Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
RESEARCH ARTICLE Open Access
Kettlebell training in clinical practice: ascoping reviewNeil J. Meigh1* , Justin W. L. Keogh1,2,3, Ben Schram1 and Wayne A. Hing1
Abstract
Background: A scoping review of scientific literature on the effects of kettlebell training. There are no authoritativeguidelines or recommendations for using kettlebells within a primary care setting. Our review objectives were toidentify the extent, range and nature of the available evidence, to report on the types of evidence currentlyavailable to inform clinical practice, to synthesise key concepts, and identify gaps in the research knowledge base.
Methods: Following the PRISMA-ScR Checklist, we conducted a search of 10 electronic databases from inception to1 February 2019. There were no exclusions in searching for publications. A single reviewer screened the literatureand abstracted data from relevant publications. Articles were grouped and charted by concepts and themesrelevant to primary care, and narratively synthesised. Effect sizes from longitudinal studies were identified orcalculated, and randomised controlled trials assessed for methodological quality.
Results: Eight hundred and twenty-nine records were identified to 1 February 2019. Four hundred and ninety-sixwere screened and 170 assessed for eligibility. Ninety-nine publications met the inclusion criteria. Effect sizes weretypically trivial to small. One trial used a pragmatic hardstyle training program among healthy college-ageparticipants. Two trials reported the effects of kettlebell training in clinical conditions. Thirty-three studies explicitlyused ‘hardstyle’ techniques and 4 investigated kettlebell sport. Also included were 6 reviews, 22 clinical/expertopinions and 3 case reports of injury. Two reviewers independently evaluated studies using a modified Downs &Black checklist.
Conclusions: A small number of longitudinal studies, which are largely underpowered and of low methodologicalquality, provide the evidence-informed therapist with little guidance to inform the therapeutic prescription ofkettlebells within primary care. Confidence in reported effects is low to very low. The strength of recommendationfor kettlebell training improving measures of physical function is weak, based on the current body of literature.Further research on reported effects is warranted, with inclusion of clinical populations and investigations ofmusculoskeletal conditions common to primary care. There is a need for an externally valid, standardised approachto the training and testing of kettlebell interventions, which better informs the therapeutic use of kettlebells inprimary care.
Keywords: Scoping review, Kettlebell, Physiotherapy, Exercise
BackgroundHistoryThe kettlebell is a round-shaped steel or cast ironweight, commonly described as resembling a cannonballwith a handle [1]. In Russia, kettlebells are a matter ofpride and a symbol of strength, with a colourful historythroughout the twentieth Century from circus strong
men to the Red Army. Use of kettlebells as measures ofweight dates back to Russia in the 1700s [2] and theword girya (kettlebell) first appears in a Russian diction-ary in 1704 [3], with excavations in Poland pre-datingearly kettlebells to the seventeenth century [4].Kettlebell sport, also referred to as Girevoy Sport origi-
nated in Eastern Europe in 1948 [5]. The InternationalUnion of Kettlebell Lifting World Championship held inOctober 2018 attracted more than 500 competitors from32 countries, testament to its popularity and growth.
© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
* Correspondence: [email protected] of Health Sciences and Medicine, Bond University, Institute of Health& Sport, Gold Coast, Queensland 4226, AustraliaFull list of author information is available at the end of the article
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 https://doi.org/10.1186/s13102-019-0130-z
http://crossmark.crossref.org/dialog/?doi=10.1186/s13102-019-0130-z&domain=pdfhttp://orcid.org/0000-0001-5125-2388http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/publicdomain/zero/1.0/mailto:[email protected]
Kettlebell sport uses competition kettlebells of standar-dised dimensions made of steel, most commonly avail-able from 8 kg to 32 kg in 2-4 kg increments. Kettlebellsport techniques are the jerk and snatch in differenttimed events.Kettlebells described as ‘traditional’ in shape are typic-
ally made from cast iron, with dimensions increasingwith weight. Kettlebells are now widely available in anarray of construction materials, from 2 kg to 92 kg. Withincreasing popularity has come diversity in use andadaptation of common exercises, however only a limitednumber of styles are widely recognised: Sport, hardstyle,juggling, and a small number of techniques associatedwith CrossFit.The popularity of kettlebells outside of Eastern Europe
and kettlebell sport can be largely attributed to Russianémigrés former World Champion Valery Fedorenko, andformer Soviet Special Forces physical training instructorand Master of Sport, Pavel Tsatsouline. Fedorenkofounded the American Kettlebell Club and Tsatsoulinethe hardstyle Russian Kettlebell Certification (RKC),which commenced training in 2001. Pavel has beenwidely credited with introducing kettlebells to the West[6] following a publication in the December 1998, Vol. 6,No. 3 Issue of MILO A Journal For Serious StrengthTraining Athletes. That was followed by Power to thePeople [7] which outlines many of the training principlesused in Enter the Kettlebell [3], and remains the founda-tion of hardstyle kettlebell training courses worldwide.Enter the Kettlebell has been the most widely cited textin academic publications where a hardstyle techniquehas been used. The six fundamental hardstyle techniquesare the Swing, Clean, Press, Squat, Snatch and Turkishget-up (TGU). Academic investigation of hardstyle train-ing represents around 50% of publications (refer toResults: report characteristics), with the two-handedkettlebell swing investigated most frequently. Neitherkettlebell sport nor hardstyle are limited to only thetechniques listed.A third person of note is former Master RKC, Kenneth
Jay. A small unpublished Bachelor of Science study com-pleted at the University of Copenhagen [8] investigatedthe VO2 and lactate effects from two weeks of dedicatedhardstyle kettlebell snatch training in a group of well-conditioned, kettlebell-trained college-age males. Jay’straining protocols later described in Viking WarriorConditioning [9] and those from Enter the Kettlebellrepresent the majority of study formats used to date.
Conceptual and contextual backgroundExercise prescription is an integral part of Physio-therapy practice [10]. Prescription of exercise asmedicine for a broad range of chronic diseases andfor relieving pain and improving musculoskeletal
function have been described [11, 12] with many atleast as effective as drug therapy [13]. The mecha-nisms of mechanotherapy in clinical practice havebeen reported [14, 15], with an understanding ofmechanobiology of musculoskeletal tissues critical toprimary care [16]. Therapists commonly seek to in-crease tissue capacity and build physical and psycho-logical resilience in their patients, from the younginjured athlete to the elderly and frail.Evidence-based Physiotherapy is an area of study, re-
search, and practice in which clinical decisions are basedon the best available evidence, integrating professionalpractice and expertise with ethical principles [17].Where high quality clinical research does not exist, goodpractice must be informed by knowledge derived fromother sources of information. When relevant and reliabledata is not available, clinicians still need to make deci-sions based upon the best available information [18].In elite sport, there is a constant need to increase
strength, power and endurance, and the kettlebell hasbecome a part of that effort [19]. Kettlebells have beenused in strength and conditioning research and injuryprevention programs for mixed martial arts [20], hand-ball [21], shot put [22], sprinting [23] and soccer [24]. Inclinical practice, kettlebells have been included in pro-grams for lower limb amputees [25], metabolic syn-drome in women [26], early treatment of breast cancer[27], for osteoporosis and fall and fracture prevention[28], home-based Physiotherapy with older adults show-ing signs of frailty and following hip fracture [29], forhealthcare workers [30] and in programs for improvinghealth-related physical fitness [31].Military and law enforcement agencies train with ket-
tlebells, reporting improvements in field performance[32]. Kettlebells have been recommended as part of theRoyal Air Force aircrew conditioning programme [33]and for simulated military task performance [34]. Thekettlebell deadlift has been recommended by the NorthAtlantic Treaty Organization to be used alongside theRanger test, which is a loaded step test, deemed to haveexcellent content validity and high inter-rater reliabilityin relation to five common physically demanding mili-tary work tasks for soldiers [35].Kettlebells have also been used to modify other com-
mon training protocols [36–38], and as a novel methodof providing valgus stress with good reliability, duringultrasound examination of the ulnar collateral ligamentof the elbow [39]. University studies have investigatedkettlebell training, including analysis of the TGU [40],for improving dynamic knee stability and performancein female netball players [41], in anterior cruciate liga-ment (ACL) injury prevention among female athletes[42], and for reducing work related musculoskeletal dis-orders of the low back [43].
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 2 of 30
Whilst kettlebells have been adopted by popular fitnessprograms such as CrossFit, the use of kettlebells remainsa relatively niche sport and knowing how to use a kettle-bell is perhaps not as intuitive as the more popular bar-bells, dumbbells and machine weights. In spite of this,kettlebells have been recommended for their ease ofteaching, cost effectiveness and being less intimidatingto use [44]. Kettlebells have already been integrated intoclinical practice but does the current body of evidencesupport their use for therapeutic purposes, and howdoes the evidence help inform clinical decision making?The aim of this review is to identify what is known
about the effect of kettlebell training from publishedacademic research, with the objective to systematicallyevaluate and critically appraise the literature and high-light areas for further investigation.
Kettlebell swing descriptorsThe ‘hip hinge’ is associated with a deadlift movementpattern and a hardstyle kettlebell swing. This has alsobeen described as a “Russian swing”, or a swing to chestheight. It can be performed with one or two hands hold-ing the bell. The two-handed overhead swing is associ-ated with a ‘squatting’ motion of the lower limbs, alsodescribed as an “American swing” and most commonlylinked with CrossFit. The ‘double-knee-bend’ pattern isassociated with kettlebell sport.
MethodsA scoping review was conducted to synthesise currentevidence of kettlebell training as it applies to therapistsworking in primary care, where movement and loadingare used clinically for therapeutic purposes. As an evolv-ing field of research, the scoping review was chosen toprovide an overview of kettlebell training, to identify keyconcepts, knowledge gaps, and types of evidence cur-rently available.
