Zebis MK, Skotte J, A ndersen CH, et al. Br J Sports Med (2012). doi:10.1136/bjsports-2011-090281 1 of 7
1Institute of Sports Science
and Clinical Biomechanics,
University of Southern
Denmark, Odense, Denmark
2Gait Analysis L aboratory,
Hvidovre Hospital, Hvidovre,
3National Research Centre
for the Working Environment,
4School of Physiotherapy,
Mette Kreutzfeldt Zebis,
University of Southern
Denmark, Institute of
Sports Science and Clinical
Biomechanics, Campusvej 55
DK-5230, Odense, Denmark;
Received 1 June 2011
Accepted 25 April 2012
Kettlebell swing targets semitendinosus and supine
leg curl targets biceps femoris: an EMG study with
Mette Kreutzfeldt Zebis,1,2 Jørgen Skotte,3 Christoffer H Andersen,3
Peter Mortensen,3 Maria Højland Petersen,4 Tine C Viskær,4 Tanja L Jensen,4
Jesper Bencke,2 Lars L Andersen3
Background The medial hamstring muscle has the
potential to prevent excessive dynamic valgus and exter-
nal rotation of the knee joint during sports. Thus, speciﬁ c
training targeting the medial hamstring muscle seems
important to avoid knee injuries.
Objective The aim was to investigate the medial and
lateral hamstring muscle activation balance during 14
selected therapeutic exercises.
Study design The study design involved single-
occasion repeated measures in a randomised man-
ner. Sixteen female elite handball and soccer players
with a mean (SD) age of 23 (3) years and no previ-
ous history of knee injury par ticipated in the present
study. Electromyographic (EMG) activity of the lateral
(biceps femoris – BF) and medial (semitendinosus –
ST) hamstring muscle was measured during selected
strengthening and balance/coordination exercises, and
normalised to EMG during isometric maximal voluntary
contraction (MVC). A two-way analysis of variance
was performed using the mixed procedure to deter-
mine whether differences existed in normalised EMG
between exercises and muscles.
Results Kettlebell swing and Romanian deadlift
targeted speciﬁ cally ST over BF (Δ17–22%, p<0.05) at
very high levels of normalised EMG (73–115% of MVC).
In contrast, the supine leg curl and hip extension speciﬁ -
cally targeted the BF over the ST (Δ 20–23%, p<0.05)
at very high levels of normalised EMG (75–87% of MVC).
Conclusion Speciﬁ c therapeutic exercises targeting
the hamstrings can be divided into ST dominant or BF
dominant hamstring exercises. Due to distinct functions
of the medial and lateral hamstring muscles, this is an
important knowledge in respect to prophylactic training
and physical therapist practice.
In recent years, a huge step is taken towards pre-
vention of serious sports injuries like anterior
cruciate ligament (ACL) injuries. Several high-
quality studies have documented that speciﬁ c
prophylactic training can reduce the incidence of
ACL injuries.1 2 The majority of validated train-
ing protocols involve balance, coordination,
strength and plyometric exercises.1 2 However, it
is not known how each exercise challenges the
neuromuscular system. In order to further opti-
mise prophylactic training, an extensive exercise
evaluation is needed based on an anatomical and
biomechanical rationale. Thus, there is a strong
need to gain knowledge about the muscle activity
pattern during exercises used in both ACL preven-
tion and rehabilitation interventions.
During forceful dynamic movements, coacti-
vation of the hamstrings is important to provide
dynamic knee joint stabilisation and to prevent
excessive ACL shear forces.3 4 Thus, the hamstri ng
muscles are considered ACL-agonists, and great
attention has been directed towards this muscle
group in the prevention of ACL injury.5
The hamstring muscles comprise the biceps
femoris muscle caput longus (BFcl) and brevis, the
semimembranosus muscle and the semitendino-
sus muscle (ST). Anatomically, the hamstrings
are a two-joint muscle group acting at the k nee
and hip joint. Differences in muscle architecture
and insertion sites between the medial and lat-
eral hamstring muscles translate into distinct
functions.6 7 Consequently, in prevention of ACL
injur y it is important to distinguish between the
different muscles comprising the hamstrings.8
The ST seems important in respect to preven-
tion of ACL injury. Contraction of the ST has the
potential to compress the medial k nee joint com-
part ment and thereby reducing the risk of dynamic
knee joint valgus.9 Dynamic knee joint valgus
during a landing phase has been shown to predis-
pose for an ACL injur y among female athletes.10
Concordantly, low-ST electromyographic (EMG)
activ ity duri ng sidecutt ing – a movement asso ciated
with the incidence of non-contact ACL injury – in
combination with high vastus lateralis (VL) activ-
ity may increase the risk for sustaining ACL inju-
ry.8 Finally, a neuromuscular programme shown to
reduce the incidence of ACL injury among female
athletes1 was found to induce a selective upregula-
tion of ST activation during sidecutting.11
Common strength exercises for the leg muscles
such as sq uats, leg presses a nd knee exte nsions show
h ig h l ev el s of mu sc le ac t iv it y in t he qu ad r ic ep s a lo n g
with a preferential higher lateral compared with
medial hamstring muscle coactivation.12 Focusing
primarily on these exercises may predispose for
knee injury. Thus, from a prophylactic perspective
it is relevant to evaluate the ST-BFcl activation bal-
ance during commonly used therapeutic exercises.
