Key words
resistance training - plyometrics - set structures - jump performance - evidence
Background
Stronger athletes can generate increased ground reaction forces, which are related
to
faster sprint times [1]
[2]
[3], change of direction speed [4], and vertical jump height [3]. Movements in field games are characterized by changes in velocity and
these changes are determined by a produced impulse [5]. For example, a strong predictor of jump height is the body’s
ability to generate maximal vertical impulse [6]
[7]. Maximal neuromuscular efforts aim at maximizing the
impulse generated, because this is decisive for the subsequent velocity and,
therefore, performance outcome. Vertical jump ability, frequently quantified using
the countermovement jump (CMJ) [8]
[9], is associated with higher level performers across a
range of team sports, such as volleyball, football, rugby, or basketball [10]
[11]
[12]
[13], as well as
individual sports, like track and field [3]
[14]. Continuous athlete improvement depends on the
amount of the trainability of numerous performance variables in athletic
populations, respectively [15].
To enhance the maximal neuromuscular performance, strength training (ST)
and/or power training (PW) are recommended [16]. For PW, low-load, high-velocity movements are performed
(e. g. plyometrics, ballistics, sprinting). This is done with body weight or
loads<30% of 1 repetition maximum (1-RM) [17]
[18]. For ST, higher loads are moved at
lower velocities during exercise (e. g. squats, leg presses, split squats,
leg extensions, leg curls) compared to PW [18]
[19]
[20]. In untrained
individuals, recent literature suggests intensities of>60% of 1-RM
to achieve strength gains [21]. A specific form of ST
is weightlifting (e. g. power cleans, clean and jerks, high pulls). These
exercises require an athlete to move heavy loads as quickly as possible. It is
suggested that 75 to 80% of 1-RM should be used for this type of strength
training [22]. Meanwhile, it is well established that
a combination of higher and lower loads during exercises improves maximum strength,
sprinting speed, and jump height [17]
[23]
[24]. When
incorporated into a strength training program, weightlifting exercises are believed
to improve force production characteristics and athletic performance, including
jumping, to a greater extent than resistance or plyometric training alone [25]. Four main set structures of combining high and low
loads in a training program exist. They are complex training (CPX), contrast
training (CT), compound training (CP), and traditional training (TT) [26]
[27]
[28]:
-
CPX: Multiple sets of a heavy resistance exercise are carried out prior to
performing sets of a lighter resistance exercise [28]
-
CT: The training is characterized by using exercises of contrasting loads,
i. e. varying heavy and light exercises set for set
-
CP: Strength and plyometric exercises are conducted on separate days [17]
[29]
-
TT: Multiple sets of lighter resistances with high velocity are carried out
before performing sets of heavy resistances with low velocity [28], because power training is considered most
effective, when exercises are performed in a fatigue-free state where the
body can produce the peak power output [30]
Power exercises should therefore be completed before ST (TT) or on another day (CP).
CPX and CT utilize a phenomenon called post-activation performance enhancement
(PAPE). PAPE occurs when one or more high-intensity voluntary conditioning
contractions results in an increased voluntary muscular performance in a subsequent
test without concurrent sign of typical post-activation potentiation (PAP) [31]. While PAP is principally attributed to the
phosphorylation of myosin light chains in type II fibers, PAPE arises as a higher
rate of force development that can mainly be elucidated by physiological responses,
such as increased muscle temperature, accumulation of intracellular water and
further mechanisms [31]
[32]
[33]. Based on this theory, CPX uses a
block-wise approach where several sets of ST alternate with several sets of PW,
while CT switches between higher load and lower load exercises in each set [26]
[34]. Currently, there
is no evidence that one of the training set structures is superior to another.
However, this may be necessary because athletes engaged in a similar competitive
environment may need different ST schemes aimed at improving the physical status of
weaker athletes and sustaining stronger athletes’ capacities during
in-season [35].
The aims of this research are thus to (1) determine the effectiveness of CPX, CT,
CP,
and TT interventions on countermovement jump performance, (2) compare their effects
to each other, (3) to ST alone, (4) to PW alone, (5) to ST and PW combined
(ST/PW), and (6) to control (CTRL), i. e. no intervention. In
addition, the aim is to examine whether the training experience or status of the
participants is associated with the intervention effects from CPX and CT, as well
as
CP and TT, because experienced or trained subjects may have less potential for
adaptation [36]. The results of this analysis may
provide strength and conditioning practitioners evidence to help them design their
training programs more effectively.
Materials and Methods
This review was conducted in accordance with Preferred Reporting Items for Systematic
Reviews and Meta-Analyses for Network Meta-Analyses (PRISMA-NMA) [37].
Literature search and study selection
The databases Web of Science, Medline (PubMed), and SPORTDiscus were
systematically searched until November 30, 2020, using the subsequent search
strings: ((((combined training OR compound training OR contrast training OR
complex training OR strength training OR weight training OR resistance training
OR weight lifting OR weightlifting OR Olympic weightlifting))) AND ((plyometric
OR plyometric training OR explosive OR explosive training OR explosive
performance OR ballistic performance OR ballistic training))). Additional
publications were obtained from reference lists of potentially eligible
articles.
Inclusion criteria
Eligible studies were randomized trials investigating the influence of ST
and/or PW on vertical jump performance in healthy subjects, published in
English or German language in a peer-reviewed journal. Specifically, studies
were included if they met the following criteria: (1) participants (male or
female) were healthy, older than 14, and younger than 50 years; (2) the training
intervention lasted at least four weeks including at least eight training
sessions in total; (3) one intervention group performed CPX, CT, CP or TT; (4)
ST was considered as training that efficiently induces a measurable growth in
muscle strength or/and hypertrophy [38].
An increase of muscular strength depends on the training load employed,
i. e. the heavier the training load the larger the maximal strength
adaptation [39]. Furthermore, the greater the
% 1-RM, the greater is the response of hypertrophy, with a maximal
increase between loads of 80 to 95% of 1-RM. However, athletes may have
different training backgrounds and, therefore, experience different muscle
hypertrophy in response to the same amount and type of ST [38]. It was reported that subjects with no
experience in ST make the most of their strength gains with mean intensities of
60% of 1-RM [21]. To consider subjects of
various levels of ST experience, studies were included that reported the effects
of different training set structures incorporating the lower extremity using an
average load>60% of 1-RM. Furthermore, moderate loads (60 to
84% 1-RM) revealed an increased power enhancement in subsequent
potentiation tasks compared to heavy loads (>85% 1-RM), with an
effect size of 1.06 versus 0.31, respectively [40].
PW was defined as explosive exercises (plyometrics, ballistic, sprint and change
of direction exercises) using an average load<30% of 1-RM [17]; (5) CPX, CT, CP, and TT were compared to each
other, a control condition, an alternative training method such as ST or PW
alone or ST and PW combined. In this regard, a combined training does not fit to
the definitions of CPX, CT, CP, TT, because combined ST and PW is characterized
by dividing ST and PW into two separate phases, e. g. four weeks of ST
using external weights for four sets, each with six repetitions, followed by
four weeks of PW using plyometric exercises [41];
(6) outcome measure was CMJ height; (7) relevant data were available.
Exclusion criteria
Studies were excluded if they (1) were non-randomized trials, (2) were trials
that examined the effect of CPX, CT, CP, TT on an underlying pathology, (3) were
trials that combined CPX, CT, CP, and TT, e. g. in a cross-over design,
and (4) reported insufficient data precluding inclusion in a network
meta-analysis.
Implementation of search
Titles and abstracts of studies were reviewed initially by two authors (MB, SB)
to screen if they might be relevant. Then, duplicates were removed ([Fig. 1]). All potential articles were assessed
against the eligibility criteria and reviewed in full text by two authors (MB,
SB) independently to determine their final relevance. If a difference of opinion
occurred, the third/senior author (MA) helped to find a consensus.
Fig. 1 Preferred Reporting Items for Systematic Reviews and
Meta-Analyses (PRISMA) flowchart of study identification and selection
according to the PRISMA guidelines [104].
