Does research support the use of bands and chains?

 

by CHRIS BEARDSLEY on OCTOBER 14, 2013

Bands and chains are one of the most popular new techniques currently used in the training practices of athletes and their use is particularly prevalent among powerlifters.

But what does the research say about using bands and chains for strength and conditioning? Here’s a review of the literature to date.

What’s the background?

Easy and hard parts of the exercise range of motion

Almost every compound exercise commonly performed in a gym has a hard part of the movement range of motion (ROM) and an easy part of the movement ROM. This is particularly evident during the lower body powerlifts. For the lower body powerlifts, the harder part of the movement ROM is generally at the bottom position (in the hole for the squat and off the floor in the deadlift).

What are strength curves?

If we were to draw a graph of the strength of a lifter against movement ROM or bar height, we would find that the strength of the lifter at lift-off was low while their strength near to lock-out was high. This graph is called a “strength curve” (see further McMaster, 2009). Since standing lower body lifts, including squat, deadlift, good morning, and lunge variations, display low strength at the bottom of the lift and high strength at the top of the lift, these exercises are said to have “an ascending strength curve”.

Why do strength curves exist?

The strength displayed by a lifter at each point in the movement ROM is basically the sum of all of the individual, interacting joint torques at each point in the lift. Each external joint torque acting on the body can be calculated as the load multiplied by the resistance moment arm, which is the horizontal distance between the line of action of the force from the joint.

If we are considering the joint torques of a lifter when lifting a barbell, we can see that the load at any given point during the movement ROM is not going to change that much, as it is going to be the barbell load plus any bodyweight not acting directly through the pivot. The combined barbell and bodyweight load is often referred to as the “system load.” So the difference in joint torques at any given point in the movement ROM most likely result mainly from changes in the external moment arm.

How much do external moment arms change?

It is quite hard to visualize the changes in external moment arms for the all of the joint torques at the same time. However, in the deadlift we might usefully look at how the external moment arm at the hip joint changes throughout the movement as a proxy for understanding the overall strength curve. In fact, the external moment arm of the system at the hip joint for the deadlift at lift-off is around 20cm but at lock-out it is just 6cm (see my write-up of deadlift biomechanics).

We can visualize this by looking at how far the butt sticks back at the bottom of a deadlift and how far the knees jut forward at the bottom of a squat and contrasting this to the top of these same movements, where they are directly underneath the barbell.

What does this mean for the muscles?

The large changes in external moment arms at the hip joint in the deadlift means that the hip extensor muscles find that they need to produce a lot less force in order to generate the equal and opposite internal hip extension torque. Hence, the lift gets a lot easier as the barbell gets higher. And while the internal hip extension torque can be affected by changes in the internal moment arms, the reality is that the hip extensor muscles basically just need to create about 3 times as much force at lift-off compared with lock-out.

In practice, this means that athletes have to work really hard to produce a peak contraction in the prime mover muscle groups at the bottom of the powerlifts, accelerate the barbell upwards and then typically coast through to the end of the lift well within their capabilities. This is more true for the squat than the deadlift because of the effects of fatigue from getting through the sticking region and because rounding can alter the biomechanics of the movement.

So how do bands and chains change this?

Bands and chains are typically added to a barbell during a compound movement in such a way that the easy part of the lift becomes harder. For bands, this generally means looping the band over the barbell from the floor (or under the bench in the case of the bench press) so that the bands begin to stretch as the bar goes upwards and are most stretched at lockout. In this way, bands and chains are a form of variable resistance training, in that they create a system in which the external load changes throughout the movement (see further McMaster, 2009).

It is also worth noting that bands or chains can be added to any exercise, not just exercises with “ascending strength curves.” For example, bench press, chin ups, and hip thrusts don’t have dramatic strength curves, but bands and chains can still be utilized to stress the end range-of-motion of those lifts.

What are variable resistance and accommodating resistance training?

Ideally, it is thought that the increase in loading provided by the bands or chains would lead to the external torque requirements remaining broadly constant throughout the exercise range of motion. If this were achieved, the exercise modality could then be termed “accommodating resistance” but where it is just an approximation it is more correctly called “variable resistance” (see further the biomechanical principles of resistance training).

***

How do bands and chains affect external loading during the powerlifts?

