Kettlebell (KB) training has become one of the most popular fitness trends to hit the United States in recent years. Described as a “cannonball with handles (11),” KBs can be pressed and swung in myriad ways with the goal of improving overall body conditioning. Anecdotal reports regarding KBs reveal that they are easy to use, require little space, are extremely portable, and are very efficient for those with little time to exercise. Also, KB training serves as an alternative to traditional weightlifting or powerlifting because exercises can be performed in all planes either unilaterally or bilaterally (11). In addition, the utilization of KBs has resulted in transference of strength and power to traditional weightlifting and powerlifting performance and, therefore, may be used to complement or augment training for athletes in these sports (9). It has also been demonstrated that KB exercises can be used to improve postural control (7), and it has been suggested that KB exercises may be incorporated into the rehabilitative programs for injured athletes and patients (2,4).
Despite the recent reemergence of KB training, there is limited scientific research on the topic. Interestingly, much of the existing research has examined the metabolic cost of the KB swing. This is due primarily to the total body nature of KB exercise. It has been demonstrated that the metabolic demands of the Swing provides a stimulus sufficient enough to meet ACSM guidelines for the development of aerobic fitness (4,5). However, at matched ratings of perceived exertion, 10 minutes of the Swing elicited a similar heart rate but a significantly lower oxygen consumption and energy expenditure than treadmill exercise (6).
Mixed results have been reported when comparing the use of KB training with traditional weightlifting (12,14). A comparison of the KB Swing with the squat and squat jump exercises revealed that mechanical demands during the KB Swing were large and indicated that KB Swings might be a useful means of developing the ability to apply force quickly, although the data also suggested that the KB Swing may not be optimal for developing strength (10). Gains in muscular strength, power, and endurance were, however, transferred well to traditional weightlifting after 10 weeks of KB training in another study (12). It has also been reported that both traditional weight training and KB training produced increases in strength and power over a 6-week period; however, increases in strength were greater in the traditional weight-training group (14).
To date, only one study has attempted to characterize the muscle activation patterns of KB exercises. McGill and Marshall (13) quantified the muscle activity and spinal loading patterns during the Swing and Snatch exercises. During the Swing exercise, modest spinal loads and peak muscle activation of 50 and 80% of maximal voluntary contraction (MVC) were recorded for back extensors and gluteal muscles, respectively. No differences were observed in muscle activation between the 2 exercises. Another study compared the mechanical characteristics of the 2-handed KB Swing and the Snatch exercises and revealed that the 2 exercises were similar, and the authors suggested that strength and conditioning professionals could use these exercises interchangeably with their athletes (9).
The purpose of this study is to further investigate the muscle activation patterns during 3 commonly performed KB exercises using electromyography (EMG). Results from this study will assist strength and conditioning professionals to discuss KB lifting technique, prescribe KB exercises, and to better develop training programs for their clients and athletes.
Experimental Approach to the Problem
This study used a repeated-measures design. All subjects performed Swings, Cleans, and Snatches with a standard cast iron KB (Power Systems Inc., Knoxville, TN, USA) during a single testing session with the order of the exercise being randomized and counterbalanced. Muscle activation (EMG) of 8 different muscles (anterior deltoid [AD], posterior deltoid [PD], biceps brachii [BB], external oblique [EO], vastus lateralis [VL], biceps femoris [BF], gluteus maximus, and lumbar erector spinae [ES]) was recorded during each of the lifts using a submaximal load and was normalized using a maximal isometric contraction.
Fourteen male subjects (mean ± SD age = 21.5 ± 2.03 years, height = 180.87 ± 3.76 cm, mass = 85.53 ± 8.11 kg, and body fat = 12.86 ± 3.32%) were recruited from the university population, forming a convenience sample. Before the study, subjects completed a health history questionnaire and signed a statement of informed consent. To qualify for this study (i.e., inclusion criteria), the males were classified as low-risk individuals as categorized by the American College of Sports Medicine (American College of Sports Medicine, 2010) risk stratification (1). Also, subjects indicated that they had completed 6 months of continuous recreational resistance exercise training for no less than 3 days a week. The exclusion criteria of the study were the following: (a) musculoskeletal problems, (b) cardiorespiratory ailments, (c) metabolic disorders, (d) blood disorders, (e) history of psychological disorders, (f) use of tobacco products, (g) consuming more than 10 alcoholic beverages per week, and (h) less than 6 months of continuous recreational training. All experimental procedures were reviewed and approved by the Institutional Review Board before initiation of the study. All subjects completed the protocol.