CMJ, countermovement jump; RCT, randomized controlled trial.
Quality assessment of included studies and treatment effect
The Cochrane Risk of Bias 2 (RoB 2) tool for randomized trials was used to assess
the included studies’ internal validity [42]. RoB 2 contains five domains that cover the main types of bias,
including randomization process, deviations from intended interventions, missing
outcome data, measurement of the outcome, and selection of the reported result.
A predefined algorithm is given with each domain, which contains several
questions that can be answered using one of four given answer choices. These
algorithms provide a guided approach to making an informed decision about the
potential risk of bias in each study. The choices of answers are (1)
"yes"; (2) "probably yes"; (3) "probably
no"; (4) "no"; and (5) "no information".
Based on the answers provided, each domain is rated as (1) "low risk of
bias"; (2) "some concerns"; or (3) "high risk of
bias". Studies were rated independently by two researchers (MB, SB). If
disagreements occurred between the two researchers, resulting in no consensus,
the third author (MA) was consulted to clear the disagreement. The Grading of
Recommendations Assessment, Development, and Evaluation (GRADE) approach was
used for rating the quality of treatment effect estimates from network
meta-analysis (NMA) [43].
Outcome measures and data extraction
After quality assessment, relevant data were extracted from the studies.
Extracted outcome data were the mean change from baseline CMJ heights and its
standard deviation (SD) in each trial arm. When the respective mean change was
not available, or the data were presented in a different way, for example median
values or another measure, the mean change of mean (±SD) CMJ heights
between preintervention and postintervention were calculated, or the available
data were converted to mean (±SD), as suggested by the Cochrane Handbook
for Systematic Reviews of Interventions (Version 6.0) [44]
[45].
When only figures were presented in the studies, the authors were first contacted
and the missing numerical information was requested. If no response was given,
the data from the figures were extracted using ImageJ (V.1.50i,
https://imagej.nih.gov/ij/). ImageJ is an image
processing program that was used to first calibrate the axis length of the
relevant figures in pixels. Then, the calibration was used to measure the
relevant axis that contained the data points of interest. This method was
applied to four studies [46]
[47]
[48]
[49]. If none of those aforementioned methods were
applicable for determining the SD, a correlation coefficient of 0.6 was used to
substitute the missing values between baseline and follow-up CMJ. This method
was applied to one study [50]. Different
correlation coefficients were applied to evaluate findings’ sensitivity
and confirm the consistency of the results. In addition, relevant study
information regarding author, year, sample description, number and gender of
participants, intervention characteristics (experimental and control groups,
duration, and frequency), and training characteristics (training duration and
frequency, volume, intensity, and exercise selection) were reported by the
researchers. Rest periods were reported in 18 out of 24 studies ([Table 1]). After the extraction, the data
included were peer-reviewed and confirmed by the senior researcher (MA). The
overall agreement on data extraction and RoB 2 assessment between the
researchers, which was calculated using the kappa statistics (κ) [51], was excellent (κ=0.81).
Table 1 Study characteristics of included
studies.
|
Author and year
|
Population
|
Training program
|
|
|
Sample size (n), sex and age (years) [mean±SD]
|
Subjects
|
Set structure (number of subjects per set structure)
|
Strength training
|
Power training
|
Frequency (times/week)
|
Training duration (weeks)
|
Results CMJ height (cm) pre-/postintervention
[mean±SD]
|
|
Alemdaroglu et al. (2013)
|
24 MW (21.6±2.3)a
|
Recreational trained students with experience in resistance
and plyometric training
|
CPX (8)
|
CPX, CT and TT: 3×6 reps of split squats, leg-press
and leg-curls at 85–90% of 1-RM; rest
periods of 1 min between sets and 2 min
between exercises
|
CPX, CT and TT: 3×6–12 reps of split jumps,
SJs and front tuck jumps
|
2
|
6
|
CPX (33.6±1.2/36.2±1.1)
|
|
CT (8)
|
CT (29.5±1.3/32.5±1.2)
|
|
TT (8)
|
TT (31.2±1.1/35.2±1.0)
|
|
Ali et al. (2019)
|
36 M (21,4±0.3)a
|
Soccer players (students) with experience in resistance
training
|
CPX (12)
|
CPX: 3×12 reps of squats, barbell lunges, lateral
lunges, and calf raises at 80% of 1-RM ST
3×12 reps of squats, barbell lunges, lateral lunges,
and calf raises at 40–80% of 1-RM; rest
periods of 1 min between sets
|
CPX: 3×12 reps of DJs, split squat jumps, lateral
hops and calf jumps
|
3
|
6
|
CPX (46.9±4.8/50.7±4.9)
|
|
ST (12)
|
ST (41.9±4.6/45.1±4.7)
|
|
CTRL (12)
|
CTRL (43.9±6.9/44.3±6.9)
|
|
Arabatzi et al. (2010)
|
36 M (20.3±2.0)
|
Active physical education students with experience in
resistance and plyometric training
|
CPX (10)
|
CPX and ST: 4–6×4–6 reps of snatches
from a squat position, high pulls, power cleans, and squats
at 75–90% of 1-RM; rest periods of
3 min between sets in CPX and ST, not reported for
PW
|
CPX and PW: 4–6×6 reps of double-leg hurdle
hops, alternated single-leg hurdle hops, double-leg hops,
and half-squats
|
3
|
8
|
CPX (34.4±8.3/39.6±8.6)
|
|
PW (8)
|
PW (31.5±6.3/36.1±6.4)
|
|
ST (9)
|
ST (34.6±7.5/39.8±6.8)
|
|
CTRL (8)
|
CTRL (33.3±5.2/35.2±5.8)
|
|
|
|
Arazi et al. (2014)
|
24 W (20.7±0.7)a
|
Untrained women with experience in resistance and plyometric
training
|
CPX (7)
|
CPX, ST and CP: 3×6 reps of squats, knee-extensions,
knee-flexions and single-leg lunges at 60% of 1-RM;
rest periods of 1 min between sets
|
CPX and PW, CP: 3×6 reps of DJs, CMJs, 10 m
zigzag drill and lunge jumps
|
2
|
6
|
CPX (25.6±4.5/30.2±4.4)
|
|
PW (8)
|
PW (26.1±3.0/31.2±3.3)
|
|
ST (7)
|
ST (23.1±4.3/27.8±5.3)
|
|
CP (7)
|
CP (24.8±3.8/31.8±2.5)
|
|
|
|
Caterisano et al. (2018)
|
28 M (18.8±0.4)
|
College footballer (NCAA Division I) with experience in
resistance training
|
TT (14)
|
TT: 2–4×3–12 reps of back squats,
front squats power snatches, power cleans, bench press and
medicine ball throws at 50–65% of 1-RM (ST)
4–8×3–9 reps of back squats, cleans
and presses, bench press and jammers with 56,82 kg;
rest periods not reported
|
TT: 3×5 reps of box jumps
|
4
|
5
|
TT (57.9±8.9/64.5±7.9)
|
|
ST (14)
|
ALT [ST] (68.1±6.9/74.9±6.6)
|
|
De Villarreal et al. (2011)
|
65 MW (20.0±2.7)a
|
Physical education students
|
CPX (14)
|
CPX: 3×3–6 reps of squats at
60–86% of 1-RM and additional lower-body
exercises with diverse intensities ST: 3 different strength
training groups (traditional, strength, power-oriented and