Before we get stuck into what happens to the biomechanics of real, live human beings during the powerlifts with bands and chains, it is useful to look at some studies that have explored the effects of bands and chains on the external load throughout the whole range of motion of the movement, as follows:

***

Shoepe (2010) investigated the tension of commonly-used elastic bands as a function of increasing length of deformation during bench presses and squats using different fixing methods for the bands. They examined 5 Flex Bands from EliteFTS of varying thickness (orange, red, black, purple, and green). They found that the manner in which the elastic band is affixed and the exercise that is performed can drastically influence the resulting additional tension that comes from the same elastic bands. The charts below show the very different tension-deformation results for the bench press and squat using the same bands and ranges of motion. The researchers also noted that the equations for modeling the tension-deformation curve were natural logarithms, suggesting that the amount of tension required for each incremental unit of band extension length actually reduced as the bands increased in length, which does not seem correct as it is contrary to most other studies performed into the tension-deformation curves of elastic bands. On the other hand, McMaster (2009) reported the same results.

McMaster (2010) investigated the tension of commonly-used elastic bands and the mass of chains as a function of increasing bar height (in 10cm increments). They examined 5 sets of Iron Woody bands (colors of yellow, red, blue, green and black corresponding to widths of 14, 22, 30, 48, and 67mm and lengths of 104 ± 2.4 cm) and 5 sets of chains (diameters of 6, 8, 10, 13, 16mm and lengths of 260 ± 2.1cm). The researchers tested the band tension and chain mass required in order to permit the bar to displace by 50cm in 10cm increments. They found that while chain mass increased broadly linearly, band tension did not follow a completely linear pattern but in fact increased exponentially at a faster rate at longer lengths. This is in contradiction to the above study by Shoepe (2010) and seems more in line with the behavior of elastic materials. Consequently, while it is possible to see that adding each 10cm of extra chain with 6, 8, 10, 13 and 16mm diameters leads to an additional 150g, 225g, 350g, 600g and 1.05kg, respectively, it is not as simple when using elastic bands. Additionally, it is noted that even the smallest elastic bands add a great deal more tension per unit increase in band length in comparison with the mass added by each chain for the same increase in vertical displacement (c.f. the vertical axis for the two charts below).

Neelly (2010) investigated the actual resistance being provided by double-looped and single chains during the back squat in a single resistance-trained male athlete. The researchers used a loaded barbell with 84.1kg, around 50% of 1RM and either 1 double-looped set of chain (9.6kg or 94.1N), 2 double-looped 2 sets of chain (19.2kg or 188.2N), 1 single set of chain (9.6kg or 94.1N) and 2 single sets of chain (19.2kg or 188.2N). Two types of chains were used to make up the total chain resistance, including heavy chains 15.9mm (5/8 inch) thick, 1.52m (5 feet) long, weighing 7.9kg and lighter chains 6.35mm (0.25 inch) thick, 1.83m (6 feet) long, weighing 1.7kg. The researchers found that the double-looped chains resulted in a much greater differential between the top and bottom loads. The ratios of the top-to-bottom loads for double-looped chains with 1 and 2 chains were 4.58 and 8.91, respectively. For the single chain conditions, the ratios were 1.73 and 1.56, respectively. This discrepancy between the ratios of the double-looped chains and the single chains was because in the single chain method nearly 50% of the weight of the heavy chain did not reach the floor and therefore did not act to vary resistance throughout the movement (i.e. the top position load does not change but the bottom position load does change).

***

What are the conclusions?

In summary, it seems that the addition of chains to a barbell leads to linearly increasing total barbell load throughout the movement but the addition of elastic bands does not. Adding bands to the barbell likely leads to exponentially increasing barbell load (pace Shoepe, 2010) with increasing bar displacement. In practice, this means that chains will be most useful for increasing loading specifically in the middle of the movement, while elastic bands will be most useful for increasing loading specifically at the end of the movement.

Additionally, it seems that the individual exercise and set-up with bands can have a very marked effect on the resultant additional tension that is provided by an elastic band. Therefore, standardization is essential when working with bands and we should try to replicate the exact same set-up every time they are used both between athletes and between sessions.