Research questionWhat evidence is available to guide therapists using ket-tlebells within a clinical therapeutic framework?
ProtocolThis scoping review was conducted by a single re-searcher (NM) using the PRISMA Extension for ScopingReviews (PRISMA-ScR): Checklist and Explanation [45].A priori protocol was not developed.
Study designThe scoping methodology proposed by Arksey andO’Malley [46] was used to map the concepts and typesof science-based evidence that exists on kettlebell train-ing. The methodology was informed by later recommen-dations [47] and guided by the Joanna Briggs Institute
framework [48, 49]. This framework includes the follow-ing steps: 1) Identify the research question by clarifyingand linking the purpose and research question, 2) iden-tify relevant studies by balancing feasibility with breadthand comprehensiveness, 3) select studies using an itera-tive team approach to study selection and data extrac-tion, 4) chart the data incorporating numerical summaryand qualitative thematic analysis, 5) collate, summarizeand report the results, including the implications for pol-icy, practice or research [50].
Information sources and literature searchA search was conducted, assisted by a health sciences li-brarian, on 10 electronic databases (CINAHL, CochraneLibrary, Embase, Medline, PEDro, ProQuest, PubMed,SportDISCUS, Web of Science, Google Scholar) frominception to 1 February 2019, using search terms “kettle-bell”, “kettle bell”, “kettlebells”, “kettle bells” in the Titleor Abstract. The search strategy was not limited bystudy design, publication type, or language. Duplicate re-cords were removed in EndNote. Backward referencesearching was performed, and additional studies wereidentified by consultation with subject matter experts.
Eligibility criteriaThe eligibility criteria were defined by the Population(therapists in primary care), Concept (prescription ofkettlebells for therapeutic purposes) and Context (evi-dence-based practice: research evidence and clinical ex-pertise). All types of study design and reviews wereincluded where kettlebells were the primary modality ofinvestigation. Any population, intervention, comparator,outcome, and setting were included, together with thesesand unpublished material from academic settings. Arti-cles/publications were excluded if, a) they were unre-lated to kettlebell training (e.g. gave historical contextonly), b) were not specific to kettlebell training (e.g. in-terventions involving kettlebells and other equipmentwhere the outcome(s) could not be attributed to thekettlebell), c) were unavailable in full text, or d) werestudies conducted on Eastern European Military popula-tions. The absence of standardised reporting guidelines(as recommended by the Enhancing the QUAlity andTransparency Of health Research network), and style ofreporting from countries of the former Soviet Union,were deemed incompatible for synthesis. The followingwere also excluded from our review: books, patents, fit-ness articles, web pages, blogs and opinion pieces fromnon-clinical or non-academic/clinical authors. Resourcelimitations precluded the translation of articles not pub-lished in English. One exception was a clinical trial ofhardstyle kettlebell training for people with Parkinson’sdisease, published in Portuguese with an English ab-stract; this was deemed to be specifically relevant to the
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 3 of 30
population, concept and context of the review and in-cluded but not translated. All levels of evidence [51]were considered.
Data abstraction and data itemsA standardised data abstraction form was not utilised. Asingle reviewer (NM) independently screened titles andabstracts for relevance and obtained full text articles ofpublications potentially relevant. As the scope and na-ture of the available evidence was not known in advance,the development of categories and grouping for mappingpurposes was developed iteratively as the data was ex-tracted and tabulated. Effect sizes were extracted wheregiven, or calculated if enough data had been provided.Cohen’s δ or standardised mean difference (SMD) wereused and magnitude of effect compared based on partici-pant’s resistance training status: untrained, recreationallytrained or highly trained [52].
Methodological quality appraisalWith the primary intent to inform clinical practice,the authors chose to critically appraise the rando-mised controlled trials using a modified Downs andBlack quality checklist [53]. This scoring system isbased on a checklist of 27 questions and has beenfound to be valid and reliable for critically evaluatingexperimental and nonexperimental studies. Thechecklist included 4 categories for evaluation: report-ing, external validity, internal validity/bias, andinternal validity/confounding [54]. Studies wereappraised by a second independent reviewer. Dis-crepancies were resolved by discussion and agree-ment reached. Quality of evidence and strength ofrecommendation was based upon the GRADEapproach [55, 56].
SynthesisData were narratively synthesised by author-defined cat-egory: (1) acute profiling, (2) athletic performance, (3)health-related physical fitness, (4) injury & rehabilitation,(5) expert/clinical option and (6) Review, with key char-acteristics and findings discussed. Publications weregrouped by nature of the study (acute vs longitudinal)and measures/outcomes. Acute profiling studies werefurther categorised by outcome: ‘sEMG’, ‘motion ana-lysis’, ‘hormonal response’, ‘cardiometabolic’, ‘mechanicaldemand’ or ‘performance’. Experiments and trials weremapped based on the population profile (age, gender,training history, kettlebell experience), types of exer-cise(s) used, style (hardstyle, sport or ‘other’), trainingformat (work-to-rest ratio, frequency, duration, inten-sity/load), measurements (sEMG, motion analysis,ground reaction force, HR, RPE, VO2), outcomes, andstudy design.
ResultsThe literature search yielded a total of 829 citations(Fig. 1). Three hundred and thirty-two records wereremoved as duplicates or not meeting the inclusioncriteria. Upon completion of the title and abstractscreening, 170 were potentially relevant and screened.Subsequently, 99 publications fulfilled the eligibilitycriteria and were included. Study flow diagram Fig. 1.Publications by category Fig. 2.
Report characteristics (extent, range, nature)The number of academic publications relating to the useof kettlebells has increased steadily since 2009 (Fig. 3)Sixty-eight (69%) of the publications were research stud-ies, including 47 (70%) measures of acute trainingresponse and 21 (31%) longitudinal investigations. Twolongitudinal trials involved clinical populations. Publica-tions were categorised as ‘acute profiling’ [47], ‘Athleticperformance’ [11], ‘Health related physical fitness’ [9],‘Injury & rehabilitation’ [4], ‘Opinion’ [22] or ‘Review [6](Fig. 4). Included in these were a Systematic Review, oneClinical Review, four Brief/Narrative Reviews, and 3 casereports from medical practitioners of injury attributed tokettlebell training. Acute profiling studies, which repre-sent almost half of the publications, were further cate-gorised based on outcomes: ‘sEMG’ [11], ‘motionanalysis’ [6], ‘hormonal response’ [3], ‘cardiometabolic’[16], ‘mechanical demand’ [6], ‘performance’ [2] or notcategorised [3] (Fig. 5).Fifty-four experiments and trials (79%) used healthy
college-age participants, with participants in 62 studies(91%) recreationally active. In fifty-five studies (> 80%),participants were novices unfamiliar with kettlebelltraining, and almost half (n = 33) explicitly used hard-style techniques and/or training principles described byTsatsouline. Only 4 investigations (2 acute, 2 longitu-dinal) involved kettlebell sport. Of the 68 experimentsand trials, 43 were published in peer-reviewed journals.The remainder were un-published conference presenta-tions [5], Theses [9], Pilot studies [3], papers acceptedfor publication [4], and University publications [4]. Re-sults described herein as significant where reported withp-values ≤0.05.
Acute profilingForty-seven studies of acute response to kettlebell train-ing were identified. Thirty-nine (83%) involved healthycollege-age participants, 7 (15%) involved adults whowere not of college-age, and 1 study did not report par-ticipant age. Twenty-one (47%) involved only males, 5(11%) involved only females, and 19 (40%) had malesand females. One study did not report gender. Only 1study had participants who were not recreationallyactive. In 34 studies (72%), participants were not
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 4 of 30
Fig. 1 Study flow diagram (PRISMA-ScR flow chart)
Fig. 2 Kettlebell publications by category
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 5 of 30
Fig. 3 Number of academic publications by year involving kettlebells to February 1, 2019
Fig. 4 Kettlebell publications by category Fig. 5 Acute profiling studies by category
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 6 of 30
Table
1Stud
ycharacteristicsinvestigatingacutesEMGwith
kettlebe
llexercise
Autho
rParticipants
Observing
(muscles
/region
)
Exercise
Hardstyle
Sport
‘other’
Load
(kg)
Com
parator
nAge
(yrs)
Weigh
t(kg)
Gen
der
Active
Kettlebe
llproficient
/no
vice
And
ersonet
al.(2016)
1625
±6
80±8
MY
novice
trun
k2H
swing
othe
r16
1Hsw
ing
Caravan
etal.(2018)
3322.2±3.5
91.8±8.0
MY
novice
serratus
anterio
rbo
ttom
s-up
carry
othe
r12
deg.