ballistic) which performed 3–4×2–6
reps of lower body exercises with different intensities;
rest periods not reported
|
CPX and PW: 4–8×5 reps of CMJs
|
3
|
7
|
CPX (40.0±8.9/43.4±8.2)
|
|
ST (39)b
|
PW (37.5±5.5/40.7±5.8)
|
|
PW (12)
|
ST
(37.4±6.7/41.1±7.1)b
|
|
Faude et al. (2013)
|
16 M (22.5±2.5)
|
Third Swiss league soccer players
|
CT (8)
|
CT: 4×4 reps of unilateral squats at 90% of
1-RM on day 1 and 2–3×5–8 reps from
two of four exercises (squats, calf raises, lateral squats,
and step-ups) at 50–60% of 1-RM on day 2;
rest periods of 4 min between sets on day 1, and of
1–2 min between sets on day 2
|
CPX: 4×5 reps of single-leg hurdle jumps on day 1 and
2–3 times two of four pairs (5
DJs+5 m sprint, 5 CMJs+1 header, 8
lateral jumps+10 zig-zag sprints, 4 bounding
jumps+3 headers) on day 2
|
2
|
7
|
CT (40.2±4.8/41.4±3.5)
|
|
CTRL (8)
|
CTRL (43.0±2.1/39.8±2.4)
|
|
Hammami et al. (2019a)
|
28 F (16.6±0.3)
|
Handball players (elite-level) with some experience in
resistance training
|
CT (14)
|
CT: 4×6–8 reps of half squats, leg-press,
isometric half squats, calf raises at 75–90%
of 1-RM; rest periods of 1–2 min between
sets
|
CT: 6 reps of hurdle jumps+10 m sprint, 6
horizontal jumps+10 m sprint, 3 single leg
hops+10 m sprint, 6 hurdle jumps with legs
extended+10 m sprint
|
2
|
10
|
CT (28.1±1.6/33.8±1.6)
|
|
CTRL (14)
|
CTRL (27.0±3.5/28.1±3.4)
|
|
Hammami et al. (2019b)
|
40 M (15.8±0.4)
|
Soccer player (elite-level)
|
CT (14)
|
CT: 3–5×3–8 reps of half squats at
70–90% of 1-RM (ascending and descending
sets); rest periods not reported
|
CT: 3–5×3 reps of CMJs (week 1–4),
3–5×1 rep of CMJ+15 m sprint
(week 5–8) PW 5–10×7–10 reps
of hurdle jumps (week 1–4), 4×7–10
reps of DJs (week 5–8)
|
2
|
8
|
CT (38.2±2.1/47.0±5.9)
|
|
PW (14)
|
PW (36.9±3.9/42.1±6.0)
|
|
CTRL (12)
|
CTRL (37.9±5.4/37.2±4.5)
|
|
Herrero et al. (2010)
|
29 M (20,9±2,1)a
|
Physical education students
|
CP (8)
|
CP: 8×10 reps of knee extension at 70% of
1-RM ST/PW 8×10 reps of knee extension with
EMS; rest periods of 3 min between sets
|
CP and ST/PW: 90 total reps of DJs and horizontal
jumps (week 1–2), 105 total reps of DJs and
horizontal jumps (week 3–4)
|
4
|
4
|
CP (37.6±1.7/38.9±3.2)
|
|
ST/PW (11)
|
ST/PW (42.5±7.0/42.8±7.3)
|
|
CTRL (10)
|
CTRL (34.7±5.8/33.4±5.0)
|
|
Juarez et al. (2009)
|
16 M (19.7±1.7)a
|
Active undergraduate students with no specific strength
training experience
|
CT (8)
|
CT: 2×4 reps of squats at 70–85% of
1-RM ST/PW 4–5×4–8 reps of
squats at 70–85% of 1-RM (week 1–4);
rest periods of 3–5 min between sets
|
CT: 2×5 reps of CMJs, hurdle jumps, or DJs combined
with 2×20 m sprints ST/PW:
4–5×5 reps of CMJs, DJs, or box jumps
combined with 4–5×2 reps of 20 m
sprints (week 5–8)
|
2
|
8
|
CT (45.9±5.8/51.3±6.6)
|
|
ST/PW (8)
|
ST/PW (46.1±5.6/48.6±6.2)
|
|
Kijowski et al. (2015)
|
18 M (21.2±1.2)
|
Male subjects with a minimum experience of 24 months in
recreational strength and power training
|
TT (9)
|
TT: 3×3 reps of squats at 90% of 1-RM; rest
periods of 2 min between drop jump sets, and
5 min between squat sets
|
TT: 5×6 reps of DJs
|
2
|
4
|
TT (46±8.0/49±6.0)
|
|
CTRL (10)
|
CTRL (47±6.0/44±5.0)
|
|
Kobal et al. (2016)
|
27 M (18.9±0.6)
|
Soccer players (first Brazilian division) without experience
in resistance or plyometric training
|
CPX (9)
|
CPX, CT and TT: 3–5×6–10 reps of
squats at 60–80% of 1-RM; rest periods of
3 min between sets and exercises
|
CPX, CT and TT: 3–5×10–12 reps of
DJs
|
2
|
8
|
CPX (37.8±4.5/42.9±4.7)
|
|
CT (9)
|
CT (37.3±4.1/42.8±3.3)
|
|
TT (9)
|
TT (37.2±4.7/42.5±4.9)
|
|
Krishna et al. (2019)
|
42 M (20,8±2,9)a
|
State and top division-level cricket fast bowlers
|
TT (21)
|
TT: 3×8–10 reps of squats, lunges, trap bar
deadlifts, single-leg hip thrusts, hamstring curl, planks
(week 1–3), front squats, step-ups, trap bar
deadlifts, single leg Romanian deadlifts, hamstring curls,
planks (week 4–6), squats, step-ups, kettlebell
swings, single leg Romanian deadlifts, hamstring curls (week
7–9), squats, lateral lunges, trap bar deadlifts,
single leg Romanian deadlifts, hamstring curls, back
extension (week 10–12) at 75% of 1-RM; rest
periods of 1 min between sets
|
TT: 3×8 reps of SJs, hurdle jumps, hurdle hops (week
1–3), broad jumps, hurdle jumps, hurdle hops (week
4–6), single-leg box jump, lateral bounds, hurdle
hops (week 7–9), box jumps, bounding, hurdle hops,
rotational vertical hops (week 10–12)
|
3
|
12
|
TT (3.7±3.6)c
|
|
CTRL (21)
|
CTRL (0.6±1.1)c
|
|
Lyttle et al. (1996)
|
33 M (22.8±5.3)a
|
Regional level athletes with no specific strength or
plyometric training in the previous year
|
CPX (11)
|
CPX: 1–3×6–10RM of bench press and
squats; rest periods not reported
|
CPX: 1–2×4 reps of medicine ball throws and
DJs PW 2–6×8 reps explosive bench press and
SJs at 30% of 1-RM
|
2
|
8
|
CPX (52.8±11.5/58.4±9.3)
|
|
PW (11)
|
PW (50.8±9.0/54.6±8.5)
|
|
CTRL (11)
|
CTRL (49.2±3.5/49.2±5.7)
|
|
Perez-Gomez et al. (2008)
|
37 M (23.9±0,3)a
|
Physical education students with no strength training for at
least six months
|
TT (16)
|
TT: 1–3×2–12 reps of leg-press,
leg-extensions, half squats and leg-curls at
50–90% of 1-RM; rest periods not
reported
|
TT 4–9×5 reps of DJs and hurdle jumps
|
3
|
6
|
TT (36±4.0/39±4.0)
|
|
CTRL (21)
|
CTRL (34±4.5/34±4.5)
|
|
Robineau et al. (2017)
|
30 M (26,3±0,5)a
|
Amateur rugby sevens players with a minimum of two-year
resistance training experience
|
CP (19)b
|
CP and ST: 3×3–10 reps of half squats,
leg-extension, deadlifts, bench press, and bench row at
70–90% of 1-RM; rest periods of
2–3 min between sets
|
CP: 4 x 30 s sprints or 2 x 8 min
30/30 s sprints (week 1–3), 6 x
30 s sprints or 2 x 10 min
30/30 s sprints (week 4–6),
8×30 sprints or 2 x 12 min
30/30 s sprints (week 7–9)
|
4
|
8
|
CP
(32.6±3.8/33.9±4.2)b
|
|
ST (11)
|
ST (31.3±4.7/34.2±4.9)
|
|
Ronnestad et al. (2008)
|
21 M (22.7±2.0)a
|
Norwegian premier league soccer players with experience in
strength training
|
CPX (8)
|
CPX and ST: 3–5×4–6RM of squats and
hip flexion exercises; rest periods of 1 min between
sets (only reported for jumps during CPX), not reported for
ST and general strength exercises
|
CPX: 2–4×10 reps of alternate leg bounds,
2×5 reps of double-leg hurdle jumps and single-leg
forward hops
|
2
|
7
|
CPX (36.0±5.6/36.7±5.3)
|
|
ST (6)
|
ST (32.9±1.9/33.9±1.4)
|
|
CTRL (7)
|
CTRL (36.0±2.3/35.7±3.7)
|
|
Santos et al.(2008)
|
25 M (14,5±0.1)a
|
Basketball players without experience in resistance or
plyometric training
|
CPX (15)
|
CPX: 2–3×10–12RM of leg-extensions,
pullovers, leg-curls, decline press, leg-press and
latissimus pulldowns; rest periods of
1–3 min between sets,
15–60 s between exercises
|
CPX: 2–3×5–15 reps of different
plyometric drills
|
2
|
10
|
CPX (29.8±5.9/33.0±6.2)
|
|
CTRL (10)
|
CTRL (30.7±5.1/28.4±4.0)
|
|
