Maroto Izquierdo Meta Analysis 2017
Maroto Izquierdo Meta Analysis 2017
Maroto Izquierdo Meta Analysis 2017
Review
a r t i c l e i n f o a b s t r a c t
Article history: Objectives: The purpose of this meta-analysis was to examine the effect of flywheel (FW) resistance
Received 19 June 2016 training with Eccentric Overload (FW-EOT) on muscle size and functional capacities (i.e. strength and
Received in revised form 24 January 2017 power) in athletes and healthy subjects, and to compare FW-induced adaptations with those triggered
Accepted 13 March 2017
by traditional resistance exercise interventions.
Available online 21 March 2017
Design: A systematic review and meta-analysis of randomised controlled trials.
Methods: A search of electronic databases [PubMed, MEDLINE (SportDiscus), Web of Science, Scopus and
Keywords:
PEDro] was conducted to identify all publications employing FW-EOT up to April 30, 2016. Outcomes
Iso-inertial
Muscle power
were analyzed as continuous outcomes using a random effects model to calculate a standardized mean
Muscle strength difference (SMD) and 95% CI. A total of 9 studies with 276 subjects and 92 effect sizes met the inclusion
Hypertrophy criteria and were included in the statistical analyses.
Results: The overall pooled estimate from the main effects analysis was 0.63 (95% CI 0.49–0.76) with a
significant (p < 0.001) Z overall effect of 9.17. No significant heterogeneity (p value = 0.78) was found. The
meta-analysis showed significant differences between FW-EOT vs. conventional resistance training in
concentric and eccentric strength, muscle power, muscle hypertrophy, vertical jump height and running
speed, favoring FW-EOT.
Conclusions: This meta-analysis provides evidence supporting the superiority of FW-EOT, compared with
traditional weight-stack exercise, to promote skeletal muscle adaptations in terms of strength, power
and size in healthy subjects and athletes.
© 2017 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
1. Introduction lier increments in muscle mass when compared with CON13,14 and
improved jumping performance.10,15 Thereby, ECC actions seem to
Eccentric (ECC) training has been extensively studied in the sci- optimize the efficacy of training.16,17
entific literature.1 In comparison with concentric (CON) exercise, The ability to produce force in the CON phase limits the
isolated ECC actions are characterized by producing higher peaks of load/weight to be used during training. As a result, and given
force2 with lower muscle activation3,4 and metabolic cost,5 as well the higher force production capacity of skeletal muscle during
as higher solicitation of Type IIx fibers,6 increased cross-education ECC actions, the loads used during traditional free weights or
effect7 and greater cortical activity.8 Furthermore, despite produc- weight-stack exercise are sub-optimal during the ECC phase of the
ing high levels of muscle damage and soreness after the initial movement.2,18 However, optimization of resistance training using
bout,9 ECC-based resistance exercise training has been associated a strictly ECC regime is rather complex and technically difficult
with effective muscle damage prevention mechanisms,10–12 ear- to apply.19 In addition, ECC actions are rarely isolated in real-
life situations, and usually appear during the stretch-shortening
cycle, inducing a greater contribution of the elastic components
in the muscle-tendon unit; the stretch-shortening cycle increases
∗ Corresponding author. the potential to produce force in the subsequent CON action due
E-mail address: smaroi@unileon.es (S. Maroto-Izquierdo).
http://dx.doi.org/10.1016/j.jsams.2017.03.004
1440-2440/© 2017 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
944 S. Maroto-Izquierdo et al. / Journal of Science and Medicine in Sport 20 (2017) 943–951
to increased storage and use of elastic energy.17 Several methods and structural muscle adaptations among athletes and healthy sub-
have been designed and proposed to offer an eccentric overload jects, and to perform a quantitatively comparison of FW-EOT vs.
