A Comparison of the Effect of Kettlebell Swings and Isolated Lumbar Extension Training on Acute Torque Production of the Lumbar Extensors

Abstract

Edinborough, L, Fisher, JP, and Steele, J. A comparison of the effect of kettlebell swings and isolated lumbar extension training on acute torque production of the lumbar extensors. J Strength Cond Res 30(5): 1189–1195, 2016—The aim of this study was to use a fatigue response test to measure the muscular fatigue (defined as a reduction in torque production) sustained by the lumbar extensors after a single set of kettlebell swings (KBS) in comparison with isolated lumbar extensions (ILEX) and a control condition (CON). The purpose of which is to measure the physiological response of KBS against an already established modality. Subsequent data provide insight of the efficacy of kettlebells swings in strengthening the lumbar muscles and lower back pain treatment. Eight physically active males participated in a repeated measures design where participants completed all conditions. There were statistically significant reductions in maximal torque, reported as strength index (SI), after both KBS and ILEX exercise. A statistically significant difference was found for reductions in maximal torque between CON and both KBS (p = 0.005) and ILEX (p = 0.001) and between KBS and ILEX (p = 0.039). Mean reduction and effect sizes were −1824 ± 1127.12 (SI) and −1.62 for KBS and −4775.6 ± 1593.41 (SI) and −3.00 for ILEX. In addition, a statistically significant difference was found between KBS and ILEX for rate of perceived exertion (p = 0.012). Data suggest that both KBS and ILEX were able to fatigue the lumbar extensors. Isolated lumbar extension was able to generate a greater level of fatigue. However, contrary to previous research, the KBS was able to elicit a physiological response, despite the lack of pelvic restraint supporting the potential to strengthen the lumbar extensors.

Introduction

The lumbar extensors are a dynamic musculature that plays a well-documented role in exercise, sport, and functional performance (^2,38). However, low-back pain (LBP) remains prevalent in athletic populations (^1,2,15,24,39). Research suggests that slower activation of the lumbar multifidus and reduced lumbar extension torque is epidemiologically prominent in persons symptomatic of chronic LBP compared with asymptomatic populations (^25,35). Consequently, it has been suggested that the deconditioning of the lumbar extensors is a risk factor for the onset of LBP (³⁵) and that conditioning them through exercise could be an effective prophylactic and rehabilitative measure (^25,31,34). A review of the research has supported the efficacy of strengthening the lumbar extensors and consequently reducing LBP symptoms using specific isolated lumbar extension (ILEX) machine-based resistance training (³⁶). However, to date, these results have not been replicated using lower-tech training methods. Thus far, attempts to strengthen the lumbar musculature using roman chair exercises and Romanian deadlifts (RDLs) have been unsuccessful (^13,26,43). This is supportive of the hypothesis that the rotation of the pelvis permitted during these exercises provides a greater training stimulus to the hip extensors and insufficient stimulus to the lumbar extensors (^13,18,34).

However, because of the relatively high cost and low availability of ILEX machines, further research is warranted to find an equally efficacious and more practical approach. McGill and Marshall (²⁸) have suggested that as a result of the bracing mechanism of the lumbar extensors during kettlebell swings (KBS), there is significant muscle activation (∼50% maximal voluntary contraction [MVC]) with loads of only 16 kg. In addition, the authors cite personal correspondence from a World International Powerlifting Federation Deadlift Champion as having used KBS in rehabilitation from a back injury and LBP. Other research has also supported the use of KBS in reducing LBP and increasing trunk extensor torque (^22,23). However, the method of assessing strength increases in these studies are limited by testing methods which allowed pelvic rotation and as such it is not known whether the KBS conditioned the lumbar extensors. Previous research considering the roman chair and RDL exercises reported strength increases for these respective exercises without strength increases of the isolated lumbar extensors (^13,26,43). In addition, previous studies have considered the use of surface electromyography (sEMG) as a method of assessing lumbar extensor activation to hypothesize training efficacy. Deadlift variations (^6,10) and roman chair exercises (^8,27) have shown activation of the lumbar muscles measured by sEMG where subsequent training interventions have failed to demonstrate a chronic strength adaptation. This might be a result of the significant limitations of sEMG including the effects of crosstalk, an issue especially prominent in the low back (^9,37).