ofabdn
DelMon
teet
al.(2017)[63]
1430.1±3.9
89.89±19.72
MY
proficient
hamstrin
g2H
swing
hardstyle
16–48
squatsw
ing+db
l.knee
bend
Dicus
etal.(2018)[64]
2121.4±0.5(F)
20.9±0.7(M
)69.7±20.4(F)
79.2±12.1(M
)M
/F
Yno
vice
delto
id+pe
c.Major
press
othe
r25%
1RM
dumbb
ell
Lim
etal.(2018)
2422.5±3.3
70.82±7.2
M/F
Yno
vice
trun
k+lower
limb
2Hsw
ing
hardstyle
3multip
le
Lyon
set
al.(2017)[65]
1421.5±2.03
85.53±8.11
MY
novice
trun
k+up
per+lower
limbs
swing,
clean+
snatch
othe
r4.5–32
clean+snatch
Rajala
etal.(2016)
921.4±1.8
67.4±9.6
FY
novice
lower
limb
2Hsw
ing
othe
r4.5
OHsw
ing
St-Ong
eet
al.(2018)[66]
1231.6±8.2
72.4±13.2
M/F
Yproficient
shou
lder
girdle
Turkishge
t-up
hardstyle
8(F)16
(M)
none
VanGelde
ret
al.(2015)[58]
2324
±1.97
61.8±8.79
(F)
80.5±10.0(M
)M
/F
Yno
vice
gluteal+
biceps
femoris
2Hsw
ing
othe
r8–16
1Hsw
ing
Wuet
al.(2019)
1922.2±1.1
76.0±13.3
MY
novice
lower
limb
squat+lung
eothe
r20
-
Zebiset
al.(2013)[67]
1623
±3
66.2±7.4
FY
novice
hamstrin
g2H
swing
hardstyle
12–16
multip
le
1Hon
e-ha
nded
,2Htw
o-ha
nded
,1RM
1repe
titionmaxim
um,O
Hov
erhe
ad,a
bdnab
duction,
dbl.do
uble
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 7 of 30
Table
2Stud
ycharacteristicsinvestigatingacutemotionanalysiswith
kettlebe
llexercise
Autho
rParticipants
Measuremen
tsExercise
Hardstyle
/Sport/
othe
r
Load
(kg)
Con
trol
/comparator
nAge
(yrs)
Weigh
t(kg)
Gen
der
Active
Kettlebe
llproficient
/no
vice
Back
etal.(2016)[68]
732
±1(exp)
30±1.41
(beg
)70.5±1.8
78.33±3.86
unknow
nY
proficient
jointsegm
entangleandvelocity
atthepe
lvis,hip,kne
e,ankle,
shou
lder,elbow
andwrist
2H swing
hardstyle
16kg
expe
rt
Bullock
etal.(2017)[57]
1526.7±4.3
77.6±13.5
M/F
Yproficient
cycletim
e+jointsegm
ent
angleandvelocity
attheankle,
knee
andhip
2H swing
othe
r12
-20
kgoverhe
ad+Indian
club
swing
Oikarinen
etal.(2016)a[69]
528–50
unknow
nM
Yno
vice
jointsegm
entangles
atthe
shou
lder,elbow
andwrist+
segm
entangleandvelocity
atthelumbar
2H swing
othe
r16
-24
kgoverhe
ad
Ross
etal.(2015)[70]
429–47
68.3–108.1
MY
proficient
horizon
taland
vertical
displacemen
t+velocity
ofthekettlebe
ll
snatch
Sport
32kg
none
Silvaet
al.(2017)b[71]
125
72M
Yno
vice
jointsegm
entankles
ofarm-trunk,thigh
-trunk
and
leg-thigh+tim
e
OH
swing
othe
run
know
nun
stablesurface
VanGelde
ret
al.(2015)[58]
2324
±1.97
61.8±8.79
(F)
80.5±10.0(M
)M
/F
Yno
vice
jointsegm
entangleat
thehip
2H swing
othe
r8-16
kg1H
swing
Zinet
al.(2018)b[72]
324
±0.82
70.2±5.18
MY
novice
jointsegm
entangles
atthehip,
knee
and
ankle
2H swing
hardstyle
4-8kg
load
a the
sis,
bconferen
cepa
per,2H
two-ha
nded
,OHov
erhe
ad,exp.e
xpert,begbe
ginn
er
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 8 of 30
kettlebell-trained i.e. novices. Twenty-three studies(50%) explicitly used hardstyle techniques, and 3 (6%)investigated kettlebell sport. Training style/techniquewas unclear or not reported in 21 studies (43%). Threestudies were given 2 category allocations [57–59] and3 were uncategorised, deemed incompatible for syn-thesis [60–62].
Acute profiling - surface electromyography (sEMG)Eleven studies investigated sEMG. Muscles and re-gions investigated with exercise(s) and load(s) usedare shown in Table 1. It appears that the TGU pro-vides a roughly equal mechanical challenge to bothshoulder girdles, one acting to stabilise the arm andkettlebell overhead and the other acting to supportthe body, through transitions from lying to kneelingand vice versa [66]. Among 14 common lower limbexercises used for therapeutic purposes, a two-handed swing was found to have the highest peaksEMG (115 ± 55%max), with greatest preferentialexcitation of the medial hamstring (Δ 22.5 ± 9.7%peak nEMG) [67]. A similar observation was notedwith mean medial activity ≈10% greater than lateralactivity across types of swing [63] with mean sEMGgreatest during the hip hinge swing (35.74 ± 16.66),although the mean difference between styles wassmall (≈4–6%). In a dataset with large variation,excitation of the hamstring muscles was also ob-served to occur before the gluteal muscle in a oneand two handed swing regardless of gender or rangeof movement [58].
Acute profiling – motion analysisSeven acute studies investigated motion of joint seg-ments or kettlebell trajectory (Table 2). Novices werefound to perform a two-handed kettlebell swing differ-ently to experts. Significant differences in joint segmentangles and angular velocities at the hip and shoulderjoint were reported during a two-handed hardstyleswing, with the order of movements reversed betweenconditions. During the up-swing (ascent), experts leadwith the hips, then the shoulders followed. In the down-
swing (descent), the arms drop first, then the hips flex.In novices, these joint segment sequences are reversed.Experts ‘hinge’ at the hips rather than squat (≈20o
greater hip flexion at the bottom of the swing and ≈15o
less knee flexion on the descent) stand up straighter(≈10o more hip extension) and ‘swing’ the bell ratherthan ‘lift’ it (≈15o less shoulder flexion at the bottom and≈20o less at the top) [68]. These findings are consistentwith Tsatsouline [3] and with what is observed inpractice.Among a cohort of 23 novices, none of the partici-
pants obtained neutral hip position while performingany of the kettlebell swings, despite the notable availabil-ity of passive hip extension ROM, and cueing during theinstructional sessions. Average terminal hip extensionlacked a mean of 9.7° (± 7.8°) from neutral for both gen-ders during the 2-handed swing. Of note, participantswere only allowed to perform a maximum of 10 repeti-tions of each swing during the instructional session [58].The kinematic similarities and differences between a
swing to chest-height, a swing overhead, and an Indianclub swing have been reported [57] although the clinicalutility of these data is unclear. Cycle time for the over-head swing was 34% longer than the shoulder heightswing and Indian club swing, with no differences in peakjoint angles between the movements reported. No iden-tifiable risk of injury from kinematic observation of thelumbar spine was identified when performing a two-handed swing to chest-height or overhead using a 16and 24 kg bell, although reliability of these data isunclear [69].Bell trajectory during a 32 kg single-arm snatch per-
formed by four elite kettlebell sport lifters was reportedto be similar between lifters and highly consistent withinlifters. Anthropometric differences were suggested tomost likely influence movement and performance effi-ciency [70]. On an unstable surface, reduction in trunkand knee flexion angles and reduced shoulder range ofmotion were reported during an overhead swing [71]; anexpected compensation strategy to increase stability.Limited low-quality data suggested a possible trend to-ward decreasing mean flexion angles at the ankle, knee
Table 3 Study characteristics investigating acute hormonal response to kettlebell exercise
Author Population Measures Exercise Hardstyle/ Sport /other
Format Load(kg)n Age (yrs) Weight
(kg)Gender Active Kettlebell
proficient/ novice
Budnar et al.(2014) [73]
10 19–30 78.7 ± 9.9 male Y novice testosterone, growthhormone, and cortisol
Swing hardstyle 30:30 × 12 16
Greenwald et al.(2016) [74]
6 24.3 ± 4.1 80.7 ± 10.2 male N novice glucose tolerance Circuit other 25 mins 9–11.3
Raymond et al.(2018) [75]
10 19–43 82.2 ± 14.6 male Y novice testosterone andcortisol
Swing hardstyle 12min30:30
8–16
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 9 of 30
and hip with increasing bell weight among males novicesusing very light weights [72].
Acute profiling – hormonal responseVery limited data is available regarding acute hormonalresponse to kettlebell training (Table 3). Changes inserum testosterone, growth hormone and cortisol havebeen observed following 12 rounds of two-handedswings with 16 kg [73]. Heavier bells had a larger effecton testosterone and cortisol when performing 12min ofswings in which workload was matched, however ca-dence was significantly different (8 kg at 42SPM Vs 16kg at 21SPM) [75]. In practice, cadence typically remainsconsistent irrespective of kettlebell weight. A single 25-min kettlebell training session had similar effects onacute post-exercise glucose tolerance to high intensityinterval running [74]. The clinical utility of these data isunclear.
