intensity RT:
healthy subjects who were aged C60 years, with a study 51 % C 1RM B 69 %; and low-intensity RT: B50 %
mean age C65 years; (b) intervention: machine-based RT 1RM [16]. In the dose–response relationship figures pre-
containing a description of at least one training variable sented in the ‘‘Results’’ section, diamonds, circles, and
(e.g., training intensity); (c) comparator: non-physically triangles symbolize high- (C70 % 1RM), moderate-
active (e.g., health education, no intervention) control (51 % C 1RM B 69 %), and low- (B50 % 1RM) intensity
groups; (d) outcome: at least one proxy of muscle strength RT groups. If exercise progression was realized over the
[e.g., 1RM, maximum voluntary contraction under iso- course of the intervention or if training variables were
metric conditions (MVC)] and/or muscle morphology reported, the average of these variables was calculated. If
[e.g., CSA (cm2, mm), volume (kg, cm3), thickness (mm)]; results of pre- and post-tests were not conclusively repor-
and (e) study design: RCTs [18]. Studies were excluded if ted, the authors of the respective studies were contacted via
they (a) did not meet the minimum requirements regarding email. Six out of 12 authors responded to our queries and
the description of training variables (e.g., period, fre- subsequently sent the missing data to calculate SMDbs.
quency, volume, intensity); (b) tested multiple repetition
maximum (e.g., 3RM); (c) did not report results ade- 2.4 Data Extraction
quately (mean and standard deviation); (d) included frail,
mobility and/or cognitively limited and/or ill subjects; The main study characteristics (i.e., cohort, age, interven-
(e) examined the effects of concurrent training (i.e., tion program, training variables, relevant outcomes) were
combined RT and endurance training); and (f) investigated extracted in an Excel template/spreadsheet.
the effects of nutritional supplements in combination with
RT. If multiple outcomes (e.g., strength properties of 2.5 Assessment of Methodological Study Quality
different muscle groups) were recorded within one study,
we chose the outcome with the highest functional rele- Evaluation of methodological study quality was conducted
vance for mobility in old age. In other words, (a) lower by two independent reviewers using the Physiotherapy
extremity muscle strength tests were preferred over upper Evidence Database (PEDro) scale [19]. The PEDro scale
extremity muscle strength tests; (b) isokinetic or dynamic includes 11 items with three items from the Jadad scale
muscle strength tests were preferred over isometric tests; [20] and nine items from the Delphi list [21]. PEDro rates
and (c) multi-joint tests (e.g., leg press) were chosen rather RCTs on a scale from 0 (low quality) to 10 (high quality),
than single-joint strength tests (e.g., leg extension/curl). In with a score of C6 representing a cut-off for high-quality
terms of muscle groups, sub-analyses were computed for studies [19]. The first item of the PEDro scale (eligibility
muscles of upper and lower extremities. Tests for the criteria were specified) is used to establish external validity
assessment of muscle strength were analyzed separately and is therefore not included in the overall score. Maher
for the 1RM and MVC. Measures of muscle morphology et al. [19] demonstrated fair-to-good inter-rater reliability,
were included if one of the following devices was used: with an intra-class correlation coefficient of 0.68 when
magnetic resonance imaging, computed tomography, dual using consensus ratings generated by two or three inde-
x-ray absorptiometry, ultrasound, or BOD POD (air dis- pendent raters.
placement plethysmograph for whole-body densitometry).
In addition, one representative part of the respective 2.6 Statistical Analyses
muscle (e.g., vastus lateralis) had to be assessed either by
muscle CSA, volume, or thickness when more than one To determine overall effects of RT on measures of muscle
muscle was tested. strength and morphology and to establish dose–response
123
relationships following RT in old adults, the between-sub- differences in the effect sizes of improvements in measures
ject standardized mean differences (SMDbs) were calcu- of muscle strength and morphology. According to Toigo
lated according to the following formula: SMDi ¼ m1i m si
2i
and Boutellier [17], RT variables were previously reported
[22], where SMDi is the standardized mean difference of insufficiently in the literature. Thus, we decided to report
one reported parameter (e.g., strength properties of dose–response relationships of each RT variable that could
quadriceps muscle), m1i and m2i correspond to the mean of maximize improvements in measures of muscle strength
the intervention and the control groups, respectively and si and morphology [17].
is the pooled standard deviation. In accordance with Hedges
and Olkin, this formula was adjusted for sample size: g ¼ 3 Results

1 4N3i 9 [23], where Ni is the total sample size of the
Our systematic literature search identified 506 potentially
intervention group and control group. SMDbs is defined as relevant studies (Fig. 1). A screening of the titles excluded
the difference between the post-test treatment and the 287 studies and then 109 duplicates were removed. The
control means divided by the pooled standard deviation, remaining 110 studies were analyzed concerning the pre-de-
with 95 % confidence intervals (CIs). If two or more studies fined eligibility criteria, and 85 of these were removed.
reported the same training variable (e.g., training volume, Finally, 25 studies with a total of 819 participants (mean
intensity, rest), weighted mean SMDbs over the studies was sample size 33 subjects) and a mean age of 70.4 years (age
calculated and presented as filled squares in the dose–re- range 60–90 years) were included in the quantitative syn-
sponse relationship figures presented in the Sect. 3. Each thesis (Table 1). Furthermore, four out of 25 studies investi-
unfilled symbol illustrates SMDbs per single training group. gated the effects of high-intensity RT compared with low-
Within-subject standardized mean difference (SMDws) was intensity RT (i.e., B50 % 1RM) [31–34]. Three studies [31,
calculated as follows: ±(mean of post-test - mean of pre- 33, 35] analyzed the effects of high-intensity RT compared
test)/SD pre-value, where SD is the standard deviation. with RT at moderate intensities (i.e., 51 % C 1RM B 69 %).
Positive SMD values indicate a favorable effect of RT as
compared with the control condition. Our meta-analysis 3.1 Overall Findings
was conducted using Review Manager version 5.3.4
(Copenhagen: The Nordic Cochrane Centre, The Cochrane 3.1.1 Effects of Resistance Training (RT) on Measures
Collaboration, 2008). The included studies were weighted of Muscle Strength
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
SMD2i
by the standard error: SEfSMDi g ¼ n1iNni 2i þ 2ðNi3:94Þ [22],
All 25 studies reported a favorable effect of RT on upper
where n1i is the sample size of the intervention group and n2i and lower extremity muscle strength. Weighted mean
is the sample size of the control group. Given that vari- SMDbs for the effects of RT on muscle strength amounted
ability (e.g., different age and muscle groups) between to mean SMDbs = 1.57 (95 % CI 1.20–1.94; I2 = 80 %,
studies was large, we decided to compute a random-effects v2 = 163.10, df = 32, p \ 0.01) (Fig. 2), which is
model to estimate the effects of RT interventions [18, 24]. indicative of a large effect. In addition, in sub-analyses, we
According to Cohen, effect size values of 0.00 to B0.49 determined the effects of RT on upper and lower body
indicate small, values of 0.50 to B0.79 indicate medium, strength tested by the 1RM. The analyses revealed
and values C0.80 indicate large effects [25]. Heterogeneity weighted mean SMDbs for the upper (mean SMDbs = 1.61;
was assessed using I2 and v2 statistics. Furthermore, a 95 % CI 0.95–2.27; I2 = 86 %, v2 = 88.52, df = 12,
random effects meta-regression was performed to examine p \ 0.01) and lower extremities (mean SMDbs = 1.76;
whether the effects of RT on measures of muscle strength 95 % CI 1.20–2.31; I2 = 87 %, v2 = 144.47, df = 19,
and morphology are predicted according to the combined p \ 0.01), corresponding to large effects. There were no
values of the different training variables using the valid studies that tested MVC in upper extremity muscles. Only
software Comprehensive Meta-analysis version 3.3.070 four studies measured leg muscle MVCs [34, 36–38]. A
(Biostat Inc., NJ, USA) [26–28]. Subcategories were cre- medium effect (mean SMDbs = 0.76; 95 % CI 0.40–1.31)
ated to extract the most important training variables of the was found for MVC of lower limbs, with non-significant
following combinations: training volume (i.e., period, fre- heterogeneity (I2 = 0 %, v2 = 2.89, df = 4, p = 0.58).
quency, number of sets per exercise, number of repetitions
per set); training intensity (i.e., intensity, time under ten- 3.1.2 Effects of RT on Measures of Muscle Morphology
sion) and rest (rest in between sets and repetitions) [29, 30].
For each subcategory, random-effects meta-regression was Nine studies examined the effects of RT on measures of
performed to identify variables that best predict the muscle morphology. An I2 value of 0 % (v2 = 7.18,
123
R. Borde et al.
Identification
Results of literature search
PubMed (n = 138), Web of Science (n = 185), Cochrane Library (n = 183)

(N = 506)
Papers excluded on basis of title (n = 287)

Screening
Potentially relevant papers remaining (n = 219)
Duplicate papers excluded (n = 109)
Potentially relevant papers remaining (n = 110)