McGill and Marshall (²⁸) also reported 80% MVC for gluteal muscles during KBS, which is supported by the correspondence of the World International Powerlifting Federation Deadlift Champion who reported increases in hip extension strength. Certainly it is evident that because of the nature of the exercise, there is pelvic rotation throughout a KBS which, based on previous research, suggests that the exercise might not be efficacious in strengthening the lumbar extensors. However, as a cheap and practical resistance training method, it is certainly an area that warrants further consideration. With this in mind, this study has used a fatigue response test (FRT) which has been evidenced to provide a measurable physiological response to specific exercise (reduction in torque production ability) by testing before and immediately after performance of the exercise and which could provide a substitute for sEMG readings (²⁰). The aim of this study was to use a FRT to measure the muscular fatigue sustained by the lumbar extensors after a single set of KBS compared with ILEX and a control condition (CON).

Methods

Experimental Approach to the Problem

This study used a FRT under 3 conditions in a repeated measures design which directly compares a preexercise and postexercise strength test to consider the fatigue of specific muscles. In the first condition participants performed KBS, in the second condition participants performed a set of ILEX, and in the third condition participants were subject to a standardized period of rest. The order of the conditions was randomized for each participant to limit the effect of repeated exposure to the FRT and of any testing learning effect.

Subjects

An a priori power analysis of effect sizes (ESs) for the primary outcome of lumbar extensor fatigue (acute change in torque production) was conducted. Recent research (²⁰) reporting the use of the FRT test for the lumbar extensors was used to determine participant numbers (n) using a treatment ES of 1.49, calculated using Cohen's d (⁷). Participant numbers were calculated using G*power (^11,12). These calculations showed that the study required at least 3 participants to meet the required power of 0.8 at an alpha value of p ≤ 0.05 for the statistical analyses proposed (see Statistical Analyses section below). Attempts were made to recruit a greater number of participants considering potential attrition rates of at least ∼50% between tests. After approval from the relevant ethics committee, 10 recreationally active males (age range 20-25 years old) were recruited from health and exercise science undergraduate degree courses (see Table 1 for participant characteristics). Inclusion criteria required participants to be asymptomatic of LBP, aged between 18 and 40 years, have no contraindicated medical or health condition, and could confirm that they were not using performance enhancing or any other medication which might affect the study. All participants completed a Physical Activity Readiness Questionnaire and informed consent and were free from injury. Two participants failed to complete the study as a result of injuries sustained away from testing.

Table 1: Participant characteristics.

Procedures

Lumbar Extension Testing

All participants attended a familiarization session, where they performed a testing session in the exact format described below. This was to reduce any learning or training effect of the testing process. All lumbar extension testing was performed using a MedX lumbar extension machine (Ocala, FL, USA; Figure 1), which has been demonstrated as valid (^16–18) and reliable (r = 0.94–0.98; 31) and the details of which have been described elsewhere (¹⁷).

Figure 1: MedX lumbar extension machine showing restraint system.

All participants were assessed for range of motion and performed a dynamic warm-up with a load equating to 90 lbs/∼41 kg and 3 submaximal isometric tests at full flexion, full extension, and a mid-range position. Maximal isometric testing was then performed at 5 joint angles (0°, 18°, 36°, 54°, and 72° of extension; 20), where participants were encouraged to build up to maximal effort over 2–3 seconds and to maintain the maximal contraction for a further 1 second. The torque produced was measured by a load cell attached to the movement arm. Participants rested for 5–10 seconds between tests at different joint angles.

Fatigue Response Testing

The FRT used the same maximal testing procedure as described previously followed by a 5-minute rest period. Participants then exercised as per 1 of the 3 conditions, separated by no less than 72 hours. The KBS condition required participants to perform KBS using a load of 16 kg (^28,30) for 60 seconds. All participants had received comprehensive tuition regarding the technical elements of a KBS at least 72 hours before any other testing or data collection. Because of the ballistic nature of the KBS, repetition maximum testing is deemed unsuitable (^22,23). The exercise time of 60 seconds was used to create parity in the time under load between KBS and ILEX conditions. The ILEX condition required participants to perform dynamic ILEX exercise at 80% of maximum torque (obtained at baseline testing) using the MedX equipment on which testing was performed. Around 8–12 repetitions were performed to momentary muscular failure at a repetition duration of 2 seconds concentric and 4 seconds eccentric as per the operating procedures for the exercise modality (^5,13,31). The CON condition allowed participants to rest for 90 seconds.