Finally, when using chains, using a greater weight of chain means that we increase the ratio of the load at the top of the movement to the load at the bottom of the movement. This principle can be used to manipulate the contribution of the chains to the total barbell load at the top of the movement, as this is thought to be a key figure for optimal adaptations.

***

How do researchers approach studying the acute effects of bands or chains in the powerlifts?

It’s important to note that it’s actually very hard to compare the acute biomechanics of the powerlifts fairly when using no bands or chains and when using bands or chains. This is because when setting up the study approach you are always going to have a discrepancy in loading either at the top or at the bottom.

Essentially, as a researcher you have a number of options, as follows: (1) you can add the resistance bands or chains to the existing weight, or (2) you can take weight off the bar and replace it with the same loading of resistance bands or chains when measured at the top position, or (3) you can take some (often half) of weight off the bar so that the loading is neither the same at the bottom nor at the top, or (4) you can try to equate another parameter across the two variants, such as mean force (which will also equate total work done by definition).

These options have the following implications:

• In option (1), the loading will be the same between conditions at the bottom but different at the top (it will be higher in the bands or chains condition). This then provides the strength coach an understanding of what will happen if they simply add bands or chains to an existing barbell load set-up (most likely during a power training session, i.e. a dynamic workout).
• In option (2), the loading will be the same between conditions at the top but different at the bottom (it will be lower in the bands or chains condition). This then provides the strength coach an understanding of what will happen if they remove weight from the barbell and replace it with bands or chains (most likely during a hypertrophy or strength session, i.e. either a repetition effort or maximal effort workout).
• In option (3), the loading is not the same at either the bottom or top and presents an option that is somewhere between (1) and (2) but because no other parameter is equated, it is hard to see how to use this practically.
• In option (4), the workload is the same in each case. This is a very interesting option, as workload is most likely to be what impacts on recovery for the athlete. Differences in power output or muscle activity between workouts with the same workload could have interesting ramifications for hypertrophic stimuli.
• In option (5), the loading is different but the relative loading (i.e. with respect to 1RM) is the same. This is another interesting approach, particularly for EMG activity studies.

Overall, there is probably no correct approach to setting up a study comparing variable and isoinertial loading protocols. In fact, each protocol provides different information about the possible choices open to strength and conditioning coaches.

Another key point that is thought to affect the biomechanics of the movements when using bands and chains is the percentage of the total barbell loading (i.e. bar + plates + either chain load or band tension at lock-out) that is made up of either chain load or band tension. As we will see, studies have used loading protocols over quite a wide range. However, in some cases, unfortunately, it is not easy to figure out exactly how much can be attributed to the variable aspect of the resistance.

***

What do studies say about the acute biomechanics of the powerlifts?

The following studies explore the acute effects of bands and chains on the powerlifts. For each study, I have provided the salient details. At the end of this section, I have analyzed the studies in a couple of ways, including (1) by reference to the options 1 – 4 outlined above, and (2) by reference to the proportion of total loading that is variable.

***

McCurdy (2008) explored the validity and reliability of a chain-loaded, free-weight bench press 1RM in 16 resistance-trained subjects (9 males and 7 females) participating in Division II college basketball. After two familiarization sessions, the subjects performed a 1RM traditional plate-loaded bench press and a 1RM chain-loaded free-weight bench press on two separate occasions. The researchers found that both the traditional bench press and chain-loaded bench press display valid and reliable 1RM results following only two familiarization sessions in male and female athletes with previous resistance training experience.

Wallace (2006) compared the acute biomechanics of Smith machine back squats performed with and without elastic bands in 10 resistance-trained subjects (4 females and 6 males) using both 60% and 85% of 1RM with and without elastic bands. In each case, in addition to a condition without bands (No-Bands) two different set-ups with elastic bands were tested, one in which 20% of the total relative load was provided by elastic bands at the top of the movement (Bands-20%) and one in which 35% of the total relative load was provided by elastic bands (Bands-35%).  The researchers found that peak power was significantly greater in Bands-20% than in No-Bands at 85% of 1RM. However, they did not identify any significant differences in respect of peak rate of force development (RFD), although there was a trend for RFD to increase between No-Bands and Bands-35%.