The aim of this study is to evaluate the level
of medial and lateral hamstring muscle activation
during selected exercises used in prophylactic and/
or rehabilitating interventions.
BJSM Online First, published on July 6, 2012 as 10.1136/bjsports-2011-090281
Copyright Article author (or their employer) 2012. Produced by BMJ Publishing Group Ltd under licence.
Zebis MK, Skotte J, A ndersen CH, et al. Br J Sports Med (2012). doi:10.1136/bjsports-2011-0902812 of 7
Sixteen young females (mean age: 23±3 year; mean height:
170.2±6.4 cm; mean weight: 66.2±7.4 kg) with no previous
history of knee injury or hamstring injury participated in the
study. The subjects were either elite handball players (n=8)
or elite soccer players (n=8), and had a training frequency of
4.7±0.7 sessions/week. Subjects had in average participated in
their respective sports for 15.6±4.1 years, and regular strength
training for 5.4±2.4 years.
The study was approved by the local ethical committee
(HC2008103) and conformed to the Declaration of Helsinki.
All subjects were informed about the purpose and content of
the project and gave written informed consent to participate
in the study.
Tes t p roc edure
On a separate day before the actual test, participants were
familiarised with the full protocol. On the day of testing par-
ticipants performed three maximal voluntary isometric mus-
cle contractions for the hamstring muscles (MVCs) with strong
verbal encouragement, and we used the highest value of the
three attempts for the normalisation. Participants warmed up
before the MVC’s. The participants warmed up during sub-
maximal execution of the respective exercises. MVC of the
knee ﬂ exors were performed to obtain maximal EMG signal
amplitude. Then three trials of each exercise were performed
in a randomised manner. The same person instructed the exer-
cises for each individual. A 3 min rest period was prescribed
between exercises to avoid fatigue.12 To ensure that fatigue
was avoided, the subject was asked to rate perceived exertion
on a Borg 10-scale,13 after the exercise protocol was accom-
plished, which showed a mean value of 2.5±1.1 corresponding
to fairly light to moderate perceived exertion.
If a trial was performed incorrectly, in respect to the instruc-
tion, it was disqualiﬁ ed and a new trial was performed.
The skin of the subject was shaved with a hand razor and
carefully cleaned with ethanol before electrode placement
on the preferred push-off leg. Bipolar surface EMG electrodes
(Neuroline 720 01-K, Medicotest A/S, Ølstykke, Den mark)
were placed according to standardised procedures14 with a
2.0 cm interelectrode distance on the BFcl and ST muscles
representing the lateral and medial hamstring muscle groups,
respectively (ﬁ gure 1A). Interelectrode resistances were
ensured to be less than 10 kΩ. Due to the more profound
nature of semimembranosus, we are not able to detect valid
EMG signals for the present muscle with the use of surface
The EMG signals were preampliﬁ ed, band-pass ﬁ ltered
(5– 450 Hz) and sampled (16-bit) with frequency of 1000 Hz.
The digitised EMG recordings were high-pass ﬁ ltered at a cut-
off frequency of 10 Hz, rectiﬁ ed and low-pass ﬁ ltered using a
second order two-way Butterworth ﬁ lter with a cut-off fre-
quency corresponding to a running average window size of
Maximal voluntary contraction
EMG activity was obtained during maximal voluntary isomet-
ric muscle contraction (MVC) for the hamstrings in a Biodex
Medical isokinetic dynamometer (System 3 Pro, Brookhaven
R&D Plaza, New York, USA). The subject was in a prone
position with a hip joint angle of 10° of ﬂ exion and the k nee
joint angle locked in a 45° position from full extension (0°) (ﬁ g-
Off-line signal processing
For each exercise trial, the ﬁ ltered EMG signals were normalised
to EMG recordings obtained during MVC. Peak levels of the ﬁ l-
tered, normalised EMG were determined as the average of three
trials performed for each examined exercise. The term peak
nEMG refers to the peak levels of the ﬁ ltered normalised EMG.
