The safest way to deadlift is with a neutral spine, right? Ask anyone with an ounce of sense who's spent time in the gym and they'll tell you that the incidence of lifters who injure themselves through rounded-back deadlifting grossly outnumber that of lifters who injure themselves with neutral-spine deadlifting.

So if back health is your primary goal, your needs are best served staying in neutral when you pull heavy loads. End of that particular discussion.

However, that's not the whole deadlifting story. Most of the world's top conventional deadlifters realize that pulling with a rounded back allows them to hoist greater loads.

This is puzzling to many, and there are countless self-conscious lifters out there who wonder why they're so much stronger with a rounded back. Many automatically assume that their body is somehow dysfunctional if they're stronger when rounded-back deadlifting, and more often than not this simply isn't the case.

I'd go as far as to argue that the vast majority of lifters, even when proficient in the hip hinge and with sound levels of core stability, are stronger when pulling with a rounded back.

Many lifters find that when the weight creeps over 90% of their 1RM, their back starts to round, and the rounding gets more pronounced as the load approaches true maximal weight. I'm one of these individuals. I wish I could pull heavy loads with a neutral spine, but unfortunately I'm unable to do so.

With what's deemed "perfect form," characterized by maintenance of the natural lordotic and thoracic spinal curvature with the extension revolving around the hip joints, I'm able to pull 425 pounds. If I allow my back to round considerably, I can pull 565 pounds. Why in the world would there be a 140-pound discrepancy between these two variations?

Let's assume that grip isn't a limiting factor for deadlifting strength. Some people indeed have a weak grip, but if this is the case, straps can be worn to allow for the lifting of maximal loads. Until around six months ago, I'd have told you that deadlift strength is limited by the strength of the glutes and hamstrings (the hip extensors). However, several things have caused me to change my stance.

First, I had a conversation with T Nation contributor Tim Henriques regarding spinal behavior in a deadlift. He conducted an informal experiment with a group of his students where he measured their spinal length while they were deadlifting using a length of string.

Without fail, the length of string was always longer when positioned in the bottom of a deadlift compared to the top. In other words, the spine is lengthened in the bottom position and shortened at the top, indicating that the deadlift is a dynamic spinal exercise when loads are heavy. Research in the literature supports Tim's findings, as does video analysis of world-class conventional deadlifters. The body wouldn't do this if the hip extensors were the weak link and the erectors were strong enough to keep an arch.

Second, I saw a video from Mike Boyle showing a guy busting out a 315-pound single-leg Romanian deadlift. However, this guy can't perform a 600+ pound bilateral RDL. At Mike's facility, MBSC, there are plenty of guys who can single-leg RDL much more than half of what they can bilateral RDL.

Sure, it takes coordination and tons of practice to achieve this, but even less-coordinated guys will find this to be true. If you take the balance-component out of the equation and test your strength with a Hammer Strength deadlift using their squat-lunge machine, then compare it to what you can achieve off one-leg, most people's single-leg strength will outperform half of their double-leg strength (they'll lift more than half the weight on one leg than they do on two legs). This is indicative of a bilateral strength deficit in the deadlift pattern. If the hip extensors were the weak link, this bilateral deficit wouldn't exist.

Third, I've come across a lot of research over the past couple of years, and piecing the research together presents a logical and compelling case. I should mention that this "weak link hypothesis" is just a theory that requires further research. I receive emails from time to time from lifters who report increased deadlift and rack pull strength from performing barbell hip thrusts, which completely contradicts my theory and lends support to the hip extensors being the weakest link.

Furthermore, I wasn't too impressed with the EMG data for the core muscles (lumbar erector, lower rectus abdominis, and internal and external oblique) during a deadlift when I measured them a while back – another blow to my weak link theory. I'm sure, like everything else in strength training, it depends on the lifter.

Furthermore, the more I learn about the human body, the more I realize that reductionism doesn't always apply, but that's another topic. Now let me discuss why the spine is stronger when flexed.

With a neutral spinal posture, you rely predominantly on the spinal extensor muscles (mostly the erector spinae, also the multifidus and a little bit of quadratus lumborum) to keep your spine in position and maintain the arches. If you're freakishly back strong and you don't have much of an imbalance between hip extensor and back strength, you can pull this off.

However, if you're not freakishly back strong, or if your hips are so damn strong that they overpower what your spinal extensors can stabilize, then the back will have to round.

