Contents lists available at ScienceDirect
Experimental Gerontology
journal homepage: www.elsevier.com/locate/expgero
Effects of different strength training frequencies on maximum strength, body
composition and functional capacity in healthy older individuals
Mari Turpelaa
, Keijo Häkkinena
, Guy Gregory Haffb
, Simon Walkera,⁎
a
Neuromuscular Research Center, Faculty of Sport and Health Sciences, University of Jyväskylä, Finland
b
Centre for Exercise and Sports Science Research (CESSR), Edith Cowan University, Joondalup, Australia
A R T I C L E I N F O
Section Editor: Christiaan Leeuwenburgh
Keywords:
Walking
Timed-up-and-go
Stair climb
Muscle mass
Aged men and women
Resistance exercise
Lower limbs
A B S T R A C T
There is controversy in the literature regarding the dose-response relationship of strength training in healthy
older participants. The present study determined training frequency effects on maximum strength, muscle mass
and functional capacity over 6 months following an initial 3-month preparatory strength training period. Onehundred
and six 64–75 year old volunteers were randomly assigned to one of four groups; performing strength
training one (EX1), two (EX2), or three (EX3) times per week and a non-training control (CON) group. Wholebody
strength training was performed using 2–5 sets and 4–12 repetitions per exercise and 7–9 exercises per
session. Before and after the intervention, maximum dynamic leg press (1-RM) and isometric knee extensor and
plantarflexor strength, body composition and quadriceps cross-sectional area, as well as functional capacity
(maximum 7.5 m forward and backward walking speed, timed-up-and-go test, loaded 10-stair climb test) were
measured. All experimental groups increased leg press 1-RM more than CON (EX1: 3 ± 8%, EX2: 6 ± 6%,
EX3: 10 ± 8%, CON: −3 ± 6%, P < 0.05) and EX3 improved more than EX1 (P = 0.007) at month 9.
Compared to CON, EX3 improved in backward walk (P = 0.047) and EX1 in timed-up-and-go (P = 0.029) tests.
No significant changes occurred in body composition. The present study found no evidence that higher training
frequency would induce greater benefit to maximum walking speed (i.e. functional capacity) despite a clear
dose-response in dynamic 1-RM strength, at least when predominantly using machine weight-training. It appears
that beneficial functional capacity improvements can be achieved through low frequency training (i.e. 1–2 times
per week) in previously untrained healthy older participants.
1. Introduction
Strength training is a widely used method to combat the deleterious
effects of aging and age-related reduced physical activity on maximum
strength, muscle mass and functional capacity. There are many combinations
of acute program variables (identified by Kraemer and
Ratamess, 2004) that can influence the overall outcome of a strengthtraining
program. These variables are; the choice of exercise(s) and
exercise order, number of sets/repetitions, inter-set and inter-exercise
rest interval, and the intensity of each exercise. The effects of several of
these variables on maximum strength and muscle mass development
have been examined over previous decades (e.g. Campos et al., 2002;
Moss et al., 1997). But one variable, training frequency, has received
little attention (Steib et al., 2010). It is important to be clear that
training frequency in the present study is limited to whole-body
strength training (rather than split programs; training one specific
muscle group per day) and the vast majority of studies using training
2–3 times per week does not allow reviews/meta-analyses to accurately
determine the effects of different frequencies on outcome variables.
Nevertheless, physical activity guidelines from bodies such as the
World Health Organization and the American College of Sports
Medicine recommend whole-body strength training for healthy individuals
above 65 years at a frequency of at least two times per week
(Ratamess et al., 2009; World Health Organization, 2010). This is despite
the little experimental evidence to support such a recommendation
regarding development of maximum strength or muscle mass, and
particularly functional capacity, in previously untrained healthy older
individuals. This is in contrast to the quite well-established evidence
base to recommend progressive loading and volume to promote
achieving these desirable outcomes (Ratamess et al., 2009).
A seminal paper investigating training frequency (one versus two
versus three times per week) on improvements in maximum strength
and functional capacity observed no difference in improvements between
groups (Taaffe et al., 1999). Also, a recent meta-analysis showed
no evidence of different strength improvements comparing frequencies
of one, two or three times per week (Silva et al., 2014). Maintenance of
http://dx.doi.org/10.1016/j.exger.2017.08.013
Received 26 April 2017; Received in revised form 11 July 2017; Accepted 8 August 2017
⁎
Corresponding author at: Room VIV223, Faculty of Sport and Health Sciences, University of Jyväskylä, FI-40014, Finland.
E-mail address: simon.walker@jyu.fi (S. Walker).
Experimental Gerontology 98 (2017) 13–21
Available online 15 August 2017
0531-5565/ © 2017 Elsevier Inc. All rights reserved.
MARK
muscle mass is another important consideration for older adults given
its role in force production and also metabolic regulation. However, to
our knowledge, no study has investigated the effect of training frequency
on muscle hypertrophy in healthy older individuals. The effect
of training frequency on muscle hypertrophy would be pertinent to
examine since most studies use either two or three times per week,
which has been shown to exert little difference (Wernbom et al., 2007),
but recent evidence suggests these frequencies are more beneficial than
one time per week (Schoenfeld et al., 2016), which does support the
physical activity guidelines.
One important methodological consideration when evaluating these
studies is the existing training status of the participants. All four original
articles that we identified in the literature investigating training
frequency in healthy older individuals used previously untrained participants
(DiFrancisco-Donoghue et al., 2007; Farinatti et al., 2013;
Padhila et al., 2015; Taaffe et al., 1999). As it is known that untrained
individuals respond more robustly to a variety training protocols, the
use of completely untrained participants may reduce any potential to
identify differences in adaptive responses in response to different
training frequencies.
Therefore, there is a need to further study the influence of training
frequency on improvements in maximum strength, muscle mass and
functional capacity in healthy older individuals that have undergone
(some) strength training prior to separation into different training frequencies.
Consequently, the purpose of the present study was to determine
whether training frequency affects improvements in maximum
strength, muscle mass and functional capacity over a 6-month period
following an initial 3-month low-intensity preparatory strength training
period.
2. Materials and methods
2.1. Participant recruitment and randomization
This study was the second arm of a randomized controlled trial
(NCT02413112). Participants were 64–75-year-old men and women.