Stevenson (2010) compared the acute biomechanics of back squats performed with and without additional elastic bands in 20 resistance-trained male volunteers (average length of time training was 10.4 ± 4.7 years). The subjects performed 3 sets of 3 reps of squats at 55% of 1RM with and without additional bands. The addition of bands produced an additional force of 20% of 1RM at lock-out but the bands were slack at the bottom position. The researchers found that the use of bands led to significantly greater peak bar velocity in the eccentric phase as well as significantly greater mean positive rate of force development (RFD) in the concentric phase. On the other hand, not using bands displayed significantly greater peak and mean bar velocity in the concentric phase (unsurprisingly, as the relative load was similar at the bottom but greater at the top when using bands).

Baker (2009) similarly investigated how replacing part of the barbell load with chains would affect bench press bar speed in 13 professional rugby league players. In one condition, the subjects performed 2 sets of 3 reps of a bench press with 60% of 1RM plus 15% of 1RM of chains at the top. In another condition, the subjects performed 2 sets of 3 reps of a bench press with 75% of 1RM without chains. The researchers found that when performing the bench press with 60% of 1RM plus 15% of 1RM of chains at the top, mean and peak concentric bar speeds were around 10% greater during both sets when compared to the bench press sets with 75% of 1RM without chains (unsurprisingly, as the relative load was smaller at the bottom but the same at the top when using chains).

Ebben (2002) investigated quadriceps and hamstrings EMG activity as well as mean and peak ground reaction forces during the 5RM back squat performed either (1) with conventional weights, (2) with conventional weights plus 10% of 1RM replaced with chains, and (3) with conventional weights plus 10% of 1RM replaced with elastic bands. The subjects were 11 NCAA Division I athletes (6 females and 5 males) with resistance-training experience. However, the researchers did not detect any differences in respect of peak ground reaction forces or EMG activity between the three different conditions.

Palmer (2011, unpublished) investigated the EMG activity of the vastus lateralis and of the hamstrings during the back squat with and without elastic resistance in 22 healthy, physically active collegiate males with at least 6 months of resistance training experience. Firstly, the researchers assessed a 1RM for the back squat without elastic resistance (No-Band), back squat with bands fixed to the bottom of the rack (Bottom Band) and back squat with bands attached to the top of the rack (Top-Band). Secondly, the researchers performed EMG testing using 3 reps with 80% of 1RM for each squat variation. They analyzed EMG activity in each 10% increment of the bar path. The researchers used Elite FTS Pro Stretch Strong Bands (2½ inches thick). The researchers found that the No-Band condition involved more barbell load than either the Bottom-Band or Top-Band conditions (317.5lb vs. 300.5lbs) but less overall load at the top of the movement (317.5 vs. 393.9 vs. 368.2lbs, respectively). They also found that the EMG activity of the vastus lateralis was greater in the 90% section of the bar path in both band conditions in comparison with the No-Band condition but the three conditions were similar in the 10% condition. The researchers therefore concluded that greater overall average muscle activity for the whole movement can be obtained by performing back squats with bands compared to without bands.

Israeltel (2010) assessed the EMG of the EMG activity of the quadriceps as well as force and power in 10 resistance-trained males during squats with elastic bands (Bands) and without elastic bands (No-Bands), where both conditions were the same in respect of mean force and total work. The subjects performed 5 reps using a 20kg barbell and elastic band resistance of 100kg at the top position, for a total of 120kg. Then, they performed the same set of squats with only free weights such that the mean force and work were matched between the two conditions (i.e. an undisclosed total load of <120kg). The researchers found that the Band squats displayed significantly greater force, power and quadriceps EMG activity during the early parts of the eccentric phase and later parts of the concentric phase (which is unsurprising, given that the bands are providing greater resistance at those points and less resistance at other points). The researchers also noted that Band squats displayed significantly greater quadriceps EMG activity values through 85 – 95% of the whole movement.

