Barbell Medicine - From Bench to Bedside

Leg Length Discrepancy (LLD, a.k.a. anisomelia) is defined as a measurable difference between the length of the lower extremities, and is often blamed for numerous issues related to pain or dysfunction due to the asymmetry. It’s important to point out the underlying flawed premise here: that a structural “problem” necessarily results in symptoms. We’ve discussed at length the flaws of such a biomedical approach in the past, and will continue to do so in the future. Pain is a complex experience that cannot be reduced to simple anatomical observations that might superficially seem problematic (see Lederman 2011).

A guiding principle to follow when discussing potential correlates to pain is “does it matter?” And if “it” matters, we need to figure out 1) when it matters, and 2) what can / should we actually do about it?

Key Points:

  1. Leg Length Discrepancy (LLD) has a high prevalence rate, with ~90% of the population displaying a <1.0 cm difference.
  2. There is very weak evidence of poor quality demonstrating correlation between a particular leg length discrepancy and symptoms; broad generalizations should not be drawn from the available data at this time.
  3. It is important to recognize that we adapt to such deviations from textbook “norm”.

Two variations of LLD have been described: “Anatomical” and “Functional”.  An “anatomical” LLD describes structural differences in femur and/or tibia length, either naturally occurring in the course of development or acquired at some point during life (e.g., due to fracture, bony disease, or joint replacement). “Functional” LLD isn’t as straightforward — and although frequently discussed, lacks supporting research evidence. Typically the discussion of functional LLD is centered around pseudoscientific, poorly defined ideas such as “pelvic torsion”, “tight” vs “loose” muscles, or “subluxations”.

How-To Measure a Leg Length Discrepancy

Many perform assessments of LLD in clinics/gyms via palpation, tape measures, or blocks (to level the pelvis). Overall, the available research shows that such approaches are ineffective. A more accurate assessment can be obtained via radiological imaging, such as X-ray or CT scanogram. Sabharwal 2008, Gibbons 2002, Cooperstein 2017

However, even though we can measure LLDs, we should ask few questions before we pathologize it and turn the finding into an issue. The questions include:

  • How common are leg length discrepancies in the general population?
  • How reliable are measurements of leg length discrepancies?
  • Do leg length discrepancies correlate well with specific signs or symptoms?

Diving into the Research


In 1978, Richard Gross was the first to try and quantify the amount of LLD necessitating intervention. At the time, he concluded any LLD less than 2 cm likely wasn’t an issue. The purpose of the article in review this month by Gordon et al was to assess the current state and quality of evidence regarding LLDs since Gross’s initial publication.

The authors performed a search for medical literature assessing the effects and treatment of leg length discrepancies. The authors excluded any level 5 evidence (case-reports, expert opinion, and personal observations), articles not written in English, and any article assessing the risk(s) of procedures resulting in LLD. The authors categorized their findings into 2 broad topics: natural history of LLD and gait analyses of patients with LLD.

Recall that the  “natural history” is typically defined as the usual trajectory of a supposed issue without intervention. In essence, these studies are examining a group of patients with LLD to assess for any associated problems.


The authors found that approximately 90% of the population has an LLD (see figure 1). On average, a 5.2 mm difference has been observed.Knutson 2005

According to Knutson et al, 10% of the population has equal lower limb lengths, ~50% has a 4 mm LLD or less, and ~90% has a 10 mm LLD or less.Knutson 2005

Given the high prevalence of LLD, we must ask whether an LLD of a particular measurement matters (i.e., is it correlated to pain and/or dysfunction)?