Smith et al. (2014)
|
28 MW (20–29)
|
Recreational athletes with regularly moderate-to-vigorous
physical activity for at least 3 months
|
CT (19)b
|
CT: 3×4–6RM of squats or kettlebell swings;
rest periods of 3 min between sets
|
CT: 3×5 reps of CMJs
|
3
|
6
|
CT
(52.5±12.4/57.6±12.9)b
|
|
CTRL (9)
|
CTRL (53.1±8.4/55.4±11.4)
|
|
Talpey et al. (2016)
|
20 M (21.1±3.5)a
|
Recreational athletes involved, football, rugby, basketball
with a minimum of one-year resistance training
experience
|
CPX (9)
|
CPX and TT: 3–4×3–8RM of half squats;
rest periods of 4 min between squat sets, and
3 min between jump sets
|
CPX and TT: 3–4×4 reps of SJs
|
2
|
9
|
CPX (43.1±3.9/46.6±5.0)
|
|
TT (11)
|
TT (41.9±5.2/45.9±4.3)
|
|
Trecroci et al. (2020)
|
18 M (14.2±0.2)a
|
Sub-elite soccer players without experience in structured and
advanced strength and plyometric training
|
CP (9)
|
CP: 3×15 reps (4–6 RIR) of front squats, hip
thrusts and alternating side lunges at 60% of
1-RM+30 s of plank, standing superman and
side plank; unclear reporting of rest periods (work to rest
ratio 1:2 plus 2 min additional rest between sets)
|
CP: 4–6×5 reps of double-leg hurdle
jumps+5 m sprint and diagonal
bounds+5 m sprint
|
2
|
5
|
CP (47.11±3.5/49.22±3.5)
|
|
CTRL (9)
|
CTRL (46.59±4.6/46.79±4.1)
|
|
Tricoli et al. (2005)
|
32 M (22±1.5)
|
Physical education students with experience in strength
training
|
TT (12)
|
TT: 4 x 6RM of half squats ST:
3–4×4–6RM of high pulls, power
cleans, clean and jerks and half squats; rest periods not
reported
|
TT: 4–10×4 reps of double-leg hurdle hops,
alternated single-leg hurdle hops, single-leg hurdle hops
and DJs
|
3
|
8
|
TT (40.2±3.9/±42.5±3.0)
|
|
ST (12)
|
ST (42.2±2.1/45.0±2.6)
|
|
CTRL (12)
|
CTRL (42.2±4.9/42.6±5.2)
|
|
Zsis (2013)
|
21 M (17.1±1.1)
|
Amateur soccer players with a minimum experience of twelve
months in conditioning programs
|
CP (7)
|
CP and ST: 3×10 reps of leg-press, knee-extensions,
squat at 80% of 1-RM; rest periods of 3 min
between sets, and 1 min between exercises
|
CP and PW 2–3×10 reps of DJs, split squat
jumps, elastic jumps, vertical and horizontal jumps
|
2
|
8
|
CP (39.8±3.3/42.4±7.1)
|
|
ST (7)
|
ST (36.6±4.5/39.1±6.6)
|
|
PW (7)
|
PW (37.3±4.3/38.6±2.3)
|
1-RM, one repetition maximum; cm, centimeter; CMJ, countermovement jump;
COD, change of direction; CPX, complex training; CTRL, control group;
DJ, drop jump/depth jump; EMS, electromyostimulation; kg,
kilogram; m, meter; min, minute(s); M, male only; MW, male and female
together; PW, power training; reps, repetition; RIR, repetitions in
reserve; RM, repetition maximum; s, seconds; SD, standard deviation; SJ,
squat jump; ST, strength training; TT, traditional training; W, female
only a: For studies not reporting pooled estimates for the sample mean
and sample standard deviation, the respective values were calculated
using the sample sizes (n1, n2), means (m1, m2) and standard deviations
(sd1, sd2) reported for the individual groups. The according equations
are pooled
mean=(m1×n1+m2×n2)/(n1+n2)
and pooled sample standard deviation=sqrt [(n1 −
1)×sd1ˆ2+(n2 −
1)×sd2ˆ2+n1×(m1 −
m)ˆ2+n2×(m2 −
m)ˆ2)/(n1+n2 − 1)]. b: Data from groups
carrying out the same set structure were combined c: Only changes from
baseline were reported
Statistical analysis
Network plots for outcome were developed using R software (V.4.0.2) to illustrate
the corresponding amount of available evidence on the different training set
structures [52]. Next, to compare the effects of
CPX, CT, CP, and TT interventions on CMJ performance with each other, to ST or
PW alone, to ST/PW as well as to control, network meta-analyses (NMA)
were performed [53]. In NMA, comparisons between
three or more interventions are possible. Furthermore, they rely on a
combination of direct and indirect comparisons, which leads to an improvement of
the precision of the treatment effect estimates [54]
[55]. Direct comparisons refer to
interventions directly compared in individual studies, whereas statistics
calculates indirect comparisons. The effects in two sets of analyses were
compared: (1) different exercise interventions on CMJ height and (2) different
exercise interventions on CMJ height in subjects, who were specifically
described as strength and/or power trained. According to recommendations
from previous literature [36]
[56]
[57], participants
were defined strength and/or power trained if they have been classified
as individuals with strength and/or plyometric training experience by
the respective study authors or if they participated in regular structured
training programs for at least 3 months prior to the intervention period.
From initial scanning, a meta-analytical approach to the data of different
exercise interventions on CMJ height in non-strength and/or power
trained subjects was considered inappropriate for the analysis, given the low
number of studies (n=6) [49]
[58]
[59]
[60]
[61]
[62] as well as the lack of sufficient study
population characteristics based on the training experience (n=5) [48]
[50]
[63]
[64]
[65]. Similar to previous studies [36]
[56]
[57], participants were defined non-strength
and/or power trained, if they were reported to be individuals with no
strength and/or plyometric training experience by the respective study
authors or if they reported no involvement in regular physical activity for at
least 3 months prior to the intervention period. If the study authors failed to
provide sufficient information on training status of the participants in their
studies, or classification of participants was unclear, the respective study was
not included into the subgroup analyses.
The network meta-analyses were completed using a random-effects model.
Random-effect models consider the variability of studies and do not require
between-study homogeneity. Therefore, they allow for differentiating the true
intervention effect between each included study [66]. With possible different true intervention effects from each
study, random-effect models give a summary effect, representing an estimation of
the mean of this distribution of true effect sizes. Mean difference (MD) and its
95% confidence intervals (95% CI) were outlined for every
intervention compared to each other and control. Participants were defined as
controls when they served as control persons without performing any intervention
in the respective studies [67]. Following the
evaluation of the interventions’ comparative effect, each one was ranked
to identify if one intervention was superior to another. The ranking was
performed by applying P-scores. P-scores are based solely on the point estimates
and standard errors of the network estimates. They compare all different
interventions and measure the probability to which extent a specific treatment
is better than another one [68].