The FRT procedure then required a second testing session 90 seconds after the first (with the experimental condition in between). This was formulated during a pilot study based on the time taken after KBS to be secured back into the MedX lumbar extension machine. The change in torque produced between the precondition and postcondition tests represents the amount of fatigue sustained by tested muscles as a result of the dynamic exercise (²⁰).

Ratings of Perceived Exertion

Immediately after the KBS or ILEX exercise session, before the postcondition lumbar extension test, each participant was required to report perceived exertion from the Borg 15-point ratings of perceived exertion (RPE) scale (³). This was used to account for the possibility that any difference in fatigue was a result of the relative intensity of effort of the dynamic exercise modalities. Although the Borg 15-point RPE scale was intended for aerobic style exercise, previous studies have used this method of RPE for resistance training when measuring a predetermined number of repetitions (^32,41,42), explosive resistance training (³²), and when measuring a set to volitional failure (⁴²). The explanation and direction of the RPE scale were derived from Noble and Robertson (²⁹). Participants were asked to assign a number from 6 to 20 that represented their overall sensation of exertion, stress, and discomfort. They were informed that 7 equates to the “lowest possible exertion,” similar in sensation to 1 repetition of 50% of their respective 1 repetition maximums (1RMs), and 19/20 represented their “maximal effort,” likened to them performing a 1RM test or a set to failure.

Statistical Analyses

Strength measured as isometric torque data was considered as a strength index (SI) provided by MedX clinical equipment. This has been reported previously (^13,14) where SI represents the area under a force curve created in each isometric test and accommodates potential increases or decreases throughout the entire strength curve for all test positions. This negates biasing data by seeking average increases or decreases or only considering specific joint angles. The independent variable considered was the training condition (CON, KBS, or ILEX) and the dependent variables included prestrength (during the first test of each condition), the absolute change in strength due to the training condition (poststrength − prestrength), and RPE. A Kolmogorov–Smirnov test was conducted to examine whether data met assumptions of normality of distribution. Where assumptions of normality were met, repeated-measures analysis of variance (ANOVA) was used to compare within participants across the independent conditions for prestrength and change in strength. Where a significant effect by condition was found, post hoc pairwise comparisons with a Bonferonni procedure were conducted to examine differences between conditions. Ratings of perceived exertion did not meet assumptions of normality and thus, a Wilcoxon signed ranks tests was used to compare across conditions for this variable. Statistical analyses were performed using IBM SPSS Statistics for Windows (version 20; IBM Corp., Portsmouth, Hampshire, United Kingdom) and p ≤ 0.05 set as the limit for statistical significance. Furthermore, 95% confidence intervals (CIs) were calculated to identify within-condition significance in addition to ES using Cohen's d (⁷) for absolute change in strength to compare the magnitude of effects between conditions. An ES of 0.20–0.49 was considered as small, 0.50–0.79 as moderate, and ≥0.80 as large.

Results

Repeated-measures ANOVA revealed no significant effect by condition for prestrength (ILEX =, KBS =, CON =; F_(2,14) = 0.525, p = 0.603). There was a significant effect by condition for change in strength (ILEX =, KBS =, CON =; F_(2,14) = 25.323, p < 0.0001). Pairwise comparisons revealed significant differences between CON and both KBS (p = 0.005) and ILEX (p = 0.001) and between KBS and ILEX (p = 0.039). Effect sizes for change in strength were small for CON (0.27) and large for both KBS (−1.62), and ILEX (−3.00) and 95% CIs suggested both KBS and ILEX produced significant changes in strength. Figure 2 shows change in strength for each condition.

Figure 2: Mean change in strength for each condition with 95% CI. CON = control; KBS = kettlebell swings; ILEX = isolated lumbar extensions.