Exclusion criteria were; (1) regular aerobic exercise (> 180 min·week−
1
), (2) any previous strength training experience, (3) Body Mass
Index > 37, (4) serious cardiovascular disease or lower limb injuries/
disease that may lead to complications during exercise or affect the
ability to perform testing and training, (5) use of walking aids, (6) use
of medication that affect the neuromuscular or endocrine systems, (7)
previous testosterone-altering treatment, and (8) smoking. Therefore,
participants were otherwise healthy apart from conditions such as Type
II diabetes, high blood pressure, and/or high cholesterol in several
cases, were not frail or obese, were not engaged in systematic fitness
training, and were able to perform strength training with no restrictions.
While the participants did not engage in aerobic exercise, it was
clear from the pre-study interviews that typical ‘Nordic’ low-intensity
physical activity (e.g. berry-picking, gardening, forestry etc.) was part
of their lifestyle – and may, in part, explain their largely healthy condition
despite not meeting recommended levels of physical activity
(WHO, 2010).
The recruitment process and exclusion of participants in shown in
Fig. 1. Prior to physician assessment, advertisement letters were posted
to 2000 65–75-year-old men and women in the Jyväskylä region and
potential participants registered to the study by completing an online
researcher-designed questionnaire (n = 454). As part of the registration
questionnaire, potential participants were asked about their current and
previous level of physical activity, medical history including any current/ongoing/permanent
conditions, current and previous medications
and also immediate family medical history. The participants were blind
to the purpose of these questions (i.e. to assess eligibility). After assessing
the eligibility of the registered individuals for lower limb injuries,
skeletomuscular diseases and physical activity levels, potential
participants were invited to an information session (n = 148). Each
participant was carefully informed of the study design and potential
risks before the study, after which they provided written consent and
attended a physician's examination (n = 116). During the physician's
examination, potential participants were interviewed by the researchers
to ensure that they were eligible to be included to the study. After
baseline testing, the participants (n = 106) were allocated an identification
number and a computer-generated random number sequencer
was used to allocate each participant into one of four groups (Fig. 1);
training one (EX1), two (EX2), three (EX3) times per week and nontraining/wait
control (CON).
During the study, one participant dropped out due to back pain
induced by the strength testing in month 3, one participant dropped out
due to re-occurrence of heart arrhythmia and one participant dropped
out due to stress-related illness. Six participants failed to attend at least
90% of the assigned training sessions for their group and were consequently
removed from the analyses (as noted in Fig. 1). Furthermore,
after data checking, several participants' electromyography and voluntary
activation level data were excluded from final analysis due to
technical faults. The study was conducted according to the Declaration
of Helsinki and was approved by the ethical committee of the University
of Jyväskylä, Finland. Baseline characteristics of the participants
in each group are shown in Table 1, with the only differences observed
between men and women in height and body mass.
Some participants were taking medication during the study that was
deemed not to interfere with their ability to participate in training or
testing. The total number users of each type of medication are listed
here; EX1: cholesterol medication (3 men + 3 women), blood pressure
medication (4 men + 3 women), blood glucose medication (1 men + 1
women), thyroid medication (1 men + 2 women), beta-blockers (1
woman); EX2: cholesterol medication (2 men + 3 women), blood
pressure medication (5 men + 6 women), blood glucose medication (2
women), thyroid medication (1 man + 2 women), beta-blockers (1
man + 1 woman); EX3: cholesterol medication (1 man + 3 women),
blood pressure medication (5 men + 5 women), blood glucose medication
(2 men), thyroid medication (1 man + 4 women), beta-blockers
(1 man + 2 women); CON: cholesterol medication (2 men + 2
women), blood pressure medication (4 men + 3 women), beta-blockers
(1 man + 1 woman).
2.2. Dynamic leg press performance
Concentric bilateral leg press one-repetition maximum (1-RM) load
(kg) was used to assess maximum dynamic strength (David Sports Ltd.,
Helsinki, Finland). Briefly, following warm-up, single repetitions with
increments of 5 kg were performed until the participants could no
longer fully extend their hips and legs (full extension = 180°). Each
trial was separated by 1.5 min. All data were relayed to a pc via an AD
converter (Micro 1401, Cambridge Electronic Design, UK) and recorded
using Signal 4.04 software (Cambridge Electronic Design, UK). Data
was sampled at 2000 Hz and filtered by a 10-Hz low-pass filter (fourthorder
Butterworth) and the best trial was used in further analyses.
2.3. Isometric knee extension and plantarflexion performance
Maximum unilateral isometric knee extension torque of the right leg
was measured using a custom-built isometric force chair. Inelastic
straps were used to secure the participant with both hip and knee angles
of 110°. Participants were instructed to kick “as fast and as hard as
possible” and maintain their maximum force for approximately 3 s. The
force signal was sampled as described in the leg press trials with the
highest force used in further analysis. Three trials were performed with
a fourth trial performed if improvements were > 5%. Thereafter, two
additional maximum isometric knee extension trials were performed
with femoral nerve stimulation delivered during the force plateau and
2 s after contraction cessation following similar procedures as Walker
et al. (2014). Rectangular pulses (400 V) of 200 μs were delivered by a
M. Turpela et al. Experimental Gerontology 98 (2017) 13–21
14
constant current stimulator (Model DS7AH, Digitimer Ltd., UK) to the
femoral nerve of the right leg through 5 cm2
self-adhesive electrodes
(Polar Trode, Niva Medical Ltd., Espoo, Finland) placed in the femoral
triangle either side of the nerve, which was identified by palpating and
identifying the femoral artery. Current intensity was gradually increased
until no further increases were observed in peak-to-peak Mwave
amplitude of VL and VM. To ensure maximal activation, an additional
20% current was used during subsequent stimulations. Single
twitches were delivered in a resting condition to determine peak-topeak
maximum M-wave amplitude. Single twitches were also delivered
about the maximum torque during isometric knee extension trials and
2 s after contraction cessation to determine voluntary activation level
according to Merton's (1954) interpolated twitch technique. Maximum
force was measured and then converted to torque by taking into account
the lever arm distance from the knee joint-center to the strain
gauge.
Maximum bilateral isometric plantarflexion torque was assessed in a
seated position by a custom-built plantarflexion device with knees
flexed to approximately 90° using similar methods to Unhjem et al.
(2015). The balls of the feet were positioned on a shelf connected to the
strain gauge (90° ankle joint-angle) and the knees were held in-place by
a cushioned board. Participants performed 3–4 isometric plantarflexion
actions following the same instructions as for the knee extension trials.