A critical tool to determine the applicability of NMA results is testing for
consistency. Consistency assumes that the treatment effects estimated from
direct comparisons do not differ from those effect estimates of indirect
comparisons. To assess each network’s consistency assumption, a global
approach that calculates the regression coefficient of each study
design’s inconsistency model was used first. Then, the Wald test, which
tests the regression coefficients’ linearity for all models, was applied
[69]. If there was agreement
(p-value>0.05), a local approach was used and side-splitting was applied
to further assess the inconsistency of each treatment. The probability of
"small study bias", where smaller studies contribute different
or greater treatment effects than more extensive trials, was evaluated using
comparison-adjusted funnel plots. This procedure was applied to comparisons,
where at least ten studies were obtainable. A frequentist framework using the R
package "netmeta" (V.1.2–1) was applied to all NMA
models. In a frequentist method, the available data are repeated infinitely
based on a general statistical theory, and the probability of significance,
known as the p-value, and the CI is calculated. Based on this statistic method,
the research hypothesis is discarded or accepted. Furthermore, the frequentist
approach is independent of external information, leading to an already defined
probability that the research hypothesis is valid within the available data.
Therefore, the choice of acceptance or rejection of the research hypothesis is
solely made based on the p-value or the CI [69].
Results
Selection process
The flow of the systematic review is presented in [Fig.
1]. The electronic database search lead to 3281 records after
duplicates (n=3447) were removed. Following the screening of titles and
abstracts, 78 full-text records were assessed for eligibility. Fifty-four
studies were excluded with reasons. Among the 24 studies included [46]
[47]
[48]
[49]
[50]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80], ten
incorporated CPX training [46]
[47]
[48]
[49]
[59]
[61]
[70]
[71]
[76]
[78], and seven included CT training [46]
[49]
[58]
[63]
[64]
[73]
[77], with five studies applying CP training [47]
[62]
[65]
[75]
[80] and eight TT training [46]
[49]
[50]
[60]
[72]
[74]
[78]
[79]. Of these 24 studies, 13 [46]
[47]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80] were considered suitable for the network,
analyzing the influence of different exercise intervention trials on CMJ height
in subjects with training experience in ST and/or PW. Six studies [49]
[58]
[59]
[60]
[61]
[62] were used for
the narrative analysis, examining the influence of different exercise
intervention trials on CMJ height in subjects without training experience in ST
and/or PW.
Characteristics of the included studies
A detailed summary of each individual study is presented in [Table 1]. The sample size in all exercise
intervention RCTs ranged from n=16 to n=65 subjects. The ages of
the subjects ranged from 14.2 to 26.3 years. Study duration ranged from four to
twelve weeks and training sessions completed in the studies ranged from eight to
36. Most of the subjects in the included studies were men (84%). The
largest number of the subjects were involved in a total of ten exercise
intervention RCTs evaluating CPX training compared to CTRL or another
intervention defined previously. Overall, 694 individuals were included from
which 346 were strength and/or power trained participants (
[Table 1]
).
Risk of bias and certainty of the evidence
Most studies were at some concerns (62.5%), with a high risk of bias in
37.5% ([Fig. 2]
[and 3]). Deviations from the intended intervention
(37.5%) and missing outcome data (33.3%) were the most common
bias sources. The certainty of the evidence for rating the quality of treatment
effect estimates was low to very low for all comparisons. The downgrading of the
comparison’s evidence was done due to the risk of bias
("serious" to "very serious") and imprecision
for all comparisons (100%) ([Table
2]). For the narrative analysis of studies that included only
non-strength and/or power trained subjects, the level of evidence was
downgraded due to the serious risk of bias limitations. Moreover, downgrading
was executed for imprecision as the overall sample size was small. In
conclusion, there is very low-quality evidence for different exercise
interventions on CMJ height in subjects without training experience.
Fig. 2 Risk of bias judgement for each study examining the effects
of different training set structures on countermovement jump height in
healthy subjects with low risk of bias, some concerns, and high risk of
bias for each domain of the Cochrane Risk of Bias 2 tool.
Fig. 3 Percentage (%) of studies examining the effects of
different training set structures on countermovement jump height in
healthy subjects with low risk of bias, some concerns, and high risk of
bias for each domain of the Cochrane Risk of Bias 2 tool.
Table 2 Certainty of the evidence using the Grading of
Recommendations Assessment, Development and Evaluation (GRADE)
approach.
|
All subjects
|
Direct evidence
|
|
Indirect evidence
|
|
Mean difference (95% CI)
|
Quality of evidence
|
Mean difference (95% Cl)
|
Quality of evidence
|
Mean difference (95% Cl)
|
Quality of evidence
|
|
Complex vs. Contrast
|
–0.40 (–1.41 to 0.61)
|
Low
|
–0.70 (–2.10 to 0.68)
|
Moderate
|
–0.50 (–1.32 to 0.31)
|
Low
|
|
Complex vs. Compound
|
2.40 (–1.31 to 6.11)
|
Moderate
|
0.26 (–1.05 to 1.58)
|
Moderate
|
0.50 (–0.73 to 1.75)
|
Low
|
|
Complex vs. Traditional
|
–0.64 (–1.04 to –0.25)
|
Moderate
|
0.64 (–0.88 to 2.17)
|
Moderate
|
–0.56 (–0.95 to –0.18)
|
Low
|
|
Complex vs. Strength
|
–0.02 (–1.87 to 1.81)
|
Moderate
|
0.79 (0.37 to 1.21)
|
Low
|
0.75 (0.34 to 1.16)
|
Low
|
|
Complex vs. Power
|
–0.10 (–2.60 to 2.38)
|
Moderate
|
–0.43 (–1.90 to 1.02)
|
Moderate
|
–0.35 (–1.61 to 0.91)
|
Low
|
|
Complex vs. Strength/Power
|
NA
|
NA
|
–1.31 (–4.65 to 2.01)
|
Low
|
–1.31 (–4.65 to 2.01)
|
Very Low
|
|
Complex vs. Control
|
3.34 (1.15 to 5.52)