Not necessarily. Knutson’s review concludes:

“In summary, childhood-onset anatomic leg-length inequality appears to have little clinical significance up to 20 mm. Several authors agree, most recently with Kakushima et al who stated: ‘Therefore, although conflicts in the literature exist, 3 cm of LLD [leg length discrepancy] can be characterized as a minimum LLD, which should be treated in the clinical practice’. This estimation of clinical significance dovetails nicely with the findings on the effects of LLI, particularly pelvic torsion. Passive structural changes – pelvic torsion, mild lumbar scoliosis, facet angulation, changes in muscle length – seem capable of compensating for anatomic LLI of up to 20 mm. Past the ~ 20 mm point, passive structural changes give way to active muscular compensatory measures.” [emphasis added]

Based on this information, 2-3 cm seems to be the minimum threshold for warranting investigation and possible intervention, and this is likely correlated more to quality of life than to pain.

Taking a closer look at the articles included in Gordon et al, we find inconclusive evidence on LLD and future issues such as joint replacement, osteoarthritis, pain, etc. The available data aren’t even capable of showing whether the long leg or the short leg are at greater risk for negative effects.

What Does Other Research Show?

Joint Replacement

Tallroth et al 2005 examined 100 participants undergoing hip-replacement surgery. 39 displayed a longer right leg and 42 with left. Nineteen participants did NOT have an LLD. Sixty-eight participants had the hip replacement performed on the longer leg (mean difference in leg length was 7.5 mm, standard deviation 4.7 mm), and 13 on the shorter leg (mean difference in leg length was 4.4 mm, standard deviation 3.2 mm). 

Tallroth et al 2017 – 193 participants initially underwent X-ray and then followed-up 29 years later to see who received joint replacements of the hip or knee. The breakdown of participants and LLDs was as follows:

  • 24 (12%) no LLD
  • 62 (32%) 1 – 4 mm LLD
  • 74 (38%) 5 – 8 mm LLD
  • 21 (11%) 9 – 12 mm LLD
  • 12 (6%) over 12 mm LLD

Of the 193 participants, only 16 (8%) went on to receive a joint replacement for the diagnosis of primary osteoarthritis. The authors go on to report 10 participants had replacement completed on their longer leg and 3 on their shorter leg. This is a low number of participants receiving joint replacement from this cohort and it would be very difficult to conclude receiving a joint replacement was primarily due to an LLD. Three participants WITHOUT an LLD went on to receive a joint replacement as well, leading us to further question LLD as a major correlate, if one at all, to the need for a future joint replacement.

Overall, this evidence isn’t sufficient to support LLD as a major correlate leading to joint replacement. 

“Degenerative” Findings

Before diving into these articles, let’s recall we have a plethora of data showing asymptomatic degenerative findings (deviation from textbook norm) in various body regions (see table 1). Many of the studies included in Gordon et al are examining the relationship between LLD and degenerative findings such as osteoarthritis or lumbar disc herniations. Even if we see a correlation between the presence of LLD and the presence of degenerative findings, this does not mean the person has symptoms or will later develop symptoms purely because of the LLD. 

Furthermore, such a biomedical premise further reduces the person experiencing symptoms to solely a biological tissue issue where we have mounting evidence has a large prevalence rate in the general population (as discussed above). 

With that said, let’s take a closer look at the included studies examining degenerative findings and LLD. 

Murray et al 2017– lumbo-pelvic x-ray imaging reviewed from 255 adults to assess LLDs and degenerative findings in the hips and lumbar spine. The authors noted an increased risk for degenerative joint disease of the hip and lower lumbar spine in patients with LLD > 5 mm. A potential issue with this study is the lack of full-length x-rays from pelvis and lower extremities. 

Ten Brinke et al 1999 – 132 participants being seen for neurological symptoms supposedly attributed to lumbar disc herniation were examined for an LLD. LLDs were measured indirectly (non-radiographically), making the results questionable. The authors report a mean LLD of 5.4 ± 5.2 mm (range of 0 – 26 mm). 104 participants demonstrated an LLD > 1 mm. 28 showed no LLD. 64 (62%) of the 104 participants with LLD displayed radiating symptoms into shorter leg. 