Godfrey (2011, unpublished) explored the relationship between tension and displacement in various elastic bands and compared the biomechanics of 3 sets of 6 reps of free-weight bench presses with and without elastic bands at 50% of 1RM in 14 well-trained male rugby players. The free-weight bench press with elastic resistance used less barbell load and, with the addition of the elastic band, 50% of 1RM was reached at the top of the concentric phase. Different band resistances were used such that the band contributed 20% or 40% of 1RM at the top (Band-20% and Band-40%). The researchers observed that there was significantly greater force displayed in the band conditions in comparison to the barbell-only condition during the middle stages (30 – 70%) of the concentric phase and they also found that peak force over a set was greater in Band-20% (but not in Band-40%) than in the barbell-only condition. They found that power was greater in Band-40% than both Band-20% and the barbell-only condition from 10 – 50% and from 90 – 100% of the concentric phase and they also noted that total set peak power was greater in Band-40% than in the barbell-only condition, although mean set power was greater for the barbell-only condition than for Band-40%. Broadly speaking, the general trend in this study is for greater power output and lower force production when using elastic resistance.

Swinton (2011) investigated the effects of chains on performance of the deadlift in 23 resistance-trained athletes. The deadlift was performed with 30, 50, and 70% of 1RM loads at submaximal velocity, maximal velocity and maximal velocity with two different chain loads adding 20 and 40% of 1RM. To roughly equate the workload between the chain and non-chain conditions, the researchers subtracted half of the chain load at the top of the movement (20 and 40%) from the barbell plates. Therefore, at the top of the movement, the chain conditions had relative loads of 40, 60 and 80% of 1RM (for the 20% condition) and 50, 70 and 90% of 1RM (for the 40% condition). The researchers found that the addition of chains led to a significantly longer acceleration phase, as well as significantly greater peak force and impulse. However, they found that the addition of chains also significantly reduced bar speed, power and rate of force development.

***

What is the acute effect of using bands or chains in the powerlifts?

What clear trends are there?

Unfortunately, across the whole body of acute studies there seem to be few clear trends. However, if there are any hints, it seems that EMG activity of the prime movers might be higher during variable resistance training with similar relative loads or during sets of similar workloads than during isoinertial training, e.g. Israeltel (2010) and Palmer (2011, unpublished) but not Ebben (2002). The lack of a finding by Ebben (2002) in this respect might be explained by the very small element of variable resistance used (11.8% of total barbell loading).

The effect of the increased EMG activity for the same level of workload or force output may mean that resistance-trained athletes are able to stimulate the muscles to adapt to a greater extent for the same amount of work performed. This may be beneficial for hypertrophy but whether this is the case or not is currently unclear.

***

How did the researchers compare bands and chains with isoinertial lifts?

In the above listed studies, Ebben (2002), Wallace (2006), Baker (2009) and Godfrey (2011, unpublished) equalized the load at the top of the lift by removing plates from the barbell and replacing them with the same amount of band tension or chain load. This meant the load was lower at the bottom of the lift in the band or chain condition. Ebben (2002) found no differences in either EMG activity or ground reaction force. Wallace (2006) found that peak power was greater using variable resistance but RFD was not (although there was a trend in that direction). Baker (2009) found that bar speeds were higher when using variable resistance. Godfrey (2011, unpublished) found that power was greater but force was lower when using variable resistance.

In general, there seems to be a trend towards higher power outputs during the variable resistance protocols in this option category (Wallace, 2006 and Godfrey, 2011, unpublished). Moreover, these higher power outputs are likely because of greater bar speeds being used (Baker, 2009), which may arise because of the lower loading at the bottom of the movement. A key limitation in these studies is that the movement to a lower load in the variable resistance condition likely shifts the loading protocol closer to the optimum load for power, as such a load is usually lower than the typical loads used with plates-only in these studies. Therefore, it is not really surprising that higher power outputs are noted. It is quite hard to see any other trends in this category of analysis.

In contrast, Stevenson (2010) equalized the load at the bottom of the lift by leaving all the plates on the barbell and just adding additional band tension. This meant the load was higher at the top of the lift in the band condition. Stevenson (2010) found that using variable resistance did lead to increased RFD in comparison to plate-only exercise. In another approach, Swinton (2011) subtracted half of the chain load at the top of the movement from the barbell plates, so the load was neither equal at the top nor at the bottom. They found that the use of chains reduced bar speed, power and RFD but increased peak force and impulse. In another variant, Israeltel (2010) equalized the two conditions by reference to work done over the course of the set. They found greater EMG activity during the variable resistance condition in comparison to the plate-only condition. Finally, Palmer (2011, unpublished) used a 1RM for each individual type of exercise separately. They found that EMG activity was higher in the variable resistance protocol. Since there are so few studies using each of these approaches to equalizing the variable and isoinertial loading protocols, it is hard to draw conclusions from these studies when considered together.