Acute profiling – mechanical demandSix acute studies investigated mechanical demands ofkettlebell exercise (Table 4). Normalised to body mass,mechanical demands of a two-handed swing with 32 kghad the largest impulse (3.0 (0.2) N.s.kg− 1) when com-pared with peak back squat at 60% 1RM (2.1 (0.2) N.s.
kg− 1) and jump squat at 40% 1RM (2.7 (0.4) N.s.kg− 1).A two-handed swing with 16 kg produced similar im-pulse to the jump squat at 0% or 60% of 1RM, and a 24kg swing produced similar impulse to the 20%1RM jumpsquat [76]. The vertical jump has been used as aproxy for measuring power output, with the swingpurported to be effective for improving activitiesassociated with explosive hip extension, such assprinting. A two-handed swing with 20% bodyweightproduced a smaller average, peak and time-to-peakrate of force development, than a vertical jump [78],suggesting a lack of specificity to improve verticaljump performance.Vertical braking force with a 24 kg bell was reported
to be approximately 25% greater during braking (down-swing) than acceleration (up-swing) during a two-handed hardstyle swing. Horizontally, the swing ap-peared to create approximately double the force andfour-times the power of a single-arm hardstyle snatchusing the same load. These data must be interpretedwith caution however, as the start of the propulsionphase was an upright standing position holding the bellin front of the thighs and included the transition fromupright standing to terminal backswing (bell betweenthe legs). Despite the difference in vertical displacement
Table 4 Study characteristics investigating mechanical demand of kettlebell exercise
Author Participants Measures Exercise Hardstyle/ Sport /other
Load(kg)
Control /comparatorn Age
(yrs)Weight(kg)
Gender Active Kettlebellproficient/ novice
Lake et al.(2012a) [76]
16 24 ± 2 90.2 ±14.4
M Y novice impulse, peak andmean force and power tocentre of mass, kettlebelldisplacement, peak andmean velocity
2Hswing
hardstyle 16 -32 kg
16, 24, 32 kg
Lake et al.(2014) [77]
22 28–41 75.2 ±14.6
M Y proficient impulse, mean force,displacement, magnitude,rate of work, phase durationsand impulse ratio
2Hswing
hardstyle 24 kg snatch
Mache et al.(2016)a [78]
25 22 ± 6 (F)23 ± 2 (M)
66.4 ± 9.2 (F)78.3 ± 8.5 (M)
M / F Y novice peak, average and timeto peak rate of forcedevelopment
2Hswing
other ≈20%BW
vertical jump
McGill et al.(2012) [79]
7 25.6 ± 3.4 82.8 ± 12.1 M Y proficient peak and averagemuscle excitation,lumbar compressionand shear force
1Hswing
hardstyle 16 kg swing withkime, snatch,bottom-up +racked carry
Mitchell etal. (2016)a
[80]
2 early 20’s 53 & 75 F Y proficient position and orientation,joints and centres of massof arm segments. Velocityand acceleration, forcesand moments of theupper limb
OHswing
other 8 -16 kg
8, 12, 16 kg
Ross et al.(2017) [5]
12 34.9 ± 6.6 87.7 ± 11.6 M Y proficient ground reaction forces,velocity and temporalmeasures of resultantkettlebell force
snatch Sport 24 kg none
aconference paper, 1H one-handed, 2H two-handed, OH overhead, BW bodyweight
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 10 of 30
http://n.s.kghttp://n.s.kghttp://n.s.kghttp://n.s.kg
of the bell (chest height vs overhead), force in the verti-cal direction was roughly equal to the swing, howeverthe swing created approximately 40% more brakingforce. Approximately 15% more Work was performed inthe down-swing during a hardstyle snatch than theswing. The swing has a significantly shorter brakingphase (0.30s Vs 0.40s), larger Impulse ratio (time undertension: 21% Vs 14%) and propulsion (26% Vs 14%) thanthe snatch [77] likely attributable to the bilateral vs uni-lateral nature of the exercises.During a single arm swing with 16 kg, a peak com-
pression force of 3195 N at the lumbar spine wasreported at the bottom of a swing, with an activebracing strategy described as the ‘kime’ increasingaverage compression by a further 1054 N [79]. Aunique property of the kettlebell swing was reportedas a posterior shear force (461 N), said to be sounusual that potential risks are unknown. The samestudy reported lumbar movement from 26o of flexionto 6o during a 2-handed swing with 16 kg.During an overhead two-handed swing, a transition
from tensile to compressive force at the shoulder wasshown to occur approximately in the upper 30% of thebells’ arc in two females, with the majority of force andpower reported to have been derived from the posteriorchain musculature [80]. A peak resultant ground reac-tion force (GRF) of 1768 N (242) was reported amongmale amateur lifters, roughly equal to 2x mean body-weight (87.7 kg ± 11.6 kg) [5].
Acute profiling – performanceTwo studies reported acute performance measures asso-ciated with kettlebell exercise (Table 5). One minute oftwo-handed swings with 16 kg was sufficient to inducefatigue (defined as a reduction in torque production) inthe lumbar extensor muscles, but was significantly lessthan an isolated lumbar extension (MedX) exercise [81].No significant interactions or main effects for any vari-able in countermovement jump performance werereported between kettlebell swings and kettlebell jumpsquats using a load equal to 20% bodyweight [82].
Acute profiling - cardiometabolic responseEstablishing whether kettlebell training has the potentialto increase aerobic capacity has been of interest toresearchers. Sixteen studies reported acute cardiometa-bolic responses to kettlebell exercise (Table 6). Theoxygen cost of completing as many two-handed swingsas possible in 12 min (197 to 333 completed) with 16 kgwas reported and compared with a graded treadmill testto exhaustion [85]. Classified as “hard” by ACSMstandards, average HR (165 ± 13 b·min− 1 = 86.8 ± 6.0%HRmax) was significantly higher than average VO2(34.31 ± 5.67 ml·kg− 1·min− 1 = 65.3 ± 9.8% VO2max). Atmatched RPE, 10 min of two-handed swings comparedwith continuous treadmill running resulted in signifi-cantly lower VO2 (34.1 ± 4.7 Vs 46.7 ± 7.3 ml·kg
− 1·min−1), METS (9.7 ± 1.3) and energy expenditure (12.5 ± 2.5Vs 17.1 ± 3.7 Kcal·min− 1). This was reported sufficientto increase aerobic capacity [88].Twelve rounds of two-handed swings produced sig-
nificant mean increases in HR with each successiveround (67 ± 0 at rest to 169 ± 5 bpm) and significantpost-exercise hypotension at 10 (~ 4 mmHg SBP, ~ 3mmHg DBP) and 30 (~ 4 mmHg SBP, ~ 3 mmHg DBP)minutes after exercise [97]. A reported resting meanHR of 67 ± 0 bpm in a group of 17 participants sug-gests challenges of reliability. Performing as manytwo-handed swings as possible in 12 min was alsoreported to be perceptually harder, with increasingfeeling of heat stress, muscle pain and higher sus-tained HR compared to a kettlebell circuit workoutcompleted at 90% 6RM [92].A Tabata-inspired kettlebell circuit using a 2:1
work:rest ratio was compared to 1:8 (30s:4min) sprintinterval cycling in “very active” males, with the kettle-bell protocol proposed to be more attractive and sus-tainable [96]. The authors concluded that the highintensity kettlebell protocol would be effective instimulating cardiorespiratory and metabolic responses,which could improve health and aerobic performance(mean VO2peak 29.1 ± 0.09 ml·kg
− 1 min− 1 = 55.7%VO2max).
Table 5 Study characteristics investigating acute performance characteristics associated with kettlebell exercise
Author Participants Observing Exercise Hardstyle/ Sport /other
Format Load(kg)n Age
(yrs)Weight (kg) Gender Active Kettlebell
proficient/ novice
Edinborough et al.(2016) [81]
10 20–25 79.94 ± 11.4 M Y novice muscular fatigue(acute torque production)
2Hswing
other 1 mincontinuous
16
Ros et al.(2016)a [82]
7 19.14 ±1.86
70.56 ± 7.25 F Y novice post-activation potentiation oncountermovement jumpperformance
2Hswing
other 5 reps, 1min rest,5 reps, 3mins rest
20%BW
athesis, 2H two-handed, BW bodyweight
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 11 of 30
Table
6Acute
cardiometabolicrespon
seto
kettlebe
llexercise
Autho
rParticipants
Observing
Exercise
Hardstyle
/Sport/
othe
r
Form
atLoad
(kg)
nAge
(yrs)
Weigh
t(kg)
Gen
der
Active
Kettlebe
llproficient
/no
vice
Chanet
al.(2018)[83,84]
1028.4±4.6
95.1±14.9
MY
proficient
VO2,HR,RER,V E,RPE
snatch
Sport
10mins(con
t.)16
Dun
canet
al.(2015)[59]
1623
±2.9
76.3±14.7
M/F
Yno
vice
HR,Bla
swing
hardstyle
2mins,40BPM
&80BPM
4–8
Farrar
etal.(2010)[85]
1020.8±1.1
77.3±7.7
MY
novice
VO2,HR
swing
hardstyle
12mins
16
Ferreira
etal.(2018)[86]
1224.3±7.0