Harvey et al 2010 – 2964 patients with full-length x-rays were followed for 30 months and examined for increase in osteoarthritis. Overall, 14.5% of participants (n=429) displayed an LLD ≥ 1 cm and 0.9% (n=27) had an LLD ≥ 2 cm. Participants with an LLD > 1 cm were further examined. Of those participants, 53% demonstrated increased OA on the short-side, while 36% showed increased OA on the long-side. Interestingly, the incident rate (new cases) of osteoarthritis were not influenced by an LLD ≥ 1 cm or ≥ 2 cm over the 30 month follow-up period. Participants demonstrating an LLD ≥ 1 cm, compared to those with < 1 cm LLD, did have an increased odds of developing knee symptoms over the 30-month follow-up period (shorter limb = 15% vs 9%, OR 1.7, 95%CI, 1.2-2.4 and longer limb = 13% vs 9%, OR 1.5, 95%CI, 1.0-2.1). An LLD ≥ 1 cm, compared to those with an LLD < 1cm, did demonstrate a 1.3 times greater odds of having progressive knee osteoarthritis in the shorter limb over the follow-up period (95% CI, 1.0-1.7, 29% vs 24%). Significance wasn’t reached for progressive knee osteoarthritis symptoms in the longer limb over the follow-up period. Oddly, an LLD ≥ 2 cm did not demonstrate a significant increase in the odds of developing progressive knee osteoarthritis in the shorter limb. Add more importantly, only 6 of 26 participants with an LLD ≥ 2 cm demonstrated progress of knee osteoarthritis – this is an interesting finding given the authors’ premise that LLD increases the risk of developing knee osteoarthritis and claim, “…leg length inequality as small as 0.5 to 1cm increased the risk of prevalent knee osteoarthritis, primarily in the shorter limb.” One would think this would be observed in what are being considered “large” LLDs more so than smaller LLDs, but this doesn’t appear to be the case. 

Low Back Pain

Defrin et al 2005 – 33 participants with persistent low back pain were examined for an LLD and given a shoe insert to “correct” the discrepancy. The fact that the cohort was labeled with persistent pain should immediately raise questions regarding the premise of this study, since the chronicity of pain symptoms drastically diminishes the correlation between identifiable tissue pathology and symptoms (See Durmez 2017). 

The authors examined LLDs with ultrasonography and then divided participants into study vs control groups. The study group (22 participants) received a shoe insert to correct the LLD and the control group (11 participants) didn’t. Baseline measurements of pain intensity and function were performed. The authors reported a significant baseline difference between groups for pain intensity, and this matters when trying to assess the efficacy of an intervention. Ideally there are no significant differences in baseline characteristics between groups that could otherwise confound results. After 12 weeks of intervention, the participants were re-examined for pain intensity and functional status. The authors conclude:

“This study suggests that the correction of an LLD of 10mm or less can significantly reduce CLBP. Shoe inserts are simple, inexpensive, and noninvasive means for alleviating CLBP and are therefore recommended to be included in the treatment of patients with LBP who have mild LLD.”

Examining the data more closely, these are rather bold conclusions. The study uses a small sample size, has baseline differences in pain intensity, and shows large standard deviations in treatment outcomes – all calling into question such claims. Even if we wanted to follow this reductionist line of thought, we have higher quality evidence demonstrating that shoe inserts lack efficacy and have low-level evidence for effectiveness. Recall, efficacy and effectiveness are not the same thing.

Efficacy, in this context, is examining how well a treatment performs under ideal or perfect circumstances. Typically, interventions are examined for efficacy in well conducted randomized controlled trials. These studies help answer the question – “Does ‘X’ intervention actually work?”. 

Effectiveness is assessing a treatments generalizability to real-world situations such as dealing with humans in clinical practice. Many treatments appear effective (reduce pain and improve function) for a myriad of reasons, often involving placebo-like contextual effects. Treatments appearing effective may lack efficacy, thus questioning their necessity in clinical practice for long-term positive outcomes. 