(EO) during resistance training, such as (1) controlling and adjust- traditional resistance exercise training in inducing gains in muscle
ing the time/velocity of CON and ECC movement during resistance mass, force and power.
training20,21 ; (2) using third-party assistance or devices for mov-
ing/rising the load during the CON phase10–12 ; and (3) employing
2. Methods
isokinetic dynamometers.22
Devices created to isolate or overload the ECC phase of the mus-
This systematic review was designed following the guidelines
cle action have emerged as an alternative method that may produce
of the Preferred Reporting Items for Systematic Reviews and
greater muscle adaptations, and therefore have been developed
Meta-Analyses Protocols (PRISMA-P).46,47 The PRISMA-P statement
and/or tested for rehabilitation and performance purposes. The
includes 26 points, grouped in 17 kinds of items checklist and it is
iso-inertial devices were originally designed by Berg & Tesch23
designed to be used as a basis for reporting systematic review of
in 1994 to counteract the deleterious effect of microgravity on
randomised trials. A review protocol was not registered for this
skeletal muscle. Such technology presents one of the most-used
review.
exercise paradigms to produce EO during resistance training. The
A systematic, computerized search of the literature in PubMed,
iso-inertial devices more frequently employed are “The Flywheel
Web of Science (including Web of Science and MEDLINE results),
Exercise Device”,23,24 “VersaPulley”,25 and “Inertial Training and
Scopus, SportDiscus and PEDro was conducted by an indepen-
Measurement System”.26 These iso-inertial devices use the fly-
dent researcher with controlled vocabulary and keywords related
wheel (FW) principle to produce unlimited resistance during the
to eccentric training, eccentric overload training and flywheel
entire range of motion. During the CON phase the force applied
training. Our search time frame was restricted to 22 years (1
unwinds a cord/strap connected to the shaft with the FW, which
January 1994 to 30 April 2016); 1994 was chosen because research
starts to rotate and store energy. Kinetic energy will increase as a
on FW technology began that year.23 We developed our search
function of the rotational speed. Once the CON action is completed,
strategy based on the lack of reviews and meta-analysis about FW-
the cord/strap rewinds and the trainee must resist the pull of the
EOT.1,48 To do this, the search strategy used by previous reviews
FW by performing a braking, ECC muscle action. By using appro-
in the field of resistance training was employed.48 The search
priate technique, i.e. resisting the inertial force gently during the
language was restricted to English, and a filter containing Medi-
first third of the ECC action, and then applying maximal effort to
cal Subject Headings (MeSH) terms was applied. First, a general
stop the movement at the end of the range of motion, EO can be
search including the terms “eccentric training”, “eccentric exer-
produced in force/power values.24,27 Then, the next CON action is
cise”, “negative work” was performed. A more specific search
immediately initiated.
included the terms of “eccentric-overload”, “eccentric-overload
The effects of inertial training using FW devices have been
training”, “inertial training”, “inertial exercise”, “isoinertial train-
extensively investigated over the past 20 years. The majority of
ing”, “flywheel”, “flywheel resistance”, “gravity-independent” and
studies assessed the effects of eccentric overload training (EOT)
“enhanced-eccentric”. These terms were chosen because they have
on lower body muscle mass in healthy and active subjects. These
been traditionally used to describe this training methodology. The
studies employed a mean workload of 4 sets of 7 maximum rep-
results of this specific search were then combined with the follow-
etitions during 5–15 weeks. Results indicate that EOT employing
ing terms: “device”, “strength”, “force”, “power”, “hypertrophy” and
FW technology induce gains of 5–13% in muscle mass,28–31 11–39%
“muscle mass”.
in maximal voluntary contraction,28–30 12–25% in 1 repetition
The reference list of all selected publications was verified to
maximum (1 RM),31,32 21–90% in ECC force,19,23,33 10–33% in mus-
retrieve relevant publications that were not identified by the com-
cle power,32,34 6–15% in jump ability,32,34,35 2–10% in running
puterized search. References of selected and included original
speed32,35,36 and up to 35% in electromyography activity.29,37,38
articles, abstracts and available conference proceedings were also
Despite EOT being associated with a high magnitude of mus-
searched, including publications, posters, abstracts or conference
cle damage and inflammatory responses during the initial phase
proceedings. To identify relevant articles, titles and abstracts of all
of training, a significant attenuation of these processes occur
selected publication after the first search were analyzed looking
shortly after,31 indicating no counterproductive effects on mus-
for training methods where the ECC phase was overloaded or rein-
cle. Additionally, these devices have been shown very effective in
forced. In the specific search, in addition to the identified citations
counteracting the decrements in muscle mass and strength during
of the first search, titles and abstracts of all recognized publications
weightlessness,39–41 as well as improving muscle function, bal-
were examined in detail. Full-text papers were recovered if the
ance, gait and functional performance in elderly42,43 and stroke
abstract provided insufficient information to establish eligibility or
patients.44,45
if the article abstract had passed the first eligibility review.