Wilcoxon signed ranks test revealed a significant difference between KBS and ILEX for RPE (Z = −2.527, p = 0.012). Rating of perceived exertion for KBS was 11.75 ± 2.5 and for ILEX was 18.38 ± 1.6.

Discussion

This study considered the fatigue sustained by the lumbar extensors in asymptomatic, physically active males after a single set of KBS in comparison with ILEX and a CON condition. Analysis suggests that the KBS condition was capable of inducing significant fatigue in the lumbar extensors, however not to the same magnitude as ILEX. Participants also reported significantly greater RPE for the ILEX condition compared with KBS. Although it is unclear whether the fatigue induced here through KBS can result in chronic adaptation, the results seem to support the findings of Jay et al. (^22,23) and suggest validity to the EMG readings previously reported (²⁸). Perhaps then, KBS could produce a conditioning effect of the lumbar musculature.

The CON condition did not induce any change in torque produced suggesting residual fatigue from the pretest did not affect the measured fatigue during the KBS and ILEX conditions. It is possible, however, that any reduction during KBS could be a result of central metabolic fatigue limiting maximal testing, as opposed to peripheral muscular fatigue. It has previously been suggested that KBS could provide sufficient stimulus to cause aerobic adaptations (^21,40), therefore this link seems plausible. If this is indeed the case, then it is unlikely that the reduced LBP as recorded by Jay et al. (^22,23) was a result of the interventions effect on the lumbar extensor musculature. Jay et al. (^22,23) did, however, also measure improved postural reaction times, which certainly could be contributory to their results.

The theorized issue of the central metabolic demands could be negated if a greater load, which could facilitate momentary muscular failure and was more comparable with ILEX, was used during KBS. However, this would require maximum testing of a ballistic exercise modality, which could be considered unfeasible because of safety and practicality issues (^22,23). Additionally, although both ecologically valid, KBS and ILEX are fundamentally different exercise modalities. Therefore, attempting to standardize the load would detract from any practical applications of the study.

If the torque reduction measured during KBS was in fact genuine muscular fatigue, then perhaps this could be of great use to practitioners. Certainly, participants did report some level of discomfort in the lower back after KBS, as evident by a mean RPE score of 11.75 ± 2.5. Although this is a subjective rating, it does lend credence to notion of KBS providing a stimulus. Because the load for KBS was not individually prescriptive, any future research considering an intervention study should use a more ecologically valid KBS workout including load, volume, and frequency.

Previous work by Fisher and et al. (¹³) suggests that RDLs at 80% 1RM were unsuited for lumbar conditioning because of the lack of pelvic restraint. The KBS prescribed during this study are similar to the RDL, only ballistic. Potentially, mechanical differences, as a result of the displaced centre of gravity during the kettlebells trajectory (²⁸), could have resulted in a shearing force across the vertebrae that required the lumbar multifidus to brace to compensate. As suggested by McGill and Marshall (²⁸), this is a factor that is possibly not present during RDLs and therefore, this could account for a stimulus in the low back musculature during KBS. As such, the use of KBS for the management of LBP has some support (^22,23,28).

Certainly, before KBS are prescribed, it would be prudent to consider safety implications. McGill and Marshall (²⁸), discussed the need for lumbar stability to cope with the aforementioned sheer forces. Additionally, 3200 Newtons of compressive force has been measured during a 16-kg KBS (²⁸). If a person cannot adequately “hip hinge,” then a compensation resulting in the repeated flexion and extension of the spine under load could increase intervertebral disc delamination and subsequent herniation risk as a result of mechanically unsound force displacement. Review articles have previously discussed injury implications as a result of repetitive and ballistic exercise modalities (^4,19). Therefore, caution is duly advocated.

The efficacy of ILEX in conditioning the isolated lumbar extensor musculature has been supported in a recent review of specificity of exercises designed for this purpose (³⁴). The ILEX was able to produce fatigue in the lumbar extensors that was statistically significant and of greater magnitude to KBS (ES of 3.00 vs. 1.62). Helmhout et al. (²⁰) recorded a smaller magnitude of difference in the 2 fatigue tests after ILEX (ES = 1.49) compared with that produced by the ILEX in this study but which was comparable to KBS. Both studies took participants to volitional failure, Helmout et al. (²⁰) however, used untrained participants rather than trained and therefore, a potential lack of motivation might have resulted in reduced effort and therefore fatigue.