Electromagnetic motion analysis
Kinematics was measured with a magnetic tracking device
(Ascension Tech Model 3D Guidance Trakstar, Ascension
Technology Corp., Burlington, VT, USA), which included a
syste m electronics un it, a mid-range transmit ter, and th ree sen-
sors (model 800). The measurement rate was 33 Hz. The trans-
mitter was positioned on a wooden stand next to the subject’s
knee and the sensors were attached on the preferred push-off
leg using double-sided tape. The sensors were positioned on
Figure 1 (A) The sur face electromyographic electrode placement on
the examined muscles; biceps femoris muscle caput longus (BFcl) and
semitendinosus (ST). (B) Maximal voluntar y contraction for the knee
ﬂ exors was obtained in an isokinetic dynamometer with the subject in
a prone position.
Zebis MK, Skotte J, A ndersen CH, et al. Br J Sports Med (2012). doi:10.1136/bjsports-2011-090281 3 of 7
the iliac crest, the distal femur and on the proximal part of
ﬁ bula. Hip and knee ﬂ exion angles were obtained from the
rotational matrix of the three sensors. Reference angles were
achieved by an initial trial in which the subject was standing
in normal upright position.
Figure 2A,B illustrate examples of the raw EMG – with cor-
responding hip joint angles – and the nEMG, respectively,
during the execution of one of the exercises (kettlebell swing;
exercise procedures are described below).
The exercise protocol included nine strength and ﬁ ve balance/
coordination exercises as described below. The exercises were
chosen based on exper iences from clinical prac tice and tra ining
as well as prevention programmes presented in the literature.
The selected balance/coordination exercises have been used in
ACL prevention programmes,1 2 and rehabilitation interven-
tions after ACL injury.16
A balance mat (40×50 cm; 7 cm thick; Alusuisse Airex, Sins,
Switzerland, 2000) was used in three of the ﬁ ve balance/coor-
One-leg jump onto balance mat (FJU)
The subject makes a one-legged forward jump onto a balance
mat. The subject must focus on stabilising the knee in the
frontal plane – the k nee over toe position – in the landing and
maintaining stability for 3 s (ﬁ gure 3A).
One-leg landing from box on balance mat (LU)
The subject steps down from a 37.5 cm high box with a one-leg
landing on a balance mat, focusing on a ﬂ exed hip and knee
posture in the landing. The subject must focus on stabilising
the knee in the frontal plane – the knee over toe position – in
the landing and maintaining stability for 3 s (ﬁ gure 3A).
One-leg drop jump on balance mat (DJU)
The subject ‘steps’ down on one leg from a 37.5 cm box onto
a balance mat and immediately jump straight up as high as
possible before ﬁ nally landing again on the balance mat on
the same leg. In the ﬁ nal landing, the subject must again focus
on stabilising the k nee in the frontal plane – the knee over
toe position – in the landing and maintaining stability for 3 s
(ﬁ gure 3A).
One-leg forward jump (FJ)
The subject jumps for ward focusing on stabilising the knee
in the frontal plane – the knee over toe position – in the land-
ing and maintaining stability for 3 s. The subject jumps back-
wards to the starting point focusing on knee stabilisation and
balance (ﬁ gure 3A).
One-leg side jump (SJ)
The subject stands on one leg and then jumps sideways
between two ﬁ xed points focusing on stabilising the knee in
the frontal plane – the knee over toe position – in the landing
and maintaining stability for 3 s (ﬁ gure 3A).
Two-hand kettlebell swings (KS)
The kettlebell swing was performed with either a 12 kg or
16 kg kettlebell according to strength level of the respective
subjects. The kettlebell weight corresponds to the weight the
subject can swing 20 times with proper technique as described
The subject stands in front of the kettlebell with their feet
parallel a shoulder width apart. By ﬂ exing the hips and knees
while keeping the spine in a neutral position, the subject
reaches down and grasps the kettlebell with both hands. The
upper body is at this point is parallel to the ﬂ oor and the knees
are slightly ﬂ exed (~10–15°). The subject forcefully swings
the kettlebell back between the legs and quickly reverses
the direction with an explosive extension of the hips swing-
ing the kettlebell out to chest level where the hips and knees
are extended and the subject is standing upright (ﬁ gure 3B).