While there has been an increase use of these technology and
All articles examining FW-EOT were eligible for full-text review.
a significant amount of research comparing the effectiveness of
An article was eligible for study inclusion if it met all of the following
eccentric overload flywheel resistance training (FW-EOT) with tra-
criteria:
ditional weight-stack exercise programs, there is no systematic
review that summarizes the results of such studies, and adequately
assess their scientific rigor. When confronting with a task of this 1. The original article was a randomised controlled trial (RCT) or
magnitude, there are inherent methodological limitations, such as clinical controlled trial (CCT) published in peer-reviewed jour-
the difficulty to isolate or enhance the ECC phase from the CON,1 nals.
and the diversity of devices employed to generate EO. To circum- 2. The article reported on athletes or physical active subjects of
vent these limitations, we performed a systematic review including either sex who had completed an EOT protocol during at least 4
only training studies using FW technology to generate the EO weeks with a minimum training frequency of 2 days per week.
vs. traditional free-weight or weight-stack training. Therefore, the 3. The study described healthy subjects without a history of injury
purpose of the current review and meta-analysis was to systemat- in the trained limb.
ically review the literature on randomised clinical trials examining 4. The manuscript included a FW-EOT intervention and a control or
the effects of FW-EOT, and how the intervention affects functional alternative intervention group, comparing training adaptations
in strength and/or power, and/or muscle mass.
S. Maroto-Izquierdo et al. / Journal of Science and Medicine in Sport 20 (2017) 943–951 945
An article was excluded if: standard deviations and number of participants allocated in each
group were not available.
1. Had subjects with any pathology or included subjects with exist-
ing, or under treatment for, musculoskeletal injuries in the
3. Results
trained limb.
2. Did not have the minimum requirements regarding in the train-
Fig. 1 shows a flow chart with the different phases of the
ing protocol (e.g, duration or frequency).
search and selection of the studies included in the review. The ini-
3. Reports focused on elderly above the age of 60 years.
tial search of electronic databases identified 264 titles, of which
4. Were not written in English.
171 were rejected for duplication issues. Four titles/articles32,50–52
were identified through manual search. Thus, 97 titles were iden-
All criteria were independently applied by two reviewers to the tified, but 48 of them were rejected because they did not meet the
full text of the articles that passed the eligibility screening of titles intervention criteria: 20 studies of bed rest, descriptive studies,
and abstracts. Any disagreement was resolved by discussion. 6 studies of elderly, 5 studies with chronic patients, 2 studies of
Two investigators independently performed quality assess- post-activation potentiation, and 3 studies with other intervention
ments of the included studies, and disagreements were resolved methodology. From a total of 49 abstracts that were screened, 24
during a consensus meeting by a third part. The methodological were excluded because they had insufficient requirements regard-
quality of individual studies was assessed using the Physiotherapy ing the training protocol, and 4 studies were excluded for lack of
Evidence Database (PEDro) scale.49 Results from individual study comparison between groups. Twenty-one full texts were reviewed,
analysis of quality were used to identify common areas of method- but only 9 studies satisfied the inclusion criteria to be considered
ological weaknesses across studies. PEDro uses 11 criteria, and for this review.26,28,30,32,34–36,52,53 The main reasons for exclusion
reviewed studies were awarded one point for each criterion that were: lack of control group (n = 4), lack of information (n = 2), com-
was clearly satisfied, for a potential maximum value of 11 points. bination FW-EOT with other training methodology (n = 2), and
Criteria included: different mechanism to produce EO (n = 4).