Eligibility
Papers excluded on basis of eligibility criteria (n = 85)
• no RCT (n = 32)
• inadequate training description (n = 17)
• no relevant outcome (n = 16)
• no healthy subjects (n = 11)
• mean age < 65 years (n = 9)
Included
Included papers (n = 25)
Fig. 1 Flow chart presenting the different steps of search and study selection. RCT randomized controlled trial
df = 10, p = 0.71) is indicative of non-existent hetero- calculated independently using the effect size of charac-
geneity, which is why no further sub-analyses were com- teristics of each training variable (Table 4).
puted (Fig. 3). We pooled weighted mean SMDbs across
the nine studies and observed a small effect (mean 3.3.1 Meta-Regression Analysis for Training Variables
SMDbs = 0.42; 95 % CI 0.18–0.66) of RT on measures of of Muscle Strength
muscle morphology.
Table 3 shows the results of the meta-regression for three
3.2 Methodological Study Quality subcategories: training volume, training intensity, and rest.
Concerning training volume, only training period predicted
Table 2 shows that the quality scores averaged 4.6 ± 1.2 (p = 0.04) the effects of RT on muscle strength. In the
points (range 2–7). This is indicative of low method- subcategory training intensity, the best predictors for the
ological study quality even though only RCTs were explanation of effects of RT on muscle strength were
included. Three studies [35, 41, 43] were identified that intensity (p \ 0.05) and time under tension (p \ 0.01). The
exceeded the pre-determined cut-off score [19] of 6 points mode of muscle action (i.e., isometric, concentric, eccen-
or higher. tric) did not influence the effects of RT (p = 0.41–0.91).
Rest in between sets (p = 0.06, trend) and in between
3.3 Dose–Response Relationships of RT repetitions did not predict strength gains.
on Measures of Muscle Strength
3.3.2 Training Period
To improve the generalizability and external validity of our
study findings, we combined the results from 25 studies On average, the training period in the 25 studies lasted
that examined lower/upper extremity muscle strength 21.2 weeks (range 6–52 weeks). Figure 4 demonstrates
based on 1RM or MVC tests. Such pooling of data was dose–response relationships for the training variable
done to explore the effects of training variables on muscle ‘‘training period’’. Mean SMDbs amounted to 1.57 (95 %
strength using meta-regression (Table 3). In addition to CI 1.20–1.94; I2 = 81 %, v2 = 163.10, df = 32,
meta-regression, dose–response relationships were p \ 0.01). The longest training intervention lasted
123
Table 1 Studies examining the effects of RT on variables of muscle strength and muscle morphology in healthy old adults
Study Sex Age N Muscles/functional Period Strength Gain in measure of Within subject Between subject Training variables
(years) movement (weeks) gain (%) muscle morphology (%) SMD (SMDws) SMD (SMDbs)
Beneka et al. M/ 66–72 M: 8/8/ Knee extension 16 1RM male 1RM male 1RM male RT: 39/week; 3 sets
[31] F Mean 8/8 HI: 11 HI: 1.36 HI vs. CG: 1.17 HI: 4–6 reps; 90 % 1RM
age: HI/MI/ MI: 8 MI: 1.14 MI vs. CG: 0.77 MI: 8–10 reps; 70 % 1RM
69 LI/
LI: 4 LI: 0.43 LI vs. CG: 0.25 LI: 12–14 reps; 50 % 1RM;
CG
CG: -2 n.s. CG: -0.16 HI vs. MI: 0.33 TUT: 6 s; 2 s con, 2–3 s
F: 8/8/
1RM 1RM female HI vs. LI: 1.03 iso, 2–3 s ecc; RIS: 120 s;
8/8
female RIR: 5 s; weight
HI/MI/ HI: 3.58 MI vs. LI: 0.60
machines
LI/ HI: 15 MI: 0.71 1RM female
CG: no intervention
CG MI: 7 LI: 0.69 HI vs. CG: 1.92
LI: 3 CG: -0.13 MI vs. CG: 0.62
CG: -1 n.s. LI vs. CG: 0.83
n.s.
HI vs. MI: 3.18
HI vs. LI: 3.49
MI vs. LI: -0.10
Charette et al. F 64–86 13/6 Leg press 12 1RM 1RM 1RM RT: 39/week; 3–6 sets; 6
[92] Mean RT: 27–106 RT: 5.92–11.00 RT vs. CG: reps;
age: CG: -2 to CG: -0.12 to 1.98–7.42 1–5 weeks: 65 % 1RM
68 11 n.s. 1.17 6–9 weeks: 70 % 1RM
10–12 weeks: 75 % 1RM;
TUT: 5 s; 2 s con, 3 s ecc;
weight machines
CG: no intervention
Daly et al. [93] M/ Mean 8/8 Upper extremity 6 1RM MRI/MV 1RM 1RM RT: 39/week;
F age: RT: -33 to RT: 1–4 NPA RT: -0.07 to RT vs. CG: -0.17 1 week: 3 sets; 8 reps; 60 %
75 14 NPA 1.00 to 0.50 1RM
CG:
CG: -19 to -3 to -1 NPA CG: -0.41 to MV 2 weeks: 3 sets; 8 reps;
28 NPA 0.11 RT vs. CG: 70 % 1RM
MV 0.36–0.52 3–6 weeks: 2 sets; 8 reps;
RT: -0.11 to 75 % 1RM;
0.13 RIS: 60–90 s; weight
CG: -0.02 to machines and free weight
-0.08 CG: no intervention
123
Table 1 continued
123
DeBeliso et al. M/ 63–83 13/17/ Lower extremity 18 1RM 1RM 1RM RT: 29/week;
[94] F Mean 13 FR: 50–67 FR: 1.40–2.33 FR vs. CG: FR: 3 sets; 9RM
age: FR/ PER: 70–81 PER: 1.08–2.09 1.33–1.80 PER: 1–6 weeks; 2 sets;
72 PER/ PER vs. CG: 15RM
CG: -5 to CG: -0.10 to
CG 1.22–1.37
25 0.72 7–12 weeks; 3 sets; 9RM
FR vs. PER: 13–18 weeks; 4 sets; 6RM;
0.07–0.21
60 min; RIS: 120–180 s;
weight machines
CG: no intervention
Fatouros et al. M 65–78 8/8 Upper/lower 16 IS IS IS RT: 39/week;
[95] Mean extremity RT: 14 RT: 1.71 RT vs. CG: 1.38 1–4 weeks: 2 sets; 13 reps;
age: CG: -1 n.s. CG: -0.08 1RM upper 55–60 % 1RM
70 5–8 weeks: 3 sets; 12 reps;
1RM upper 1RM upper RT vs. CG: 3.65
60–70 % 1RM
RT: 114 RT: 6.65 1RM lower
9–12 weeks: 3 sets; 10 reps;
CG: 1 n.s. CG: 0.02 RT vs. CG: 4.88
70–80 % 1RM
1RM lower 1RM lower
13–16 weeks; 3 sets; 8 reps;
RT: 77 RT: 7.23 80 % 1RM; 45–50 min;
CG: 3 n.s. CG: 0.20 TUT: 7.5 s; 2–3 s con, 2 s
iso, 2–3 s ecc; RIS: 120 s;
RIR: 5 s; weight
machines
CG: no intervention
Fatouros et al. M 65–78 14/12/ Upper/lower 24 1RM upper 1RM upper 1RM upper RT: 39/week; 2–3 sets;
[33] Mean 14/ extremities HI: 73 HI: 3.52 HI vs. CG: 2.71 8–15 reps
age: 10 HI: 80 % 1RM
MI: 48 MI: 2.25 MI vs. CG: 1.93
71 HI/MI/ MI: 60 % 1RM
LI: 34 LI: 1.77 LI vs. CG: 1.38
LI/
CG: 2 n.s. C: 0.10 HI vs. MI: 0.78 LI: 40 % 1RM;
CG
1RM lower 1RM lower HI vs. LI: 1.44 TUT: 7.5 s; 2–3 s con,
2–3 s iso, 2–3 s ecc;
HI: 63 HI: 4.94 MI vs. LI: 0.63
HI RIS: 360 s
MI: 53 MI: 5.45 1RM lower
MI RIS: 240 s
LI: 38 LI: 4.86 HI vs. CG: 4.10
LI RIS: 120 s;
CG: -2 n.s. C: -0.18 MI vs. CG: 3.75
RIR: 3–5 s; weight
LI vs. CG: 3.34
machines
HI vs. MI: 0.62
CG: no intervention
HI vs. LI: 1.81
MI vs. LI: 1.22
R. Borde et al.
Table 1 continued
Granacher et al. M/ 60–80 20/20 Lower extremity 13 MVC MVC MVC RT: 39/week; 3 sets; 10
[36] F Mean RT: 27 RT: 1.24 RT vs. CG: 1.15 reps; 80 % 1 RM; 60-min
age: sessions; RIS: 120 s;
CG: -4 n.s. CG: -0.16
67 weight machines;
CG: no intervention
Henwood and M/ 65–84 22/22 Upper/lower 8 1RM upper 1RM upper 1RM upper RT: 29/week; 3 sets; 8
Taaffe [40] F Mean extremities RT: 2 n.s. – RT: 0.06–0.54 RT vs. CG: reps; 75 % 1RM; 60-min
age: 25 3.62–5.02 sessions; RIS: 60 s; TUT:
CG: -0.30 to -
70 6 s; con: 3 s, ecc: 3 s;
CG: -3 to 0.09 1RM lower
weight machines
-14 n.s. 1RM lower RT vs. CG:
CG: no intervention
1RM lower RT: 0.35–1.06 4.30–7.66
RT: 11–27 CG: -0.22 to
CG: -10 to 0.07
3 n.s.
Hortobagyi M/ 66–83 9/9/9 Leg press 10 MVC MVC MVC RT: 3 9/week;
et al. [34] F Mean HI/LI/ HI: 24 n.s. HI: 1.06 HI vs. CG: 0.89 HI: 5 sets; 4–6 reps; 80 %
age: CG LI: 28 n.s. LI: 1.00 LI vs. CG: 0.67 1RM
72 LI: 5 sets; 8–12 reps; 40 %
CG: 2 n.s. CG: -0.10 HI vs. LI: 0.03
n.s. 1RM;
IS IS
IS TUT: 3 s; 1–2 s con, 1–2 s
HI: 38 n.s. HI: 1.17
ecc; RIS:
LI: 29 n.s. LI: 0.84 HI vs. CG: 0.86
120 s; weight machines
CG: 1 n.s. CG: -0.02 LI vs. CG: 0.37
CG: no intervention
1RM 1RM HI vs. LI: 0.45
n.s.
HI: 35 n.s. HI: 1.05
1RM
LI: 33 n.s. LI: 0.78
HI vs. CG: 1.05
CG: 3 n.s. CG: -0.10
LI vs. CG: 0.52
HI vs. LI: 0.41
n.s.
123
Table 1 continued
123
Hunter et al. M/ 61–77 14/14/ Knee extension/ 25 1RM BP/FFM 1RM 1RM RT: 3 9/week; 2 sets; 10
[96] F Mean 14 elbow flexion HI: 13–24 HI: 4 HI: 0.43–0.74 HI vs. CG: reps; 45-min session;
age: HI/VI/ 0.85–1.13 RIS: 120 s; weight
VI: 10–28 VI: 4 VI: 0.21–0.75
66 CG machines
CG: -6 to CG: 1 n.s. CG: -0.18 to - VI vs. CG:
0.05–0.67 HI: 80 % 1RM
-2 n.s. 0.04
HI vs. VI: VI: 50, 65, 80 % 1RM
FFM
0.61–0.96 n.s. across the 3 sessions per
HI: 0.19 week
FFM
VI: 0.17 CG: no intervention
HI vs. CG: 0.38
CG: 0.03
VI vs. CG: -0.23
HI vs. CG: 0.71
n.s.
Judge et al. [43] M/ C75 28/27 Lower extremity 13 1RM 1RM 1RM RT: 39/week; 3 sets; 12
F Mean RT: 12 RT: 0.64 RT vs. CG: 0.11 reps; 75 % RM; 45-min
age: session; TUT: 4 s; 2 s
CG: -3 n.s. CG: -0.05
80 con, 2 s ecc; RIS:
120–180 s; RIR: 1–2 s;
weight machines
CG: no intervention
Kalapotharakos M/ 60–74 11/12/ Upper/lower 12 1RM upper CT/CSA 1RM upper 1RM upper RT: 39/week; 3 sets;
et al. [35] F Mean 10 extremities HI: 66 HI: 10 HI: 2.73 HI vs. CG: 2.11 HI: 8 reps; 80 % 1RM
age: HI/MI/ MI: 43 MI: 7 MI: 1.62 MI vs. CG: 1.47 MI: 15 reps; 60 % 1RM;
65 CG
CG: -1 n.s. CG: -1 n.s. CG: -0.04 HI vs. MI: 0.50 TUT: 6 s; 2 s con, 2 s iso,
1RM lower 1RM lower 1RM lower 2 s ecc; RIS: 120 s; RIR:
2–3 s; weight machines
HI: 78 HI: 3.13 HI vs. CG: 2.51
CG: no intervention
MI: 44 MI: 1.45 MI vs. CG: 1.51
CG: 0 n.s CG: 0.02 HI vs. MI: 0.97
CSA CSA
HI: 0.34 HI vs. CG: 0.38
MI: 0.37 MI vs. CG: 0.34
CG: -0.02 HI vs. MI: 0.10
Kalapotharakos M 61–75 9/9 Lower extremity 10 1RM 1RM 1RM RT: 39/week; 3 sets; 15
et al. [71] Mean RT: 24 RT: 0.83 RT vs. CG: 1.50 reps; 60 % 1RM; 60-min
age: session; RIS: 120 s;
CG: 0 n.s. CG: 0.01
68 weight machines
CG: no intervention
R. Borde et al.
Table 1 continued
Lovell et al. M/ 70–80 12/12 Leg extension 16 1RM CT/LM 1RM 1RM RT: 39/week; 3 sets; 6–10
[97] F Mean RT: 90 RT: 7 RT: 5.97 RT vs. CG: 4.33 reps; 70–90 % 1RM;
age: RIS: 120 s; weight
CG: -1 n.s. CG: 1 n.s. CG: -0.07 LM
74 machines
LM RT vs. CG: 0.10
CG: no intervention
RT: 0.14
CG: 0.03
Miszko et al. M/ 65–90 13/15 Lower extremity 16 1RM upper 1RM upper 1RM upper RT: 39/week; 3 sets; 6–8
[98] F Mean RT: 14 RT: 0.28 RT vs. CG: 0.33 reps;
age: CG: -1 n.s. CG: 0.01 1RM lower 1–8 weeks: 50–70 % 1RM
72 9–16 weeks: 80 % 1RM;
1RM lower 1RM lower RT vs. CG: 0.53
RT: 23 RT: 0.43 TUT: 4 s; 4 s con; weight
machines ? free weights
CG: 5 n.s. CG: 0.11
CG: no intervention
Morse et al. M 70–82 13/8 Lower extremity 52 MVC MRI/MV MVC MVC RT: 39/week (2 9 group
[99] Mean (ankle) RT: 0 n.s. RT: 15 RT: 0.00–1.29 RT vs. CG: 0.89 based, 1 9 home based);
age: -25 BD-1.51 2 - 3 sets; 8 - 10 reps;
CG: 2 n.s. CG: -0.09 to
74 80 % 1RM; rubber
CG: -2 to 0.35 MV
bands, weight machines
5 n.s. MV RT vs. CG: 1.03
CG: no intervention
RT: 1.53
CG: 0.22
Pinto et al. [41] F 60–69 19/17 Lower extremity 6 1RM US/MT 1RM 1RM RT: 29/week;
Mean RT: 22 RT: 11–21 RT: 1.16 RT vs. CG: 1.33 1–3 weeks: 2 sets; 15–20
age: CG: -1 n.s. CG: -5 to 7 n.s. CG: -0.04 MT reps
66 4–6 weeks: 3 sets; 12–15
MT RT vs. CG:
0.52–0.99 reps;
RT: 0.59–0.90
RIS: 120 s
CG: -0.38 to
0.24 CG: no intervention
Pyka et al. [39] M/ 61–78 8/6 Upper/lower 52 1RM upper 1RM upper 1RM upper RT: 39/week; 3 sets; 8
F Mean extremities RT: 23–51 RT: 3.30–5.38 RT vs. CG: reps; 65–75 % 1RM;
age: 4.69–6.12 60-min sessions; TUT:
CG: -4 to CG: -1.35 to
68 5 s; 2 s con, 3 s ecc; RIS:
-12 n.s. -0.63 1RM lower
60 s; weight machines
1RM lower 1RM lower RT vs. CG:
CG: no intervention
RT: 27–62 RT: 4.50–9.51 5.87–7.67
CG: -3 to CG: -1.45 to
-12 n.s. -0.32
123
Table 1 continued
123
Raso et al. [42] F 60–77 14/9 Trunk/lower 52 1RM N/A/FFM 1RM 1RM RT: 39/week; 3 sets; 12
Mean extremity RT: 48 RT: -3 n.s RT: 4.73 RT vs. CG: 2.20 reps; 55 % 1RM; 60-min
age: sessions; TUT: 4 s; 1–2 s
CG: 5 n.s. CG: -2 n.s. CG: 0.67 FFM
68 con, 2–3 s ecc; RIS:
FFM RT vs. CG: 0.20 120 s; weight machines
RT: -0.22 CG: no intervention
CG: -0.20
Reeves et al. M/ 65–79 9/9 Lower extremity 14 MVC MVC MVC RT: 39/week; 2 sets; 10
[37] F Mean RT: 15 RT: 0.32 RT vs. CG: 0.52 reps; 70–75 % 1RM;
age: TUT: 5 s; 2 s con, 3 s
CG: -12 CG: -0.45 NPA
71 ecc; RIS: 180 s; weight
n.s.
machines
CG: no intervention
Rhodes et al. F 65–75 20/18 Upper/lower 52 1RM upper 1RM upper 1RM upper RT: 3 9/week; 3 sets; 8
[100] Mean extremity RT: 9 n.s. – RT: 0.55–1.70 RT vs. CG: reps; 75 % 1RM; 60-min
age: 25 0.60–1.25 sessions; TUT: 6 s; 2–3 s
CG: 0.02–0.09
69 con, 3–4 s ecc
CG: 0–2 1RM lower 1RM lower
n.s. CG: no intervention
RT: 0.83–2.62 RT vs. CG:
1RM lower 1.28–2.85
CG: -0.21 to
RT: 19–54 0.06
CG: -4 to
1 n.s.
Strasser et al. M/ C70 15/14 Upper/lower 26 1RM upper 1RM upper 1RM upper RT: 39/week; 3–6 sets;
[72] F Mean extremities RT: 24–31 RT: 0.61–0.76 RT vs. CG: 10–15 reps; 60–70 %
age: 1.00–1.40 1RM
CG: 3 n.s. CG: 0.10–0.12
74 1RM lower CG: no intervention
1RM lower 1RM lower
RT: 15 RT: 0.47 RT vs. CG: 0.77
BD
CG: 9 n.s. CG: 0.35
Tracy et al. [38] M/ 65–80 11/9 Knee extension 16 MVC MVC MVC RT: 39/week; 3 sets; 10
F Mean RT: 26 RT: 0.81 RT vs. CG: 0.27 reps; 80 % 1RM
age: CG: -1 n.s. CG: -0.05 1RM CG: no intervention
74
1RM 1RM RT vs. CG: 0.53
RT: 27 RT: 0.67
CG: 2 n.s. CG: 0.05
R. Borde et al.
Table 1 continued
Vincent et al. M/ 60–83 22/24/ Upper/lower 24 1RM CT/FFM 1RM 1RM RT: 39/week; 1 set;
[32] F Mean 16 extremities, HI: 18 HI: 0.4 n.s. HI: 0.42 HI vs. CG: 0.66 HI: 8 reps; 80 % 1RM
age: HI/LI/ trunk
LI: 17 LI: -3.6 n.s. LI: 0.45 LI vs. CG: 0.49 LI: 13 reps; 50 % 1RM;
68 CG (total strength)
CG: -1 n.s. CG: -1 n.s. CG: -0.04 HI vs. LI: 0.25 RIS: 120 s; weight
FFM n.s. machines
HI: 0.02 FFM CG: no intervention