Maximum force was measured and then converted to torque by taking
into account the lever arm distance from the ankle joint-center to the
strain gauge.
2.4. Functional capacity
Four maximal walking tests performed were included in the assessment
of functional capacity; (1) 7.5 m forward walk, (2) 7.5 m
backward walk, (3) timed up-and-go test (TUG), and (4) loaded 10-stair
climb test. Participants were instructed to perform the tests “as fast as
possible without compromising safety”. Each test was recorded by
photocells except TUG, which used a contact mat positioned under the
chair to determine rise from and return to the chair. The participants
were not allowed to use their arms to assist in the chair rise or return.
During the 10-stair climb test, the participants carried one bag of 5 kg
(women) or 10 kg (men) and were instructed to maintain and extended
elbow position and prevent arm-swinging during the ascent (Fig. 2).
The best performance from two acceptable trials was used in the analyses,
and the sum result from both directions was used for TUG.
Fig. 1. Study flowchart from the point of physician assessment. M = men, W = women, EMG = electromyography, VA = voluntary activation level.
Table 1
Participant characteristics and performance at baseline (mean ± SD).
EX1 EX2 EX3 Control
Sex (M/W) 11/13 9/14 11/14 11/9
Age (years) 70 ± 3 69 ± 3 70 ± 3 69 ± 2
Body mass
(kg)
76 ± 15 81 ± 15 82 ± 16 74 ± 11
Height (m) 1.67 ± 0.09 1.68 ± 0.07 1.67 ± 0.10 1.68 ± 0.09
Body mass
index
(kg/m2
)
27 ± 3 29 ± 5 29 ± 4 26 ± 3
Fat mass (kg) 25.9 ± 6.4 29.1 ± 9.3 28.3 ± 8.1 22.5 ± 6.5⁎
Fat-free mass
(kg)
47.0 ± 11.9 47.9 ± 10.6 48.7 ± 10.6 48.5 ± 10.1
1-RM load
(kg)
104.0 ± 34.7 115.4 ± 36.5 111.6 ± 37.0 119.5 ± 29.7
KE MVC (Nm) 153 ± 57 155 ± 49 147 ± 44 157 ± 45
PF MVC (Nm) 158 ± 61 159 ± 44 160 ± 40 160 ± 44
Forward walk
(s)
3.1 ± 0.5 2.9 ± 0.6 3.0 ± 0.5 2.7 ± 0.4
Backward
walk (s)
4.3 ± 1.2 4.2 ± 1.5 4.5 ± 1.4 3.7 ± 0.8
TUG (s) 9.4 ± 1.5 9.3 ± 1.7 9.4 ± 1.3 8.6 ± 0.8
Stair climb (s) 3.5 ± 0.9 3.4 ± 0.7 3.5 ± 0.7 3.1 ± 0.4
M = men, W = women, 1-RM = one-repetition maximum, KE MVC = maximum isometric
knee extension torque, PF MVC = maximum isometric plantarflexion torque,
TUG = timed-up-and-go.
⁎
P = 0.039 between Control and EX2.
M. Turpela et al. Experimental Gerontology 98 (2017) 13–21
15
2.5. Quadriceps electromyography measurement and analysis
Electrode locations for electromyography (EMG) recordings were
marked by indelible ink tattoo to allow accurate replacement during all
test sessions. Bipolar Ag/AgCl electrodes (5 mm diameter, 20 mm interelectrode
distance, common mode rejection ratio > 100 dB, input
impedance > 100 MΩ, baseline noise < 1 μV rms) were positioned
following shaving and skin abrasion on the vastus lateralis (VL) and
medialis (VM) of the right leg according to SENIAM guidelines. Raw
EMG signals were sampled at 2000 Hz and amplified at a gain of 500
(sampling bandwidth 10–500 Hz). Raw signals were sent from a hipmounted
pack to a receiving box (Telemyo 2400R, Noraxon, Scottsdale,
USA), then were relayed to an AD converter (Micro1401, Cambridge
Electronic Design, UK) and recorded by Signal 4.04 software
(Cambridge Electronic Design, UK). Offline, EMG signals were bandpass
filtered at 20–350 Hz and root mean square was obtained from
approx. 65° to full leg extension (i.e. 180°) during 1-RM trials with
values from VL and VM averaged (VL + VM/2) from the best trial and
used in further analysis.
2.6. Body composition
Participants fasted overnight for 12 h and were instructed to drink
0.5 l of water 1 h before measurements. After determination of height
by a fixed wall-mounted scale, participants underwent full body scanning
by dual-energy X-ray absorptiometry (DXA) in minimal clothing
(LUNAR Prodigy Advance with encore software version 9.3, GE medical
systems, USA). The legs were separated by a polystyrene block and
secured by inelastic straps about the ankles. Total body fat mass and fatfree
mass, as well as fat-free mass of the legs was determined using
software-generated analysis.
2.7. Muscle cross-sectional area
Muscle cross-sectional area (CSA) measurements of the right leg
were taken 1–2 days prior to dynamic leg press performance tests and
6–7 days after the final training session to account for any exerciseinduced
swelling. CSA of the vastus lateralis, vastus intermedius, gastrocnemius
medialis and lateralis was assessed by B-mode axial-plane
ultrasound (model SSD-α10, Aloka Co Ltd., Tokyo, Japan) using a
10 MHz linear-array probe (60 mm width) coated with water-soluble
transmission gel with the extended-field-of-view mode (23 Hz sampling
frequency). This method has been used during several of our training
studies (Walker et al., 2014, 2015) and has been shown to be valid
(Ahtiainen et al., 2009). Indelible ink tattoos on the medial and lateral
sides of the target muscles ensures accurate replacement of scanning
track. Oriented in the axial-plane, the probe was moved manually with
a slow and continuous movement from medial to lateral along a marked
line on the skin. Great care was taken to diminish compression of the
muscle tissue. Images were obtained throughout the movement. As the
orientation of each image relative to adjacent images is known, the
software builds a composite image. Four panoramic CSA images were
taken at; (1) 50% femur length from the lateral aspect of the distal
diaphysis to the greater trochanter and (2) 30% lower limb length from
the lateral articular cleft between the femur and tibia condyle to the
lateral malleolus following the methods used by Rosenberg et al.
(2014). Upon visual inspection of the composite images three were
selected to undergo further analysis. CSA was determined by manually
tracing along the border of each muscle using Image-J software (version
1.37, National Institute of Health, USA). The mean of the two closest
values for each muscle were taken as the CSA result.