Kettlebell swings (KS) exercise has been used to relief neck/
shoulder and low-back pain among adults with musculoskel-
etal pain symptoms.17
Figure 2 The raw electromyographic (EMG) and the nEMG, respectively, during two different single kettlebell swings. (A) A typical example
of raw EMG activity during kettlebell swing. Grey line represents semitendinosus (ST), black line represents biceps femoris (BF) and dotted line
represents hip joint angle. (B) A typical example of the nEMG during kettlebell swing. Grey line represents ST and black line represents BF.
Zebis MK, Skotte J, A ndersen CH, et al. Br J Sports Med (2012). doi:10.1136/bjsports-2011-0902814 of 7
Figure 3 The examined exercises. (A) FJU, one-leg jump onto balance mat, LU, one-leg landing from box on balance mat, DJU, one-leg drop
jump on balance mat, FJ, one-leg forward jump, SJ, one-leg side jump. (B) SPL, supine pelvis lifts, KS, two-hand kettlebell swings, NH, nordic
hamstring lowers, SuLC, supine one-leg curls. (C) HE, hyperextensions off table, HEB, hyperex tensions off table with barbell, RD, Romanian
deadlift, SeLC, seated leg curl in Biodex isokinetic dynamometer, PrLC, prone leg curl in biodex isokinetic dynamometer.
Nordic hamstring lowers (NH)
The subject is kneeling on a balance mat while the partner
holds the ankles. The subject leans the upper body slowly for-
ward, using the hamstrings to resist falling forward as long as
possible (ﬁ gure 3B). Nordic hamstring lowers (NH) exercise is
used both in ACL prevention programmes,2 and in hamstring
injury prevention interventions.18 19
Supine one-leg curls (SuLC)
The subject is lying supine on the ﬂ oor with the arms by their
sides. The hips are extended and lifted off the ﬂ oor and one leg
is straightened. The other leg – which is the working leg – is
kept on the ﬂ oor with the foot on a carpet tile which can slide
easily on the ﬂ oor. The subject now slides the foot backwards
and forwards (ﬁ gure 3B).
Zebis MK, Skotte J, A ndersen CH, et al. Br J Sports Med (2012). doi:10.1136/bjsports-2011-090281 5 of 7
Supine pelvis lifts (SPL)
The subject is lying supine on the ﬂ oor with arms by the sides.
The knee of the working leg is bent with the foot kept on the
ﬂ oor. The other leg is bent and crossed over the working leg.
The hips are lifted off the ﬂ oor until knees, hips and shoulders
are in a straight line (ﬁ gure 3B).
The supine pelvis lifts (SPL) and SuLC exercises are believed
to strengthen and rehabilitate the hamstring and gluteus mus-
cles, and are commonly used for patients with hip and knee
Romanian deadlift (RD)
The subject is standing close to the barbell (weight correspond-
ing to 12 R M) with the feet parallel a shoulder width apart. By
ﬂ exing the hips and pushing them backwards while keeping
the upper body straight, the subject reaches down and grasps
the bar. At this point, the knees are slightly ﬂ exed (~10–15°).
The bar is lif ted by extending at hips and k nees until standing
upright (ﬁ gure 3C). Deadl ift has been suggested as an effective
closed kinetic chain exercise for strength athletes to be used
during knee rehabilitation.20
Hyperextensions off table (HE)
The subject is placed prone with the legs ﬁ xated on a table
and the hip and upper body off the table’s edge. The body is
lowered by bending the hip until a stretch is felt in the ham-
strings. The upper body is raised again until the hip is fully
extended and the upper body is kept straight throughout the
entire movement (ﬁ gure 3C).
Hyperextensions off table with barbell (HEB)
The exercise is performed as above but with a barbell (weight:
13.3 kg) placed in a wide grip in the hands for added resistance
(ﬁ gure 3C).
Seated leg curl (SeLC)
The subject is seated in a Biodex Medical isokinetic dynamom-
eter (System 3 Pro, Brookhaven R&D Plaza) with 80° of ﬂ ex-
ion in the hip joint (refer ﬁ gure 3C). With the back of her calf,
she pushes against the padded arm of the dynamometer and
performs maximal voluntar y concentric contraction at 60° per
second. The range of motion (ROM) is set at 80° starting from
a 10° ﬂ exion in the knee to 90° ﬂ exion.
Prone leg curl (PrLC)
The subject is placed prone in a Biodex Medical isokinetic
dynamometer (System 3 Pro, Brookhaven R&D Plaza) with no
ﬂ exion in the hip joint (ﬁ gure 3C). With the back of her calf,
she pushes against the padded arm of the dynamometer and
performs maximal voluntar y concentric contraction at 60° per
second. The ROM is set at 80° starting from a 10 degree ﬂ exion
in the knee to 90° ﬂ exion.