Each subject reported to the laboratory on 2 occasions before the experimental trial with all sessions being separated by 7 days. In the first session, subjects' demographic information was collected, followed by familiarization with the exercise protocol. Additionally, they were asked to abstain from exercise 48 hours before the testing session. Subjects reported little or no KB training; so, the proper lifting technique was taught by a Certified Strength and Conditioning Specialist (CSCS) who is an experienced KB instructor.
In session 2, subjects were allowed ample time to practice so that they felt comfortable with each lift. Subjects worked with the researcher to determine a load for each individual exercise that could be performed for 8–10 repetitions with a good technique. If subjects could not achieve 8 repetitions, then a lighter KB was selected. If the subject could perform more than 10 repetitions, then a heavier KB was selected. An 8 to 10 repetition maximum was used to control relative intensities across subjects, decrease the risk of injury, and to reflect the skill level of the participants. Subjects were not permitted to proceed onto the data collection phase of the experiment until they consistently displayed a proper lifting technique per the approval of the CSCS instructor. Kettlebell weights used in the study ranged from 4.5 to 32 kg. Subjects' experimental loads varied across the swing (23.23 ± 4.21 kg), clean (21.68 ± 5.21 kg), and snatch (18.43 ± 4.31 kg).
During the third visit, before the experimental trial, each subject warmed up by light pedaling on a stationary bike for 10 minutes. Next, the subject's skin was prepared by shaving, abrading, and cleaning with a cotton ball soaked in a 70% isopropyl alcohol solution. Eight separate bipolar surface (2.0-cm center-to-center) electrode (Noraxon Dual Electrodes, silver/silver chloride) arrangements were placed on the right side of the body over the muscle bellies of the AD, PD, BB, VL, gluteus maximus (GM), BF, lumbar ES, and contralateral EO according to the recommendations of Cram and Kasman (3). The EO electrodes were placed on the left side of the body, as the left EO acts as a stabilizer for the muscles on the right side of the body. The electrodes for the AD muscle were placed on the anterior aspect of the arm, 4 cm below the clavicle, and approximately parallel to the muscle fibers. The electrodes for the PD muscle were placed 2 cm inferior to the lateral border of the spine of the scapula, and angled at an oblique angle toward the arm so that they run parallel to the muscle fibers. The electrodes for the BB muscle were placed over the longitudinal axis 1/3 the distance from the fossa cubit to the acromion process, starting at the fossa cubit. The electrodes for the VL muscle were placed over the lateral portion of the muscle approximately 33% of the distance between the superior, lateral border of the patella to the anterior superior iliac spine (ASIS), and angled to approximate the pennation of the muscle fibers. The electrodes for the GM muscle were placed 6 cm lateral to the gluteal fold, 50% between the sacral vertebrae and the trochanter, and obliquely angled toward the hip to run parallel to the muscle fibers. The electrodes for the BF muscle were placed on the lateral aspect of the thigh 67% of the distance between the trochanter and popliteal fossa, starting at the trochanter. The belly of the BF muscle was identified by muscle palpation while holding the subject's leg at 90° and having the subject flex against tester resistance. The electrodes for the ES muscle were placed 3 cm lateral to the L3 spinous process. Electrodes were also placed over the left EO muscle 50% between the ribs and the ASIS, immediately superior to the ASIS, and at an oblique angle to run parallel to the muscle fibers. The reference electrode was placed over the lateral clavicle, approximately 2 cm from the sternoclavicular joint. Interelectrode impedance was kept below 2000 Ω by shaving the area and careful skin abrasion. The EMG signal was preamplified (gain ×1,000) using a differential amplifier (MyoResearch XP; Noraxon EMG and Sensor Systems, Scottsdale, AZ, USA, bandwidth 10–500 Hz).