The dissimilarities in the extent of fatigue between the ILEX and KBS conditions can potentially be attributed to the ability to isolate the lumbar extensors during ILEX and the continued muscular tension until failure that is a feature of the modality. The ballistic nature of the KBS suggests that momentum is used to move the kettlebell after an initial contraction, rather than prolonged muscular tension. As such, the musculature used during the action can recover as the kettlebell reaches the apex of its trajectory, as per EMG findings (²⁸). Therefore, as hypothesized, because of the isolated and controlled motion of the ILEX, a higher level of fatigue is possible.

It is established that a once weekly set, to volitional failure, of ILEX can improve lumbar extension torque (^13,17,31,33). Therefore, perhaps it can be concluded that the fatigue sustained as a result of ILEX was indeed sufficient to cause a chronic adaptation. However, the exercise and the isometric tests were performed on the same ILEX machine. Therefore, because of the similarity of the movements, the KBS condition was disadvantaged. This could account for the difference in magnitude between the 2 conditions. Indeed, this is an issue that is present throughout the literature in this field (^13,17,31,33). That said, considering the ability of ILEX to alleviate symptoms of chronic LBP (^33,36) and further that some of the research has shown ILEX to induce strength increases in other exercise modes (e.g., RDLs; 13), practically speaking, this may not be a concern.

Despite the apparent effectiveness of ILEX, the machines remain limited in their use. Previous research has highlighted kettlebells as potential tools in improving lower extremity strength, power, and, as previously discussed, inducing an aerobic stimulus (^21–23,40). Whereas, ILEX is only capable of facilitating isolated exercise and isometric testing for the lumbar extensors. Furthermore, comparatively, ILEX devices are more costly than kettlebells. Kettlebells then, remain a far more comprehensive tool for fitness. Research, albeit limited, does suggest that KBS could be used to prevent LBP and alleviate symptoms. The mechanisms involved are not fully understood, but contrary to what was hypothesized, this study does offer data that suggest KBS could offer a stimulus to the lumbar extensor complex, in turn offering a modality of preventing the deconditioning of the musculature through exercise.

Typical ILEX protocols already prescribe relative loadings, whereas this prescription does not exist for kettlebells. This study used a standardized 16-kg kettlebell for all participants regardless of repetition maximum. This was based on the prescribed weight used in sEMG (²⁸) and kettlebell studies relating to LBP (^22,23). As such, the results from this investigation can be better related to previous data. This, of course, limits any standardization between conditions and therefore, conclusions should be made with this in mind. Future research should consider a comparative intervention study using more individually prescriptive loads and more commonly used volumes and frequencies of training for KBS.

Conclusions

In conclusion, this study sought to measure the fatigue sustained in the lumbar extensors as a result of both KBS and ILEX exercise. It was hypothesized that because of the lack of pelvic restraint and limited time under tension, KBS would not be able to provide a fatigue response or, at least, not to the same extent as ILEX. Kettlebell swings did seem to be able to fatigue the lumbar extensors despite no pelvic restraint. It is possible that mechanical factors associated with the exercise were sufficient to load and fatigue the lumbar extensors. Although the magnitude of fatigue was not comparable to that sustained during ILEX, KBS represent a more versatile and cost-effective alternative. Certainly, future research should consider KBS to determine whether the measured fatigue can indeed produce a strengthening effect to the lumbar musculature over a chronic training intervention.

Practical Applications

The results from this study support the notion that KBS might be used in LBP treatment and prevention by conditioning the lumbar extensors through exercise. This is surely of note to the conditioning coach and to the personal trainer who might encounter LBP as part of their practice, in both the athlete and the lay person. However, it should be considered that one proposed mechanism that negates the need for pelvic stabilization is the bracing effect by the lumbar multifidus in response to a sheering force. Epidemiologically speaking, those with LBP tend to have reduced activation of the muscles in the low back. Therefore, in application, it might be prudent for the practitioner to progress a client through bracing and mobility exercises before introducing KBS as an integrated movement.