The mean methodological quality of the studies in PEDro scale
was 7 ± 0.5, out of 11, with scores ranging from 6 to 8. All studies
1. Eligibility criteria reported.
were categorized as high quality. The most common flaws were
2. Random assignment.
the lack of blinding of participants, therapists and assessors. It
3. Concealed allocation.
should be considered however, that blinding of participants in these
4. Groups similar at baseline regarding most important prognostic
studies is a difficult requisite to satisfy. Inter-rater reliability was
indicator.
significantly high (ICC = 0.98).
5. Blinding of participants.
The main characteristics of the studies included in the review
6. Blinding of therapists who administered the therapy.
regarding participants, interventions and results are illustrated in
7. Blinding of assessors who measured key outcome.
Table 1. After adjusting for dropouts, the total number of partici-
8. Measures of at least one key outcome were obtained from more
pants in the studies included was 276. Of these 276 participants,
than 85% of initial participants.
165 performed FW-EOT. The rest (n = 111) served as controls or
9. All participants received treatment or control condition as allo-
performed resistance training without EO. Demographic data were
cated.
provided for all studies; the estimated mean age of the experimen-
10. Results of between-group statistical comparisons are reported.
tal and control groups were 25.8 ± 8.5 and 23.2 ± 10.6, respectively.
11. Study provides point measures and measures of variability for
The distribution of sex among studies was not proportional, with
at least one key outcome.
only one study including female subjects,28 with a total of 3 women
and 273 men in both the experimental and control groups.
Although PEDro does not provide specific instructions to classify Training interventions ranged from 4 to 10 weeks with a mean
studies according to the score obtained, we have used the criteria frequency of 2.33 ± 0.72 sessions per week, with a mean total of
established by others.1,16 Thus, a study was considered of high qual- 14 ± 2 sessions per study. The total number of sets (3–6) and rep-
ity when the score was greater than 5, of moderate quality when etitions (6–8) per session differed across studies. Inertial Training
the score was 5 or 4, and of low quality when the study was scored and Measurement System (ITMS) studies used a training load of
3 or less. 15–20 s per set (between 8 and 11 repetitions).26,34,52 All studies
The meta-analysis was conducted using to RevMan 5.3 (The employed FW devices to produce EO. Exercise devices employed
Nordic Cochrane Centre, The Cochrane Collaboration, Copen- were FW leg press,32 YoYo squat,35 Multi-Gym FW,31 FW supine
hagen 2012. Free to download at http://tech.cochrane.org/revman/ leg curl35,36 and FW prone leg extension device.28,30,50,53,54 Also,
download) to determine the efficacy of FW-EOT in increasing three studies employed ITMS.26,34,52 The muscle groups trained
strength, power and muscle mass, as well as jump ability and were knee extensors,28,30,34,50,53,54 knee flexors36 or both.31,32,35
speed. Data were pooled in different subgroups according to five Also, two studies of upper limb targeting shoulder abductors26
physical variables: Strength (CON, ECC or both), muscle power and elbow flexors and extensors52 were included. In the flywheel
(isotonic or isokinetic conditions only in CON movements), hyper- devices, the workload is provided by the inertia of a rotating mass,
trophy (increases in muscle size, volume or thickness in muscle which in turns depends on its geometrical and physical proper-
groups subjected to training), jump ability (CMJ height) and speed. ties, and may vary depending on the total number of flywheels
Outcomes were analysed as continuous outcomes using a random installed.28,50 The most common moment of inertial employed was
effects model to calculate a standardized mean difference (SMD) between 0.07 and 0.145 kg m2 .28,30,32,53 Given the fact that muscles
and 95% CI. A p value less than 0.05 indicated a statistical signif- are capable of achieving higher absolute forces when contracting
icance for an overall effect, and a p value less than 0.1 indicated eccentrically as compared with concentrically,1 it is important to
statistical significance for heterogeneity between studies. When quantify the intensity of the training during FW exercise. In most of
the articles selected did not provide sufficient data for the anal- the studies28,30,32,35,36,53 subjects were asked to push with maximal
ysis, authors were contacted to obtain relevant data. Studies were effort through the entire CON action. At the end of this CON action,
not included in the meta-analyses if summary statistics of means, the FW strap winds back due to inertial forces, which initiate the
946 S. Maroto-Izquierdo et al. / Journal of Science and Medicine in Sport 20 (2017) 943–951
Fig. 1. Flow chart illustrating the different phases of the search and selection of the studies included in the review (PRISMA).