|
Moderate
|
2.78 (to 1.82 to 3.74)
|
Low
|
2.87 (1.99 to 3.74)
|
Low
|
|
Contrast vs. Compound
|
NA
|
NA
|
0.00 (–1.38 to 1.38)
|
Moderate
|
0.00 (–1.38 to 1.38)
|
Low
|
|
Contrast vs. Traditional
|
–0.91 (–1.88 to 0.06)
|
Low
|
1.67 (0.27 to 3.06)
|
Moderate
|
–0.06 (–0.86 to 0.73)
|
Low
|
|
Contrast vs. Strength
|
NA
|
NA
|
0.24 (–0.56 to 1.05)
|
Low
|
0.24 (–0.56 to 1.05)
|
Very Low
|
|
Contrast vs. Power
|
–3.60 (–7.20 to 0)
|
Moderate
|
–0.37 (–1.89 to 1.14)
|
Low
|
–0.85 (–2.25 to 0.54)
|
Low
|
|
Contrast vs. Strength/Power
|
–2.90 (–8.24 to 2.44)
|
Moderate
|
–1.14 (–5.39 to 3.11)
|
Moderate
|
–1.82 (–5.15 to 1.50)
|
Low
|
|
Contrast vs. Control
|
5.25 (3.54 to 6.96)
|
Moderate
|
2.36 (1.10 to 3.62)
|
Low
|
3.37 (2.36 to 4.39)
|
Low
|
|
Compound vs. Traditional
|
NA
|
NA
|
–0.05 (–1.26 to 1.14)
|
Moderate
|
–0.05 (–1.26 to 1.14)
|
Low
|
|
Compound vs. Strength
|
0.02 (–1.48 to 1.52)
|
Moderate
|
0.63 (–1.34 to 2.61)
|
Moderate
|
0.24 (–0.95 to 1.44)
|
Low
|
|
Compound vs. Power
|
–1.45 (–2.94 to 0.04)
|
Moderate
|
1.02 (–1.64 to 3.70)
|
Moderate
|
–0.85 (–2.16 to 0.44)
|
Low
|
|
Compound vs. Strength/Power
|
–1.00 (–5.18 to 3.18)
|
Moderate
|
–3.22 (–8.66 to 2.21)
|
Moderate
|
–1.82 (–5.14 to 1.48)
|
Low
|
|
Compound vs. Control
|
2.22 (–0.16 to 4.61)
|
Moderate
|
3.87 (2.31 to 5.42)
|
Moderate
|
3.38 (2.07 to 4.68)
|
Low
|
|
Traditional vs. Strength
|
0.20 (0.05 to 0.34)
|
Low
|
–1.08 (–2.42 to 0.25)
|
Moderate
|
0.18 (0.03 to 0.33)
|
Low
|
|
Traditional vs. Power
|
NA
|
NA
|
–0.91 (–2.14 to 0.30)
|
Moderate
|
–0.91 (–2.14 to 0.30)
|
Low
|
|
Traditional vs. Strength/Power
|
NA
|
NA
|
–1.88 (–5.20 to 1.43)
|
Low
|
–1.88 (–5.20 to 1.43)
|
Very Low
|
|
Traditional vs. Control
|
3.04 (1.84 to 4.25)
|
Moderate
|
3.78 (2.65 to 4.90)
|
Moderate
|
3.43 (2.61 to 4.26)
|
Low
|
|
Strength and/or power trained
|
|
Complex vs. Contrast
|
–0.40 (–2.02 to 1.22)
|
Low
|
–1.51 (–3.78 to 0.74)
|
Low
|
–0.78 (–2.10 to 0.54)
|
Very Low
|
|
Complex vs. Compound
|
2.40 (–1.51 to 6.31)
|
Moderate
|
0.25 (–1.83 to 2.33)
|
Moderate
|
0.72 (–1.11 to 2.56)
|
Low
|
|
Complex vs. Traditional
|
–0.87 (–1.88 to 0.13)
|
Moderate
|
0.05 (–2.15 to 2.25)
|
Low
|
–0.71 (–1.62 to 0.20)
|
Low
|
|
Complex vs. Strength
|
–0.09 (–2.23 to 2.05)
|
Moderate
|
0.77 (–0.62 to 2.16)
|
Low
|
0.51 (–0.65 to 1.68)
|
Low
|
|
Complex vs. Power
|
0.20 (–3.00 to 3.42)
|
Moderate
|
–0.62 (–2.69 to 1.45)
|
Moderate
|
–0.37 (–2.12 to 1.37)
|
Low
|
|
Complex vs. Control
|
2.38 (–0.33 to 5.11)
|
Moderate
|
3.38 (1.58 to 5.18)
|
Moderate
|
3.08 (1.58 to 4.58)
|
Low
|
|
Contrast vs. Compound
|
NA
|
NA
|
–0.05 (–2.07 to 1.96)
|
Low
|
–0.05 (–2.07 to 1.96)
|
Very Low
|
|
Contrast vs. Traditional
|
–1.00 (–2.60 to 0.60)
|
Low
|
2.02 (–0.14 to 4.18)
|
Moderate
|
0.07 (–1.21 to 1.35)
|
Low
|
|
Contrast vs. Strength
|
NA
|
NA
|
–0.26 (–1.71 to 1.18)
|
Low
|
–0.26 (–1.71 to 1.18)
|
Very Low
|
|
Contrast vs. Power
|
–3.60 (–7.41 to 0.21)
|
Moderate
|
–0.36 (–2.53 to 1.80)
|
Low
|
–1.15 (–3.04 to 0.72)
|
Low
|
|
Contrast vs. Control
|
5.47 (3.44 to 7.51)
|
Moderate
|
1.94 (–0.27 to 4.16)
|
Low
|
3.86 (2.36 to 5.36)
|
Low
|
|
Compound vs. Traditional
|
NA
|
NA
|
0.01 (–1.75 to 1.78)
|
Low
|
0.01 (–1.75 to 1.78)
|
Very Low
|
|
Compound vs. Strength
|
0.00 (–1.70 to 1.71)
|
Low
|
–1.55 (–5.78 to 2.68)
|
Moderate
|
–0.21 (–1.79 to 1.37)
|
Low
|
|
Compound vs. Power
|
–1.48 (–3.25 to 0.29)
|
Moderate
|
1.11 (–3.19 to 5.41)
|
Moderate
|
–1.10 (–2.74 to 0.54)
|
Low
|
|
Compound vs. Control
|
NA
|
NA
|
3.81 (1.81 to 5.80)
|
Moderate
|
3.81 (1.81 to 5.80)
|
Low
|
|
Traditional vs. Strength
|
0.26 (–0.85 to 1.37)
|
Low
|
–1.54 (–3.46 to 0.37)
|
Moderate
|
–0.19 (–1.16 to 0.77)
|
Low
|
|
Traditional vs. Power
|
NA
|
NA
|
–1.08 (–2.78 to 0.60)
|
Moderate
|
–1.08 (-2.78 to 0.60)
|
Low
|
|
Traditional vs. Control
|
3.04 (0.19 to 5.89)
|
Moderate
|
4.04 (2.39 to 5.69
|
Moderate
|
3.79 (2.36 to 5.22)
|
Low
|
|
Non-strength and/or power trained
|
|
GRADE
|
Risk of bias
|
Inconsistency
|
Indirectness
|
Imprecision
|
Publication bias
|
Quality
|
|
Rating
|
Serious
|
No
|
No
|
Yes
|
NA
|
Very Low
|
95% CI, 95% confidence interval; NA, not applicable.
Comparative effects on CMJ height
Across all subjects, the analysis showed that all interventions were superior to
control: CPX (MD=2.87, 95% CI: 1.99 to 3.74), CT
(MD=3.37, 95% CI: 2.36 to 4.39), CP (MD=3.38,
95% CI: 2.07 to 4.68), TT (MD=3.43, 95% CI: 2.61 to
4.26) ([Table 2]). Although the P-score
indicated that ST (0.8892), TT (0.6998), CP (0.6937), and CT (0.6923) were the
best interventions for improving CMJ height in all subjects ([Table 3a]), only CPX was inferior to TT and to
ST (MD=0.56, 95% CI: 0.18 to 0.95, and, MD=0.75
95% CI: 0.34 to 1.16, respectively) while TT was inferior to ST
(MD=0.18, 95% CI: 0.03 to 0.33). No further specific
intervention was superior to another ([Table
4a]). In strength and/or power trained subjects, all
interventions were effective in increasing CMJ height compared with control: CPX
(MD=3.08, 95% CI: 1.58 to 4.58), CT (MD=3.86,
95% CI: 2.36 to 5.36), CP (MD=3.81, 95% CI: 1.81 to
5.80), TT (MD=3.79, 95% CI: 2.36 to 5.22) ([Table 2]). The P-score indicated that the best
treatments for improving CMJ height in subjects with training experience were CT
(0.7443), TT (0.7394), CP (0.7124), and ST (0.6297) ([Table 3b]). The NMA revealed that no specific
intervention was superior to another intervention except for CPX, which was
inferior to ST (MD=0.51, 95% CI: 0.65 to 1.68) ([Table 4b]). The network-graph is presented in
[Fig. 4].
Fig. 4 Network meta-analysis demonstrating available evidence
comparing (a) the influence of different interventions on
countermovement jump (CMJ) height and (b) the influence of
different interventions on CMJ height in strength and/or power trained
subjects. The nodes represent different interventions and the lines
connecting the nodes represent direct head-to-head randomized controlled
trials comparing the interventions. The thickness of the lines and the
size of the dots are proportional to the number of trial comparisons and
the number of participants in the treatment arms, respectively.