LI: -0.12 HI vs. C: 0.17
CG: -0.05 LI vs. C: 0.22
HI vs. LI: -0.06
NPA
Vincent et al. M/ 60–72 10/10 Total body strength 24 1RM CT/FFM 1RM 1RM RT: 39/week; 1 set; 8–13
[73] F Mean RT: 16 RT: 4 n.s. RT: 1.35 RT vs. CG: 0.08 reps; 50–80 % 1RM;
age: weight machines
CG: -2 n.s. CG: 1 n.s. CG: -0.15 FFM
69 CG: no intervention
FFM RT vs. CG: 1.30
RT: 0.57
CG: 0.13
1RM one-repetition maximum, BD baseline differences (p [ 0.05), BP BOD POD (air displacement plethysmograph for whole-body densitometry), CG control group, con concentric, CSA
cross-sectional area, CT computed tomography, ecc eccentric, F female, FFM fat-free mass, FR fixed repetitions, HI high-intensity, IS isokinetic strength, iso isometric, LI low-intensity, LM
lean mass, M male, MI moderate-intensity, MRI magnetic resonance imaging, MT muscle thickness, MV muscle volume, MVC maximal voluntary contraction, N/A not available, NPA no
p values available, n.s. not significant, PER periodized repetitions, reps repetitions, RIR rest in between repetitions, RIS rest in between sets, RT resistance training, SMDbs difference between
the post-test treatment and the control means divided by the pooled standard deviation with 95 % confidence intervals, SMDws difference of mean of post-test and mean of pre-test divided by
standard deviation of pre-value, TUT total time under tension, US ultrasonography, VI variable intensity
123
R. Borde et al.
Fig. 2 Effects of RT on measures of muscle strength. CG control repetition training group, Random random effects model, RT resis-
group, CI confidence interval, FR fixed repetition training group, HI tance training, SE standard error, SMD standardized mean difference,
high-intensity training group, IV inverse variance, LI low-intensity Weight weight attributed to each study due to its statistical power
training group, MI moderate-intensity training group, PER periodized
Fig. 3 Effects of RT on measures of muscle morphology. CG control training, SE standard error, SMD standardized mean difference,
group, CI confidence interval, HI high-intensity training group, IV Weight weight attributed to each study due to its statistical power
inverse variance, LI low-intensity training group, MI moderate-
intensity training group, Random random effects model, RT resistance
123
Table 2 Physiotherapy Evidence Database (PEDro) scores of the 25 included studies
Authors Eligibility Random Concealed Baseline Blind Blind Blind Adequate follow-up Intention-to- Between-group Point estimates Score
criteria allocation allocation comparability subjects therapists assessor dropout \15 % treat analysis comparisons and variability
Beneka et al. - ? - ? - - - ? - ? ? 5
[31]
Charette et al. ? ? - ? - - - - - ? ? 4
[92]
Daly et al. [93] - ? ? - - - ? ? - ? - 5
DeBeliso et al. - ? - ? - - - - - ? ? 4
[94]
Fatouros et al. - ? - ? - - - ? - ? ? 5
[95]
Fatouros et al. - ? - ? - - - ? - ? ? 5
[33]
Granacher et al. - ? - - - - - - - - ? 2
[36]
Henwood and ? ? - ? - - - ? - ? ? 5
Taaffe [40]
Hortobagyi - ? - ? - - - ? - ? ? 5
et al. [34]
Hunter et al. ? ? - ? - - - ? - ? ? 5
[96]
Judge et al. ? ? - ? - - ? ? ? ? ? 7
[101]
Kalapotharakos - ? - ? - - - - - ? ? 4
et al. [71]
Kalapotharakos - ? - ? - - ? ? - ? ? 6
et al. [35]
Lovell et al. ? ? - ? - - - ? - ? ? 5
[97]
Miszko et al. - ? - - - - - - - ? ? 3
[98]
Morse et al. - ? - - - - - - - ? ? 3
[99]
Pinto et al. [41] ? ? - ? - - ? ? ? ? ? 7
Pyka et al. [39] - ? - ? - - - ? - ? ? 5
Raso et al. [42] ? ? - ? - - - - - ? ? 4
Reeves et al. - ? - ? - - - - - - ? 3
[37]
Rhodes et al. ? ? - ? - - - ? - ? ? 5
[100]
123
R. Borde et al.
Score
50–53 weeks and revealed the largest mean SMDbs, with a
4.6
value of 2.34.
5
5
4
Point estimates
and variability
3.3.3 Training Frequency
Twenty-five studies were included in this sub-analysis, and

the mean training frequency was 2.9 sessions per week,
?
?
?
?
with a mean SMDbs of 1.57 (range two to three sessions per
Between-group
week; 95 % CI 1.20–1.94; I2 = 79 %, v2 = 163.10,

comparisons
df = 32, p \ 0.01). That is, two and three training sessions

per week produced large effects on measures of muscle
strength, with mean SMDbs of 2.13 (two sessions) and 1.49
?
?
?
(three sessions).
treat analysis
Intention-to-
3.3.4 Number of Sets and Repetitions

-
-
-
In the 25 studies included in this sub-analysis, the number

of sets per exercise averaged 2.9 (range one to five sets)
Adequate follow-up
and the number of repetitions per set averaged 10.0 (range

dropout \15 %
five to 16 repetitions). Mean SMDbs for number of sets and

repetitions per exercise were 1.57 (95 % CI 1.20–1.94;
I2 = 80 %, v2 = 163.10, df = 32, p \ 0.001) and 1.61
(95 % CI 1.22–1.99; I2 = 81 %, v2 = 161.71, df = 31,
-
-
?
p \ 0.01), indicative of large effects. Two to three sets per

assessor
exercise (mean SMDbs = 2.99) and seven to nine repeti-

Blind
tions (mean SMDbs = 1.98) resulted in the largest

-
?
-
improvements in muscle strength.