Subjects then performed three 5-second trials of a maximal voluntary isometric contraction (MVIC) against manual resistance from the researcher for each of the 8 muscles. All MVIC trials were performed by the same researcher and were based on standard muscle-testing techniques (8). For the BB, the subject was seated with the elbow flexed to 90° and the forearm supinated. With one hand, the researcher stabilized the distal end of the posterior humerus at the epicondyles, while the hand provided resistance to the anterior distal end of the forearm while the subject attempted to flex the elbow. With the subject seated, the AD was tested with the glenohumeral joint abducted to 70° with 20° of flexion and the humerus in slight external rotation. The research stabilized the posterior scapula with one hand and provided downward resistance to the middle portion of the humerus while the subject attempted to abduct the shoulder. The position for the PD was identical to the AD, except the humerus was abducted to 70° with 20° of extension. For the ES, the subject was placed prone on an examination table with the hands behind the head. With the researcher stabilizing the lower extremities, the subject raised the trunk from the table and held the position. Because of the risk of injury, no manual resistance was applied. The VL was tested with the subject seated and the knee in full extension. The researcher used one hand to stabilize the upper leg and provided resistance with the other hand proximal to the subject's ankle. For the BF, the subject lay prone on an examination table with the knee flexed to 70° and the hip externally rotated to 20°. The researcher stabilized the lateral hip with one hand and resisted knee flexion by placing the other hand proximal to the ankle. The EO was tested with the subject supine on an examination table with the hands behind the head. The researcher stabilized the lower extremities while the subject flexed and rotated the trunk. To minimize the risk of injury, this position was held and no manual resistance was provided. For the GM, the subject was positioned supine on an examination table. With the knee flexed to 90° and the hip extended off the surface of the table, the researcher stabilized the posterior, lateral aspect of the low back. The researcher's other hand provided resistance to the posterior thigh while the subject attempted to extend the hip. A 60-second rest period between trials was administered to avoid muscle fatigue. After all of the MVIC trials were complete, a 5-minute rest period was provided before the experimental trials.
Next, the subjects completed 5 single repetitions for each of the 3 exercises (Swings, Cleans, and Snatches). Exercises were randomized with a 1-minute rest between each repetition, and 2 minutes was allotted between the different lifts. The velocity of each repetition was self-paced, and the exercise order was randomized. Completion of an exercise condition occurred when 5 successful repetitions were accomplished. If the investigator deemed a trial unsuccessful, subjects continued the protocol until 5 successful repetitions were completed.
The Swing, Snatch, and Clean, are touted as whole-body ballistic exercises. They involve the lower body, core, and upper body musculature; and they are initiated with great force, and they culminate with momentum. Thus, subjects were instructed to use their whole bodies and to explosively move the KB. It should also be noted that these KB exercises have a greater concentric component and a lesser eccentric component because of their ballistic nature.
The Swing was initiated with the KB in the right hand and feet shoulder-width apart (Figures 1 and 2). Starting in a squatting position with a stabilized neutral spine, subjects were cued to move the KB in the sagittal plane by rapidly extending their hips and knees. Subjects used the momentum gathered from the lower extremity to carry the KB to the chest level before it was returned to the initial starting position. The lifting technique for the Snatch was similar to that of the Swing except that the bell was swung into a snatch position and caught overhead (Figures 3 and 4). The subjects were instructed to absorb the force of the bell by flexing the hips and knees as they performed the catch. Subjects were instructed to keep the elbow extended, but not locked, and to hold the KB overhead for approximately 2 seconds before returning the load to the starting position.
The Clean was initiated in a squat position with the feet slightly wider than the shoulders and a stabilized neutral spine. The subjects were instructed to reach down and grasp the KB with the right hand. From this position, a simultaneous integrated effort involving the lower body, core, and right upper extremity elevated the KB by pulling it up close to the body so that the elbow is high with the shoulder abducted and the elbow flexed (Figures 5 and 6). The momentum generated through this pull is used to facilitate the catch. Once the bell was pulled high, the elbow and hand “trade places” by quickly dropping the elbow as the shoulder externally rotates. The bell, driven by its momentum, flips over the hand, and it was caught posterior to the vertical forearm. As the hand and elbow trade places and the bell flips, subjects were cued to absorb the force of the bell by flexing the hips and knees. Subjects were instructed to stand and recover (for 2 seconds) by extending the hips and knees, flipping the bell back to the front, and trading hand and elbow position so that the bell may drop down to the original starting point.
Electromyography data were collected using the Noraxon Telemyo 2400T system (Noraxon USA Inc., Scottsdale, AZ, USA). The EMG signal was telemetered to a receiver that contained a differential amplifier with an input impedance of 10 MΩ and a common mode rejection ratio of 130 dB. An amplifier gain of 1,000 was used, and the signal-to noise-ratio was less than 1 μV root mean square of the baseline. The EMG signals were then filtered with a bandpass Butterworth filter at 15 Hz and 500 Hz. The receiver was interfaced with a Latitude C840 computer (Dell, Round Rock, TX, USA). Disposable 4 × 2.2 self-adhesive Ag/AgCl electrodes were used for data collection. A sampling rate of 1,000 Hz was used for all testing. Noraxon MyoVideo version 1.7 was used in conjunction with a DCR-TRV 140 digital 8 video camera (Sony Corp, Tokyo, Japan) to time match EMG data to each repetition of every KB lift. Electromyography files were then accessed and processed using Noraxon MyoResearch XP version 1.07.