2.8. Strength training program
The experimental group performed whole-body strength training
either one (EX1), two (EX2) or three (EX3) times per week for 6 months
on non-consecutive days and each session was supervised by experienced
gym instructors. The 6-month program was divided into two 3month
mesocycles (see Supplementary material). All exercises were
performed on commercially available weight-stack equipment (Precor
Vitality Series™, Precor Inc., UK) apart from several free-weight exercises
in mesocycle 2. The primary goal of mesocycle 1 was to increase
maximum strength and muscle mass. The primary goal of mesocycle 2
was to increase maximum strength and muscle activation/power.
Intensity for all upper and lower limb exercises was approximately
70–90% 1-RM with power training performed using 30–80% 1-RM
loads. Multiple sets (2–5) were performed with repetition ranges of
4–12 and inter-set rest of 1-3 min depending on the training goal (loads
used during training are depicted in Fig. 3). All participants were required
to perform at least 1 set to concentric failure (with the exception
of power training). All participants were required to complete at least
90% of all allocated training sessions prior to testing. Participants in the
non-training control group were instructed to maintain their normal
physical activity throughout the study period. All participants recorded
their daily leisure-time physical activity levels in diaries throughout the
6-month period and 3-day (including one weekend day) diet diaries
were collected during each mesocycle. The recording of physical activity
followed procedures of Waller et al. (2013).
2.9. Statistical analyses
All data are presented as means and standard deviations ( ± SD).
All statistical methods were performed using IBM SPSS statistics 24
software. The Kolmogorov-Smirnov test was used to test normality and
Levene's test was used to analyze homogeneity of variance. One-way
ANOVA was used to assess potential differences at baseline. To determine
whether sex influenced adaptation to the training programs,
analysis of covariance was used with sex as the covariate. Since there
were no statistically significant time × sex interactions, both men and
women in each group were combined for further analyses (results for
Fig. 2. Experimental set-up and procedures for the functional
capacity tests. TUG = timed-up-and-go.
M. Turpela et al. Experimental Gerontology 98 (2017) 13–21
16
men and women separately are shown in the Supplementary material).
Repeated measures analysis of variance (ANOVA; 4 group × 2 time)
was used to determine significant time and time × group effects of the
intervention. Bonferroni post hoc tests used to determine significant
differences within-group over time, while one-way ANOVA with
Bonferroni post hoc tests were used to determine whether the relative
changes (Δ%) over time were different between-groups. Effect sizes
(Hedges' g) were calculated for the differences in relative change between
the experimental and control groups, where small (< 0.3),
medium (0.3–0.8), and large (> 0.8) effect sizes were identified.
Pearson's product moment correlation coefficient was used to determine
possible relationships between outcome measures. Statistical
significance was accepted when P < 0.05. Reliability for the performance
measures between the familiarization session and baseline
measures were; 1-RM 0.97 and 5.5%, peak power 0.94 and 11.2%,
maximum isometric knee extension force 0.89 and 9.6%, maximum
isometric plantarflexion force 0.87 and 9.7%, forward walk 0.82 and
6.3%, backward walk 0.81 and 8.3%, TUG 0.89 and 3.2%, 10-stair
climb 0.96 and 3.2%, and CSA 0.94 and 4.2% for Intra-class correlation
coefficient (r) and coefficient of variation (%), respectively.
3. Results
3.1. Loads used during the strength training intervention
All training loads used throughout the study are presented in the
Supplementary material for the leg press, knee extension and chest/
bench press exercises. Fig. 3 shows the highest loads used during each
week for the leg press exercise in all training groups.
3.2. Neuromuscular performance and muscle activation
Statistically significant main effects for time (F = 25.8, P < 0.001)
and time × group interaction (F = 12.7, P < 0.001) were observed in
leg press 1-RM, where post hoc tests revealed that EX2 (6 ± 6%,
P < 0.001) and EX3 (10 ± 8%, P < 0.001) increased strength significantly
over the intervention period. At month 9, between-group
differences were observed for all experimental groups versus control
and between EX3 and EX1 (P = 0.007, 95% confidence interval = 1.3
to 12.4%, g = 0.89). The same results were observed when 1-RM was
expressed relative to BM (Fig. 4). There were no significant (time × group)
main effects for maximum isometric knee extension (Fig. 4) or
plantarflexion.
Statistically significant main effects for time (F = 40.8, P < 0.001)
and time × group interaction (F = 4.4, P = 0.007) were observed in
muscle activity of the quadriceps (VL + VM/2) during leg press 1-RM,
where post hoc tests revealed that all experimental groups increased
significantly over the intervention period (EX1: 25 ± 29%, P = 0.003;
EX2: 31 ± 37%, P = 0.002; EX3: 43 ± 31%, P < 0.001). Betweengroup
differences were observed for EX3 and control (P = 0.005, 95%
confidence interval = 7.9 to 63.2%, g = 1.18). However, once normalized
to maximum M-wave amplitude there were no significant main
effects, and within-group changes were no longer statistically significant.
Furthermore, there were no significant main effects for voluntary
activation level assessed by the twitch interpolation technique
during isometric knee extension trials.
3.3. Functional capacity
Statistically significant main effects for time and trends for
time × group interaction were observed for backward walk (F = 9.9,
P = 0.002, and F = 2.7, P = 0.053, respectively), TUG (F = 19.9,
P < 0.001, and F = 2.6, P = 0.056, respectively) and loaded 10-stair
climb (F = 26.8, P < 0.001, and F = 2.7, P = 0.051, respectively). A
significant main effect for time (F = 35.9, P < 0.001) was observed
for forward walk. From month 3 to 9, EX1 improved forward walk
(−7 ± 8%, P = 0.002) and TUG (−4 ± 4%, P < 0.001) performance
only. EX2 and EX3 improved performance in all functional capacity
tests from month 3 to 9; forward walk (EX2: −4 ± 6%,
P < 0.001; EX3: −5 ± 6%, P = 0.001), backward walk (EX2:
−4 ± 8%, P = 0.02; EX3: −8 ± 9%, P = 0.001), TUG (EX2:
−2 ± 3%, P = 0.011; EX3: −3 ± 6%, P = 0.033) and loaded 10stair
climb (EX2: −5 ± 4%, P < 0.001; EX3: −4 ± 5%,
P = 0.001). At month 9, between-group differences were observed for
EX3 and control in backward walk (P = 0.047, 95% confidence interval
= −19.7 to −0.1%, g = 0.72, Fig. 5) and EX1 versus control in
TUG (P = 0.029, 95% confidence interval = −8.2 to −0.3%,
g = 0.95, Fig. 5).