reversed ECC action. During the first third of the ECC action, subjects hypertrophy (SMD 0.57; 95% CI 0.25–0.9), vertical jump perfor-
were instructed to resist gently, and thereafter to apply maximal mance (SMD 0.46; 95% CI 0.09–0.83) and running speed (SMD 0.41;
breaking force to stop the movement. Then, the next CON action 95% CI 0.0–0.82) (Fig. 2).
was instantly initiated. In these studies, CON and ECC power during
training was assessed with an encoder or by means of rate per-
4. Discussion
ceived exertion.32 Some studies adjusted the inertia to the higher
CON power output.35 Only one study failed to use any methodology
This is the first systematic review that analyses the efficacy
to quantify the load during training.36
of eccentric overload flywheel resistance training as a method to
Individual main effect analysis, overall pooled estimate and
improve strength, muscle power and muscle mass. In addition, the
measures of heterogeneity are illustrated in Fig. 2. This figure
statistical approach employed, allows comparing the efficacy of
depicts SMD and 95% CI of the included studies for strength, mus-
FW-EOT vs. traditional resistance exercise training in eliciting func-
cle power, muscle hypertrophy, jump height and running speed
tional and structural muscle adaptations in athletes and healthy
adaptations following EOT. The overall pooled estimate from the
subjects. The search performed yielded 9 studies that met the
main effects analysis was 0.63 (95% CI 0.49–0.76) with a signifi-
inclusion criteria.26,28,30,32,34–36,52,53 Using data from these sources,
cant (p < 0.001) Z overall effect of 9.17. No significant heterogeneity
strong effects towards higher gains after FW training compared
(heterogeneity p value = 0.78) was found.
with other methodologies, in terms of force, muscle power and
Meta-analyses of subgroups demonstrated significance dif-
hypertrophy, as well as jump ability and speed were found. The
ferences in training-induced adaptations favouring FW-EOT vs.
heterogeneity among studies in the subgroup analysis reported in
control group in both CON and ECC strength (SMD 0.66; 95% CI
the current investigation was mainly caused by the different meth-
0.44–0.89), muscle power (SMD 0.8; 95% CI 0.53–1.07), muscle
ods of evaluation employed in each investigation. Interestingly, the
S. Maroto-Izquierdo et al. / Journal of Science and Medicine in Sport 20 (2017) 943–951 947
Table 1
Characteristics of the studies included.
de Hoyo et al.35 33 junior soccer Knee extensors and Exp (n = 18): Half-squat and leg Significant increases in CMJ and
players (under 17 to flexors curl FW device running speed in exp. No
under 19) 10 weeks (15 sessions) differences in control group
3–5 sets of 6 reps/session
Maroto-Izquierdo et al.32 29 professional Knee extensors and Exp (n = 15): Leg press FW device Exp showed significant increases in
handball players flexors 6 weeks (15 sessions) 1 RM, muscle power in all % of 1
4 sets of 7 reps/session RM, SJ and CMJ, running speed,
agility T-test and muscle mass.