Raw EMG data were full-wave rectified and smoothed using a moving window (50 ms) with a linear algorithm. The middle 3 seconds of the MVICs were used for data analysis, allowing subjects 1 second to reach full muscle activation and eliminating the potential effects of fatigue during the last second. For each subject, the mean EMG during the MVIC trials was averaged for each of the 8 muscles. Electromyography data for the 8 muscles were then averaged during the KB Swing, KB Snatch, and KB Clean. The mean EMG activity for the 8 muscles from the KB exercises was normalized as a percentage of the MVIC (%MVIC). Data were exported to Excel (version 2010; Microsoft Corp, Redmond, WA, USA) and imported to SPSS (version 20 for Windows; SPSS, Inc., Chicago, IL, USA) for analysis.
For each dependent variable, the average muscle activation expressed as a %MVIC of each muscle was calculated for the independent variables of Swing, Snatch, and Clean. Analysis of variance with repeated-measures and t-test post hoc analysis were used to compare the effects of and differences between the 3 KB exercises across the 8 different muscles. The alpha level was set at p ≤ 0.05 for all comparisons.
Table 1 provides the muscle activation values for the 8 muscles during the 3 KB exercises. The results of the statistical analysis revealed significant differences for 3 of the 8 muscles. Pairwise comparisons revealed that for the ES (F2,26 = 12.015; P < 0.001), the Swing (60.89 ± 24.34%) elicited greater muscle activation than did the Snatch (38.38 ± 17.67%); for the EO, there was greater muscle activation (F2,26 = 11.196; P < 0.001) during the Clean (23.36 ± 10.06%) and Snatch (20.71 ± 7.72%) compared with the Swing (15.59 ± 5.91%); and the VL was significantly more active (F2,16 = 5.786; P = 0.008) during the Swing (56.81 ± 27.37%) compared with the Clean (40.9 ± 36.14%). Effect sizes computed as partial eta squared were 0.701 (ES), 0.630 (EO), and 0.630 (VL). Statistical power ranged between 0.952 and 0.990 for all statistically significant findings. There were no significant differences in the muscle activation of the AD (F2,26 = 0.2224; P = 0.801), PD (F2,26 = 1.764; P = 0.191), BB (F2,26 = 2.79; P = 0.08), BF (F2,26 = 1.588; P = 0.224), and GM (F2,26 = 0.160; P = 0.853).
Results of this study revealed that the Swing, Snatch, and Clean do involve muscular contributions from the lower extremity, core, and upper extremity. Contributions across the 3 exercises from the 8 muscles tested were somewhat similar, but not the same. Conversely, McGill and Marshall (13) compared only the Swing and Snatch exercises and found no differences in muscle activation between the 2 exercises.
The ES are important for postural stabilization and dynamic movement of the spine. In this study, the ES made a greater contribution during the Swing than during the Snatch. The contralateral EO ostensibly serves an important stabilizing function of the spine mostly in the frontal plane during these 3 lifts. The contribution of the contralateral EO was greater during the Snatch and Clean than during the Swing. This can probably be attributed to subtle differences in the KB movement paths for these lifts. During both the Snatch and Clean, the KB tended to stay on the right side of the body causing the contralateral EO to work harder to counteract a lateral flexion to the right of the spine. During the Swing, the KB tended to be brought up along the centerline of the body in front of the chin causing less lateral flexion of the spine.
The VL is important for effectuating and controlling knee activity. In this study, the VL made a greater contribution during the Swing than during the Clean. Because knee extension plays a meaningful role in both lifts, it is not readily apparent why this was so. It is possible that the recreationally trained subjects with minimal KB experience had not fully developed the capacity to fully engage their lower bodies during the Clean.
No other statistically significant differences among the 3 lifts for the 8 muscles tested were revealed. It is possible that advanced KB lifters may manifest different motor unit recruitment patterns and different patterns of intermuscular coordination. In this study, the KB lifts started from a static position so that individual repetitions of each exercise were separated by a pause. Typically, these KB exercises are done continuously without pause. This difference may have affected EMG readings and muscle contributions. This study involved young adult male subjects; female subjects may yield different fiber activation patterns. Future research should explore KB muscle activation patterns using different lifts with varying intensities, more diverse subjects, and longitudinal designs. Research should also focus on comparing KB exercises with similar lifts such as cleans and snatches performed with dumbbells and barbells.