3.4. Body composition and muscle mass
There were no significant main effects for total fat mass, total lean
mass or lean leg mass (Fig. 4), or CSA for any muscle assessed.
3.5. Habitual physical activity and nutritional intake
Habitual physical activity external to the prescribed intervention
was similar between all groups (average minutes per week = EX1:
136 ± 125, EX2: 121 ± 110, EX3: 108 ± 109, control: 172 ± 69).
Walking was the most common mode of physical activity, and this did
not differ between groups either in total time or in proportion of total
physical activity (EX1: 107 ± 122, EX2: 80 ± 70, EX3: 65 ± 77,
Fig. 3. Highest (weekly) load used during the leg press
exercise (mean ± SD). Every fourth week was used as a
‘tapering week’ with lighter loads used. Progressive resistance
training was used during mesocycle 1, which began
again at the beginning of mesocycle 2. The last four weeks
training was power training using loads of approx. 30–60%
1-RM.
M. Turpela et al. Experimental Gerontology 98 (2017) 13–21
17
Fig. 4. Changes in maximum strength and muscle mass for each group during the study. 1-RM load (top left), lean leg mass (top right), 1-RM normalized to body mass (bottom left),
maximum isometric knee extension torque (bottom right). In the bottom panels, each participant's data is marked by an × and the line indicates the group mean. Significant within-group
differences are marked with P < 0.05, significant between-group differences are marked above the data points with the number corresponding to the group to which it differs.
Fig. 5. Changes in time to complete all functional capacity tests from month 3 to 9. 7.5 m forward walk (top left), 7.5 m backward walk (top right), timed-up-and-go test (bottom left),
loaded 10-stair climb (bottom right). Each participant's data is marked by an × and the line indicates the group mean. Significant within-group differences are marked with P < 0.05,
significant between-group differences are marked above the control group's data points with the number corresponding to the experimental group to which it differs.
M. Turpela et al. Experimental Gerontology 98 (2017) 13–21
18
control: 122 ± 65 min·week− 1
; EX1: 69 ± 33, EX2: 70 ± 32, EX3:
60 ± 36, control: 70 ± 26% of total physical activity).
Results from the 3d diet diaries showed that there were no statistically
significant differences between any group for total carbohydrate
(EX1: 218 ± 60, EX2: 220 ± 81, EX3: 205 ± 74, control:
213 ± 61 g), protein (EX1: 88 ± 23, EX2: 89 ± 25, EX3: 82 ± 21,
control: 86 ± 35 g) or fat (EX1: 75 ± 24, EX2: 77 ± 27, EX3:
84 ± 20, control: 84 ± 43 g) intake or when these totals were normalized
to body mass (e.g. protein g·kg− 1
= EX1: 1.2 ± 0.3, EX2
1.2 ± 0.3, EX3: 1.1 ± 0.2, control: 1.2 ± 0.4).
3.6. Correlation analyses
All neuromuscular performance and functional capacity outcome
measure change scores were assessed to determine bivariate correlation.
The only variable-pair to demonstrate statistical significance was
the change in leg press 1-RM and the change in loaded 10-stair climb
from month 3 to 9 (r = −0.22, P = 0.036, n = 88, Fig. 6). There were
no relationships between habitual physical activity level and change in
neuromuscular performance or functional capacity.
4. Discussion
The main findings of the present study were that; 1) training frequency
influenced maximum 1-RM gain in a dose-dependent manner
but did not result in improved strength as measured by maximum isometric
unilateral knee extension or isometric seated plantarflexion
torque, 2) all experimental groups improved in some or all tests of
functional capacity, but higher training frequency did not provide additional
benefit, and 3) the present study did not observe any changes in
body composition or muscle mass nor observe differences in muscle
activity that might explain possible differences in 1-RM performance.
Taaffe et al. (1999) were the first to study the effects of training
frequency in older individuals (one versus two versus three times per
week). In all 8 exercises included in their study, equivalent improvements
were made in 1-RM using a program of 3 sets at 80% of 1-RM for
24 weeks. Furthermore, equivalent improvements were made in a chair
sit-to-stand test, while improvements in a 6 m backward walking test
(that approached statistical significance) also displayed no betweengroup
differences. The results of the present study are in contrast to
those of Taaffe et al. (1999) for 1-RM development but are similar regarding
functional capacity tests. One factor for the conflicting 1-RM
data may have been that the participants in the present study had undergone
a 3-month preparatory strength training period where 1-RM
had already increased approximately 12% (unpublished data), and so
this would reduce the potential for further improvement in EX1 and
EX2. Interestingly, reducing training frequency from two to one time
per week did not adversely affect the gain in 1-RM strength in older
women (Walker et al., 2017). When investigating maximum strength
only, DiFrancisco-Donoghue et al. (2007) did not observe different 1RM
gain when training with a frequency of one or two times per week
and, with the exception of the chest press exercise, Padhila et al. (2015)
did not observe differences in 1-RM gain with two or three times per
week. Therefore, it seems that there is potentially little difference between
training frequencies in the magnitude of maximum strength gain
over the first few months of initiating strength training in older people.
The only study that we are aware of to show clear differences between
training frequencies was a study by Farinatti et al. (2013). Using
a training-specific test (but not maximum strength test per se) a doseresponse
was observed for 10-RM improvement in the knee extension
and bicep curl exercises (3× > 2× > 1×) but not for the bench
press or calf raise. While TUG performance was not influenced by
training frequency, chair sit-to-stand and 2 m walking speed improved
more after two and three times compared to one time per week
(Farinatti et al., 2013). In comparison with the present study's results,
the (shorter distance) 2 m walking speed test was in contrast but the
TUG data were similar.
Overall, studies that have investigated the effect of training frequency
on strength improvement have produced conflicting results and
we are unable to determine a definitive conclusion. In our study at
least, men and women responded to the training in a similar manner,
since using sex as a covariate did not influence the findings. It may be
simply that the initiation of strength training itself in previously untrained
subjects is the main factor for improved functional capacity,
rather than any single program design variable. Also, it is interesting
that the magnitude of strength improvement seems not to be a major
factor and this should be investigated in more detail in future.