If you stay arched, here's what "holds the bridge up," or in other words, what keeps the spine in neutral position:

  • Contractile elements of the spinal extensor muscles. Basically, the actin and myosin crossbridges. It's worth mentioning that the erector spinae have greater moment arms (better leverage) when the spine is in neutral compared to when the spine is flexed.
  • Intraabdominal pressure (IAP). When the diaphragm contracts, it pushes down on the "inner core," and muscles such as the pelvic floor, multifidus, and TVA resist this pressure, which pushes back against the spinal column and contributes to the extension torque. The heavier the loads, the greater the IAP production requirements.
  • Co-contraction for stability. Muscles surrounding the core such as the rectus abdominis and obliques will contract to provide spinal stability. It should be noted that these muscles actually produce a flexion torque when they contract, meaning that they'd encourage rounding if not balanced out by increased erector contraction. But the net result from co-contraction is a more stable spine.

Figure 1

Obviously, the vast majority of support in neutral-spine deadlifting comes from contracting muscles – probably supporting 98% of the tension.

If you round your back, now you get active plus passive support. Sure, the bridge isn't truly "held up" when rounding, but the spine does indeed stabilize in a rounded position. Here are the contributors to the stabilization of a flexed spine:

  • Contractile elements of the spinal extensor muscles. The back extensors can produce more force when flexed compared to when extended (even though their leverage is diminished, the muscles are stronger in flexed positions). This is especially true for the small but surprisingly strong posterior core muscles called the multifidi. It should be noted that when fully rounded, most of the spinal extensors become myoelectrically silent, which has been coined the "flexion-relaxation phenomenon," or FRP. Essentially, the erectors don't have to contract much since all the tension is being supported by passive structures, but this doesn't happen until full-flexion.
  • Passive elements of the erector spinae muscles. Even if a muscle doesn't contract, it can provide considerable passive tension. Imagine stretching out the meaty thoracic extensors – they could still provide some decent tension even if they weren't contracted. Furthermore, if an active muscle moves into an eccentric contraction (is stretched while activated), the titin filaments will contribute heavily toward passive force contribution as they behave like springs.
  • Thoracolumbar fascia. The thoracolumbar fascia (TLF) provides spinal extensor torque when stretched (when the back rounds). This contribution is increased when the TLF is stretched while muscles attaching to it are activated. Contraction of the erectors underneath "inflate" the TLF while contraction of muscles such as lats, glutes, external oblique, low traps, transverse abdominis (TVA), quadratus lumborum, and internal oblique pull the various layers tight and help prevent further flexion of the spine.
  • Spinal ligaments, joints, and discs. As the spine rounds over considerably, it begins to "hang" on the spinal structures. The more you round, the more slack they pick up, but they don't contribute heavily to the movement until you're really rounded over near full flexion.
  • Increased intra-abdominal pressure (IAP). It's worthy of mention that the body can produce more IAP in spinal flexion than it can when extended.
  • Increased co-contraction strength. It should also be noted that the anterior core is stronger in flexion – the rectus abdominus, transverse abdominis (TVA), and external oblique (but not the internal oblique) are all stronger when the spine is flexed, indicating that core stability through co-contraction is stronger in slight flexion.

Still, the majority of contribution in rounded-back deadlifting comes from active muscle contribution, but now you're getting considerable contribution from passive structures and you've got more stabilization through IAP and co-contraction. And this is what allows for the hoisting of heavier loads.

In my case, the factors mentioned above contribute to allowing me to pull 140 more pounds in a conventional deadlift. I'm sure that some of the factors contribute much more heavily than others, but I'm not sure as to the exact proportions.

My guess is that the increased strength is mostly because the spinal extensors are simply stronger in flexion, and that the increased IAP and core strength in flexion help prevent the lumbar spine from flexing too far forward and therefore protecting against injury.

As mentioned, the safest spinal posture is a neutral position, with no incidents of flexion or hyperextension in any segments. However, studies show that powerlifters flex their spines considerably, but they avoid injury simply because they avoid end-ranges of lumbar flexion.

Furthermore, if you examine where most of the rounding occurs in powerlifters, a vast majority takes place in the upper back (the thoracic spine), and this strategy is safer compared to the other way around (mostly lumbar rounding).

If you choose to go ahead with rounded-back deadlifting, make sure that your lumbar spine doesn't approach end-range flexion.