74.1±9.4
M/F
Yno
vice
BPsw
ing
othe
r20
mins,18
s/min
unknow
n
Fortne
ret
al.(2014)[87]
1418–25
66.5±6.9(F)
81.7±3.0(M
)M
/F
Yno
vice
VO2,HR,Bla,RPE
swing
othe
r4mins,20:10
4.5(F)8
(M)
Hulseyet
al.(2012)[88]
1319–27
73.0±9.2
M/F
Yno
vice
VO2,HR,METS,RER,
ventilatio
n,kcal,RPE
swing
othe
r10
mins,35:25
8(F)16
(M)
Martin
etal.(2012)a[89]
828.5±5.5
86±15
MY
novice
HR,BP
swing
hardstyle
12mins
16
Santilloet
al.(2016)a[90]
1518–35
54.3–101.8
M/F
Yno
vice
Bla
swing
hardstyle
volitionalfailure
(RPE
15or
95%
HR m
ax)
14–16(F)
20–24(M
)
Schn
ettleret
al.(2009)a
[91]
1029–46
59.9–116.6
M/F
Yproficient
VO2,HR,kcal,RPE
snatch
hardstyle
5mins,15:15
12–20
Schreibe
ret
al.(2014)a
[92]
1029–46
59.9–116.6
M/F
Yproficient
HR,Temp,
Bla,pain,
thermalsensation,RPE
snatch
hardstyle
5mins,15:15
12–20
Šentija
etal.(2017)[93]
1125.9±4.0
73.1±21.1
M/F
Yno
vice
VO2,HR,Bla,ventilatio
nsw
ing
hardstyle
toexhaustio
nUnkno
wn
Thom
aset
al.(2014)a[94,
95]
1023
±4.4
75.45±20.1
M/F
Yno
vice
VO2,HR,BP,RER,M
ETS,
grip
streng
th,RPE
swing+
deadlift
hardstyle
3x10mins,EM
OM
8(F)16
(M)
Thom
aset
al.(2014)[94,
95]
1025.3±4.3
60.98±16.1
(F)
93.5±16.1
(M)
M/F
Yno
vice
VO2,HR,BP,RER,kcal,RPE
swing+
deadlift
hardstyle
3x10mins,EM
OM
8(F)16
(M)
Wesleyet
al.(2017)[1]
1830
±9.6
68.2±9
FY
proficient
HR,Bla,RPE
swing
hardstyle
15:15,×10
8,12,16
Williamset
al.(2015)[96]
821.5±
0.86
2.95
±11.62
MY
novice
VO2,HR,RER,TV,
breathingfre
quen
cy,V
E,kcal
circuit
othe
r20:10,4mins,×3
10–22
Won
get
al.(2017)b[97]
1723
±1
74.1±4.9
M/F
Yno
vice
HR,BP
swing
hardstyle
30:30,×12
8(F)16
(M)
a the
sis,
baccepted
forpu
blication,
EMOM
everyminuteon
theminute,
RPErate
ofpe
rceivedexertio
n,BP
Mbe
atspe
rminute
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 12 of 30
At a controlled work rate of 20 two-handed swings(40SPM) and 10 sumo deadlifts performed every mi-nute on the minute, versus continuous cycling on anergometer at 80 rpm, no significant differences werereported in any physiological (HR, VO2), subjective(RPE) or metabolic (RER, MET) response [94]. HRand RPE were significantly higher using the same30-min kettlebell protocol when compared withtreadmill walking at matched VO2, with no differ-ence in RER, kcal.min− 1 and BP [95]. Both studieshad male and female participants, with very largevariation in anthropometrics.No post-exercise hypotensive response was observed
in normotensive individuals performing two-handedswings for 20 min [86]. A statistically significant at-tenuation in BP reactivity compared to control wasreported, immediately following a cold pressor test,however the clinical utility of this phenomenon inpractice is unclear. Significant reductions in post-ex-ercise BP 120 min post-exercise were also reportedfollowing 12-min of discontinuous two-handed swings(88 to 486 swings completed), compared with a ket-tlebell circuit of 6 exercises among hypertensive orpre-hypertensive males [89]. Comparisons of effectare limited due to exposure bias, and a decrease ofonly 4 mmHg to reach clinical significance.Jay’s VO2 snatch cadence test (cMVO2) [9] was
modified by Chan [83] to simulate a kettlebell sportevent and measure VO2 over 10 min. Increasingsnatch cadence with 16 kg and multiple arm changes,was compared with a graded rowing ergometer withincreasing power output. HR was comparable (≈175 ±8-10b.min− 1) but mean peak oxygen consumption(37.5 ± 43.5 Vs 45.7 ± 6.6) respiratory exchange ratio(1.10 ± 0.060 Vs 1.18 ± 0.047) and minute ventilation(132.7 ± 19.2 Vs 157.1 ± 20.1) were significantly lower.VO2 response to the cMVO2 test was also reportedto be significantly lower than the Bruce treadmillprotocol (40.3 ± 2.2 Vs 49.7 ± 6.6 ml·kg·min− 1) amonga small mixed gender group with very large variation
in anthropometrics [91]. Mean VO2 was 31.6 ± 3.71ml·kg·min− 1 however the range in HR (128-180 bpm),%VO2max [3, 69–83, 85, 86, 88, 89, 92, 94–97] andRPE [10–18] suggest these data may be poorlyreliable.An incremental kettlebell swing test (IKT) using in-
creasing bell weight showed a strong correlation inpeak oxygen uptake with the incremental treadmilltest (3.27 ± 0.67 Vs 3.99 ± 0.71 LO2·min
− 1, r = 0.92)[93]. Mean peak values for the IKT were significantlylower for VO2, HR, BLa and VE. It was reported thatin most subjects, muscle fatigue rather than cardio-re-spiratory factors caused exhaustion in the IKT test.Clinical utility, validity and reliability of the IKT arecurrently unknown.With respect to the modifiable factors of swing cadence,
bell weight and rest periods, increases in kettlebell weight(8 kg,12 kg,16 kg) or cadence (32,40,48spm) were reportedto significantly increase cardiometabolic demand (HR, RPE& BLa) [1]. It should be noted that kettlebell proficient par-ticipants reported that a cadence of 32spm was “unnaturallyslow”, with the ballistic hip hinge eliminated and dynamicswinging motion becoming a static resistive motion. Re-searchers suggested that the resultant shoulder-dominantexercise likely inflated the physiological variables. Inaddition, swings have been reported to become perceptuallyharder with increasing bell weight [59], and reduced restperiods have significantly increased metabolic responsewhen volume-matched with low load kettlebells [87]. Theeffect of different recovery strategies on lactate clearancefollowing two-handed swings to volitional failure has beenreported [90]. A statistically significant difference in clear-ance time and post-recovery performance occurred at ≈9-mins post-exercise, which is unlikely to be clinicallymeaningful.
Acute profiling – ‘uncategorised’Three publications (Table 7) were not categorised as theoutcomes were unique and incompatible for synthesis.Small effect size reductions in pain pressure threshold
Table 7 Study characteristics investigating change in pain pressure threshold and task-related predictive test of a bilateral carry
Author Participants Observing Exercise Hardstyle/ Sport /other
Format Load(kg)n Age (yrs) Weight
(kg)Gender Active Kettlebell
proficient/ novice
Keilman et al.(2017) [60]
60 25.12 ±2.86
70.49 ±13.32
M / F unknown novice pain pressurethreshold
2Hswing
hardstyle 8 rounds, 20:10
8 (♀) 12(♂)
Beck et al.(2016) [61]
73 43.4 ± 9.7(F)40.9 ±10.2 (M)
67.2 ± 9.6(F)90.3 ±12.4 (M)
M / F Y n/a carry distance tovolitional failure
farmer’swalk
n/a at 4.5 and 5.0km/hr
2 × 22
Beck et al.(2017) [62]
67 24–59 82.9 ±15.7
M / F Y n/a carry distance tovolitional failure
farmer’swalk
n/a at 4.5 km/hr 2 × 22
2H two-handed
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 13 of 30
Table
8Stud
ycharacteristicsinvestigatinglong
term
physiologicalrespo
nseto
kettlebe
lltraining
Autho
rParticipants
Observing
Exercise
Hardstyle
/Sport/
othe
r
Duration
Form
atFreq
/wk
Load
(kg)
Con
trol/
comparator
Rand
omised
nAge
(yrs)
Weigh
t(kg)
Gen
der
Active
Kettlebe
llproficient
/no
vice
Che
net
al.(2018)
[98]
3366.7
±5.3
66.3
±12.1
FN
novice
skeletalandappe
ndicular
musclemass,bo
dyfatmass,
sarcop
eniainde
x,pu
lmon
ary
functio
n,grip
streng
th,
isom
etric
back
streng
th,chron
icproinflammatorycytokine
concen
tration
×11:swing,
deadlift,
goblet
squat,squat
lung
e,row,singlearm
row,b
icep
scurl,tricep
sextension,tw
o-arm
military
press,Turkish
getup
hardstyle
860 mins,5
ex’s,3
sets,8–
12reps,
2–3min
rest
260–
70%
1RM
Con
trol
Y
Marcelino
etal.
(2018)
2664.94±9.29
74.85
±13.94
M/F
(13M,
4F)
Nno
vice
timed
upandgo
,sitandlift,
elbo
wflexion
,6-m
inwalk,lower
limbpe
aktorque,BergBalance
Scale,staticpo
sturalstability
(COPdisplacemen
t)
bottom
-up,
firststages
ofTurkishge
t-up
,farm
erwalk,go
blet
squat,de
adliftand
swing
hardstyle
15–
–60–
85%
1RM
body
building
and
stretching
exercises
N
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 14 of 30
Table
9Stud
ycharacteristicsinvestigatingmeanlong
-term
perfo
rmance
improvem
entfro
mkettlebe
lltraining
Autho
rParticipants
Observing
Exercise
Hardstyle
/Sport/
othe
r
Duration
(weeks)
Form
atFreq
/wk
Load
(kg)
Effect
Effect
size
Con
trol/
comparator
Rand
omised
nAge
(yrs)
Weigh
t(kg)
Gen
der
Active
Kettlebe
llproficient
/no
vice
Ambrozyet
al.
(2017)
[100]
4027–32
unknow
nF
Yno
vice
hand
speed,
flexibility,explosive
streng
th,stren
gth
endu
rance,agility,
VO2,Wingate
test
circuit
(unkno
wn)
othe
r8
60mins
3un
know
nVO
2+6.31
ml.kg−
1 .min−1
Average
Power+8.30
W Maxim
umPo
wer+
20.53W
Bent
arm
hang
+13
s
Trivial(δ=0.27,
p≤0.05)
Trivial(δ=0.10,
p≤0.05)
Trivial(δ=0.15,
p≤0.05)
Mod
erate(δ=
1.36,p
≤0.05)
Minim
ally
active
Y
Beltz
etal.