***

How much load was assigned to the variable component?

The following list shows how the studies approach the variable component, arranged in broadly increasing order of relative contribution from variable resistance [type of variable resistance noted in brackets]:

• Ebben (2002) used 10% of 1RM as part of a 5RM (estimated as 85% of 1RM), which is 11.8% of total barbell loading. [Both with bands and with chains]
• Baker (2009) used 15% of 1RM as top of a 60% of 1RM to make up 75% of 1RM, which is 20% of total barbell loading. [Chains].
• Palmer (2011, unpublished) used 23.7% and 18.4% of total barbell loading. [Bands]
• Wallace (2006) used 20% and 35% of the total barbell loading. [Bands]
• Stevenson (2010) used 20% of 1RM on top of a 55% of 1RM to make up 75% of 1RM, which is 26.7% of total barbell loading. [Bands]
• Godfrey (2011, unpublished) used 20% and 40% of 1RM in a 50% of 1RM exercise, which is 40% and 80% of total barbell loading. [Bands]
• Swinton (2011) used chains of 20% in total loading protocols of 40, 60 and 80% of 1RM, which is 50%, 33% and 20% of total barbell loading, and chains of 40% in total loading protocols of 50, 70 and 90% of 1RM, which is 80%, 57% and 44% of total barbell loading. [Chains].
• Israeltel (2010) used 83.3% of total barbell loading. [Bands]

Again, this analysis is not hugely forthcoming. Even contrasting those studies that displayed a beneficial effect on power output (i.e. Wallace, 2006 and Godfrey, 2011, unpublished) with the single study that found a negative effect on power output (i.e. Swinton, 2011) is not particularly enlightening.

***

What conclusions can we draw?

In summary, the current state of the research in respect of the acute effects of bands and chains on the biomechanics of the powerlifts is not particularly cohesive. The only partially supported trends are for increased power outputs in comparison to barbell loads equated at the top of the movement and increased EMG activity during exercises performed with bands or chains in comparison with isoinertial exercises.

Future research would perhaps benefit from using approaches such as equated workloads or equated relative loadings (of 1RM) rather than trying to equate either the top or bottom barbell load. Equated workloads might help coaches decide between protocols where recovery is a limited quantity.

***

What chronic studies have been performed using bands or chains in the powerlifts?

This for me is the key section out of the whole literature review. It is all very interesting looking at what happens acutely in the powerlifts during different types of training modality but if there are big differences in the training effects then that is what we should really be focusing on! Here’s the low-down:

***

McCurdy (2009) investigated the effects of chain-loaded and plate-loaded off-season bench press training on strength, shoulder pain and muscle soreness in 28 Division II baseball players with 4.8 ± 2.7 years of resistance-training experience. The researchers allocated the subjects to either a chain-loaded or a plate-loaded training group, who exercised 2 days per week for 9 weeks. The chain-loaded bench press had no plates so the total barbell load was made up of the barbell and the chains. The researchers found that both training groups displayed significant increases in strength scores in both chain-loaded and plate-loaded bench press tests but there were no significant differences between the groups. However, there was a trend for the chain-loaded group to improve to a greater extent in the chain-loaded bench press. Also, the researchers did not detect any significant difference between the groups in respect of the levels of shoulder pain and soreness post-workout. However, they did note a strong trend for shoulder pain to be greater in the plate-loaded training group, which they felt would have been significant in a larger sample size.