The above mentioned strength exercises – hyperextensions
off table (HE), seated leg curl (SeLC) and prone leg curl (PrLC)
– are commonly used in resistance training by both novice
and experienced individuals,21 as wel l as in rehabilitation af ter
ACL injur y.16
Before the main analyses, Shapiro-Wilk testing for normal-
ity was performed. For the hamstrings a two-way analysis
of variance was performed in SAS (version 9, SAS Institute,
Cary, North Carolina, USA) using the mixed procedure to
determine whether differences existed in peak nEMG, and
knee and hip joint angle at peak n EMG between exercises
and muscles. Factors included in the model were muscle (ST
and BF) and exercise (14 exercises), as well as the interaction
between these. When a signiﬁ cant ma in effect was fo und , pos t
hoc analyses were performed to locate differences between
muscles, expressed as delta values (Δ). Intraday reliability was
assessed by intraclass correlation coefﬁ cients (ICC).
The critical p value was set to 0.05.
A pr io ri po we r a na l ys is sho we d t h at 16 su bj ec t s i n t hi s pa ired
design were sufﬁ cient to obtain a statistical power of 80% at
a min imal relevant difference between exercises of 10% with
an α level of 5%.22 Results are reported as group means (±SD)
unless otherwise stated.
Intraday reliability (ICC)
The intraday reliability of peak nEMG was calculated as ICC
for the three attempts of each exercise, and ranged from 0.783
to 0.941 (p<0.05).
Level of hamstring muscle activation
Testing of main effects showed a signiﬁ cant effect of the type
of exercise on nEMG (F=21, p<0.0001). For the BFcl high lev-
els of muscle activation (ie, above 60% nEMG) were observed
during – in ascending order – DJU, FJU, SPL, HE, NH, PrLC,
KS, HEB, SeLC, SuLC.
For the ST, high levels of muscle activation were observed
during – in ascending order – FJU, HE, RD, SPL, NH, PrLC,
SeLC, HEB, SuLC, KS. Levels of muscle activation are sum-
marised in table 1.
Medial-lateral hamstring activation balance
Testing of main effects showed a signiﬁ cant muscle by exer-
ci se int erac tion on peak nEMG (F=1.88 , p<0.05). Post hoc anal-
yses revealed a preferential activation of the ST over the BFcl
(Δ±SD) during RD (Romanian Deadlift) (Δ 17.3±9.7% peak
nEMG, p<0.05) and Kettle Ball Swings (KS) (Δ 22.5±9.7% peak
nEMG, p<0.05). A preferential activation of the BFcl over the
ST was observed during supine leg curl (Δ 22.9±9.4% nEMG,
p<0.05) and hip extension (Δ 20.4±8.8% nEMG, p<0.05). For
the remainder of exercises there was no signiﬁ cant difference
between n EMG of the ST compared with the BFcl (ﬁ gure 4).
Knee and hip angles at peak nEMG
During the FJU and DJU exercises, peak normalised ST EMG
occurred at a more extended knee joint angle compared
with BFcl EMG (table 1). For the remainder of the exercises
no signiﬁ cant differences in knee or hip joint angles at peak
nEMG were observed.
The main ﬁ nding of our study was that speciﬁ c therapeutic
exercises targeting the hamstrings can be divided into ST
dominant and BFcl dominant hamstring exercises.
KS and RD targeted the ST speciﬁ cally and at intensit y levels
able to stimulate muscle strength gains.23 The ST dominance
of the two exercises may partly be explained by the fact that
ST is, in contrast t o BFcl, pa rallel ﬁ bered with long-ﬁ bre lengths
and a high number of sarcomeres in series. This arrangement
increases the total shortening capacity and absolute velocity of
contraction for the ST muscle.24 Thus the potential to shorten
at long distances is excellent for the ST muscle. During execu-
tion of KS and RD, the hamstrings are extensively stretched
with the highest load in the most stretched position – that is
hip ﬂ exed. Thus, we recommend these exercises when aiming
Zebis MK, Skotte J, A ndersen CH, et al. Br J Sports Med (2012). doi:10.1136/bjsports-2011-0902816 of 7
at enh anced ability to recruit ST dur ing forcef ul movements. In
respect t o a transfer effect to real-life spor ts activit y, t he KS may
be superior to RD, due to a high-velocity training component.