(2013)
[101]
1722.1±
2.80
(M)
21.5±
3.93
(F)
77.7±
10.86(M
)64.5±
12.56(F)
M/F
Yno
vice
streng
th,aerob
iccapacity,b
ody
compo
sitio
n,flexibility,balance,
andcore
streng
th.
circuit
hardstyle
845–
60min
2un
know
nDynam
icbalance(PL)
+7.2cm
VO2+5.0
ml.kg−
1 .min−
1 Legpress+
41.7kg
Grip
streng
th+1.7kg
Pron
eplank
+45
s
Mod
erate
(SMD=1.14,
p≤0.05)
Trivial(δ=0.33,
p≤0.05)
Mod
erate(δ=
0.82,p
≤0.05)
Small(δ=0.42,
p≤0.05)
Very
large(δ=
1.32,p
≤0.05)
YN
Che
net
al.
(2018)
[98]
3366.7±
5.3
66.3±
12.1
FN
novice
skeletaland
appe
ndicular
muscle
mass,bo
dyfatmass,
sarcop
eniainde
x,pu
lmon
aryfunctio
n,grip
streng
th,
isom
etric
back
streng
th,chron
icproinflammatory
cytokine
concen
tration
×11:swing,
deadlift,go
blet
squat,squat
lung
e,row,
sing
learm
row,
biceps
curl,
tricep
sextension,
two-
arm
military
press,Turkish
getup
hardstyle
860
mins,
5ex’s,3
sets,8–
12reps,
2–3min
rest
260–70%
1RM
Sarcop
enia
inde
x−0.26
Handg
ripstreng
th(L)
+4.26
kgHandg
ripstreng
th(R)
+4.71
kgBack
streng
th+6.37
kgPeak
expiratory
flow
+0.79
L/s
Med
/Lrg
(δ=
0.79,p
≤0.05)
Very
large(δ=
1.2,p≤0.05)
Large(δ=0.88,
p≤0.05)
Large(δ=0.83,
p≤0.05)
Med
/Lrg
(δ=
0.79,p
≤0.05)
YY
Elbadryet
al.
(2018)
[102]
4020.17±
0.4
75±2.9
FY
novice
Standing
LJ,softball
throw,g
rip,LL
streng
th,h
ammer
throw
unknow
nothe
r8
60mins
3un
know
nStanding
long
jump+
11.37
Softball
throw
+6.94
Handg
ripstreng
th(L)
+12.72
Handg
rip
Med
ium
(δ=
0.66,p
≤0.05)
Hug
e(δ=2.74,
p≤0.05)
Hug
e(δ=2.92,
p≤0.05)
Large(δ=0.94,
p≤0.05)
Med
ium,(δ=
Yun
know
n
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 15 of 30
Table
9Stud
ycharacteristicsinvestigatingmeanlong
-term
perfo
rmance
improvem
entfro
mkettlebe
lltraining
(Con
tinued)
Autho
rParticipants
Observing
Exercise
Hardstyle
/Sport/
othe
r
Duration
(weeks)
Form
atFreq
/wk
Load
(kg)
Effect
Effect
size
Con
trol/
comparator
Rand
omised
nAge
(yrs)
Weigh
t(kg)
Gen
der
Active
Kettlebe
llproficient
/no
vice
streng
th(R)
+6.82
Static
streng
th(BS)
+6.06
Static
streng
th(LS)
+9.76
Perfo
rmance
(ham
mer)+
22.27
0.64,p
≤0.05)
Very
large,(δ
=1.37,p
≤0.05)
Very
large(δ=
1.20,p
≤0.05)
Falatic
etal.(2015)
1719.7±
1.0
64.2±8.2
FY
novice
aerobiccapacity
(VO2)
Jay’sMVO
2snatch
protocol
hardstyle
420
mins,
15:15
312
VO2m
ax+
2.3ml.kg−
1 .min−1
Med
ium
(SMD=0.72,
p<0.05)
circuit
weigh
ttraining
N
Holmstrupet
al.
(2016)
[103]
1818–25
unknow
nF
Yno
vice
Sprin
tpe
rform
ance
swing
hardstyle
810:10×8
29.1–13.7
Verticaljump
+2.4cm
Sprin
t-1s
Small(SM
D=
0.48,p
>0.05)
Trivial(SM
D=
0.04,p
>0.05)
NN
Jayet
al.(2011)
[104]
4044
±8
68±11
M/F
Nno
vice
self-repo
rted
pain
(neck/shou
lders,
back),streng
th(back,
trun
k,shou
lder),VO
2
swing
hardstyle
810–
15min,
30:60to
30:30×
10
38–16
Neck/
shou
lder
pain
VAS−
1.7
Low
back
pain
VAS−
1.6
Back
extension
MVC
+19.6
Nm
Trun
kflexion
MVC
+12.0
Nm
Shou
lder
elevation
MVC
+7.0
Nm
VO2m
ax+
2.9ml.kg−
1 .min−1
Small(SM
D=−
0.94,p
≤0.05)
Small(SM
D=−
0.89,p
≤0.05)
Trivial=
(SMD=
0.47,p
≤0.05)
Trivial(SM
D=
0.29,p
>0.05)
Trivial(SM
D=
0.39,p
>0.05)
Trivial(SM
D=
0.39,p
>0.05)
YY
Jayet
al.(2013)
[105]
4044
±8
68±11
M/F
Nno
vice
posturalreactio
nto
externalpe
rturbatio
n,verticaljump
perfo
rmance
swing
hardstyle
810–
15min,
30:60to
30:30×
10
38–16
Verticaljump
+1.5cm
Stop
ping
time-109
ms
Med
ium
(SMD=1.5,p>
0.05)
Large(SMD=
−2.87,p
≤0.05)
YY
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 16 of 30
Table
9Stud
ycharacteristicsinvestigatingmeanlong
-term
perfo
rmance
improvem
entfro
mkettlebe
lltraining
(Con
tinued)
Autho
rParticipants
Observing
Exercise
Hardstyle
/Sport/
othe
r
Duration
(weeks)
Form
atFreq
/wk
Load
(kg)
Effect
Effect
size
Con
trol/
comparator
Rand
omised
nAge
(yrs)
Weigh
t(kg)
Gen
der
Active
Kettlebe
llproficient
/no
vice
Lake
etal.
(2012)
2118–27
72.58±
12.87
MY
novice
halfsquat1RM
and
verticaljumphe
ight
swing
hardstyle
630:30×
122
12–16
Halfsquat+
18kg
Verticaljump
+3cm
Small(SM
D=
0.82,p?)
Small(SM
D=
0.60,p?)
Manocchiaet
al.(2013)[106]
3740.8±
12.9
76.6±
14.4
M/F
Yno
vice
transfersof
streng
thandpo
wer+
muscularen
durance
>20
othe
r10
60mins
2un
know
nBarbellclean
&jerk
+4.2
kg Barbell
benchpress
+14.2kg
Mod
erate
(SMD=1.17,
p<0.05)
Small(SM
D=
0.63,p
<0.05)
Maulit
etal.
(2017)
[107]
3123.1±
2.3
83.9±
13.8
MY
novice
streng
thandpo
wer
swing
othe
r4
4sets×5
reps
to6×4
210–
12.5%
MTP
Deadlift
1RM
+8.2kg
Verticaljump
+1.1cm
Mod
erate
(SMD=1.17,
p>0.05)
Small(SM
D=
0.63,p
>0.05)
Ooraniyan
etal.
(2018)
[108]
3018–23
unknow
nM
Yno
vice
lower
limbstreng
th,
muscularstreng
thpistol
squat,
biceps
curl,row,
front
raise
othe
r6
unknow
n3
unknow
nLower
limb
streng
thMuscular
streng
tha m
easures
not
describ
eda
Large(SMD=
0.40,p
≤0.05)
Small(δ=1.02,
p≤0.05)
Mod
erate
(SMD=1.02,
p≤0.05)
Large(δ=0.99,
p≤0.05)
Ottoet
al.
(2012)
[44]
3019–26
78.99±
10.68
MY
novice
streng
th,p
ower,and
body
compo
sitio
nsw
ing,
accelerated
swing,
goblet
squat
othe
r6
1–3sets
(3×6,
4×4,4×
6then
4×6,6×
4,4×6)
216
Verticaljump
+0.18
cmBack
squat+
5.58
kgPo
wer
clean
+3.35
kg
Trivial(SM
D=
0.05,p
≤0.05)
Trivial(SM
D=
0.18,p
≤0.05)
Trivial(SM
D=
0.18,p
≤0.05)
Parasuraman
etal.(2018)[109]
4518–25
unknow
nM
Yno
vice
Max
push-ups,p
lank
endu
rance
vario
usothe
r6
45mins
3un
know
nMax
push-
ups+4
trun
ken
durance+
57s
KB>CON,p
≤
0.05
Smith
etal.
(2014)
[110]
2820–29
unknow
nM
/F
Yno
vice
post-activation
potentiatio
nand
verticaljump
swing
othe
r8
×4–6
squat+
×5CMJS
320–36
Verticaljump
+4.62
cmVery
small(δ=
−0.04,p
>0.05)
Kram
eret
al.
(2015)
[111]
2718–26
80.0(M
)63.9(F)
M/F
Yno
vice
anaerobicpo
wer
vario
us(×5)
othe
r4a
15:45×
5×2
39–16
MeanPo
wer
Peak
Power
Rate
ofFatig
ue
Trivial,p>0.05
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 17 of 30
Table
9Stud
ycharacteristicsinvestigatingmeanlong
-term
perfo
rmance
improvem
entfro
mkettlebe
lltraining
(Con
tinued)
Autho
rParticipants
Observing
Exercise
Hardstyle
/Sport/
othe
r
Duration
(weeks)
Form
atFreq
/wk
Load
(kg)
Effect
Effect
size
Con
trol/
comparator
Rand
omised
nAge
(yrs)
Weigh
t(kg)
Gen
der
Active
Kettlebe
llproficient
/no
vice
Kruszewskie
tal.(2017)[112]
2421.2±
1.2
89.83±
9.62
MY
novice
40yrdrun,agility,C
Mjump,
standing
long
jump,
benchpress
vario
usSport
32Com
plex
416–72
Training
effectsno
trepo
rted
Wadeet
al.