therapists
Blind
3.3.5 Training Intensity

-
-
-
-
subjects
Twenty-four studies were included in this sub-analysis, and

Blind
training intensity was classified as high (C70 % 1RM),

-
-
-
moderate (51 % C 1RM B 69 %), and low (B50 % 1RM)

comparability
[16]. The sub-analysis revealed a mean intensity of 69 % of

Baseline
the 1RM (range 40–90 % 1RM) across studies. Figure 5

illustrates dose–response relationships for training inten-
?
?
?
sity, with a mean SMDbs of 1.63 (95 % CI 1.21–2.05;

I2 = 82 %, v2 = 157.81, df = 28, p \ 0.01). The largest
Concealed
allocation
? indicates a ‘‘yes’’ score, - indicates a ‘‘no’’ score
effects on measures of muscle strength were found for

intensities of 70–79 % of the 1RM (mean SMDbs = 1.89).
-
-
-
-
allocation
3.3.6 Time Under Tension per Repetition

Random
Time under tension is an important variable to induce

?
?
?
adaptations in muscle strength and morphology [17]. In 14

Eligibility
studies, the total time under tension averaged 5.7 s per

criteria
repetition (range 3–7.5 s; mean SMDbs = 1.60; 95 % CI

1.09–2.10; I2 = 82 %, v2 = 102.65, df = 18, p \ 0.01).
?
-
-
-
Table 2 continued
The largest effect was shown for 6 s, with a mean SMDbs

Tracy et al. [38]
Vincent et al.
Vincent et al.
Strasser et al.
of 3.61. Figure 6 shows the dose–response relationships for

Mean score
the training variable ‘‘time under tension’’. In addition, the

Authors
[72]
[73]
[32]
mean time under tension was 2.3 s for isometric (range

2–2.5 s; SMDbs = 2.48; 95 % CI 1.36–3.32; I2 = 83 %,
123
Table 3 Meta-regression for training variables of different subcategories to predict RT effects on muscle strength
Coefficient Standard error 95 % lower CI 95 % upper CI Z value P value
Training volume
Training period 0.0316 0.0155 0.0012 0.0619 2.04 0.04
Training frequency 0.0900 0.3315 -0.5598 0.7397 0.27 0.79
Number of sets 0.1142 0.1810 -0.2406 0.4690 0.63 0.53
Number of repetitions per set 0.0219 0.0585 -0.0927 0.1366 0.37 0.71
Training intensity
Training intensity 0.0182 0.0052 0.0084 0.0288 3.57 0.01
Time under tension 0.3154 0.1094 0.1010 0.5297 2.88 0.01
Rest
Rest in between sets 0.0095 0.0051 -0.0006 0.0196 1.85 0.06
Rest in between repetitions 0.1600 0.2255 -0.282 0.6019 0.71 0.48
CI confidence interval, RT resistance training
Table 4 Training variables with largest mean SMDbs

Training variables Measures of muscle strength Measures of muscle morphology
Highest value Mean SMDbs Highest value Mean SMDbs
Training period [weeks] 50–53 2.34 50–53 0.59a

Training frequency [sessions per week] 2 2.13 3 0.38
Number of sets per exercise 2–3 2.99 2–3 0.78a
Number of repetitions [per set] 7–9 1.98 7–9 0.49
Training intensity [% of 1RM] 70–79 1.89 51–69 0.43
Time under tension (total) [s] 6.0 3.61 6 0.36a
a
Time under tension (isometric mode) [s] 2.0 2.70 2.0 0.36a
Time under tension (concentric mode) [s] 2.5 3.44 2.0 0.36a
Time under tension (eccentric mode) [s] 3.0 2.98 2.0 0.36a
a
Rest in between sets [s] 60 4.68 120 0.30
Rest in between repetitions [s] 4 3.72a 2.5 0.36a
The content of this table is based on individual training variables with no respect for interaction between training variables
SMDbs between-subject standardized mean difference, 1RM one-repetition maximum
a
Based on less than three studies
v2 = 47.19, df = 8, p \ 0.01), 2.2 s for concentric (range or repetitions. The mean rest time between sets was 132 s
1.5–4.0 s; SMDbs = 2.18; 95 % CI 1.26–2.54; I2 = 84 %, (range 60–360 s; mean SMDbs = 1.87; 95 % CI
v2 = 101.94, df = 16, p \ 0.01), and 2.5 s for eccentric 1.35–2.38; I2 = 84 %, v2 = 138.61, df = 22, p \ 0.01),
actions (range 1.5–3.5 s; SMDbs = 2.28; 95 % CI and between repetitions (five studies) it was 3.9 s (range
1.36–2.79; I2 = 87 %, v2 = 123.06, df = 16, p \ 0.01). 1.5–5 s; mean SMDbs = 2.24; 95 % CI 1.52–2.31;
During the isometric mode, a time under tension of 2.0 s I2 = 83 %, v2 = 47.19, df = 8, p \ 0.01). Figure 7
with a mean SMDbs of 2.70 appears most effective. In the shows the dose–response relationships for the training
concentric and eccentric modes, times under tension of variable ‘‘rest in between sets’’. Eleven out of 17 studies
2.5 s (mean SMDbs = 3.44) and 3.0 s (mean used 120 s of rest in between sets, resulting in a mean
SMDbs = 2.98) seem to be most effective. SMDbs of 1.57. With reference to the results of two
studies [39, 40], a rest in between sets of 60 s appears to
3.3.7 Rest Time (Rest in Between Sets and Repetitions) be most effective to increase muscle strength (mean
SMDbs = 4.68) (Fig. 7). A rest time between repetitions
Based on data from 17 studies, we computed dose–re- of 4.0 s seems to be most effective, coupled with a mean
sponse relationships regarding rest time between sets and/ SMDbs of 3.72.
123
R. Borde et al.
Fig. 4 Dose-response 7
relationships for training period
and measures of muscle strength
following resistance training. 6
Each unfilled symbol illustrates
the SMDbs per single study.
Filled black squares represent
the weighted mean SMDbs of all 5
studies. Diamonds, circles, and
triangles symbolize high-,
Standardized mean difference

moderate-, and low-intensity
4
resistance training groups,
respectively. SMDbs between-
subject standardized mean
difference 3
0
50 6
6-9 7
10-13 8
14-17 9
18-21 10
… 11
24-27 12
… 13
50-53
Training period (weeks)
relationships for training
intensity and measures of
muscle strength following 6
resistance training. Each
unfilled symbol illustrates the
SMDbs per single study. Filled
black squares represent the 5
weighted mean SMDbs of all
triangles symbolize high-, 4

subject standardized mean 3
difference
0
0 ≤ 50 51-69 70-79 80-89 ≥ 90
Training intensity (%)
123
relationships for total time

under tension and measures of
muscle strength following 6
resistance training. Each
unfilled symbol illustrates the
SMDbs per single study. Filled
black squares represent the 5
weighted mean SMDbs of all


triangles symbolize high-, 4
subject standardized mean 3
difference
0
0 3 4 5 6 7 8
Time under tension (s)
relationships for rest in between

sets and measures of muscle
strength following resistance 6
training. Each unfilled symbol
illustrates the SMDbs per single
study. Filled black squares
represent the weighted mean 5
SMDbs of all studies. Diamonds,

circles, and triangles symbolize

high-, moderate-, and low- 4
intensity resistance training
groups, respectively. SMDbs
between-subject standardized
mean difference 3
0
0 50 100 150 200 250 300 350 400
Rest in-between sets (s)
123
R. Borde et al.
3.4 Dose–Response Relationships of RT 3.4.5 Training Intensity