In contrast, the Supine Leg Curl (SuLC) and hip extension
(HE) speciﬁ cally targeted BFcl. Hamstring muscle strain inju-
ries are common in sports involving sprinting and jumping.25
Exercises that mostly affect BFcl may be important in respect
to prevention of hamstring muscle strain injury since most
acute strains involve BFcl.26 –28 The NH exercise has been
shown to decrease the rate of overall, new and recurrent acute
hamstring injuries,19 and has been established as an important
prophylactic exercise.29 Muscle strain injuries most frequently
occur during eccentric muscle contractions.30 Although the
medial and lateral hamstrings were equally targeted during
the Nordic Hamstring (NH) exercise, the muscles act solely
Figure 4 Medial-lateral hamstring activation balance. *Denotes
signiﬁ cant difference in activation of the semitendinosus and the
biceps femoris muscle caput longus during Romanian deadlift (RD),
kettlebell swings (KS), supine leg curl (SuLC) and hip extension (HE),
Table 1 Peak normalised EMG (nEMG) activity of ST and BFcl and concomitant joint angles during the
examined exercises. Expressed as mean (± SD). *Denotes signiﬁ cant dif ference between ST and BFcl,
Peak nEMG (% of max )
Knee Joint An gle at pe ak
Hip Joint Angl e at peak
ST BFcl ST BFcl ST BFcl
Hip extensions 67 (32) 87* (30) 0 (4) 1 (5) 15 (10) 12 (11)
Hip ext ension s+Barbell 92 (47) 105 (29) 1 (5) 1 (5) 17 (11) 15 (10)
Kettlebell swings 115 (55) 93* (31) 7 (8) 7 (8) 72 (28) 70 (27)
Nordic hamst ring 82 (23) 91 (21) 67 (7) 63 (14) 10 (7) 11 (9)
Supine leg curl 98 (20) 121* (33) 40 (23) 27* (23) 3 (6) 3 (10)
Romanian deadlift 73 (32) 56* (20) 0 (5) 0 (5) 66 (23) 65 (23)
Seate d leg cur l 91 (23) 106 (58) 46 (2 8) 44 (27) 80 (0) 80 (0)
Prone l eg curl 83 (40) 92 (36) 28 (25) 37 (27) 10 (0) 10 (0)
Supine pelvis lift 80 (22) 75 (20) 66 (15) 65 (16) 12 (12) 8 (13)
Sidewise one-leg ju mp 44 (19) 57 (23) 27 (9) 32 (12) 30 (16) 33 (17)
Forw ard one -leg jump 45 (18) 55 (21) 22 (12) 28 (8) 22 (16) 27 (14)
Forward one -leg jump onto
65 (22) 70 (21) 23 (13) 31* (8) 23 (13) 27 (15)
One- leg landing from box on
48 (25) 46 (18) 31 (12) 31 (11) 26 (24) 26 (24)
One- leg drop jump on
57 (21) 64 (19) 23 (16) 37* (11) 24 (24) 34 (23)
BFcl, biceps femoris muscle ca put lon gus; EMG, electromyography; ST, semi tendinosus.
eccentrically which makes this exercise particularly important
in relation to prevention of strains.
Isolated leg curls using training machines are widely used
and recommended in clinical rehabilitation after knee injury.
Leg curl in a prone position (PrLC) has been reported to equally
target the medial and lateral hamstring muscle, and the SeLC
has been reported to speciﬁ cally target the ST.31 32 Using a
Biodex isokinetic dynamometer we did not detect any differ-
ence between the prone and seated leg curl (SeLC) in respect to
the level of ST and BFcl activation. Thus, differences may exist
between isokinetic and isotonic training devices.
The main purpose of the balance/coordination exercises is
developing a proper tech nique in the one-legged landi ng phase.
All the examined balance/coordination exercises displayed an
equivalent degree of ST and BFcl activation, and may be pro-
claimed ‘universal’ hamstring exercises. Thus, the exercises
may be important when aiming at establishing – and modi-
fying existing – motor programmes. Although the latency of
the ACL ligamentomuscular reﬂ ex arc (>100 ms) seems too
long to provide a protective mechanism per se for the ACL
in acute situations,33 afferent feedback from the ACL poten-
tially plays an important role in the updating and formation
of preprogrammed motor patterns for optimising knee joint
stabilisation.34 Interestingly, two of the supervised balance/
coordination exercises – FJU and DJU – displayed peak ST and
BFcl nEMG at different knee joint angles. In respect to execu-
tion of the two exercises, ST EMG peaked at more extended
knee joint angles – that is earlier in the landing phase – which
may represent a protection mechan ism against dynamic knee
joint valgus. Thus, the present exercises may further optimise
preprogra mmed motor patterns in respect to t he timing of neu-
romuscular activity. In fact, a re-programming of movement
pattern and neuromuscular activit y has been found due to
coordination/balance and jump exercises,11 when performed
as outlined in a study shown to reduce the incidence of ACL
injur y among female athletes.1
After ACL injury, the ligament is typically reconstructed by
harvesting the ST tendon.35 Although the regeneration of the
Zebis MK, Skotte J, A ndersen CH, et al. Br J Sports Med (2012). doi:10.1136/bjsports-2011-090281 7 of 7
ST tendon has been conﬁ rmed in MRI studies, the volume of
the ST in the reconstructed limb is signiﬁ cantly smaller than
in the normal limb.6 Thus, atrophy and shortening of the ST
after its tendon has been harvested, as well as insufﬁ cient com-
pensation from the semimembranosus and biceps femoris, due
to architectural and functional differences, makes it even more
important, in terms of rehabilitating physical therapy, to intro-
duce exercises wit h prefe renti ally hig h level s of acti vation of t he
medial hamstri ng. However, to avoid adverse effects in the early
phase of rehabilitation, caution must be taken when introduc-
ing high-intensit y training among ACL reconstructed patients.