Control showed smaller training
effect in 1 RM, SJ and running
speed
Naczk et al.34 58 male physical Knee extensors 2 training groups: T0 (no Significant increases in muscle
education students additional load) and T10 (10 kg force, muscle power, CMJ, SJ and
additional load). 2 control groups muscle mass
ITMS No significant differences between
5 weeks (15 sessions) training groups
3 sets of 15 s/set
Naczk et al.26 46 male physical Shoulder abductors Exp (n = 33) divided in 3 groups: T0 All groups showed significant
education students (no additional load), T5 (5 kg training effects in torque and
additional load) and T10 (10 kg power
additional load) No significant differences between
ITMS training groups
4 weeks (12 sessions)
3 sets of 20 s/set
Naczk et al.52 38 male physical Elbow flexors and Exp (n = 20) divided in 2 groups: TF Greater improvements in elbow
education students extensors (7.5 rad s−1 ) and TS (5.76 rad s−1 ) flexor force and power in TF than
ITMS in TS
5 weeks (15 sessions) Elbow extensor force and power
3 sets of 15 s/set. Flexion and increased significantly only in TF
extension, right and left arms
worked separately
Norrbrand et al.30 15 men (mean age 39) Knee extensors 5 weeks (12 sessions) QF volume improved significantly
Exp (n = 7). Seated knee extension in both groups, greater gains in exp
FW device
4 sets of 7 reps/session
Control (n = 8). Standard seated
weight stack machine
4 sets of 7 RM/session
Norrbrand et al.53 17 men (mean age 39) Knee extensors Same protocol than Norrbrand Both groups improved significantly
et al.30 MVC and strength. Higher EMG
activity in ECC in exp
Tesch et al.28 10 middle-age Knee extensors Unilateral (left) leg extension CON and ECC strength, maximal
(30–53 year) men and exercise in a seated knee extension isometric strength and QF volume
women FW device increased significantly
5 weeks (12 sessions)
4 sets of 7 reps/session
% 1 RM, percentage repetition maximum; AE, aerobic exercise; CMJ, countermovement jump; CSA, cross-sectional area; con, concentric; ecc, eccentric; Ext, experimental
group; MVC, maximal voluntary contraction; QF, quadriceps; RE, resistance exercise; RF, rectus femoris; SJ, squat jump; TF, faster training group; TS, slower training group;
VL, vastus lateralis; VI, vastus intermedius.
a
Number of participants at the end of the studies.
b
Only results for the outcomes of interest are provided.
only studies assessing FW-EOT adaptations in the upper limbs were gravity-independent devices (i.e. FW), show robust increments in
the ones reporting the largest effect sizes (ES between 0.52 and force production.27,31,33,35 The current systematic review confirm
1.75) of all investigations analyzed.26,52 these data, showing that resistance training offering brief episodes
During traditional CON-ECC resistance exercise performed at of EO employing FW devices, is associated with significantly greater
maximal intensity, the ECC phase is clearly under-loaded (e.g. about improvements in maximal dynamic strength when compared to
40–50%).55 Multiple studies using different protocols have shown traditional resistance exercise programs. These results are sup-
the critical role of the ECC muscle action to improve contractile ported by Fernandez-Gonzalo et al.,31 who reported increases
characteristics and muscle size in humans.3,55,56 Therefore, it seems of 25% in maximal dynamic strength in both men (ES = 1.78)
reasonable to believe that applying eccentric overload during resis- and women (ES = 0.62). Furthermore, those studies assessing both
tance exercise could increase force production capability.19 Indeed, CON and ECC force, showed larger increments in ECC than CON
studies employing eccentric overload training by means of inertial force.36,51 Also, Martinez-Aranda and Fernandez-Gonzalo have
948 S. Maroto-Izquierdo et al. / Journal of Science and Medicine in Sport 20 (2017) 943–951
Fig. 2. Forest plot with meta-analysis of standardized mean difference showing comparison of eccentric overload training versus control/weight training on strength (N),
muscle power (W), muscle hypertrophy (cm3 , mm or kg), jump height (cm) and running speed (s).