The KB exercises performed in this study involve dynamic motion that could cause noise within the EMG data. Precautions to prevent noise were taken by the researchers, including careful preparation of the skin for maximal adhesion of the surface electrodes. The researchers monitored the recorded EMG signal for obvious signs of noise created by motion of the EMG leads. Additionally, the EMG data were smoothed with a moving window, reducing the effects of low level noise. Despite these precautions, it is possible that some noise within the data exists. This should be viewed as an acceptable limitation given the exercises involved in this study.
Finally, variety in training is important for both physiological and psychological reasons. Kettlebells represent an alternative training modality for fitness enthusiasts. Knowledge concerning KB training has been mostly limited to theoretical analysis and anecdotal commentary and experience. Little rigorous empirical research has been done using KBs. This study establishes that the KB Swing, Snatch, and Clean are whole-body exercises. This study also demonstrated that although the lifts are similar, they are not exactly the same. Three muscles were determined to have made different contributions during the 3 lifts. Therefore, we conclude that the Swing, Snatch, and Clean, place different demands on the ES, contralateral EO, and the VL. Kettlebells represent an authentic alternative for lifters, and the Swing, Snatch, and Clean are indeed whole-body exercises, and they are similar, but not redundant.
The results of this investigation demonstrate that KBs may be an effective and alternative method for conditioning the whole body. Kettlebell conditioning has been used to improve strength, power, and endurance (10,12) and may also serve as an interesting option for dynamic warm-up. Kettlebell training has also been shown to produce transference of strength and power to traditional weightlifting (12). Most KB exercises, like the 3 evaluated in this study, require large muscle groups to be used in an explosive fashion. This may increase the likelihood that strength and power will be transferred across lifting styles. Finally, coaches and practitioners lacking the budget and the necessary space to perform explosive barbell exercises with their athletes may consider substituting the performance movements with KB exercises such as the Swing, Snatch, and Clean.
No grant support was used to perform this study. The results of this study do not constitute endorsement of any products or techniques by the authors or National Strength and Conditioning Association.
1. American College of Sports Medicine, Thompson WR, Gordon NF, Pescatello LS. ACSM's Guidelines for Exercise Testing and Prescription (8th ed.). Philadelphia, Pennsylvania: Lippincott Williams & Wilkins, 2010.
2. Brumitt J, En Gilpin H, Brunette M, Meira EP. Incorporating kettlebells
into a lower extremity sports rehabilitation program. N Am J Sports Phys Ther 5: 257–265, 2010.
3. Cram JR, Kasman GS, Holz J. Introduction to Surface Electromyography. New York, NY: Aspen Publishing, 1998.
4. Crawford M. Kettlebells
: Powerful, effective exercises and rehabilitation tools. J Am Chiropr Assoc 48: 7–10, 2011.
5. Farrar RE, Mayhew JL, Koch AJ. Oxygen cost of kettlebell swings. J Strength Cond Res 24: 1034–1036, 2010.
6. Hulsey CR, Soto DT, Koch AJ, Mayhew JL. Comparison of kettlebell swings and treadmill running at equivalent rating of perceived exertion values. J Strength Cond Res 26: 1203–1207, 2012.
7. Jay K, Jakobsen MD, Sundstrup E, Skotte JH, Jorgensen MB, Andersen CH, Pedersen MT, Andersen LL. Effects of kettlebell training on postural coordination and jump performance: A randomized controlled trial. J Strength Cond Res 27: 1202–1209, 2013.
8. Kendal FP, McCreary EK, Provance PG, Rodgers MM, Romani WA. Muscles: Testing and Function (5th ed.). Baltimore, MD: Lippincott Williams & Wilkins, 2005.
9. Lake JP, Hetzler BS, Lauder M. Magnitude and relative distribution of kettlebell snatch force-time characteristics. J Strength Cond Res 28: 3063–3072, 2014.
10. Lake JP, Lauder MA. Kettlebell swing training improves maximal and explosive strength. J Strength Cond Res 26: 2228–2233, 2012.
11. Lake JP, Lauder MA. Mechanical demands of kettlebell swing exercise. J Strength Cond Res 26: 3209–3216, 2012.
12. Manocchia P, Spierer DK, Lufkin AKS, Minichiello J, Castro J. Transference of kettlebell training to strength, power, and endurance. J Strength Cond Res 27: 477–484, 2013.
13. McGill SM, Marshall LW. Kettlebell swing, snatch, and bottoms-up carry: Back and hip muscle activation, motion, and low back loads. J Strength Cond Res 26: 16–27, 2012.
14. Otto WH III, Coburn JW, Brown LE, Spiering BA. Effects of weightlifting vs. kettlebell training on vertical jump, strength, and body composition. J Strength Cond Res 26: 1199–1202, 2012.