(2016)
[113]
1730.4±
6.9
81.4±
12.4
M/F
Yno
vice
USA
Ffitne
ss,spe
ed,
power,and
agility
swing
othe
r10
5×2
mins,30s
rest
36–14
Body
fat
(2.5%)
Max
push-
ups[6]
Max
sit-up
s[3]
1.5-mile
run
(12s)
Max
grip
streng
th(3.1
kg)
Proagility
(0.1s)
Verticaljump
(1.3cm
)40-yarddash
(0.4s)
Small(δ=
−0.37,p
≤0.05)
Very
small(δ=
−0.03,p
>0.05)
Small(δ=−
0.47,p
>0.05)
Med
ium/Large
(d=−0.79,p
>0.05)
Small(δ=0.10,
p>0.05)
Small(δ=−
0.40,p
>0.05)
Small(δ=0.36,
p>0.05)
Small(δ=−
0.40,p
≤0.05)
a discontinuo
us
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 18 of 30
(PPT) have been reported in lumbar and hip muscula-ture following a Tabata-inspired (2:1) work:rest ratio,using a low-load, load-volume protocol [60]. A bilateralkettlebell carry was shown to be highly predictive ofstretcher carry performance among Australian Army sol-diers [61] with lean leg mass determined to be the mostinfluential physical characteristic [62].
Long-term physiological responseTable 8 shows the outcomes from two randomised con-trolled trials using pragmatic hardstyle kettlebell trainingwith older adults. Large effect sizes were reported in amixed-gender group with Parkinson’s disease following15 weeks of training [99]. Significant improvements werereported for the Timed Up and Go, Sit and Lift, elbowflexion and lower limb strength and torque measurescompared to the Non-Periodic Activities Group whichperformed bodybuilding and stretching exercises. Very en-couraging medium to large effect size increases in hand-grip strength, back strength and sarcopenia index werereported in a good-quality RCT in women with sarcopenia[98]. Improvements in axial skeletal muscle mass and sar-copenia index were maintained at four weeks after cessa-tion of training, with signification reductions in the samemeasures occurring in matched controls.
Long-term performance improvementTable 9 shows the long-term performance improvementsfrom kettlebell training. Changes in postural reactiontime following kettlebell training have been reported ina good-quality RCT [105]. A basic low-volume, low-in-tensity program of kettlebell swings performed twice aweek for 8 weeks, resulted in a large (109ms) reductionin reaction time to perturbation. In a separate publica-tion with the same participant demographics, relative re-ductions in mean musculoskeletal pain intensity of 57and 46% in the low-back and neck/shoulder regions re-spectively were also reported [104].Moderate increases in upper limb endurance (bent
arm hang time) have been reported in a moderate-qual-ity trial [100]. Large improvement in trunk endurance(prone plank time), moderate improvements in dynamicsingle leg balance and leg press strength, and small im-provement in grip strength, were also reported from amoderate-quality hardstyle kettlebell circuit performedtwice a week for 8 weeks with young, heathy, active par-ticipants [101]. When compared to weightlifting training[44], there was no statistically significant difference be-tween groups for the power clean, and only a small ef-fect size difference in back squat strength wasstatistically significant. Changes in half squat strengthand vertical jump height have been reported [114] al-though the effect size was small for a trained population.Small improvements in mean bench press 1RM and
moderate mean improvement in barbell clean and jerkwere reported from a comprehensive (> 20 exercises)pragmatic hardstyle kettlebell program performed twicea week for 10 weeks [106].No conclusions can be drawn from a Pilot study com-
paring complex training protocols, in which kettlebellswings were compared with barbell back squats [110].No statistically significant difference was found betweengroups on vertical jump performance, although reportedto be “practically significant”. No significant differencewas found in vertical jump and sprint performance inrecreationally active females, although the training vol-ume was described as inadequate [103]. Kettlebell swingsusing a ‘kettleclamp’ was reported to increase power andstrength when compared with explosive deadlift training,however these conclusions were not supported by thedata presented [107]. Limitations in study design preventany conclusions being made from kettlebell trainingwhen compared with battle ropes [109, 111], on thephysical performance of American Football players usingkettlebell sport [112], male handball players [108], col-lege females performing a hammer throw [102], or inmilitary fitness training [113].
Injury and rehabilitationBased on a large differential in vastus lateralis (VL) to semi-tendinosus (ST) sEMG pre-activity during standardisedside-cutting manoeuvres, a single case study was describedof a female soccer player, retrospectively identified post-in-jury as a high risk of ACL rupture [115]. Risk was charac-terized by reduced sEMG pre-activity for the ST andelevated sEMG pre-activity for the VL, with a high-riskzone defined as one SD above the mean VL-ST difference[116] (Table 10). Ten months post-ACLR and standardpost-surgical rehabilitation, the player was deemed ready toreturn to play despite persistence of the high-risk neuro-muscular pattern. Based on the author’s previous work[67], a low-volume intervention (< 1000 swings,
‘functional training’, proposed to “mirror the challengesone faces in day to day activities” [121].The TGU was described for patient self-management,
to teach “the motor control needed for daily activities,occupation, and sports” [122] and specifically for inte-grating mobility, stability, symmetry (left, right, front,back), coordination, balance and strength [123], as atherapeutic exercise for injury prevention and perform-ance enhancement [124], as a strength and conditioningtool for a variety of athletes [125], and as a componentof kettlebell training to develop strength and power[126]. Only one article written for instructional purposesillustrates each of the ‘big 6’ techniques as descried byTsatsouline [127]. Five kettlebell exercises have beenindividually described with proposed clinical or perform-ance benefits; a modified swing [128], thruster [129],arm bar [130], reverse lunge with overhead press [131]and a lunge clean [132].In sports, the use of kettlebells within program design
has been described as a safe and effective modality thatenhances the training experience [19] and was discussedin a point/counterpoint for inclusion in strength andconditioning [133]. A sample periodised program for theclean and jerk and snatch exercises within an athlete’sgeneral conditioning for kettlebell sport has also beenoffered [134].Three case reports of injury have been published.
An onset of De Quervain’s tenosynovitis was attrib-uted to repetitive trauma to the extensor pollicisbrevis tendon [135], exercise induced non-traumaticrhabdomyolysis without complication [136], and aradial stress fracture [137]. Each case report appearsto outline a training load error which may haveaccounted for the injury however, the potential influ-ence of training load was not identified in any case.Broad risk management strategies appropriate for fit-ness professionals have also been described [138].
Quality of evidence strength of recommendationsTwo reviewers independently evaluated randomisedcontrolled trials using a modified Downs & Black
quality assessment checklist [53]. Trials were ex-cluded from quality assessment for the following rea-sons: i) single participant [115], ii) effects could notbe attributed to only the kettlebell [113], iii) the trialwas discontinuous [111] and iv) pre-interventiondata was not captured [112]. The quality scores areillustrated in Fig. 6.