Consequently, the underlying adaptations leading to improved functional
capacity (i.e. walking ability and chair rise), as well as main
training program variable(s) to achieve these adaptations are yet to be
elicited.
One methodological consideration that may influence findings of
intervention studies is the choice of strength performance test. As
Buckner et al. (2016) noted, strength training for maximum lifts (i.e. 1RM)
comprises an element of (task-specific) skill. This may, in some
cases, falsely represent functionally relevant increases in strength. As
observed in the present study, there was a linear dose-response relationship
between training frequency and improved 1-RM performance
but this difference was diminished when testing isometrically. It should
be noted that, at least in the present study, training was conducted
bilaterally using (predominantly) weight-stack devices whereas the
isometric knee extension action was tested unilaterally. This non
training-specific test in itself would reduce the likelihood of observing
training-induced improvements (Abernethy and Jurimäe, 1996; Baker
et al., 1994). Similar observations have been made when training dynamically
and testing isometrically when investigating the effect of
training frequency on back extension in older adults (Graves et al.,
1990) and even when comparing low- to high-intensity strength
training (Mitchell et al., 2012; Van Roie et al., 2013). Nevertheless,
overall it could be proposed that the small-to-moderate differences in
improved 1-RM between training frequencies observed in various studies
may be a consequence of greater practice rather than ‘true’ strength
gain.
The exact cause of disparity in 1-RM improvement between groups
is difficult to discern. All experimental groups demonstrated an increase
in muscle activity as assessed by surface EMG amplitude. This measure
may or may not represent neural adaptation to strength training (for
review see Farina et al., 2014), but it was a systematic change in all
groups, whereas increased 1-RM performance was not. Correlation
analysis revealed that there was a weak but statistically significant relationship
between the change in 1-RM load and the change in EMG
amplitude (r = 0.23, P = 0.041, n = 78), which explained only ~5%
of the variance. Perhaps improvements in inter-muscle coordination
Fig. 6. Relationship between the change in 1-RM load (i.e. maximum dynamic strength)
and the change in loaded 10-stair climb time from month 3 to 9. Each participant's data is
marked by an ×.
M. Turpela et al. Experimental Gerontology 98 (2017) 13–21
19
during specific parts of the lift, which would be very challenging to
determine experimentally, may have led to improved 1-RM performance
and could be considered part of the improved task-specific ‘skill’
of 1-RM performance.
To the authors' knowledge, this is the first study to investigate the
influence of training frequency on muscle hypertrophy in healthy older
individuals. Whereas the present study's 3 months of preparatory
training induced gains in muscle mass (approx. 2% increase in lean leg
mass, P < 0.05, unpublished data), neither reduced or maintained or
increased training frequency led to further increases in muscle mass
during the subsequent 6-month intervention period of the present study
(Fig. 4). In young subjects, higher training volume using high- to
moderate-loads has been shown to be a key element of training for the
development of muscle hypertrophy (Campos et al., 2002; Ratamess
et al., 2009), but training frequency has been rarely examined directly,
and there appears to be no differences between a training frequency of
2 and 3 times per week in novice trainers (Wernbom et al., 2007). Given
that the training program in the first 3-month preparation period was
not optimized to increase maximum strength or muscle mass it was
expected that further increases would have been made during the
subsequent 6-month intervention. Certainly there is a large body of
evidence showing that initial gains in muscle mass continue over the
first year of training (e.g. Häkkinen et al., 2002; Pyka et al., 1994;
Taaffe et al., 1996). However, there are also rare observations where
older individuals do not gain muscle mass at all, such as the ~70 year
old men in Häkkinen et al. (1998).
One possible explanation for the lack of prolonged development in
muscle mass in the present study may be that older individuals are
thought to be “anabolically resistant” to the effect of strength training,
may not demonstrate hypertrophy to the extent of younger individuals
(Kraemer et al., 1999) and may require greater protein intake to maximize
hypertrophic potential (Moore et al., 2015). Unfortunately for the
development of muscle mass, Mero et al. (2013) showed that older
individuals habitually consumed less protein (and calories in general)
than young individuals, and this was also observed in the present study.
The self-reported protein intake of the participants in the present study
(range 0.6–1.8 g·kg−1
body mass) perhaps indicates a sub-optimal environment
for muscle hypertrophy despite the more optimized training
program for strength and hypertrophy. Consequently, with no control
of diet and/or supplementation during training to optimize protein
intake, the present study does not fully allow the effect of training
frequency on muscle hypertrophy in healthy older individuals to be
determined.
Since there was no improvement in voluntary activation level
during unilateral isometric knee extension and no increase in quadriceps
muscle mass, it is not surprising that maximum knee extension
torque did not increase during the present study. A lack of improved
voluntary activation level could be a result of using a non-specific test,
as discussed above, or it may indicate that no neural adaptation occurred
during the present study. This would seem to be supported by
the finding that surface EMG amplitude was no longer significantly
changed due to training once normalized to the maximum M-wave
amplitude. Alternatively, it may be that the use of single-pulse twitches
was not sensitive enough to determine training-induced changes
(Herbert and Gandevia, 1999). It would be recommended that more
detailed measurements assessing muscle activation would be included
in future studies.
Interestingly, relationships between increased strength, muscle activity/activation
or muscle mass and improved functional capacity were
not observed. In fact, the only statistically significant (negative) relationship
was observed between change in leg press 1-RM and change
in loaded 10-stair climb, and this would be classed as a weak relationship
with only 5% of the variance explained. In this regard, it
appears that performing strength training (regardless of training frequency)
is important to improve functional capacity even in healthy
older individuals, but the actual magnitude of strength gain has little or
no influence on improved maximum walking speed.
The decision to recommend strength training two-three times per
week may have originated from the meta-analysis of Rhea et al. (2003)
who were the first to investigate the effect of training frequency on
maximum strength development. However, due to the integration of
studies utilizing different participant groups, the authors cautioned that
“additional reviews are needed to verify the application of the dose-response
trends to those populations” (p. 458). Subsequently, a meta-analysis in
healthy adults over 55 years did not observe the same dose-response
relationship for maximum strength (Silva et al., 2014). Regarding
muscle hypertrophy, a recent meta-analysis did observe a benefit of
higher training frequency but this included studies with age-ranges
throughout the lifespan (Schoenfeld et al., 2016). It may well be that
the present number of intervention studies examining training frequency
does not allow meta-analysis techniques to accurately evaluate
its influence on strength, muscle mass, and functional capacity. For
example, 17 out of 21 treatments included in study by Silva et al.