Let's say that each lumbar segment can flex 13 degrees. Make sure you only go to around 8 degrees or so and keep 5 degrees "in the tank." Just keeping two degrees in the tank (11 degrees of lumbar flexion in a particular motion segment) has been shown to cut the disc stress in half compared to full rounding (13 degrees of lumbar flexion). But even 50% of maximal disc stress is too much – keep 5 degrees in the tank and you'll be much better off.

Of course, it's nearly impossible to consciously control the spine at the segmental level during heavy deadlifting, but just ensure that you "lock up" your lumbar spine plenty shy of full flexion. You can toy around with video-analysis to help you learn what your full lumbar ROM is and then figure out how to stay in the mid-ranges when deadlifting.

As you gain proficiency in rounded-back deadlifting you'll learn to:

  • Set up with the scapulae pulled apart (protracted).
  • Set up with slight lumbar rounding, and a majority of rounding coming from the upper back.
  • Brace the core and create considerable IAP to help lock down the lumbar spine.
  • Maintain whatever spinal position is chosen and not allow the position to buckle.

This last point is incredibly important. Here's a quote from Konstantin Konstantinovs, one of the world's best deadlifters (he's Latvian and raw deadlifted 939 pounds without a belt):

"That I will break my back I heard from as long as I started to deadlift, or for the last 16 years. I have always deadlifted with a rounded back. My legs have always been lagging in development, but results in deadlift have always been increasing. My back is prone to injuries only when I squat with a heavy weight, but when I deadlift, my back remains in the same rounded position throughout the lift, irrespective of whether I can lift the weight or not, and this protects it from injury. But you need very strong abs if you want to deadlift like that."

Now what? If you're stronger when deadlifting with a neutral spine, then lucky you. This article doesn't apply to you, so carry on. But if you're stronger when deadlifting in a flexed spinal position, here are some strategies to consider:

Stick to submaximal training with "perfect posture."

Stay between 60-80% of 1RM and make sure your spinal position never breaks down. Avoid taking the sets anywhere close to failure or you'll inevitably end up rounding your back.

This strategy won't beat you up so much, so you'll be able to deadlift more frequently. You can go with a mixture of explosive and controlled reps. For example, 3 x 3 smooth, controlled reps, followed by 5 x 1 explosive singles. You could also do a heavier day (something like 3 x 3-5 with 80% of 1RM) along with a dynamic day (something like 5 x 1 with 60% or 1RM). This is the safest strategy.

Rely predominantly on barbell plus accommodating resistance.

Bands and chains may provide a good compromise – the load will be lighter at the bottom of the lift, which is the portion that's most difficult for the spine and hips, but the top of the lift will be heavier. This will allow for good spinal posture throughout the entire ROM while still overloading the system, especially up top as the band or chain tension really kicks in.

You'll need to use considerable band/chain resistance to really off-set the loading curve, such as 50% barbell resistance and 50% band or chain resistance. This practice also allows for a greater percentage of the duration of the lift to be spent in acceleration. This too is a very safe strategy.

Go heavy but pull back on the reigns a little bit with "sub-optimal posture."

Try to use ideal form with warm-up sets. As the weight gets heavier, allow for controlled rounding. Push the sets hard but avoid taking them to failure to avoid hyper-flexion.

Don't ever truly "max out." Only go to 95% of 1RM. This is what I've done for much of my lifting career. I allow the spine to flex, but I control the flexion and don't ever push my sets to complete failure or truly max out. If I did, I'd inevitably end up rounding too far forward and damaging my spine. This is more risky, but can be done successfully if you're consistent and diligent.

Go balls to the wall and accept the risk with "horrible posture."

Throw caution to the wind. Let the spine round as far as need-be. Go to complete failure and lift the heaviest loads possible. If this is the strategy you choose, I can guarantee you that you'll eventually suffer severe injury and have a lifting career hampered by constant pain and injury to show for it. This is the most unsafe option, and unfortunately there are lifters out there who go this route and must learn the hard way to change their ways. Take it from an experienced lifter and avoid this route and live to train another week – if not you'll regret it, mark my words.

Most people are stronger when rounding their back during deadlifts. This is natural and shouldn't be thought of as "dysfunctional." You can choose to stick with lighter loads and deadlift in neutral spinal position, or to roll the dice and allow for some rounding. If you go this route, make sure you build up the awareness and discipline to avoid fully flexing the lumbar spine while lifting heavy loads.

Bret Contreras is considered by many to be the world’s foremost expert on glute training. He has turbo-charged the fitness industry by introducing effective new exercises and training methods for optimal glute development. Follow Bret Contreras on Twitter