on Measures of Muscle Morphology
Eight studies that reported training intensities were classified as
3.4.1 Meta-Regression Analyses for Training Variables high (C70 % 1RM), moderate (51 % C 1RM B 69 %), and
of Muscle Morphology low (B50 % 1RM) [16]. Mean intensity across studies was
71 % of the 1RM (range 50–80 % of 1RM), with a mean
Due to the low number of studies, we performed meta-re- SMDbs of 0.38 (95 % CI 0.13–0.64; I2 = 0 %, v2 = 6.61,
gression only for the subcategory ‘‘training volume’’. The df = 9, p = 0.68). Exercise at a moderate intensity between 51
regression analysis revealed that no variable within the and 60 % of the 1RM produced the greatest effects on measures
training volume subcategory (i.e., period, frequency, num- of muscle morphology, with a mean SMDbs of 0.43 (four
ber of sets, number of repetitions) produced significant studies). One study showed the same effect (SMDbs = 0.43) on
effects (p = 0.52–0.94) on measures of muscle morphology. muscle volume using an intensity of 70–79 % of 1RM.
3.4.2 Training Period 3.4.6 Time Under Tension per Repetition
Pooled data from nine studies revealed a mean training period Based on two studies, the total time under tension averaged
of 24.0 weeks (range 6–52 weeks), with a mean SMDbs of 0.42 5.3 s, with a mean SMDbs of 0.31 (range 4–6 s; 95 % CI -
(95 % CI 0.18–0.66; I2 = 0 %, v2 = 7.18, df = 10, 0.18 to 0.80; I2 = 0 %, v2 = 0.10, df = 2, p = 0.95). The
p = 0.71). With reference to the results of one study [41], a largest effect occurred at 6 s, with a mean SMDbs of 0.36
training period of 6 weeks appeared to be most effective to (one study). Considering specific muscle action modes,
improve measures of muscle morphology, with an SMDbs of only one study [35] reported time under tension during
0.66. Of note, the results of the two studies that used isometric muscle actions and two studies [35, 42] reported
50–53 weeks as a training period showed a slightly lower effect time under tension for concentric and eccentric muscle
on measures of muscle morphology (mean SMDbs = 0.59). actions. The mean time under tension was 2.0 s for the
isometric mode (SMDbs = 0.36; 95 % CI 1.13–4.27;
3.4.3 Training Frequency I2 = 75 %, v2 = 7.98, df = 2, p = 0.02), 1.8 s for the
concentric mode (range 1.5–2 s; SMDbs = 0.31; 95 % CI
Our sub-analysis included nine studies and revealed a mean -0.18 to 0.80; I2 = 0 %, v2 = 0.10, df = 2, p = 0.95),
training frequency of 2.9 training sessions per week (range and 2.2 s for the eccentric mode (SMDbs = 0.31; 95 % CI
two to three sessions per week), with a mean SMDbs of 0.42 -0.18 to 0.80; I2 = 0 %, v2 = 0.10, df = 2, p = 0.95).
(95 % CI 0.18–0.66; I2 = 0 %, v2 = 7.18, df = 10, The most effective time under tension appears to be 2.0 s
p = 0.71). The results of one study [41] suggested the largest for isometric, concentric, and eccentric muscle actions
improvement in measures of muscle morphology with two (SMDbs = 0.36; one study), respectively.
(SMDbs = 0.66) compared with three sessions per week
(mean SMDbs = 0.38). Of note, eight out of nine studies 3.4.7 Rest Time (Rest in Between Sets and Repetitions)
examined the effects of three training sessions per week.
In each of the six studies, the mean rest time was 120 s
3.4.4 Number of Sets and Repetitions between sets. Only one study [35] provided detailed
information regarding rest time between repetitions (2.5 s).
Based on nine studies, the average number of sets per The mean SMDbs was 0.30 for rest in between sets (95 %
exercise was 2.3 (range one to three sets). On average, 10.6 CI 0.04–0.57; I2 = 0 %, v2 = 1.74, df = 7, p = 0.97) and
repetitions (range eight to 16 repetitions) were performed 0.36 for rest in between repetitions (95 % CI -0.24 to
per set. The mean SMDbs for number of sets as well as 0.96; I2 = 0 %, v2 = 0.00, df = 1, p = 0.95).
repetitions per exercise was 0.54 (95 % CI 0.30–0.78;
I2 = 0 %, v2 = 7.25, df = 10, p = 0.70) and 0.42 (95 %
CI -0.32–0.90; I2 = 0 %, v2 = 0.08, df = 1, p = 0.77), 4 Discussion
indicative of moderate and small effects, respectively. Two
to three sets per exercise (mean SMDbs including two To the best of our knowledge, this is the first systematic
studies = 0.78) and seven to nine repetitions (mean literature review and meta-analysis that provides an inte-
SMDbs = 0.49; six studies) resulted in the largest grated overview of the general effectiveness of RT on
improvements in measures of muscle morphology based on measures of muscle strength and morphology in healthy
findings of more than one study. One study conducting RT old adults. The results from the 25 eligible RCTs suggest a
with 16–18 repetitions per set reported an SMDbs of 0.66. large and systematic training effect of RT on muscle
123
strength (Fig. 2) and a small effect on measures of muscle in CSA with aging and that the change in quadriceps
morphology (Fig. 3). We also performed a meta-regression muscle area only explains about 6–8 % of the between-
analysis to determine how such training variables as vol- subject variability in the change in knee extensor torque.
ume, intensity and rest modify the RT effects on measures This implies that the loss in muscle strength with age
of muscle strength and morphology. Additional dose–re- (dynapenia) is more related to impairments in neural acti-
sponse relationships of each training variable were com- vation and/or reductions in the intrinsic force-generating
puted independently from the other training variables capacity of skeletal muscle [3]. Based on these findings, it
(Table 4). Moreover, we discuss the findings with refer- seems plausible to argue that primarily neural adaptations
ence to the relevant literature concerning the general account for training induced improvements in muscle
effects and dose–response relationships following RT in strength, with improvements in measures of muscle mor-
healthy old adults. If no age-group specific information was phology playing a minor role, particularly during the early
available in the literature, we extended our search and phase of RT [56]. This may explain the observed larger
discussion to findings regarding the effects of RT in heal- gains in muscle strength compared with measures of
thy young adults. muscle morphology [2, 7].
Despite the large effect of RT on muscle strength, there
was still considerable variation in the magnitude of adap-
4.1 Effects of RT on Measures of Muscle Strength tations between studies. Methodological issues may also
and Morphology in Healthy Old Adults contribute to the large variability. For example, the mag-
nitude of response varies between body regions (upper vs.
In healthy old adults, RT improved muscle strength sub- lower limbs) or muscle groups. Adaptations to RT can be
stantially (13–90 %; 25 studies) and measures of muscle highly specific, as training-induced changes in CSA can
morphology to a smaller extent (1–21 %; nine studies). The differ between vastus lateralis and vastus medialis and can
results seem to suggest that the various forms of RT also be muscle-length specific [57]. Another factor con-
reviewed here have a greater potential to improve healthy tributing to the large variation in the response to RT is the
old adults’ ability to generate maximal voluntary force age of the subjects, which ranged widely, between 60 and
compared with the potential to improve measures of mus- 90 years. Spontaneous physical activity is much higher for
cle morphology (mean SMDbs = 1.57 vs. 0.42). These seniors at age 65 vs. 85, with some older individuals
findings are in line with the results of two meta-analyses, making as few as 100–200 steps per day [58]. The obser-
which examined the effects of RT on muscle strength [12] vations from a large cross-sectional study that in some
and size [44] in healthy as well as frail and/or disabled healthy old cohorts there could be accelerated muscle
middle-aged and/or old adults (range 50–95 years) and strength loss even as early as age 60–69 just further
reported increases in muscle strength and size of 24–33 % strengthen the argument for prescribing RT for old adults
and 1.5–16 %, respectively [13–16]. Recent imaging, aging healthily [1].
magnetic brain stimulation, and peripheral nerve stimula-
tion studies seem to lend support to the emerging hypoth- 4.2 Dose–Response Relationships of RT to Increase
esis that life-long RT could be an important non- Muscle Strength
pharmaceutical intervention to slow the age-related neural
dysfunction through which muscle strength loss can be The previous section established a large overall effect of
reduced [45–54]. This prediction is corroborated by in vitro RT on maximal voluntary strength in healthy old adults.
evidence suggesting that age and disuse do not affect We further performed meta-regression to identify training
intrinsic upper- and lower-limb skeletal muscle function variables that affected strength gains after conducting RT.
even in the oldest-old. While age does affect in vivo whole To specify the characteristic of each training variable with
muscle function, which is exacerbated by disuse [55], RT the largest effect on muscle strength, we conducted addi-
could effectively counteract the age-related strength loss. tional analyses of independently computed dose–response
The effectiveness of RT was investigated by the present relationships.
and several previous reviews [12–16]. Further, Delmonico
et al. [2] conducted a 5-year longitudinal study with well- 4.2.1 Training Volume (Period, Frequency, Number
functioning men and women (N = 1678) between the ages of Sets, Number of Repetitions)
of 70 and 79 years at baseline and measured knee extensor
torque using an isokinetic device and mid-femur CSA Of the four training variables within training volume, meta-
using computer tomography at the beginning of the study regression identified training period only to have a signif-
and after 6 years. It was found that decreases in isokinetic icant effect on muscle strength. The longest training period
leg muscle torque were two to five times greater than losses produced the largest increases in voluntary muscle strength
123
R. Borde et al.
(mean SMDbs = 2.34; 50–53 weeks). This result is based analyses of dose–response relationships of the number of
on only four studies, as in the majority of the studies the sets per exercise revealed an inverse U-shape, with the
intervention duration ranged from 6 to 26 weeks. Curi- largest effect (mean SMDbs = 2.99) being prevalent in RT
ously, RT as short as 6–9 weeks was only slightly less protocols that applied two to three sets. However, it seems
effective than RT of 50–53 weeks to improve muscle that there is no difference between single versus multiple
strength (mean SMDbs = 2.27; two studies). This obser- sets in short-term RT (6 weeks) in old adults [64]. More-
vation suggests that RT is a suitable intervention to combat over, these results suggested that during the early phase of
weakness in healthy old adults because the nervous system RT, number of sets was not the primary variable respon-
exhibits a rapid responsiveness to mechanical overload [4, sible for increases in muscle strength and thickness in old
30, 49, 51, 59]. In agreement with our findings, a current adults [64]. In addition, ‘‘number of sets’’ appears not to
meta-analysis that included 15 studies confirmed the out- result in neural adaptations because no differences were
come of the general analysis that ‘‘training period’’ is the found in electromyography activation of quadriceps mus-
only significant variable (p \ 0.01) to improve muscle cles between groups of old women (60–74 years) that
strength based on results of meta-regression [15]. These trained using single or multiple sets [64]. But although the
authors reported that long (24–52 weeks) versus short musculoskeletal system is adapted through the stimulus of
training periods (8–18 weeks) are more effective. In a single set to failure, multiple sets appear to be required to
addition, Kennis et al. [60] investigated detraining effects add continued strength gains [65]. Multiple versus single
following 1 year of RT on different variables of muscle number of sets seemingly has a higher impact on muscle
strength in old adults (60–80 years). After 7 years of strength in combination with longer training periods. In this
detraining, initially strength-trained participants still context, Radaelli et al. [66] examined the effects of one set,
exhibited improved muscle strength characteristics com- three sets, and five sets of RT applied over a period of
pared with the control group. However, the authors pointed 6 months (three sessions per week) on measures of upper-
out that RT cannot attenuate the age-related decline in and lower-limb muscle strength and muscle thickness in
muscle strength and therefore suggested the application of young untrained men age 24 years. Multiple versus single
lifelong RT. These findings are in accordance with ACSM sets improved muscle strength and muscle thickness par-
recommendations [61]. ticularly of the upper body more effectively, especially
In contrast to the results of meta-regression, additional with five sets of RT. In addition, two non-RCTs investi-
analyses of dose–response relationships indicated large gated the impact of one set or three sets per exercise on
differences between two training sessions per week (mean measures of muscle strength in old adults aged 60–80 years
SMDbs = 2.13) and three training sessions per week (mean [67, 68]. Only the study examining a longer training period
SMDbs = 1.49). Because studies that administered two (20 vs. 12 weeks) found a significant effect of three-set
sessions per week were also of short duration (6–9 weeks), versus one-set training on peak torque and maximum vol-
learning effects and neuronal adaptions must have con- untary contraction of the knee extensors in elderly subjects
tributed strongly to the effects associated with two versus aged 65–78 years [68]. Together, there is a paucity of data
three sessions per week [4, 30, 49, 51, 59]. In support of from high-quality RCTs concerning the effects of training
our meta-regression data, DiFrancisco-Donoghue et al. [62] frequency on muscle strength, especially in the elderly.
reported similar increases in muscle strength after 9-week- Finally, concerning the training variable ‘‘number of
long programs consisting of one and two weekly sessions repetitions’’, the largest effects in strength gains occurred
in healthy old adults age 65–79. Furthermore, Taaffe et al. when old adults used seven to nine repetitions per set
[63] conducted a 24-week RT intervention with three dif- (mean SMDbs = 1.98). Despite that the ‘‘number of repe-
ferent training frequencies (one to three sessions per week) titions’’ within a set in RT could provide a distinct physi-
in old adults aged 65–79 years. The authors concluded that ological stimulus for strength gains—with lower repetitions
a weekly or biweekly RT is equally effective to enhance predicted to be more effective [69]—our systematic search
muscle strength as compared with three sessions per week. identified no study that specifically examined the effects of
Of note, our findings must be interpreted with caution different repetitions per set on variables of muscle strength.
because the range of training frequencies was narrow (two This can most likely be explained by the fact that the
to three sessions per week). Finally, the current meta- variable ‘‘number of repetitions’’ is often used as an indi-
analysis confirms the conclusion reached by expert opinion cator of training intensity, which is why previous research
in the ACSM position stand that recommended RT fre- efforts focused on ‘‘training intensity’’ rather than ‘‘number
quencies of at least two sessions per week [61]. of repetitions’’. In fact, it has been reported that a given
Our analyses revealed little or no effect of the training percentage of the 1RM determines the realized number of
variables ‘‘number of sets per exercise’’ and ‘‘number of repetitions within a set until failure [15]. For that reason,
repetitions per set’’ on strength gains. The additional lower repetitions resulted in higher training intensity that
123
induced greater acute neuromuscular fatigue accompanied per repetition seemed to be also important [17]. However,
by greater hormonal responses [70]. the mode of muscle action (isometric, concentric, eccen-
tric) had no effect on strength gains (p = 0.41–0.91). Our
4.2.2 Training Intensity (Intensity, Time Under Tension) search identified 14 studies that reported information on
muscle action-specific time under tension per repetition
In support of the meta-regression results that training during RT (isometric: four studies, range 2.0–2.5 s; con-
intensity (p \ 0.01) predicted the effects of RT on muscle centric: 14 studies, range 1.5–4 s; eccentric: 13 studies,
strength, the largest effect of RT (intensity mean range 1.5–3.5 s). The most effective time under tension
SMDbs = 1.89) on 1RM strength occurred when strength amounted to 2.0 s (mean SMDbs = 2.70), 2.5 s (mean
training intensity was set at 70–79 % of 1RM (range SMDbs = 3.44), and 3.0 s (mean SMDbs = 2.98) for iso-