In regards to shoe lifts  – Campbell et al completed a systematic review of the literature in 2018 on adults experiencing musculoskeletal symptoms. The authors concluded: 

“We sought evidence to answer fundamental questions for guiding clinical treatment of LLD for common painful musculoskeletal conditions. In the setting of mechanical LBP, hip, and knee OA, correction of LLD using a shoe lift may reduce pain, improve function and increase ROM; however, these benefits remain uncertain due to very low-quality evidence. We were unable to make evidence-based conclusions regarding the magnitude or proportion of LLD that should be corrected. More rigorous, high-quality studies evaluating which LLD-associated conditions benefit from shoe lift correction, shoe lift correction strategy, and relevant patient outcomes are required to guide clinical treatment. An appropriate comparison group would be helpful in this regard.”

The absence of a clear level of LLD necessitating intervention, and the lack of appropriate RCTs on the usage of shoe inserts negates our ability to draw any pragmatic conclusions.

Adaptive Responses to Leg Length Discrepancies

Khamis and Carmeli et al 2018 –  7 healthy participants were recruited to go through a simulated LLD gait analysis. Participants were equipped with shoes simulating LLDs of 5, 10, 15, 20, 30, and 40 mm. The authors found at 5 mm LLD compensatory strategies were employed. At 10 mm the authors noted significant differences in compensatory gait strategies involving shortening of the longer limb, lengthening of shorter limb, or both.

Song et al 1997 – 35 children recruited with LLD. The reasons for the LLD ranged from idiopathic (unknown), congenital, fracture, and dislocation. The included children displayed an LLD ranging from 0.6 to 11.1 cm. The authors found children with small mean discrepancies of 1.6 cm, no compensatory gait strategies were observed. However, children with larger mean LLD of 6.5 cm displayed a compensatory gait strategy of toe walking. Eight children displaying LLDs between 2 and 15.8 cm demonstrated pelvic obliquity (drop of pelvis on short side). Finally, 9 children displayed increased range of pelvic obliquity as a compensatory gait strategy.

Aiona et al 2014 demonstrated compensated gait strategies for LLD. Forty-five children recruited with an LLD > 2 cm. The average LLD was 4.6 cm (range 2 – 12.2 cm) – this is considered a large discrepancy compared to what we’ve been discussing. The LLDs were due to a variety of issues: Legg-Calve-Perthes disease, hip dysplasia, growth plate abnormalities due to trauma or infection, and congenital shortening of the femur or tibia to name a few. The authors also included 20 children in a control group for comparative gait analysis. Various biomechanical compensations were observed in the study group, from a single kinematic deviation to several adaptations (see table 2).

Interestingly, the authors found very similar gait velocities between groups. The control group had an average walking velocity of 1.3 ± 0.2 m/s (range 0.9 to 1.6 m/s) vs. the study group with an average walking velocity of 1.2 ± 0.2 m/s (range 0.8 to 1.5 m/s). The authors state:

“Our study demonstrated a variety of gait compensations for LLD. The magnitude of the difference and the location of the difference appear to be important factors in determining the compensation strategy. If the discrepancy was >7 cm, all patients used a combination strategy. As the differences became less, a greater variety of pattern choice was noted. For example, if >4 cm but <7 cm, only 3/11 chose multiple strategies, with the most frequent isolated pattern being pelvic tilt.”

The primary takeaway from the above articles is that compensation is evident in gait based on the size of the LLD. However, whether these compensations lead to any negative outcomes over the longer term is not known. We also don’t know if a certain size LLD warrants intervention given these compensatory strategies.

What’s the Take-Home Message?

Leg length discrepancies are readily identifiable in the general population, with some sources reporting an average difference of approximately 5 mm. Furthermore, screening for LLDs would not be advisable given the large prevalence rate in the asymptomatic general population and lack of strong correlation to symptomatic development or future negative effects. There is substantial controversy over a threshold LLD necessitating intervention, and after review of the available evidence, this remains unknown.