Rhea (2009) compared the effects of heavy-load-low-velocity, light-load-high-velocity and variable resistance (with elastic bands) training methods on peak power and strength in 48 NCAA Division I athletes over a 12-week training intervention. The researchers randomly assigned the subjects to 1 of 3 training groups, who all used free weight compound resistance training exercises (back squats, power cleans, deadlifts, dumbbell walking lunges and Romanian deadlifts) as well as plyometrics and sprints (20 – 40-m sprints, front/side hurdle jumps, depth jumps (0.3 – 0.6m), split jumps and bounding). The heavy-load-low-velocity group performed the back squats with 75 – 85% of 1RM. The light-load-high-velocity group performed the highest load possible while still maintaining a bar speed of 0.6 – 0.8m/s. The variable resistance group performed the back squats with 50% of 1RM plus elastic bands, similarly at a resistance that enabled a bar speed of 0.6 – 0.8m/s. The researchers found that the increase in power in the variable resistance group trended towards being greater than the increase in the light-load-high-velocity group and was significantly greater than the increase displayed by the heavy-load-low-velocity group (17.8% vs. 11.0% vs. 4.48%, respectively). They found that the increases in external load (i.e. strength) were similar in the variable resistance and heavy-load-low-velocity groups (9.44% and 9.59%) but were much smaller in the light-load-high-velocity group (3.20%). The researchers therefore concluded that variable resistance training may be able to concurrently improve power to a similar degree to light-load-high-velocity training and improve strength to a similar degree to heavy-load-low-velocity training, thereby combining the best of both worlds.

Ghigiarelli (2009) compared the effects of three different bench press training programs, involving either traditional isoinertial bench press training, bench press training with elastic bands, and bench press training with chains in 36 healthy football players. The researchers randomly allocated the subjects into 3 groups: elastic band, weighted chain and isoinertial groups. The subjects performed upper body exercise 2 times per week for 7 weeks, for 6 sets of 3 reps and an intent to move the barbell as quickly as possible. Before and after the intervention, the researchers tested strength and power during the 5RM speed bench press. The researchers found that all of the groups improved strength significantly but there were no significant differences between groups. The researchers noted a trend for the elastic band group to display greater increases in both power and strength in comparison with the other groups.

Anderson (2008) compared the effects of elastic-band resisted and isoinertial bench press and back squat training on strength and power adaptations in 44 young, resistance-trained basketball, hockey and wrestling athletes. The researchers randomly allocated the subjects to either an isoinertial training or an elastic band training group for a 7-week intervention. The researchers deducted the elastic band tension from the bar weight for the elastic-band training group and the band tension accounted for approximately 20% of the total barbell load. The researchers found that the elastic-band training group displayed a significantly greater increase in back squat 1RM (16.47 ± 5.67 vs. 6.84 ± 4.42kg) and 1RM bench press (6.68 ± 3.41 vs. 3.34 ± 2.67kg) in comparison with the isoinertial training group. Similarly, the researchers found that the elastic-band training group displayed a significantly greater increase in counter-movement jump power output (68.55 ± 84.35 vs. 23.66 ± 40.56W).

Joy (2013) assessed the effects of adding variable resistance to the back squat and bench press exercises in one out of every four workouts over a 5-week period in 14 NCAA Division II male basketball players. The researchers divided the athletes into either a variable resistance or an isoinertial training group. The variable resistance group added 30% of their 1RM (N.B. this is much greater than 30% of the total barbell load) as band tension to their prescribed weight for one session per week. The researchers assessed the effects of this intervention on rate of power development, peak power, strength, body composition, and vertical jump height. The researchers observed that rate of power development increased by significantly more in the variable training group in comparison with the isoinertial group. No other significant differences between the groups were observed.

Dobbs (2010, unpublished) compared the effects of a 5-week period of isoinertial and variable resistance back squat training on 4RM back squat, vertical jump height and 30m sprint time in 20 male high school athletes. The variable resistance was provided by elastic bands and made up 25 ± 5% of the total load. The researchers found that the variable resistance training group displayed a significantly greater increase in predicted 1RM back squat (based on 4RM) and vertical jump performance than the isoinertial training group.

***

What is the chronic effect of using bands or chains in the powerlifts?