40–90 % 1RM, Fig. 5). Our systematic search identified metric, concentric, and eccentric muscle actions, respec-
six studies that directly compared RT protocols of different tively. But to the best of our knowledge, there is no study
intensities [31–35]. This analysis showed that high-inten- that compared the effects of contraction duration on
sity RT produced the largest effects on muscle strength in strength gains. The meta-analysis of Roig et al. [76] allows
comparison to moderate- (high vs. moderate mean us at least some insight into muscle action-specific adaptive
SMDbs = 0.60) or low-intensity (high vs. low mean processes in healthy adults aged 18–65 years. These
SMDbs = 0.88) training regimes. Also, moderate-intensity authors stated that separate eccentric muscle actions pro-
RT produced a larger effect on muscle strength compared duce larger gains in muscle strength and morphology
with low-intensity RT (moderate vs. low mean compared with concentric muscle actions. However, these
SMDbs = 0.93). The effects of moderate- and low-inten- findings have to be interpreted with caution because in
sity RT compared with a passive control group had a mean several cases, isotonic RT is applied, which consists of
SMDbs of 1.75 and 1.02 in favor of RT [31, 33–35, 42, 71– concentric and eccentric muscle actions, so that informa-
73]. tion on muscle action-specific time under tension is needed.
Previous meta-analyses suggested similar effects of It has previously been hypothesized that a longer eccentric
high-intensity RT (C70 % 1RM) compared with moderate- phase results in improved training efficiency because
[e.g., mean SMDbs (high vs. moderate) = 0.62] and low- eccentric loads affect the protein synthesis and muscle
intensity [e.g., mean SMDbs (high vs. low) = 0.88] RTs activation and thus muscle hypertrophy and strength [77,
[12, 14, 15] on muscle strength in healthy old adults. These 78]. The results concerning time under tension are limited
findings are in accordance with the ACSM position stand by the low number of studies and by a lack of direct
that states higher intensities result in greater strength gain determination of the muscle action duration effects on
in old adults [61]. Nevertheless, recent reviews rated the strength gains. For example, no study has performed RT
importance of training intensity as a training variable to be with longer contraction duration than 7.5 s per muscle
of minor relevance if no other training variables (i.e., time action. Based on our and previous findings [17], we rec-
under tension, rest time) were considered [15, 74]. Training ommend that authors report time under tension, measured
intensity defined as the individual percentage of 1RM, or estimated, as this seems an important variable underly-
appears not to be as sensitive as the rate of perceived ing gains in muscle strength and muscle morphology.
exertion using, for instance, the OMNI resistance exercise
scale [75]. In other words, the number of repetitions con- 4.2.3 Rest (Rest in Between Sets and Repetitions)
ducted at a given percentage of 1RM differs inter-indi-
vidually because of training status, and intra-individually Meta-regression revealed that rest between sets (p = 0.06)
because of the muscle groups trained [75]. Therefore, the and repetitions did not modify the effects of RT on muscle
1RM represents a method to regulate training intensity that strength. Of the two specific studies that examined dose–
should always be combined with information about the response relationship with respect to rest in between sets,
time under tension [17, 74]. one using 60 s produced the largest mean SMDbs of 4.68 in
Total time under tension had a strong effect (p \ 0.01) healthy old adults. The overall analysis is limited by a
on strength gains, with 6 s per repetition producing the uniform use of 120-s rest in between sets, resulting in a
largest effect size (mean SMDbs = 3.61; 14 studies, range mean SMDbs of 1.57 (Fig. 7). The recent study of Vil-
3–7.5 s). The time under tension is an important variable lanueva et al. [79] investigated the effects of short (60-s)
for mechano-biological adaptations, because different vs. long (240-s) rest intervals between sets on muscle
times under tension affect different metabolic changes as strength and lean body mass after 8-week RT (39/week,
well as motor unit (MU) recruitment and MU firing rates 2–3 sets, 4–6RM) in 22 old men aged 66 years. The find-
occurring during RT [17]. Furthermore, temporal distri- ings revealed that short rest intervals between sets resulted
bution of isometric, concentric, and eccentric muscle action in significant greater increases in leg press 1RM
123
R. Borde et al.
(p \ 0.001) and in lean body mass (p = 0.001). Moreover, examine the effects of training volume on measures of
it is suggested that less rest times produced greater levels of muscle morphology. We found that variation in the volume
fatigue, providing a stimulus which resulted in increases in of RT had no effect on measures of muscle morphology. A
muscle strength [17, 79, 80]. Furthermore, Willardson [81] training period of 6 weeks and using 16–18 repetitions per
hypothesized in a narrative review that shorter rests in set during RT is ineffective for muscle hypertrophy. We
between sets are associated with a more prominent interpret this unexpected result [41] as an abnormality
hypertrophic effect. In addition, there is information in the caused by the choice of unusual training variables (6 weeks
literature stating that the duration of rest in between sets of training; 16–18 repetitions per set), producing an SMDbs
has to be configured to the training goal. Based on different of 0.66 [41]. Nevertheless, a cumulative analysis of the
metabolic and hormonal loads, a narrative review sug- remainder of the studies revealed the following specific
gested that rest in between sets of 180–300 s is suitable for effects on healthy old adults’ muscle morphology when
improvements in maximal strength, 1–2 min for gains in conducting RT with a training period of 50–53 weeks
hypertrophy and 30–60 s for improvements in muscle (mean SMDbs = 0.59), a training frequency of three ses-
endurance [30, 82]. sions per week (mean SMDbs = 0.38), a training volume of
The training variable ‘‘rest time between repetitions’’ two to three sets per exercise (mean SMDbs = 0.78), seven
was computed independently to elucidate dose–response to nine repetitions per set (mean SMDbs = 0.49), a training
relationships, and the results indicated that a 4.0-s rest in intensity of 51–69 % of the 1RM (mean SMDbs = 0.43), a
between repetitions seems to be most effective to increase total time under tension of 6 s (mean SMDbs = 0.36), a
muscle strength (mean SMDbs = 3.72). However, this time under tension of 2.0 s for isometric, concentric, and
finding is preliminary because it is based on one study with eccentric muscle actions (mean SMDbs = 0.36 each),
three training groups only. Nevertheless, the variable ‘‘rest respectively, a rest between sets of 120 s (mean
in between repetitions’’ seems to be a significant mechano- SMDbs = 0.30), and a rest between repetitions of 2.5 s
biological determinant of myocellular oxygen homeostasis (mean SMDbs = 0.36). In general, our findings agree with
[17]. Therefore, it needs to be specified in RT protocols. results reported previously [13, 84, 85]. The meta-analysis
None of the five included studies reported the reason for of Peterson et al. [13] suggested that RT with a mean
the duration of rest used between repetitions. Furthermore, training period of 21 weeks (three training sessions per
no other study compared the effects of in between repeti- week), an intensity of 75 % of the 1RM, two to three sets
tions rest on strength gains at any age. Basically, the effi- and ten repetitions with a 110-s rest in between sets was
ciency of RT (i.e., duration of a single training session) is effective to significantly increase lean body mass in old
influenced by the amount of rest in between repetitions. adults (weighted pooled estimate 1.1 kg; 95 % CI 0.9–1.2).
However, longer rest times between repetitions prolong the The narrative reviews of Mayer et al. [84] and Petrella and
time of a single training session and may thus make Chudyk [85] also illustrated dosage of training variables to
training less efficient. On the other hand, longer rest times prevent the loss of muscle mass. These authors recom-
between repetitions might be particularly beneficial in old mended the following RT variables to prevent the loss of
adults because acute deteriorations in postural control were muscle mass in old age: training period of 8–12 weeks,
reported following one bout of high-intensity RT exercise three training sessions per week, training intensities of
(four sets) [83]. Longer rest times during RT exercises may 60–80 % of the 1RM, three to four sets and eight to 12
affect postural control to a lesser extent by reducing the repetitions per exercise. These recommendations are con-
acute risk of falling during training [83]. This review sistent with the results of the present meta-analysis. How-
provided for the first time information on how to effec- ever, we consider our findings preliminary with regard to
tively implement rest in between repetitions in RT proto- the effects of RT on measures of muscle morphology
cols for old adults. Based on the low number of studies because our systematic search identified only nine eligible
(five studies) and the results of meta-regression, these studies for inclusion in our quantitative sub-analyses and
findings should be interpreted with caution and further meta-regression could not be performed for all
studies are needed. subcategories.
4.3 Dose–Response Relationships of RT to Improve 4.4 Limitations and Strengths of this Review
Measures of Muscle Morphology
Even though the present review has identified the numer-
To the best of our knowledge, no systematic review or ical characteristics of the dose–response relationships, it is
meta-analysis has examined whether changes in muscle a major limitation that such analyses fail to provide insights
morphology would scale according to RT dose in healthy into the physiological stimulus for increasing old adults’
old adults. Due to a low number of studies, we could only muscle strength and muscle size. This is a particularly
123
relevant issue because the number of theories concerning cohorts, RT protocols, and study findings. In addition, large
the stimulus for strength gains involves fatigue [80], total heterogeneity was found across studies, which implies a
work [34, 59, 86], hypoxia [87, 88], and time under tension large variability in the tested muscle strength variables
[89] and these factors are often also cited as concurrently (i.e., tests for upper- and lower-extremity muscles) and the
acting as stimulus for muscle hypertrophy [3, 90]. investigated cohorts (i.e., large age ranges from 60 to
The ultimate aim was to establish a possible combina- 90 years).
tion of a set of RT variables that provides an effective Despite these limitations, this systematic review and
training stimulus for slowing age-related muscle strength meta-analysis is the first to provide an adequate overview
and muscle mass loss. To investigate the effects of training of RT effects on measures of muscle strength and muscle
variables on muscle strength and morphology, subcate- morphology in one meta-analysis. The present meta-anal-
gories were created on the basis of best applicability for ysis analyzed sedentary old adults who commenced RT to
practitioners and clinicians. Afterwards, a meta-regression mitigate the age-related loss of muscle strength and mass.
was performed to find best predictors for effects of RT on In addition, we were able to extract crucial training vari-
measures of muscle strength and muscle morphology. ables, such as volume, intensity, and rest, and their dose–
Indeed, we constructed a dose–response relationship from response relationships for clinicians and practitioners
individual RT variables as additional analyses. The vari- seeking to implement an effective RT in healthy old adults.
ables may be most effective in improving measures of Furthermore, we undertook the first attempt to provide
muscle strength and morphology, but it is unclear if the dose–response relationships for other important training
interaction between the so-specified variables would still variables such as time under tension and rest in between
remain ‘optimal’. We recognize the limitation that our sets and repetitions, albeit these were calculated indepen-
results may not represent one such general dose–response dently of other training variables.
relationship. Modeling of training variables can, however,
address this issue; holding a set of RT variables constant
while changing the effects of one specific variable could 5 Conclusion
determine the unique effects of each training variable [91].
With regard to training volume, the training effects have to This systematic literature review and meta-analysis showed
be interpreted with caution because of the difficulty in that the effects of RT on measures of muscle morphology
quantifying training volume if more than one exercise per (mean SMDbs = 0.42) were much smaller compared with
muscle is performed (e.g., leg press and knee extension/ the effects on muscle strength (mean SMDbs = 1.57) in
curl). Furthermore, due to the nature of meta-analysis, we healthy old adults. The dose–response relationship analyses
focused on those strength outcomes with the highest showed that training period (50–53 weeks, p = 0.04),
functional relevance (e.g., dynamic before isometric intensity (70–79 % 1RM, p \ 0.01), and time under ten-
strength tests). Thus, our findings are outcome specific and sion (6 s, p \ 0.01) can significantly and independently
cannot necessarily be transferred to different strength out- modify the RT effects on muscle strength in healthy old
comes that were not computed in the present study. adults. Data for other variables were insufficient to draw
The methodological quality of the included studies is firm conclusions. It seems that 60 s of rest between sets
rather low because only three out of 25 studies reached the (p = 0.06; two studies), a training frequency of two ses-
pre-determined cut-off score of 6 points on the PEDro scale sions per week, a training volume of two to three sets per
that stands for high-quality studies. Of note, possible sys- exercise, seven to nine repetitions per set, and 4.0 s
tematic errors cannot be eliminated because important between repetitions appear to be the training variables that
points (e.g., blinding of subjects or therapists) for internal could have the greatest and most rapid effects on improv-
validity were not considered in all included studies. Fur- ing maximal voluntary strength in healthy old adults.
thermore, our findings of effects of RT on measures of RT with the following parameters seems to be effective
muscle morphology have to be considered as preliminary to improve measures of muscle morphology: a training
because our systematic search identified only nine studies period of 50–53 weeks, a training frequency of three ses-
based on our selected inclusion criteria. Another limitation sions per week, a training volume of two to three sets per
is that many studies failed to report the training variables. exercise, seven to nine repetitions per set, a training
Further, information regarding subject characteristics were intensity from 51 to 69 % of the 1RM, a total time under
often incomplete (e.g., training status, age, health status) tension of 6.0 s, a rest of 120 s between sets and 2.5 s
and results were inconclusively reported (e.g., means and between repetitions. Practitioners, clinicians, and therapists
standard deviation) so that in several cases we were not should consult these findings with caution and only as an
able to compute SMDs. Future studies should present initial attempt for a comprehensive analysis to characterize
detailed information and data sets on the investigated RT variables for improving healthy old adults’ muscle
123
R. Borde et al.
morphology. Future studies should particularly focus on 14. Steib S, Schoene D, Pfeifer K. Dose-response relationship of
the detailed description of training variables (e.g., time resistance training in older adults: a meta-analysis. Med Sci
Sports Exerc. 2010;42(5):902–14.
under tension) to allow in-depth analysis of dose–response 15. Silva NL, Oliveira RB, Fleck SJ, et al. Influence of strength
relationships following RT in healthy, mobility limited, training variables on strength gains in adults over 55 years-old:
and/or frail old adults. a meta-analysis of dose-response relationships. J Sci Med Sport.
2013;17(3):337–44.
Compliance with Ethical Standards 16. Raymond MJ, Bramley-Tzerefos RE, Jeffs KJ, et al. Systematic
review of high-intensity progressive resistance strength training
This work was supported by a grant from the German Research of the lower limb compared with other intensities of strength
Foundation (MU 3327/2-1). Ron Borde, Tibor Hortobágyi, and Urs training in older adults. Arch Phys Med Rehabil. 2013;94(8):