Table 3 P-score for CMJ height, a) when comparing all
subjects, and b) when comparing strength and/or power
trained subjects.
|
a) all subjects
|
|
Intervention
|
P-score
|
|
Strength
|
0.8892
|
|
Traditional
|
0.6998
|
|
Compound
|
0.6937
|
|
Contrast
|
0.6923
|
|
Complex
|
0.4022
|
|
Power
|
0.3321
|
|
Strength/power
|
0.2647
|
|
Control
|
0.0259
|
|
b) strength and/or power trained subjects
|
|
Intervention
|
P-score
|
|
Contrast
|
0.7443
|
|
Traditional
|
0.7394
|
|
Compound
|
0.7124
|
|
Strength
|
0.6297
|
|
Complex
|
0.3776
|
|
Power
|
0.2962
|
|
Control
|
0.0004
|
Table 4 Treatment effects for CMJ height. Change from
baseline when comparing a) all subjects, and b) when comparing
strength and/or power trained subjects.
|
a) all subjects
|
|
Control
|
|
|
|
|
|
|
|
|
2.87 (1.99 to 3.74)
|
Complex
|
|
|
|
|
|
|
|
3.37 (2.36 to 4.39)
|
0.50 (–0.31 to 1.32)
|
Contrast
|
|
|
|
|
|
|
3.38 (2.07 to 4.68)
|
0.50 (–0.73 to 1.75)
|
0.00 (–1.38 to 1.38)
|
Compound
|
|
|
|
|
|
3.43 (2.61 to 4.26)
|
0.56 (0.18 to 0.95)
|
0.06 (–0.73 to 0.86)
|
0.05 (–1.14 to 1.26)
|
Traditional
|
|
|
|
|
3.62 (2.79 to 4.45)
|
0.75 (0.34 to 1.16)
|
0.24 (–0.56 to 1.05)
|
0.24 (–0.95 to 1.44)
|
0.18 (0.03 to 0.33)
|
Strength
|
|
|
|
2.52 (1.17 to 3.86)
|
–0.35 (–1.61 to 0.91)
|
–0.85 (–2.25 to 0.54)
|
–0.85 (–2.16 to 0.44)
|
–0.91 (–2.14 to 0.30)
|
–1.10 (–2.33 to 0.12)
|
Power
|
|
|
1.55 (2.79 to 4.45)
|
–1.31 (–4.65 to 2.01)
|
–1.82 (–5.15 to 1.50)
|
–1.82 (–5.14 to 1.48)
|
–1.88 (–5.20 to 1.43)
|
–2.07 (–5.39 to 1.24)
|
–0.96 (–4.40 to 2.47)
|
Strength/Power
|
|
b) strength and/or power trained subjects
|
|
Control
|
|
|
|
|
|
|
|
|
3.08 (1.58 to 4.58)
|
Complex
|
|
|
|
|
|
|
|
3.86 (2.36 to 5.36)
|
0.78 (–0.54 to 2.10)
|
Contrast
|
|
|
|
|
|
|
3.81 (1.81 to 5.80)
|
0.72 (–1.11 to 2.56)
|
–0.05 (–2.07 to 1.96)
|
Compound
|
|
|
|
|
|
3.79 (2.36 to 5.22)
|
0.71 (–0.20 to 1.62)
|
–0.07 (–1.35 to 1.21)
|
–0.01 (–1.78 to 1.75)
|
Traditional
|
|
|
|
|
3.60 (2.20 to 4.99)
|
0.51 (0.65 to 1.68)
|
–0.26 (–1.71 to 1.18)
|
–0.21 (–1.79 to 1.37)
|
–0.19 (–1.16 to 0.77)
|
Strength
|
|
|
|
2.70 (0.85 to 4.56)
|
–0.37 (–2.21 to 1.37)
|
–1.15 (–3.04 to 0.72)
|
–1.10 (–2.74 to 0.54)
|
–1.08 (–2.78 to 0.60)
|
–0.89 (–2.44 to 0.66)
|
Power
|
|
Note: Mean differences (MD) with their 95% confidence intervals
from the network meta-analysis are shown; A negative MD value favors the
upper-left treatment for any cell, and a positive MD value favors the
lower-right treatment. Relative treatment effect differences are shown
in bold type.
Overall, both the four different approaches, CPX, CT, CP, and TT, as well as the
solely strength or power training interventions, achieved similar changes from
baseline in comparison to control conditions across all analyses ([Fig. 5]). There was no evidence of
inconsistency between direct and indirect comparisons in both networks, where
data from direct and indirect evidence were available. "Small study
bias" could not be assessed due to the low number of trials.
Fig. 5 Findings of network meta-analyses. Change from baseline
countermovement jump (CMJ) height (cm) and 95% confidence
interval (CI) achieved with training interventions as compared with
control (no training). (a) influence of different interventions
on CMJ height and (b) influence of different interventions on CMJ
height in strength and/or power trained subjects.
Six studies [49]
[58]
[59]
[60]
[61]
[62] investigated the effectiveness of different interventions on CMJ
height in subjects without training experience. Three of them compared either
TT, CPX, or CP with CTRL [60]
[61]
[62]. All of them
supported the use of one of these interventions to increase CMJ height when
compared to CTRL [TT (mean±SD): 3.0±3.6 cm; CTRL:
0.0±4.0 cm; CPX: 3.2±6.4 cm; CTRL:
-2.3±6.4 cm; CP: 2.1±3.1 cm; CTRL:
0.2±3.9 cm]. One study [49]
compared CPX with CT and TT. Results suggested that each intervention was
effective to increase CMJ height, but no intervention was superior (TT:
5.3±4.3 am; CPX: 5.1±4.1 cm; CT:
5.5±3.4 cm). One study [59]
compared CPX with PW and CTRL. CPX and PW were equally effective when compared
to CTRL, but no intervention was superior (CPX: 5.6±9.5 cm; PW:
3.8±7.8 cm; CTRL: 0.0±4.6 cm). One study [58] compared CT with solely ST or PW and showed
that both interventions increased CMJ height, but CT was superior (CT:
5.4±5.6 cm; ST or PW: 2.5±5.3 cm).
Discussion
The purpose of this network meta-analysis was to determine the effects of CPX, CT,
CP, and TT in comparison to each other, ST and/or PW alone, and control
conditions, on CMJ performance. The results of 24 RCTs, including 694 healthy
subjects, were incorporated. The analyses indicated that individuals performing
either CPX, CT, CP, or TT significantly increased CMJ height compared to those of
the controlled conditions (no training). However, compared to ST and/or PW
alone or to each other, all interventions yielded similar improvements in both sets
of NMAs.
To the best of the authors’ knowledge, the present study is the first formal
evaluation of the comparative effects of different exercise interventions on CMJ
height. Previous systematic reviews and meta-analyses have examined the CMJ height
effects of CPX, CT or their combination [17]
[23]
[24]
[81]
[82]. Similar to these
reviews, the different training set structures in the present analysis varied
broadly based on the number of exercises, volume, intensity, and duration. In a
recent review and meta-analysis, Marshall et al. [28]
suggested, that CT, CPX and TT are all useful to particularly target athletic
properties. To increase force, the exercise should be carried out with an augmented
level of fatigue leading to training close to failure. This can be induced by
completing multiple sets of a comparable lighter exercise prior to the heavy
exercise sets, which is defined as TT. Enhancing velocity of the lighter exercise
can be achieved, e. g. by combining it with a heavier exercise in a contrast
pair to generate a PAPE effect.
The present findings were in accordance with the meta-analyses by Pagaduan et al.
[23], Pagaduan et al. [24], and Freitas et al. [81], revealing
that both CPX and CT administer an appropriate training stimulus to improve CMJ
height when compared to control conditions. A recent meta-analysis by Bauer et al.
[17] with a large number of trials strengthens the
present findings that no differences between CPX or CT and TT or CP, and
alternative training methods such as ST or PW exist to improve CMJ performance. It
should be noted that authors investigated combined CPX and CT compared to a
combination of TT, CP, ST, and PW.