One limitation of the present study is that the inherent vari-
ance in surface EMG measurements makes it difﬁ cult to eval-
uate small differences in muscle activity between exercises.
Furthermore, as the present study evaluated the exercises
among female athletes, we cannot conclude that the ﬁ ndings
presented can be transferred directly to male athletes.
In conclusion, speciﬁ c therapeutic exercises targeting the ham-
strings can be divided into ST dominant (Kettleball Swing and
Romanian Deadlift) or BF dominant (Supine Leg Curl and hip
extension) hamstring exercises. In respect to the prevention
of ACL injur y, the ability to activate ST during forceful and
explosive movements could be en hanced by ST dominant
exercises like the KS. However, further research is needed to
examine the transfer effect to real-life sports activity, and to
test the efﬁ cacy of the exercises in the prevention of injuries.
Correction notice This paper has been cor rected since it was published Online
First. The fourth author’s name was incorrect, her name is Maria Højland Petersen.
Contributors MKZ, LLA, PM, JB and CHA contributed to conception and design.
JS, MKZ and LLA contributed to analysis and interpretation of data. TCV, TLJ and
MHP recruited all participants and supervised the exercise protocol. MKZ draf ted
the article and all authors revised it critically for important intellectual content and
gave ﬁ nal approval of the version submitted.
Competing interests None.
Patient consent Obtained.
Ethics approval The study was approved by the local ethical commit tee
Provenance and peer review Not commissioned; externally peer reviewed.
1. Myklebust G, Engebretsen L, Braekken IH, et al. Prevention of anterior cruciate
ligament injuries in female team handball players: a prospective intervention study
over three seasons. Clin J Sport Med 2003;13:71–8.
2. Olsen OE, Myklebust G, Engebretsen L, et al. Exercises to prevent lower limb injuries
in youth sports: cluster randomised controlled trial. BMJ 2005;330:449.
3. Draganich LF, Vahey JW. An in vitro study of anterior cruciate ligament strain
induced by quadriceps and hamstrings forces. J Orthop Res 1990;8:57–63.
What this paper adds
The present study ﬁ nds that speciﬁ c therapeutic exercises ▶
targeting the hamstrings can be divided into semitendinosus
dominant or biceps femoris dominant hamstring exercises.
Due to the distinct functions of the medial and lateral ham-
string muscles, this is an important knowledge in injury pre-
vention and rehabilitation programmes.
4. More RC, Karras BT, Neiman R, et al. Hamstrings–an anterior cruciate ligament
protagonist. An in vitro study. Am J Sports Med 1993;21:231–7.
5. Hewett TE, Zazulak BT, Myer GD, et al. A review of electromyographic activation
levels, timing differences, and increased anterior cruciate ligament injury incidence in
female athletes. Br J Sports Med 2005;39:347–50.
6. Makihara Y, Nishino A , Fukubayashi T, et al. Decrease of knee ﬂ exion torque in
patients with ACL reconstruction: combined analysis of the architecture and function
of the knee ﬂ exor muscles. Knee Surg Sports Traumatol Arthrosc 2006;14:310–17.
7. Onishi H, Yagi R, Oyama M, et al. EMG-angle relationship of the hamstring muscles
during maximum knee ﬂ exion. J Electromyogr Kinesiol 2002;12:399–406.
8. Zebis MK, Andersen LL, Bencke J, et al. Identiﬁ cation of athletes at future risk of
anterior cruciate ligament ruptures by neuromuscular screening. Am J Sports Med
9. Buchanan TS, Kim AW, Lloyd DG. Selective muscle activation following rapid varus/
valgus perturbations at the knee. Med Sci Sports Exerc 1996;28:870–6.