Granacher declare that they have no conflicts of interest. 1458–72.
17. Toigo M, Boutellier U. New fundamental resistance exercise
determinants of molecular and cellular muscle adaptations. Eur J
Open Access This article is distributed under the terms of the
Appl Physiol. 2006;97(6):643–63.
Creative Commons Attribution 4.0 International License (http://
18. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement
creativecommons.org/licenses/by/4.0/), which permits unrestricted
for reporting systematic reviews and meta-analyses of studies
use, distribution, and reproduction in any medium, provided you give
that evaluate health care interventions: explanation and elabo-
appropriate credit to the original author(s) and the source, provide a
ration. PLoS Med. 2009;6(7):e1000100.
link to the Creative Commons license, and indicate if changes were
19. Maher CG, Sherrington C, Herbert RD, et al. Reliability of the
made.
PEDro scale for rating quality of randomized controlled trials.
Phys Ther. 2003;83(8):713–21.
20. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of
reports of randomized clinical trials: is blinding necessary?
Control Clin Trials. 1996;17(1):1–12.
References 21. Verhagen AP, de Vet HC, de Bie RA, et al. The Delphi list: a
criteria list for quality assessment of randomized clinical trials
1. Danneskiold-Samsoe B, Bartels EM, Bulow PM, et al. Isokinetic for conducting systematic reviews developed by Delphi con-
and isometric muscle strength in a healthy population with sensus. J Clin Epidemiol. 1998;51(12):1235–41.
special reference to age and gender. Acta Physiol (Oxf). 22. Deeks J, Higgins J. Statistical algorithms in Review Manager 5.
2009;197(Suppl 673):1–68. Melbourne: The Cochrane Collaboration; 2010.
2. Delmonico MJ, Harris TB, Visser M, et al. Longitudinal study of 23. Hedges LV, Olkin I. Statistical methods for meta-analysis. San
muscle strength, quality, and adipose tissue infiltration. Am J Diego: Academic Press; 1985.
Clin Nutr. 2009;90(6):1579–85. 24. Lau J, Ioannidis JP, Schmid CH. Summing up evidence: one
3. Manini TM, Clark BC. Dynapenia and aging: an update. answer is not always enough. Lancet. 1998;351(9096):123–7.
J Gerontol A Biol Sci Med Sci. 2012;67(1):28–40. 25. Cohen J. Statistical power for the behavioral sciences. Hillsdale:
4. Aagaard P, Suetta C, Caserotti P, et al. Role of the nervous system Erlbaum; 1988.
in sarcopenia and muscle atrophy with aging: strength training as 26. Bax L, Yu LM, Ikeda N, et al. A systematic comparison of
a countermeasure. Scand J Med Sci Sports. 2010;20:49–64. software dedicated to meta-analysis of causal studies. BMC Med
5. Rubenstein LZ. Falls in older people: epidemiology, risk factors Res Methodol. 2007;7:40.
and strategies for prevention. Age Ageing. 2006;35(Suppl 27. Egger M, Smith GD, Altman DG. Systematic reviews in health
2):ii37–41. care. 2nd ed. Meta-analysis in context. London: BMJ Books;
6. Akima H, Kano Y, Enomoto Y, et al. Muscle function in 164 2001.
men and women aged 20–84 yr. Med Sci Sports Exerc. 28. Thompson SG, Sharp SJ. Explaining heterogeneity in meta-anal-
2001;33(2):220–6. ysis: a comparison of methods. Stat Med. 1999;18(20):2693–708.
7. Frontera WR, Hughes VA, Fielding RA, et al. Aging of skeletal 29. Kraemer WJ, Häkkinen K. Strength training for Sport. Oxford:
muscle: a 12-yr longitudinal study. J Appl Physiol. 2000;88(4): Blackwell Science; 2002.
1321–6. 30. Baechle TR, Earle RW, Wathen D. Resistance training. In:
8. Frontera WR, Bigard X. The benefits of strength training in the Baechle TR, Earle RW, editors. Essentials of strength training
elderly. Sci Sport. 2002;17(3):109–16. and conditioning. Champaign: Human Kinetics; 2000. p. 381–
9. Frontera WR, Meredith CN, O’Reilly KP, et al. Strength con- 409.
ditioning in older men: skeletal muscle hypertrophy and 31. Beneka A, Malliou P, Fatouros I, et al. Resistance training
improved function. J Appl Physiol. 1988;64(3):1038–44. effects on muscular strength of elderly are related to intensity
10. American College of Sports M. ACSM’s guidelines for exercise and gender. J Sci Med Sport. 2005;8(3):274–83.
testing and prescription. Baltimore: Lippincott,Williams & 32. Vincent KR, Braith RW, Feldman RA, et al. Resistance exercise
Wilkins; 2009. and physical performance in adults aged 60 to 83. J Am Geriatr
11. Burns PB, Rohrich RJ, Chung KC. The levels of evidence and Soc. 2002;50(6):1100–7.
their role in evidence-based medicine. Plast Reconstr Surg. 33. Fatouros IG, Kambas A, Katrabasas I, et al. Resistance training
2011;128(1):305–10. and detraining effects on flexibility performance in the elderly
12. Peterson MD, Rhea MR, Sen A, et al. Resistance exercise for are intensity-dependent. J Strength Cond Res. 2006;20(3):
muscular strength in older adults: a meta-analysis. Ageing Res 634–42.
Rev. 2010;9(3):226–37. 34. Hortobagyi T, Tunnel D, Moody J, et al. Low- or high-intensity
13. Peterson MD, Sen A, Gordon PM. Influence of resistance strength training partially restores impaired quadriceps force
exercise on lean body mass in aging adults: a meta-analysis. accuracy and steadiness in aged adults. J Gerontol A Biol Sci
Med Sci Sports Exerc. 2011;43(2):249–58. Med Sci. 2001;56(1):B38–47.
123
35. Kalapotharakos VI, Michalopoulou M, Godolias G, et al. The 56. Sale DG. Neural adaptation to resistance training. Med Sci
effects of high- and moderate-resistance training on muscle Sports Exerc. 1988;20(5 Suppl):S135–45.
function in the elderly. J Aging Phys Act. 2004;12(2):131–43. 57. Häkkinen K, Pakarinen A, Kraemer WJ, et al. Selective muscle
36. Granacher U, Gruber M, Gollhofer A. Resistance training and hypertrophy, changes in EMG and force, and serum hormones
neuromuscular performance in seniors. Int J Sports Med. during strength training in older women. J Appl Physiol.
2009;30(9):652–7. 2001;91(2):569–80.
37. Reeves ND, Narici MV, Maganaris CN. In vivo human muscle 58. Davis MG, Fox KR, Hillsdon M, et al. Objectively measured
structure and function: adaptations to resistance training in old physical activity in a diverse sample of older urban UK adults.
age. Exp Physiol. 2004;89(6):675–89. Med Sci Sports Exerc. 2011;43(4):647–54.
38. Tracy BL, Byrnes WC, Enoka RM. Strength training reduces 59. Vaczi M, Nagy SA, Koszegi T, et al. Mechanical, hormonal, and
force fluctuations during anisometric contractions of the hypertrophic adaptations to 10 weeks of eccentric and stretch-
quadriceps femoris muscles in old adults. J Appl Physiol (1985). shortening cycle exercise training in old males. Exp Gerontol.
2004;96(4):1530–40. 2014;58:69–77.
39. Pyka G, Lindenberger E, Charette S, et al. Muscle strength and 60. Kennis E, Verschueren SM, Bogaerts A, et al. Long-term impact
fiber adaptations to a year-long resistance training program in of strength training on muscle strength characteristics in older
elderly men and women. J Gerontol. 1994;49(1):M22–7. adults. Arch Phys Med Rehabil. 2013;94(11):2054–60.
40. Henwood TR, Taaffe DR. Short-term resistance training and the 61. Chodzko-Zajko WJ, Proctor DN, Fiatarone Singh MA, et al.
older adult: the effect of varied programmes for the enhance- American College of Sports Medicine position stand. Exercise
ment of muscle strength and functional performance. Clin and physical activity for older adults. Med Sci Sports Exerc.
Physiol Funct Imaging. 2006;26(5):305–13. 2009;41(7):1510–30.
41. Pinto RS, Correa CS, Radaelli R, et al. Short-term strength 62. DiFrancisco-Donoghue J, Werner W, Douris PC. Comparison of
training improves muscle quality and functional capacity of once-weekly and twice-weekly strength training in older adults.
elderly women. Age (Dordr). 2014;36(1):365–72. Br J Sports Med. 2007;41(1):19–22.
42. Raso V, Benard G, da Silva Duarte AJ, et al. Effect of resistance 63. Taaffe DR, Duret C, Wheeler S, et al. Once-weekly resistance
training on immunological parameters of healthy elderly exercise improves muscle strength and neuromuscular perfor-
women. Med Sci Sports Exerc. 2007;39(12):2152–9. mance in older adults. J Am Geriatr Soc. 1999;47(10):
43. Judge JO, Whipple RH, Wolfson LI. Effects of resistive and 1208–14.
balance exercises on isokinetic strength in older persons. J Am 64. Radaelli R, Wilhelm EN, Botton CE, et al. Effects of single vs.
Geriatr Soc. 1994;42(9):937–46. multiple-set short-term strength training in elderly women. Age
44. Stewart VH, Saunders DH, Greig CA. Responsiveness of muscle (Dordr). 2014;36(6):9720.
size and strength to physical training in very elderly people: a 65. Fleck SJ, Kraemer WJ. Designing resistance training programs.
systematic review. Scand J Med Sci Sports. 2014;24(1):e1–10. 3rd ed. Champaign: Human Kinetics; 2003.
45. Chen WT, Chou KH, Liu LK, et al. Reduced cerebellar gray 66. Radaelli R, Fleck SJ, Leite T, et al. Dose response of 1, 3 and 5
matter is a neural signature of physical frailty. Hum Brain Mapp. sets of resistance exercise on strength, local muscular endurance
2015;36(9):3666–76. and hypertrophy. J Strength Cond Res. 2014;29(5):1349–58.
46. Heetkamp J, Hortobagyi T, Zijdewind I. Increased bilateral 67. Abrahin O, Rodrigues RP, Nascimento VC, et al. Single- and
interactions in middle-aged subjects. Front Aging Neurosci. multiple-set resistance training improves skeletal and respiratory
2014;6:5. muscle strength in elderly women. Clin Interv Aging.
47. Boisgontier MP. Motor aging results from cerebellar neuron 2014;9:1775–82.
death. Trends Neurosci. 2015;38(3):127–8. 68. Galvao DA, Taaffe DR. Resistance exercise dosage in older
48. Clark BC, Law TD, Hong SL. Editorial: ‘‘From brain to body: adults: single- versus multiset effects on physical performance
the impact of nervous system declines on muscle performance in and body composition. J Am Geriatr Soc. 2005;53(12):
aging’’. Front Aging Neurosci. 2015;7(66):1–2. 2090–7.
49. Clark BC, Taylor JL, Hong SL, et al. Weaker seniors exhibit 69. Campos GE, Luecke TJ, Wendeln HK, et al. Muscular adapta-
motor cortex hypoexcitability and impairments in voluntary tions in response to three different resistance-training regimens:
activation. J Gerontol A Biol Sci Med Sci. 2015;70(6):1112–9. specificity of repetition maximum training zones. Eur J Appl
50. Strotmeyer ES, de Rekeneire N, Schwartz AV, et al. Sensory Physiol. 2002;88(1–2):50–60.
and motor peripheral nerve function and lower-extremity 70. Walker S, Hulmi JJ, Wernbom M, et al. Variable resistance
quadriceps strength: the health, aging and body composition training promotes greater fatigue resistance but not hypertrophy
study. J Am Geriatr Soc. 2009;57(11):2004–10. versus constant resistance training. Eur J Appl Physiol.
51. Ward RE, Boudreau RM, Caserotti P, et al. Sensory and motor 2013;113(9):2233–44.
peripheral nerve function and longitudinal changes in quadriceps 71. Kalapotharakos V, Smilios I, Parlavatzas A, et al. The effect of
strength. J Gerontol A Biol Sci Med Sci. 2015;70(4):464–70. moderate resistance strength training and detraining on muscle
52. Yao WX, Li J, Jiang Z, et al. Aging interferes central control strength and power in older men. J Geriatr Phys Ther.
mechanism for eccentric muscle contraction. Front Aging 2007;30(3):109–13.
Neurosci. 2014;6:86. 72. Strasser B, Keinrad M, Haber P, et al. Efficacy of systematic
53. Bayram MB, Siemionow V, Yue GH. Weakening of cortico- endurance and resistance training on muscle strength and endur-
muscular signal coupling during voluntary motor action in ance performance in elderly adults–a randomized controlled trial.
aging. J Gerontol A Biol Sci Med Sci. 2015;70(8):1037–43. Wien Klin Wochenschr. 2009;121(23–24):757–64.
54. Papegaaij S, Taube W, Baudry S, et al. Aging causes a reor- 73. Vincent HK, Bourguignon C, Vincent KR. Resistance training
ganization of cortical and spinal control of posture. Front Aging lowers exercise-induced oxidative stress and homocysteine
Neurosci. 2014;6:28. levels in overweight and obese older adults. Obesity (Silver
55. Venturelli M, Saggin P, Muti E, et al. In vivo and in vitro Spring). 2006;14(11):1921–30.
evidence that intrinsic upper- and lower-limb skeletal muscle 74. Marschall F, Buesch D. Positionspapier für eine
function is unaffected by ageing and disuse in oldest-old beanspruchungsorientierte Trainingsgestaltung im Krafttraining.
humans. Acta Physiol (Oxf). 2015;215(1):58–71. Schweiz Z Sportmed Sporttraumatol. 2014;62(1):24–31.
123
R. Borde et al.
75. Robertson RJ, Goss FL, Rutkowski J, et al. Concurrent valida- 88. Scott BR, Slattery KM, Sculley DV, et al. Hypoxia and resis-
tion of the OMNI perceived exertion scale for resistance exer- tance exercise: a comparison of localized and systemic methods.
cise. Med Sci Sports Exerc. 2003;35(2):333–41. Sports Med. 2014;44(8):1037–54.
76. Roig M, O’Brien K, Kirk G, et al. The effects of eccentric versus 89. Atha J. Strengthening muscle. Exerc Sport Sci Rev. 1981;9:
concentric resistance training on muscle strength and mass in 1–73.
healthy adults: a systematic review with meta-analysis. Br J 90. Hortobagyi T. The positives of negatives: clinical implications
Sports Med. 2009;43(8):556–68. of eccentric resistance exercise in old adults. J Gerontol A Biol
77. Cermak NM, Snijders T, McKay BR, et al. Eccentric exercise Sci Med Sci. 2003;58(5):M417–8.
increases satellite cell content in type II muscle fibers. Med Sci 91. Arandjelovic O. A mathematical model of neuromuscular
Sports Exerc. 2013;45(2):230–7. adaptation to resistance training and its application in a com-
78. Hortobagyi T, Hill JP, Houmard JA, et al. Adaptive responses to puter simulation of accommodating loads. Eur J Appl Physiol.
muscle lengthening and shortening in humans. J Appl Physiol 2010;110(3):523–38.
(1985). 1996;80(3):765–72. 92. Charette SL, McEvoy L, Pyka G, et al. Muscle hypertrophy
79. Villanueva MG, Lane CJ, Schroeder ET. Short rest interval response to resistance training in older women. J Appl Physiol.
lengths between sets optimally enhance body composition and 1991;70(5):1912–6.
performance with 8 weeks of strength resistance training in 93. Daly M, Vidt ME, Eggebeen JD, et al. Upper extremity muscle
older men. Eur J Appl Physiol. 2015;115(2):295–308. volumes and functional strength after resistance training in older
80. Rooney KJ, Herbert RD, Balnave RJ. Fatigue contributes to the adults. J Aging Phys Act. 2013;21(2):186–207.
strength training stimulus. Med Sci Sports Exerc. 1994;26(9): 94. DeBeliso M, Harris C, Spitzer-Gibson T, et al. A comparison of
1160–4. periodised and fixed repetition training protocol on strength in
81. Willardson J. A brief review: factors affecting the length of the older adults. J Sci Med Sport. 2005;8(2):190–9.
rest interval between resistance exercise sets. J Strength Cond 95. Fatouros IG, Taxildaris K, Tokmakidis SP, et al. The effects of
Res. 2006;20(4):978–84. strength training, cardiovascular training and their combination
82. Bird SP, Tarpenning KM, Marino FE. Designing resistance on flexibility of inactive older adults. Int J Sports Med.
training programmes to enhance muscular fitness: a review of 2002;23(2):112–9.
the acute programme variables. Sports Med. 2005;35(10): 96. Hunter GR, Wetzstein CJ, McLafferty CL Jr, et al. High-resis-
841–51. tance versus variable-resistance training in older adults. Med Sci
83. Donath L, Kurz E, Roth R, et al. Does a single session of Sports Exerc. 2001;33(10):1759–64.
high-intensity interval training provoke a transient elevated 97. Lovell DI, Cuneo R, Gass GC. The effect of strength training
risk of falling in seniors and adults? Gerontology. 2015; and short-term detraining on maximum force and the rate of
61(1):15–23. force development of older men. Eur J Appl Physiol.
84. Mayer F, Scharhag-Rosenberger F, Carlsohn A, et al. The 2010;109(3):429–35.
intensity and effects of strength training in the elderly. Dtsch 98. Miszko TA, Cress ME, Slade JM, et al. Effect of strength and
Arztebl Int. 2011;108(21):359–64. power training on physical function in community-dwelling
85. Petrella RJ, Chudyk A. Exercise prescription in the older athlete older adults. J Gerontol A Biol Sci Med Sci. 2003;58(2):171–5.
as it applies to muscle, tendon, and arthroplasty. Clin J Sport 99. Morse CI, Thom JM, Mian OS, et al. Muscle strength, volume
Med. 2008;18(6):522–30. and activation following 12-month resistance training in
86. Alegre LM, Aguado X, Rojas-Martin D, et al. Load-controlled 70-year-old males. Eur J Appl Physiol. 2005;95(2–3):197–204.
moderate and high-intensity resistance training programs pro- 100. Rhodes EC, Martin AD, Taunton JE, et al. Effects of one year of
voke similar strength gains in young women. Muscle Nerve. resistance training on the relation between muscular strength
2015;51(1):92–101. and bone density in elderly women. Br J Sports Med.
87. Manimmanakorn A, Hamlin MJ, Ross JJ, et al. Effects of low- 2000;34(1):18–22.
load resistance training combined with blood flow restriction or 101. Judge JO, Whipple RH, Wolfson LI. Effects of resistive and
hypoxia on muscle function and performance in netball athletes. balance exercises on isokinetic strength in older persons. J Am
J Sci Med Sport. 2013;16(4):337–42. Geriatr Soc. 1994;42(9):937–46.
123
View publication stats