(2014) comprised training of three times per week with majority of
studies using 2–3 times per week. This clearly does not allow valid
evaluation of a training frequency of one time per week, specifically
over the influence of other program variables, such as intensity, volume,
number of exercises per muscle group etc., on outcome measures.
Therefore, scientists do not currently have sufficient evidence to inform
policy makers as to recommendable training frequencies for young or
older individuals.
Finally, although habitual physical activity was tracked (daily)
throughout the present study, it was done so subjectively using diaries.
This method of course is non-blinded, i.e. the participants can see their
daily activity level and this may influence the results, and also is open
to typical errors of subjective reporting. Future studies may wish to
implement objective measures of tracking physical activity in order to
verify this finding. Nevertheless, since there were no differences between
the groups and that there were no significant relationships between
habitual physical activity and any performance outcome measure,
it appears that this potential confounding factor did not likely
influence the data. In other words, we can be confident that functional
capacity did not improve due to physical activity performed externally
to the prescribed intervention. Therefore, in general it seems that the
act of performing strength training, and not necessarily the increase in
strength or muscle mass, is important to improve functional capacity in
older individuals, and in order to prescribe individualized training
programs for older adults the precise underlying factor should be
identified.
5. Conclusions
The present study found no evidence that higher training frequency
would induce greater benefit to maximum walking speed, functional
capacity or muscle mass. There is a clear dose-response in dynamic
bilateral 1-RM strength but this was not observed during unilateral
isometric tests of maximum strength. It appears that in healthy older
individuals with some (albeit limited) experience in strength training
beneficial functional capacity improvements can be achieved through
low frequency training (i.e. 1–2 times per week). This study suggests
that a sufficient frequency for whole-body strength training to improve
strength, muscle mass and functional capacity are not one-in-the-same
and perhaps should be noted in physical activity guidelines for healthy
older individuals.
Acknowledgements
The authors would like to thank Mrs. Pirkko Puttonen for her assistance
in analysis and all of the dedicated participants and student
assistants that took part in this study. We also acknowledge the great
work by MSc. Jaakko Forssell during training, testing and analyses. This
work was funded by a personal grant to Dr. Simon Walker by the
M. Turpela et al. Experimental Gerontology 98 (2017) 13–21
20
Ministry of Education and Culture, Finland (OKM/56/626/2014).
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.exger.2017.08.013.
References
Abernethy, P.J., Jurimäe, J., 1996. Cross-sectional and longitudinal uses of isoinertial,
isometric, and isokinetic dynamometry. Med. Sci. Sports Exerc. 28, 1180–1187.
Ahtiainen, J.P., Hoffren, M., Hulmi, J.J., Pietikäinen, M., Mero, A.A., Avela, J., Häkkinen,
K., 2009. Panoramic ultrasound is a valid method to measure changes in skeletal
muscle cross-sectional area. Eur. J. Appl. Physiol. 108, 273–279.
Baker, D., Wilson, G., Carlyon, B., 1994. Generality versus specificity: a comparison of
dynamic and isometric measures of strength and speed-strength. Eur. J. Appl. Physiol.
68, 350–355.
Buckner, S.L., Jessee, M.B., Mattocks, K.T., Mouser, J.G., Counts, B.R., Dankel, S.J.,
Loenneke, J.P., 2016. Determining strength: a case for multiple methods of measurement.
Sports Med. http://dx.doi.org/10.1007/s40279-016-0580-3.
Campos, G.E., Luecke, T.J., Wendeln, H.K., Toma, K., Hagerman, F.C., Murray, T.F., Ragg,
K.E., Ratamess, N.A., Kraemer, W.J., Staron, R.S., 2002. Muscular adaptations in
response to three different resistance-training regimens: specificity of repetition
maximum training zones. Eur. J. Appl. Physiol. 88, 50–60.
DiFrancisco-Donoghue, J., Werner, W., Douris, P.C., 2007. Comparison of once-weekly
and twice-weekly strength training in older adults. Br. J. Sports Med. 41, 19–22.
Farina, D., Merletti, R., Enoka, R.M., 2014. The extraction of neural strategies from the
surface EMG: an update. J. Appl. Physiol. 117, 1215–1230.
Farinatti, P.T.V., Geraldes, A.A.R., Bottaro, M.F., Lima, M.V.I.C., Albuquerque, R.B.,
Fleck, S.J., 2013. Effects of different resistance training frequencies on the muscle
strength and functional performance of active women older than 60 years. J. Strength
Cond. Res. 27, 2225–2234.
Graves, J.E., Pollock, M.L., Foster, D., Leggett, S.H., Carpenter, D.M., Vuoso, R., Jones, A.,
1990. Effect of training frequency and specificity on isometric lumbar extension
strength. Spine 15, 504–509.
Häkkinen, K., Kallinen, M., Izquierdo, M., Jokelainen, K., Lassila, H., Mälkiä, E., Kraemer,
W.J., Newton, R.U., Alen, M., 1998. Changes in agonist-antagonist EMG, muscle CSA,
and force during strength training in middle-aged and older people. J. Appl. Physiol.
84, 1341–1349.
Häkkinen, K., Kramer, W.J., Pakarinen, A., Triplett-McBride, T., McBride, J.M., Häkkinen,
A., Alen, M., McGuigan, M.R., Bronks, R., Newton, R.U., 2002. Effects of heavy resistance/power
training on maximal strength, muscle morphology, and hormonal
response patterns in 60-75-year-old men and women. Can. J. Appl. Physiol. 27,
213–231.
Herbert, R.D., Gandevia, S.C., 1999. Twitch interpolation in human muscles: mechanisms
and implications for measurement of voluntary activation. J. Neurophysiol. 82,
2271–2283.
Kraemer, W.J., Ratamess, N.A., 2004. Fundamentals of resistance training: progression
and exercise prescription. Med. Sci. Sports Exerc. 36, 674–688.
Kraemer, W.J., Häkkinen, K., Newton, R.U., Nindl, B.C., Volek, J.S., McCormick, M.,
Gotshalk, L.A., Gordon, S.E., Fleck, S.J., Campbell, W.W., Putukian, M., Evans, W.J.,
1999. Effects of heavy-resistance training on hormonal response patterns in younger
vs. older men. J. Appl. Physiol. 87, 982–992.