recently shown that the EO produced in terms of force/torque dur- muscle atrophy and neuromuscular deconditioning during unload-
ing FW exercise (% above CON) is ∼25% for both men and women.57 ing has been analyzed in multiple occasions during the last 15
This EO seems to be maintained over the duration of the training years, eliminating or greatly reducing loss in muscle mass dur-
period. In addition, anecdotal data seem to indicate that the use of ing long term bed rest studies,39–41,58 and even increasing muscle
lower inertia26,34 or high movement velocity52 during FW training volume during unilateral lower limb suspension.41 The results
may induce greater force gains. However, this hypothesis needs to of the current meta-analyses indicate that not only FW-EOT is
be tested in future investigations. effective in increasing muscle mass, but also that is more potent
Originally, FW devices were created to counteract the mus- than traditional resistance training in inducing hypertrophy adap-
cle mass and strength losses during space-like, microgravity tations (p < 0.001). The high variability in the effect size of the
situations.39–41,58,59 Thus, the efficacy of FW training to reduce the different studies in terms of muscle hypertrophy after EOT employ-
S. Maroto-Izquierdo et al. / Journal of Science and Medicine in Sport 20 (2017) 943–951 949
ing FW vs. traditional resistance exercise training could be partly in this variable showed the lowest significance of all outcomes
explained by the different methods used to evaluate changes in analysed (p < 0.05), FW-EOT training appeared more effective to
muscle mass. While some studies used bioelectrical impedance34,52 improve running speed than traditional resistance training32,35,36
or ultrasonography,32 only 2 studies employed magnetic resonance (ES between 0.3 and 1.45 in FW-EOT groups and −0.08 and 0.55
imaging,28,30 which is considered the “gold-standard” for muscle in control groups). These results are supported by Tous-Fajardo
mass assessment.1 Thus, even though all studies showed benefits et al.,64 not included in the meta-analysis, who showed greater
on muscle mass, magnetic resonance imaging studies were those improvements in soccer-specific performance tasks such as change
showing the lower effect (ES = 0.23–0.34). In addition to the stud- of direction and linear running speed, after a combined FW-EOT and
ies included in the meta-analysis, two additional investigations29,31 vibration training protocol vs. conventional training (plyometric,
reported significant gains in muscle mass in young subjects after speed and weight-loaded training). Adding to this, from a molecu-
FW-EOT employing magnetic resonance imaging and DXA, respec- lar perspective, it seems EOT favours increments of mRNA levels of
tively. Interestingly, it seems that adding aerobic exercise 6 h50 or genes preferentially expressed in fast glycolytic fibers, ultimately
15 min51 prior to FW-EOT elicits even greater muscle hypertrophy inducing a faster muscle phenotype. Such changes may help the
(ES = 0.57 and 0.27, respectively) than FW-EOT alone (ES = 0.31 and muscle to become better suited for explosive, high-speed actions.65
0.14, respectively). Finally, the largest effect of FW-EOT training in
muscle mass were found in a study describing FW-induced adap-
tations in well-trained athletes (ES = 0.6 in FW-EOT group versus 5. Conclusion
ES = 0.38 in weight training group).32 Indeed, the EO produced dur-
ing FW training is generally greater in athletes with experience in Although free weights and weight-stack machines are the most
FW training, highlighting the importance of an appropriate tech- popular modes of resistance training, the results of this systematic
nique to maximize the benefits of this training paradigm.27 In review indicate that brief episodes of eccentric overload induced
addition, a proper technique during FW training may impact the by flywheel devices, and performed at high intensity are associated
working angle, where the main gains in strength and muscle mass with greater improvements in both concentric and eccentric force,
occur.60,61 This is an important factor to consider in future studies muscle power and muscle hypertrophy in healthy and well-trained
because muscle mass gains are greater when the ROM employed subjects. In addition, eccentric overload training appeared to be
during training is longer.62 more effective than traditional resistance exercise in promoting
The greatest overall effect of subgroups analysis was found increases in capacities highly related to athletic performance, such
for the variable “muscle power” (Z = 5.75; p < 0.001. SMD 0.8 W; as vertical jump height and running speed. The efficacy of eccen-
95% CI 0.53–1.07). Thus, regardless of the specific exercise device tric overload flywheel training to promote functional and structural
or muscle group analysed, FW-EOT induced very robust adapta- adaptations is possibly mediated by the capacity to achieve higher
tions in terms of muscle power, which are greater than those seen forces during the eccentric muscle action, which maximizes the
after traditional resistance exercise training. Apart from the stud- stretch-shortening cycle and thus, the capacity to produce greater
ies included in the meta-analysis, the first investigation assessing force in the subsequent concentric action. However, these gains
FW-induced adaptations in muscle power at different loads (i.e. could even be greater considering the possibilities offered by FW
40–80% of 1 RM) was Fernandez-Gonzalo et al.,31 who interestingly, technology. Thus, it appears at high speed and with light moments
showed that power increments at high percentages of 1 RM were of inertia is the most effective training protocol to induced muscle
greater for men than women. Expanding on these data, FW-EOT adaptations employing FW. Finally, muscle power is the training
induced power gains at high loads (70–90%) of 1 RM in athletes.32 outcome that experienced the greatest increase after a period of
These major improvements in muscle power may be explained by FW-EOT in healthy subjects and athletes. Future studies should
the particular characteristics of the FW training to induce EO, since analyse FW resistance training with eccentric overload programs
the EO is usually applied in the last portion of the range of motion using and comparing different training devices, volumes, intensi-
during the ECC action. That is, the EO is produced mainly at a joint ties and exercise modes to fine-tune FW training routines.