Research

Uploaded by

Copyright:

Available Formats

Research

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Research

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author proﬁles for this publication at: https://www.researchgate.

Dose–Response Relationships of Resistance Training in Healthy Old Adults:

Article in Sports Medicine · September 2015

Ron Borde Urs Granacher

SEE PROFILE SEE PROFILE

Metastability in Sport Biomechanics View project

The user has requested enhancement of the downloaded ﬁle.

Dose–Response Relationships of Resistance Training in Healthy

Abstract Study Eligibility Criteria The initial search identified 506

PubMed (n = 138), Web of Science (n = 185), Cochrane Library (n = 183)

Papers excluded on basis of title (n = 287)

Potentially relevant papers remaining (n = 219)

Duplicate papers excluded (n = 109)

Potentially relevant papers remaining (n = 110)

Papers excluded on basis of eligibility criteria (n = 85)

Included papers (n = 25)

HI: 0.02 FFM CG: no intervention

3.3.3 Training Frequency

Twenty-five studies were included in this sub-analysis, and

week; 95 % CI 1.20–1.94; I2 = 79 %, v2 = 163.10,

df = 32, p \ 0.01). That is, two and three training sessions

3.3.4 Number of Sets and Repetitions

In the 25 studies included in this sub-analysis, the number

and the number of repetitions per set averaged 10.0 (range

five to 16 repetitions). Mean SMDbs for number of sets and

p \ 0.01), indicative of large effects. Two to three sets per

exercise (mean SMDbs = 2.99) and seven to nine repeti-

tions (mean SMDbs = 1.98) resulted in the largest

improvements in muscle strength.

3.3.5 Training Intensity

Twenty-four studies were included in this sub-analysis, and

training intensity was classified as high (C70 % 1RM),

moderate (51 % C 1RM B 69 %), and low (B50 % 1RM)

[16]. The sub-analysis revealed a mean intensity of 69 % of

the 1RM (range 40–90 % 1RM) across studies. Figure 5

sity, with a mean SMDbs of 1.63 (95 % CI 1.21–2.05;

? indicates a ‘‘yes’’ score, - indicates a ‘‘no’’ score

effects on measures of muscle strength were found for

3.3.6 Time Under Tension per Repetition

Time under tension is an important variable to induce

adaptations in muscle strength and morphology [17]. In 14

studies, the total time under tension averaged 5.7 s per

repetition (range 3–7.5 s; mean SMDbs = 1.60; 95 % CI

The largest effect was shown for 6 s, with a mean SMDbs

of 3.61. Figure 6 shows the dose–response relationships for

the training variable ‘‘time under tension’’. In addition, the

mean time under tension was 2.3 s for isometric (range

Table 4 Training variables with largest mean SMDbs

Training period [weeks] 50–53 2.34 50–53 0.59a

Standardized mean difference

Training period (weeks)

triangles symbolize high-, 4

Training intensity (%)

relationships for total time

weighted mean SMDbs of all

Standardized mean difference

Time under tension (s)

relationships for rest in between

SMDbs of all studies. Diamonds,