In summary, using bands or chains in the powerlifts may lead to similar increases in strength to heavy-load isoinertial training but greater increases in power and reduced joint soreness, as follows:

• Joint and muscle soreness – McCurdy (2009) found that using a chain-loaded bench press in training displayed a strong trend towards reduced shoulder pain post-workout in comparison with standard plate-loaded bench press training.
• Strength gains – McCurdy (2009) found that using a chain-loaded bench press in training displayed similar improvements in 1RM isoinertial bench press performance as isoinertial plate-loaded bench press training. Rhea (2009) found that using elastic-band resisted back squat training displayed a similar increase in isoinertial back squat strength as isoinertial back squat training. Ghigiarelli (2009) also found that both elastic-band and weighted-chain bench press training displayed similar increases in isoinertial bench press strength as isoinertial bench press training. Dobbs (2010) found that band-resisted back squat training displayed greater increases in isoinertial back squat strength than isoinertial back squat training. Anderson (2008) found that elastic-band resisted back squat and bench press training displayed greater increases in isoinertial back squat and isoinertial bench press strength than isoinertial back squat and isoinertial bench press training.
• Power gains – Rhea (2009) found that using elastic-band resisted squats displayed a greater increase in power in comparison with isoinertial heavy-load back squats and a similar increase in power in comparison with isoinertial light-load back squats. Ghigiarelli (2009) also found that elastic-band bench press training displayed a trend towards greater increases in power in comparison with isoinertial bench press training. Dobbs (2010) found that band-resisted back squat training displayed greater increases in vertical jump height than isoinertial back squat training. Anderson (2008) found that elastic-band resisted back squat training displayed greater increases in counter-movement jump power output than isoinertial back squat training.
• Rate of power development gains - Joy (2013) found that adding in one workout a week with elastic bands led to greater gains in rate of power development than a standard isoinertial training program for the bench press and back squat.

Overall, there seems to be a much better agreement among the chronic studies than among the acute studies, which is quite surprising. In fact, for a selection of training studies in a specific area, this is quite good by most standards.

In many ways, this is very reassuring. There is nothing quite so promising as an intervention that works even when we can’t figure out exactly why it works. On the other hand, I am always suspicious of interventions that have great theoretical merit according to the generally accepted mechanisms of the time but which don’t really deliver the goods when put to the test in long-term studies.

In general, therefore, it appears that both elastic band-resisted and weighted-chain training in the powerlifts can lead to similar or greater increases in strength and greater increases in power outputs in resistance-trained subjects in comparison with isoinertial training in the same movements. Additional benefits of elastic band-resisted and weighted-chain training may include lower joint soreness and greater rates of power development. However, whether such training modalities improves rate of force development (RFD) is unclear.

***

What are the limitations?

The main limitations of the literature at present are as follows:

1. It is unclear whether the effects of the variable resistance training modality are beneficial because they provide a novel stimulus to the resistance-trained athletes. This is particularly important for the powerlifts, as athletes are likely to have performed the isoinertial equivalents extensively for many years and may therefore see little further improvement over the course of a study. Indeed, this could be one of the mechanisms behind the reported huge success of the technique in experienced powerlifters at Westside.
2. It is unclear whether the use of variable resistance training of this kind leads to greater hypertrophy in comparison with isoinertial training. Studies in isolated joint exercises indicate that it produces comparable results (e.g. O’Hagan, 1995) but this may not be the same with the big, compound movements. If it was found that hypertrophy was different between these modalities, this could also provide a possible explanation for the apparent success of this training technique in powerlifters, given the importance of muscle mass for that population.
3. It is not clear to what extent the variable resistance training produces different regional muscular adaptations to isoinertial training. Indeed, Bloomquist (2013) recently found that partial squats produced different regional muscular gains to full squats.
4. It is unclear to what extent training with a focus on building strength later in the strength curve is directly applicable to geared powerlifting, where the suits provide elastic support at the bottom of the movement. It is possible that some of the improvements in powerlifting performance that are ascribed to the use of variable resistance arrive as a result of this effect.
5. It is unclear to what extent the faster descents in the eccentric phase that are typically observed during variable resistance training contribute to an increased stretch-shortening cycle effect. This may be an effect that contributes to increased acute power outputs, increased bar speed and possibly rate of force development.
6. Many of the chronic training studies used both isoinertial and variable resistance training modalities together in the variable training groups. This means that the variable resistance training group had a greater level of training variety than the isoinertial group, which may be beneficial for adaptations.

In summary, it is currently very unclear whether the apparent improvements in performance achieved through variable resistance training in combination with the powerlifts is caused by the novelty of the stimulus, the additional variety of the movements used, the sports-specificity of the movement ROM, the greater stretch-shortening cycle stimulus, the greater acute power outputs, or the greater hypertrophy stimulus (regional or total muscle).

***