angle close to 70◦ during leg press32 and elbow flexion exercise,52
or at 90◦ during knee extension,28,30,34,53 just before the subsequent
CON action is initiated. This exercise technique would maximize the Practical implications
stretch shortening cycle, allowing for greater production of force
during the first part of the CON action,32 and therefore, a higher Iso-inertial flywheel resistance exercise, a technology originally
velocity during the entire movement (i.e. increased power). developed to combat muscle deconditioning during spaceflight, is
The significant improvements in vertical jump after FW train- a powerful tool to induce muscle force and power increments, as
ing corroborate the efficacy of this training model to induce positive well as muscle hypertrophy in healthy subjects and athletes.
power adaptations in healthy and well-trained subjects. Thus, the The available scientific literature indicates flywheel resis-
articles included in the current meta-analysis indicate that vertical tance exercise, calling for brief episodes of eccentric overload,
jump improvements induced by FW-EOT are higher than tradi- triggers significantly greater skeletal muscle adaptations (i.e.
tional resistance training, as long as the jump modality used to strength, power and muscle mass) compared to traditional, gravity-
assess jumping ability includes an eccentric phase (i.e. a counter dependent resistance exercise paradigms.
movement jump, CMJ). Of note however, De Hoyo et al.63 showed Eccentric overload flywheel resistance exercise boosts
that under specific circumstances, i.e. FW exercise (VersaPulley performance-related abilities (i.e. jump height and running
device) performed in the horizontal plane vs. free weights using speed). The magnitude of these adaptations is greater than those
guided (vertical) squat exercise, gains in jumping height were lower reported after weight-stack or free weight resistance training,
for FW than traditional training (ES = 0.4 and 0.9, respectively). which highlights the efficacy of iso-inertial exercise training.
Therefore, vertical FW training is recommended to increase jump-
ing ability.
The selection of running speed as an outcome in this meta- Acknowledgment
analysis is based on the fact that it is an important performance
factor in many sports, and it is usually related with vertical Sergio Maroto-Izquierdo is supported by the Ministry of Educa-
jump and multi-joint strength training. Although the findings tion of Spain (FPU014/05732).
950 S. Maroto-Izquierdo et al. / Journal of Science and Medicine in Sport 20 (2017) 943–951
61. Noorkoiv M, Nosaka K, Blazevich AJ. Neuromuscular adaptations associ- 64. Tous-Fajardo J, Gonzalo-Skok O, Arjol-Serrano JL et al. Change of direction speed
ated with knee joint angle-specific force change. Med Sci Sports Exerc 2014; in soccer players is enhanced by functional inertial eccentric overload and vibra-
46(8):1525–1537. tion training. Int J Sports Physiol Perform 2016; 11:66–73.
62. McMahon GE, Morse CI, Burden A et al. Impact of range of motion during ecolog- 65. Friedmann-Bette B, Bauer T, Kinscherf R et al. Effects of strength training with
ically valid resistance training protocols on muscle size, subcutaneous fat, and eccentric overload on muscle adaptation in male athletes. Eur J Appl Physiol 2010;
strength. J Strength Cond Res 2014; 28(1):245–255. 108(4):821–836.
63. de Hoyo M, Sanudo B, Carrasco L et al. Effects of traditional versus horizontal
inertial flywheel power training on common sport-related tasks. J Hum Kinet
2015; 47:155–167.