Mero, A.A., Hulmi, J.J., Salmijärvi, H., Katajavuori, M., Haverinen, M., Holviala, J.,
Ridanpää, T., Häkkinen, K., Kovanen, V., Ahtiainen, J.P., Selänne, H., 2013.
Resistance training induced increase in muscle fiber size in young and older men. Eur.
J. Appl. Physiol. 113, 641–650.
Merton, P.A., 1954. Voluntary strength and fatigue. J. Physiol. 29, 553–564.
Mitchell, C.J., Churchward-Venne, T.A., West, D.W., Burd, N.A., Breen, L., Baker, S.K.,
Phillips, S.M., 2012. Resistance exercise load does not determine training-mediated
hypertrophic gains in young men. J. Appl. Physiol. 113, 71–77.
Moore, D.R., Churchward-Venne, T.A., Witard, O., Breen, L., Burd, N.A., Tipton, K.D.,
Phillips, S.M., 2015. Protein ingestion to stimulate myofibrillar protein synthesis
requires greater relative protein intakes in healthy older versus younger men. J.
Gerontology Med. Sci. 70, 57–62.
Moss, B.M., Refnes, P.E., Abildgaard, A., Nicolaysen, K., Jensen, J., 1997. Effects of
maximal effort strength training with different loads on dynamic strength, crosssectional
area, load-power and load-velocity relationships. Eur. J. Appl. Physiol. 75,
193–199.
Padhila, C.S., Ribeiro, A.S., Fleck, S.J., Nascimento, M.A., Pina, F.L.C., Okino, A.M.,
Venturini, D., Barbosa, D.S., Mayhew, J.L., Cyrino, E.S., 2015. Effect of resistance
training with different frequencies and detraining on muscular strength and oxidative
stress biomarkers in older women. Age 37, 104. http://dx.doi.org/10.1007/s11357-
015-9841-6.
Pyka, G., Lindenberger, E., Charette, S., Marcus, R., 1994. Muscle strength and fiber
adaptations to a year-long resistance training program in elderly men and women. J.
Gerontol. 49, M22–M27.
Ratamess, N.A., Alvar, B.A., Evetoch, T.K., Housh, T.J., Kibler, B.W., Kraemer, W.J.,
Triplett, T.N., 2009. American College of Sports Medicine position stand. Progression
models in resistance training for healthy adults. Med. Sci. Sports Exerc. 41, 687–708.
Rhea, M.R., Alvar, B.A., Burkett, L.N., Ball, S.D., 2003. A meta-analysis to determine the
dose response for strength development. Med. Sci. Sports Exerc. 35, 456–464.
Rosenberg, J.G., Ryan, E.D., Sobolewski, E.J., Scharville, M.J., Thompson, B.J., King,
G.E., 2014. Reliability of panoramic ultrasound imaging to simultaneously examine
muscle size and quality of medial gastrocnemius. Muscle Nerve 49, 736–740.
Schoenfeld, B.J., Ogborn, D., Krieger, J.W., 2016. Effects of resistance training frequency
on measures of muscle hypertrophy: a systematic review and meta-analysis. Sports
Med. 46, 1689–1697.
Silva, N.L., Oliveira, R.B., Fleck, S.J., Leon, A.C.M.P., Farinatti, P., 2014. Influence of
strength training variables on strength gains in adults over 55 years-old: a metaanalysis
of dose-response relationships. J. Sci. Med. Sport 17, 337–344.
Steib, S., Schoene, D., Pfeifer, K., 2010. Dose-response relationship of resistance training
in older adults: a meta-analysis. Med. Sci. Sports Exerc. 42, 902–914.
Taaffe, D.R., Pruitt, L., Pyka, G., Guido, D., Marcus, R., 1996. Comparative effects of highand
low-intensity resistance training on thigh muscle strength, fiber area, and tissue
composition in elderly women. Clin. Physiol. 16, 381–392.
Taaffe, D.R., Duret, C., Wheeler, S., Marcus, R., 1999. Once-weekly resistance exercise
improves muscle strength and neuromuscular performance in older adults. J. Am.
Geriatr. Soc. 47, 1208–1214.
Unhjem, R., Lundestad, R., Fimland, M.S., Mosti, M.P., Wang, E., 2015. Strength traininginduced
responses in older adults: attenuation of descending neural drive with age.
Age 37, 9784. http://dx.doi.org/10.1007/s11357-015-9784-y.
Van Roie, E., Delecluse, C., Coudyzer, W., Boonen, S., Bautmans, I., 2013. Strength
training at high versus low external resistance in older adults: effects on muscle
volume, muscle strength, and force-velocity characteristics. Exp. Gerontol. 48,
1351–1361.
Walker, S., Peltonen, H., Sautel, J., Scaramella, C., Kraemer, W.J., Avela, J., Häkkinen, K.,
2014. Neuromuscular adaptations to constant vs. variable resistance training in older
men. Int. J. Sports Med. 35, 69–74.
Walker, S., Peltonen, H., Häkkinen, K., 2015. Medium-intensity, high volume “hypertrophic”
resistance training did not induce improvements in rapid force production in
healthy older men. Age 37, 9786. http://dx.doi.org/10.1007/s11357-015-9786-9.
Walker, S., Serrano, J., Van Roie, E., 2017. Maximum dynamic lower-limb strength was
maintained during 24 weeks reduced training frequency in previously sedentary
older women. J. Strength Cond. Res published online. http://dx.doi.org/10.1519/
JSC.0000000000001930.
Waller, B., Munukka, M., Multanen, J., Rantalainen, T., Pöyhönen, T., Nieminen, M.T.,
Kiviranta, I., Kautiainen, H., Selänne, H., Dekker, J., Sipilä, S., Kujala, U.M.,
Häkkinen, A., Heinonen, A., 2013. Effects of a progressive aquatic resistance exercise
program on the biochemical composition and morphology of cartilage in women with
mild knee osteoarthritis: protocol for a randomized controlled trial. BMC
Musculoskelet. Disord. 14, 82.
Wernbom, M., Augustsson, J., Thomee, R., 2007. The influence of frequency, intensity,
volume and mode of strength training on whole muscle cross-sectional area in humans.
Sports Med. 37, 225–264.
World Health Organization, 2010. Global Recommendations on Physical Activity for
Health. (Geneva, Switzerland. ISBN 978 92 4 159 997 9).
M. Turpela et al. Experimental Gerontology 98 (2017) 13–21
21