10. Hewett TE, Myer GD, Ford KR, et al. Biomechanical measures of neuromuscular
control and valgus loading of the knee predict anterior cruciate ligament injury risk in
female athletes: a prospective study. Am J Sports Med 2005;33:492–501.
11. Zebis MK, Bencke J, Andersen LL, et al. The effects of neuromuscular training on
knee joint motor control during sidecutting in female elite soccer and handball players.
Clin J Sport Med 2008;18:329–37.
12. Andersen LL, Magnusson SP, Nielsen M, et al. Neuromuscular activation in
conventional therapeutic exercises and heavy resistance exercises: implications for
rehabilitation. Phys Ther 2006;86:683–97.
13. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc
14. Hermens HJ, Freriks B, Disselhorst-Klug C, et al. Development of recommendations
for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol
15. Brereton LC, McGill SM. Frequency response of spine extensors during rapid
isometric contractions: effects of muscle length and tension. J Electromyogr Kinesiol
16. Risberg MA, Holm I, Myklebust G, et al. Neuromuscular training versus strength
training during ﬁ rst 6 months after anterior cruciate ligament reconstruction: a
randomized clinical trial. Phys Ther 2007;87:737–50.
17. Jay K, Frisch D, Hansen K, et al. Kettlebell training for musculoskeletal and cardiovascular
health: a randomized controlled trial. Scand J Work Environ Health 2011;37:196–203.
18. Arnason A, Andersen TE, Holme I, et al. Prevention of hamstring strains in elite
soccer: an intervention study. Scand J Med Sci Sports 2008;18:40–8.
19. Petersen J, Thorborg K, Nielsen MB, et al. Preventive effect of eccentric training on
acute hamstring injuries in men’s soccer: a cluster-randomized controlled trial.
Am J Sports Med 2011;39:2296–303.
20. Escamilla RF, Francisco AC, Kayes AV, et al. An electromyographic analysis of sumo
and conventional style deadlifts. Med Sci Sports Exerc 2002;34:682–8.
21. Fleck SJ, Kraemer WJ. Designing Resistance Training Programs. Human Kinetics
22. Andersen LL, Kjaer M, Andersen CH, et al. Muscle activation during selected strength
exercises in women with chronic neck muscle pain. Phys Ther 2008;88:703–11.
23. Kraemer WJ, Adams K, Cafarelli E, et al. American College of Sports Medicine
position stand. Progression models in resistance training for healthy adults.
Med Sci Sports Exerc 2002;34:364–80.
24. Berne RM, Levy MN. Skeletal Muscle Physiology. In: Physiology. Third edition.
Farrell R, eds. St. Louis, Missouri: Mosby – Year Book, 1993.
25. Petersen J, Thorborg K, Nielsen MB, et al. Acute hamstring injuries in Danish elite
football: a 12-month prospective registration study among 374 players. Scand J Med
Sci Sports 2010;20:588–92.
26. De Smet AA, Best TM. MR imaging of the distribution and location of acute
hamstring injuries in athletes. AJR Am J Roentgenol 2000;174:393–9.
27. Garrett WE Jr, Rich FR, Nikolaou PK, et al. Computed tomography of hamstring
muscle strains. Med Sci Sports Exerc 1989;21:506–14.
28. Malliaropoulos N, Papacostas E, Kiritsi O, et al. Posterior thigh muscle injuries in
elite track and ﬁ eld athletes. Am J Sports Med 2010;38:1813–19.
29. Copland ST, Tipton JS, Fields KB. Evidence-based treatment of hamstring tears.
Curr Sports Med Rep 2009;8:308–14.
30. Garrett WE Jr. Muscle strain injuries: clinical and basic aspects. Med Sci Sports
31. Oliver GD, Dougherty CP. Comparison of hamstring and gluteus muscles
electromyographic activity while performing the razor curl vs. the traditional prone
hamstring curl. J Strength Cond Res 2009;23:2250–5.
32. Tesch P. Muscle Meets Magnet. Stockholm: PA TESCH AB, 1993.
33. Dyhre-Poulsen P, Krogsgaard MR. Muscular reﬂ exes elicited by electrical
stimulation of the anterior cruciate ligament in humans. J Appl Physiol
34. Johansson H, Sjölander P, Sojka P. Receptors in the knee joint ligaments and their
role in the biomechanics of the joint. Crit Rev Biomed Eng 1991;18:341–68.
35. Lind M, Menhert F, Pedersen AB. The ﬁ rst results from the Danish ACL reconstruction
registry: epidemiologic and 2 year follow-up results from 5,818 knee ligament
reconstructions. Knee Surg Sports Traumatol Arthrosc 2009;17:117–24.