As mentioned before, ST alone was ranked as one of the best interventions while PW
achieved similar effects for improving the CMJ height compared to other
interventions both in the complete analysis and the exclusive analysis of subjects
with training experience. These results are not surprising since both ST and PW
alone can lead to an increase in muscular power and therefore to an improvement of
vertical jump performance [83]
[84]
[85]. A growing body of evidence
suggests muscular strength as the fundamental component to increase the
athlete’s performance, especially in terms of power production, velocity,
and rate of force development, which is defined as the ability to produce large
forces in a short time [86]
[87]
[88]
[89].
While these power gains become less distinct when higher muscular strength is
achieved, some evidence suggests that squatting at least two times of a
subject’s bodyweight might be a good indicator of an optimal lower body
strength standard, which may lead to more benefits of power exercises like
plyometrics [3]
[90]
[91]
[92]. Therefore, youth
athletes and subjects without training experience might prioritize ST to build a
solid base before focusing on PW or incorporating power exercises in their training
plans, respectively [92]
[93]. Taking these considerations into account, stronger and more strength
and/or power trained subjects might benefit more from CPX, CT, CP, and TT
than their weaker counterparts. This could not be shown in the present
study’s exclusive analysis of strength and/or power trained
subjects, where the combined set structures were not superior to ST or PW alone.
Implications for research
The current research suggests several factors influencing the level of the
PAPE effect in training practice. These factors include ideal parameters of
conditioning activity, such as the optimal type of exercise, optimal intensity
and volume, and rest periods [94]. Thereby,
intensities range from plyometric body-weight to supramaximal loads. Moreover,
individual characteristics of the subjects, involving training experience, the
type of muscle fibers, muscle strength, and fatigue resistance are crucial.
There is inconsistent evidence when strength training load and volume, as well
as recovery periods are being discussed. PAPE effects occur over a wide range of
intensities, with loads around 80 to 90% 1-RM being the most
investigated [94]. In the included studies of the
present analysis, loads over 80% 1-RM were primarily used in strength
and/or power trained subjects ([Table
1]). Studies with non-strength and/or power trained
subjects usually used loads≤80% 1-RM. Evidence revealed that
stronger and more strength and/or power trained athletes show
considerably larger potentiation effects than their weaker and less trained
counterparts [33]
[40], suggesting that differences in strength and training experience
contribute to PAPE effects. If non-strength and/or power trained
subjects would show smaller PAPE effects, their CMJ height may be smaller
compared to strength and/or power trained subjects, regardless of the
type of training intervention. However, due to the low number of studies with
non-strength and/or power trained subjects and the lack of sufficient
study population characteristics based on the training experience, analysis of
non-strength and/or power trained subjects could only be carried out
narratively. The present analysis of strength and/or power trained
subjects showed no different effects on CMJ heights between the compared
training interventions, concluding that all the analyzed regimes could be
recommended to increase CMJ height, which is similar to findings reported
recently [28]. A growing body of evidence
indicates that training experience influences the outcomes achieved by ST and PW
[95]
[96]. While
non-strength and/or power trained subjects may getting stronger and more
athletic primarily through the neural adaptations, the influence of muscular
strength on an athlete’s performance may diminish when strength
and/or power trained subjects already maintain high strength levels
[92]. Future research should provide precise
study population characteristics to distinguish between the effects the
combination of ST and PW has on non-strength and/or power trained
subjects and strength and/or power trained subjects.
The training volume is considered the number of performed repetitions, which may
vary depending on movement speed, in a set or a complete session in addition to
the time under tension [97]
[98]. The number of repetitions and sets of the
interventions varied between studies in the present analysis, leading to the
assumption that strength and power training volume may not be the main factor
for improving CMJ height. The effects of PAPE on resistance training volume
remain unclear, because they have only been investigated for the upper body
[99], demonstrating the need for studies of
the lower extremity.
The recovery periods of the CPX and CT interventions in the present study
differed from current research recommendations on potentiation. It was reported,
that it should last at least five minutes [33]
[40]. Formerly, a recovery duration
of eight to twelve minutes after the conditioning activity was reported to
generate the greatest PAPE effect [100]. The
included studies on CPX and CT interventions used average rest periods of about
two minutes, ranging from 30 seconds to five minutes [46]
[47]
[48]
[49]
[58]
[59]
[61]
[63]
[64]
[70]
[71]
[73]
[77]
[78]. Therefore,
athletes potentially elicited lower PAPE levels, which in turn could have been
limiting their potential to adapt to CPX and CT interventions. However, in a
recent study it was found that a PAPE recovery time course of one minute after
squat sets within different contrast resistance training schemes revealed no
adverse effect on subsequent drop jump performance when compared to a four
minute recovery duration [101].
Limitations
This study has limitations that needed to be mentioned. First, the
studies’ quality was low with the possible risk of bias, which leads to
limitations in terms of the informative value of the present findings. With
high-quality studies, the certainty that the effect estimates from the analyses
are the true effects is supposed to be higher. Then, altering the estimate would
be less probable if more studies would be added. Unfortunately, large
high-quality studies are rare in the field of sports science. This limitation
was attempted to be overcome by using the GRADE approach. Since the quality of
the evidence was low to very low in the present study, future modifications of
the treatment rankings and the effect sizes should be considered.
Second, although the different interventions were well defined, the frequency,
intensity, as well as exercises, total number of exercises, and training loads
differed. This may have led to different adaptations of resistance training, as
it is load specific, with higher loads leading to greater strength gains [18]. The optimal range is between 80 and
95% of 1-RM [39]. However, ST has been
defined as exercises involving the lower extremity with an average
load>60% of 1-RM, as maximal strength gains are mainly achieved
with training loads>60% of 1-RM [21]. Some interventions incorporated weightlifting exercises into
strength training programs. Since power output is one of the primary goals of
weightlifting, loads are selected with this goal in mind. Power is a product of
force and velocity, and there is an inverse relationship between the two. In
fact, it has generally been found that power output during weightlifting
exercises is greatest at loads of 70 to 85% of 1-RM for snatch or clean
exercises [18]
[22].
Although typical loads of 70 to 85% of 1-RM are on the low end of the
optimal range for strength gains (80 to 95% of 1-RM), several studies
suggest that when compared, both traditional strength training and weightlifting
can lead to similar improvements in strength, power, and overall fitness [22]
[102].
Studies in the present analysis reporting intensities on the lower spectrum
(average training load of 60%) are those with non-strength
and/or power trained subjects. This is in accordance with Rhea et al.
[96], who reported that subjects with no
experience in resistance training make the most of their strength gains with
mean intensities of 60% of 1-RM. However, the influence of training
frequency remains unclear. It seems that, especially in non-strength
and/or power trained individuals, higher frequencies in terms of
training volume are likely to result in greater muscle strength gains [103].
Conclusion
The present network meta-analysis confirms that CPX, CT, TT, and CP have a
beneficial effect on CMJ performance compared to control condition (no
intervention). However, none of these interventions seem to be superior compared to
each other, or to strength or power training alone, or to strength and power
training combined, in non-strength and/or power trained subjects as well as
strength and/or power trained subjects. These conclusions can only be drawn
from low to very low-quality evidence and should therefore be interpreted
cautiously. Nonetheless, the present findings are mainly important for
practitioners because the choice of how ST and PW exercises could be incorporated
in
one training regime might be decided on individual preference. Furthermore, coaches
and athletes can potentially switch the approaches and bring greater variety to
their training programs. More high-quality research on combined ST and PT should be
conducted to confirm and possibly extend this systematic review results.
Future studies should focus on longer intervention durations with clear distinction
of non-strength and/or power trained participants and strength
and/or power trained subjects. Further work is also needed to understand how
PAPE can be maximized in terms of optimal load and volume, and recovery periods. In
summary, the present findings support the combination of ST and PW to improve
CMJ height in healthy subjects.