DiscussionWe conducted a scoping review which included 99 pub-lications. The current body of evidence is represented bya small number of longitudinal studies, which are largelyunderpowered and generally of low methodological qual-ity [56]. Three publications [104, 105, 108] from twostudies had participants randomised to an interventionor inactive control. With high risk of bias, confidence inreported effects is low. Further research is very likely tohave an important impact on our confidence in the esti-mate of effect. Trial descriptions of exercise interven-tions are suboptimal and no publication has used theConsensus on Exercise Reporting Template (CERT)[139] with only 1 RCT pre-registered. The validity of re-ported outcomes likely to have clinical utility, have yetto be established with repeated trials. Largely based onhealthy college-age participants, the current body ofresearch has limited applicability to clinical or high-per-formance athletic populations.Our findings highlight a growing research interest in
the effects of kettlebell training since 2009. There havebeen no adverse events reported during clinical trials,and no clear or quantifiable risk of harm from kettlebelltraining has been identified. It is unclear if the absenceof reported adverse events is a true representation ofkettlebell training or a limitation in reporting. Only 1publication [127] illustrates how to perform each of thefundamental hardstyle exercises. Clinicians unfamiliarwith kettlebell training wanting to prescribe them fortherapeutic purposes, would be wise to consult withtrained practitioners. Anecdotal reports of delayed onsetmuscle soreness, bruising and discomfort from repetitiveimpact force to the forearm among novices are not
Table 10 Single case study characteristics investigating kettlebell swings 10-months post-ACLR surgery
Author Participants Observing Exercise Hardstyle/ Sport /other
Duration(weeks)
Format Freq/wk
Load(kg)
Effect
n Age(yrs)
Weight(kg)
Gender Active Kettlebellproficient /novice
Zebis(2017)[115]
1 21 unknown F Y novice differential invastus lateralis tosemitendinosuspre-activity(% of max EMG)Counter movementjump
swings hardstyle 6 3–5 sets,×20 reps,20s rest(×10sessions)
2 16–20 ST - 23%↑ 61%BF - 26%↓ 17%-0.3 cm
ST semitendinosus, BF biceps femoris
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 20 of 30
unusual. As a dynamic skilled activity using a freeweight, it is advisable for a novice to receive appropriateinstruction to mitigate avoidable error in execution orinappropriate loading.Kettlebells are increasingly being used to perform ex-
ercises typically associated with other equipment, suchas the one-arm bent over row and sumo squat. In thesecases, the tool simply becomes a weight with a handleand the exercise (or potentially the outcome) not uniqueto the equipment. There may be instances where this ismore desirable or necessary within a clinical context,however this becomes generalised exercise prescriptionand ‘training using a kettlebell’ rather than kettlebelltraining. Kettlebells are also being used to augment trad-itional exercises, such as hanging kettlebells by elasticbands to the end of an Olympic bar during a squat orbench press [37], which bear no resemblance to kettle-bell training.The differences between kettlebell sport and hardstyle
could be summed up by a statement made by ValeryFedorenko in 2013, “It’s not about 5 or 10 sets of 10, its1 set of 100; that’s the principle” [140]. In contrast, Jaydescribed hardstyle training as “intermittent, high-pow-ered work at maximal or supramaximal intensity in thecorrect ratio of work and rest” [9]. There are similaritiesand differences between kettlebell sport and hardstyle.For the primary care therapist or strength and
conditioning specialist, there is no indication that onetechnique or style is better, more appropriate, or moreeffective than any other. Recommendation would mostlikely be based upon the provider’s experience with orexposure to kettlebells, and the person’s values, expecta-tions and preference about an exercise program theymay wish to engage in. Only 4 studies published inEnglish have investigated kettlebell sport. Two in-volved acute biomechanical analysis of kettlebellexercises [5, 70], 1 involved the development of a ket-tlebell snatch protocol for kettlebell sport that couldbe used in the laboratory [83], and 1 was a Universitystudy showing medium to huge effect size changes instanding long jump, strength and throw performance,although with high risk of bias the results are unreli-able [102].The U.S. Department of Energy “Man Maker” protocol
was described by Tsatsouline as “alternate sets of kettle-bell swings to a comfortable stop, with a few hundredyards of easy jogging for active recovery”. Performed twicea week for an arbitrary time of 12 min, it was recom-mended that people also complete 2 days each week of5 min continuous TGUs. The program would continueuntil they could perform 100 single-arm swings < 5 minsand 10 TGUs < 10 mins at a target weight. In the re-search literature, the Man Maker challenge was firstcited by Farrar [85] as a “popularly recommended
Fig. 6 Modified Downs & Black quality assessment of Randomised Controlled Trials
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 21 of 30
kettlebell workout”, however the study protocol usedwas 12 min of continuous two-handed swings. The same12-min continuous format was subsequently used tomeasure blood pressure response [89] and later com-pared to a high-resistance circuit workout [92]. Whilsthardstyle techniques were cited, these studies illustratean evolution in the literature away from the principlesand practices described by Tsatsouline, based upon re-searcher’s interpretation of training practices.Due to the variety of ways in which an exercise pre-
scription could possibly include kettlebell exercises forclinical and athletic populations, it is vital that the exer-cise professional has a clear idea of the acute stressesimposed on the body by this form of exercise prior to itsutilisation. An initial understanding of these acutestresses is being provided by studies assessing the acutehormonal, kinetic, kinematic, cardiometabolic and elec-tromyographic responses to kettle bell exercise in arange of populations.Surface electromyography (sEMG) is a popular re-
search tool which records the electrical potential of skel-etal muscle, with a wide variety of clinical andbiomedical uses. Within rehabilitation sciences, EMGsignals are collected as participants perform the activityunder investigation, frequently using different loadingconditions. Common methodology involves the com-parison of EMG amplitudes, with researchers makingconclusions based on the neuro- and electrophysio-logical correlation with muscle force. Hypotheses maybe made regarding potential longitudinal adaptations inthe characteristics and performance of skeletal muscle,such as strength and hypertrophy. However, conclusionscannot be made about muscle activation, force andmechanisms of force production, or inferences madefrom longitudinal outcomes based solely on sEMG amp-litude [141] The use of unconventional exercises [64, 65]adds further complexity to the interpretation. With exe-cution of a swing influenced by so many variables, it islikely that the differences between swing types may notbe clinically meaningful, although considered importantwithin their own discipline.The difference in movement pattern between expert
and novice performing a two-handed hardstyle swing[68] is consistent with 1o of hip extension observed atthe top of a swing in kettlebell-trained subjects [79] andwith what trainers report in practice. The skill acquisi-tion of a hardstyle swing appears clear and consistent,however its utility in clinical practice is unclear. The ob-served difference between expert and novice is likely toapply to other kettlebell exercises, thus the experience ofparticipants in research studies should be consideredwhen assessing validity of findings, and the generalisabil-ity of outcome data to other populations. Other factorslikely to influence outcomes include kettlebell specific
differences such as training style, bell weight and swingcadence, and factors common to other training modal-ities such as work-to-rest ratio, peripheral and central fa-tigue. Each of these should be assessed when prescribingkettlebell exercises and their relative importance estab-lished for clinical populations on a case-by-case basis.There is no indication that one type or style of swing
has greater clinical utility than another. No data suggeststhat someone performing a swing counter to the pre-scriptive hardstyle pattern, is at increased risk of harm.When performing a hardstyle swing in practice, muchemphasis is placed on the production of power (in thehorizontal plane) and of developing “power-endurance”[3] however, no published data currently exists whichquantifies or validates these claims. The potential forusing movement(s) associated with kettlebell training fortherapeutic purposes has not been investigated.Although limited, ground reaction force data is clinic-
ally helpful, particularly where the mechanical demandsof a kettlebell swing are compared to other commonlyused exercises, or where some objective quantifiableloading of tissues is indicated. Large increases in groundreaction force relative to bodyweight [5] may be of inter-est to clinicians where manipulation of lower limb loadis needed, such as with symptomatic knee osteoarthritis.The load influence from kettlebell training on specificjoints, or with musculoskeletal conditions more gener-ally, remains unknown and warrants investigation.Lumbar motion, compression and shear force data
during a kettlebell swing offer meaningful information,albeit limited to a single study [79]. These data are en-couraging, that in the absence of spinal pathology,mechanical loads through the lumbar spine during a 16-kg two-handed hardstyle kettlebell swing are low andnot indicative of increased risk of harm. Indeed, com-pression loads were reported below the National Insti-tute for Occupational Safety and Health action limit, andhalf that of lifting 27 kg on an Olympic bar. Resultantspine loads were described as “quite conservative” and“not be problematic”. How these forces might changewith increasing kettlebell weight is not known and clini-cians should be cautious not to assume they remain low.Biomechanical modelling identified a unique posteriorshear force in the lumbar spine during a kettlebell swing.Whether this is a consistent feature across individualsremains to be seen, and the potential effect on patho-logical presentations such as spondylolisthesis, a parsinterarticularis defect, or osteoporosis is unknown. Morecommon resistance training exercises such as a barbelldeadlift produce an anterior shear force at the level ofL4/5, with forces of much larger magnitude reportedamong competitive power lifters [142]. Until further datais available, clinicians would be wise to use caution ifconsidering a kettlebell swing with someone who has a
Meigh et al. BMC Sports Science, Medicine and Rehabilitation (2019) 11:19 Page 22 of 30
significant or unstable lumbar spine pathology. Add-itionally, among kettlebell-trained subjects, the lumbarspine was reported to flex approximately half full range(up to 26o) at the bottom of the swing [79].Greater time-under-tension (impulse) may support the
premise of enhancing power endurance, however theclinical utility of impulse when compared with otherforms of resistance exercise is unclear [76]. The manipu-lation of resistance training variables is widely consid-ered an essential strategy to maximise muscularadaptations, and guidelines exist in relation to volumeload to maximise muscle hypertrophy. No consensushowever currently exists for a metric of volume load inresistance training [143] and kettlebell weight is likely tobe well below an intensity threshold sufficient to stimu-late anabolism. Any difference in impulse per repetitioncompared with back squat and jump squat [76] are un-likely to be clinically meaningful when compared withthe kettlebell weight and number of repetitions per-formed in a training session. Further research is requiredto better understand the mechanical demands of kettle-bell training, which may involve several hundreds of rep-etitions and multiple exercises.The clinical utility of reduced torque in the lumbar ex-
tensor muscles following swings is unclear [81]. Consist-ent with temporal and kinetic data [78], no significantdifference in countermovement jump performance [82]suggests that kettlebell swings are unlikely to provideany meaningful benefit to jump performance. Change inpain pressure threshold may be used in clinical practice,however there is no suggestion that this phenomenonwould be unique to a kettlebell swing, or that change inpain pressure threshold following kettlebell swings [60]has a clinically meaningful effect. Loaded carries are alsonot unique to kettlebells, so the utility of carry data forclinical practice in relation to the specific prescription ofkettlebell exercises remains limited [61, 62]. Kettlebellcarries however (rack, bottoms-up, overhead, suitcase)have been proposed as good exercise to increase trunkstiffness and reduce “energy leakage” when transmittingpower generated by the hips, to sporting and daily livingtasks involving pushing, pulling, lifting, carrying, andtorsional exertions [144]. These principles do have clin-ical utility but have not been investigated.Kettlebell training appears to induce a cardiometa-
bolic response sufficient to improve cardiovascularfitness [1, 59, 84, 85, 87, 88, 91–96] provided thatthe dose (kettlebell weight, volume load and work:rest ratio) is appropriate for the individual and suffi-cient to provide a supraphysiological load. Effectshave often been over-reported, and reliable clinicallymeaningful effects remain to be demonstrated in ahigh quality randomised controlled trial. Many of thesame investigations have also demonstrated kettlebell
training produces a lower peak VO2 when otherphysiological and metabolic variables are matched[83, 88, 93, 95]. These data are consistent with sug-gestion that hardstyle kettlebell training is not themost effective form of exercise for improving cardio-vascular capacity. Physiological mechanisms for thepressor response (disproportionately elevated HRwhen compared to oxygen consumption during re-sistance training) have been proposed, however theseclaims have not been validated in practice [8].A basic kettlebell swing protocol has shown to pro-
duce a similar cardiometabolic demand to other formsof physical activity such as walking [95] and cycling [94].For someone who is home-bound with a ca