Many clinicians and coaches anecdotally report LLDs being causative for pain or performance issues, however, there is a paucity of evidence linking a particular LLD to either. Given what we know about the complexities of pain, this would be an extremely reductionist approach that at this time is indefensible based on current evidence. This means that if a patient reports in clinic with low back pain, it would not be recommended to check for an LLD. Low back pain has a high prevalence in the 4th decade of life and yet many clinicians check for LLDs to see whether it is related to the patient’s symptoms. However, if the patient is in their 40s, a finding of an LLD would certainly have been adapted to at this point. Even if we wanted to follow this line of clinical reasoning, the typical recommendation is then a shoe insert – an intervention greatly lacking in high quality evidential support.

Regarding LLDs and performance, at this time it would appear we adapt to a range of LLDs. Is there potentially a point where an LLD becomes problematic for a patient, sacrificing quality of life? Potentially, but we’ve yet to find a generalizable, predictive rule. Such scenarios are likely much more related to non-idiopathic situations involving congenital deficits, fracture, post-joint replacement, tumor, infection, etc. We even have examples of high-level athletes with LLDs. Gordon et al. discuss Usain Bolt having a reported 1.3 cm LLD and video analysis demonstrating “…right leg striking the ground with 13% more force and his left leg spending 14% more time on the ground.” This makes it difficult to single out such an isolated biomechanical finding and claim it as a “problem” necessitating fixing. Overall, Gordon et al conclude:

“In conclusion, the evidence for the effect of leg length discrepancy and the amount of leg length discrepancy that we should be treating is quite poor and probably has advanced little since Gross’s initial survey of pediatric orthopaedic surgeons.”

What does this mean for us as clinicians and coaches? 

We should ensure we aren’t making problems out of normative findings in the general population, and we should understand that either way, we are adaptable. We certainly should NOT be screening for LLDs in routine practice. Finally, attempting to make claims about an LLD’s relationship with pain or performance issues would not only be overly reductionist but stands in opposition to what evidence is currently showing us. There may be cases where a non-idiopathic LLD is an issue, but for the majority of clinicians and coaches, these cases are likely an exceedingly small fraction of real-world encounters.


  1. Gordon JE, Davis LE. Leg Length Discrepancy: The Natural History (And What Do We Really Know). Journal of pediatric orthopedics. 2019; 39(Issue 6, Supplement 1 Suppl 1):S10-S13. 
  2. Lederman E. The fall of the postural-structural-biomechanical model in manual and physical therapies: exemplified by lower back pain. Journal of bodywork and movement therapies. 2011; 15(2):131-8. 
  3. Sabharwal S, Kumar A. Methods for Assessing Leg Length Discrepancy Clin Orthop Relat Res. 2008; 466(12):2910-2922.
  4. Gibbons P, Dumper C, Gosling C. Inter-examiner and intra-examiner agreement for assessing simulated leg length inequality using palpation and observation during a standing assessment Journal of Osteopathic Medicine. 2002; 5(2):53-58.
  5. Cooperstein R, Lucente M. Comparison of Supine and Prone Methods of Leg Length Inequality Assessment Journal of Chiropractic Medicine. 2017; 16(2):103-110.
  6. Knutson GA. Anatomic and functional leg-length inequality: a review and recommendation for clinical decision-making. Part I, anatomic leg-length inequality: prevalence, magnitude, effects and clinical significance. Chiropr Osteopat. 2005;13:11. Published 2005 Jul 20. doi:10.1186/1746-1340-13-11.
  7. Tallroth K, Ylikoski M, Lamminen H, Ruohonen K. Preoperative leg-length inequality and hip osteoarthrosis: a radiographic study of 100 consecutive arthroplasty patients. Skeletal radiology. 2005; 34(3):136-9. 
  8. Tallroth K, Ristolainen L, Manninen M. Is a long leg a risk for hip or knee osteoarthritis?. Acta Orthop. 2017;88(5):512–515. doi:10.1080/17453674.2017.1348066
  9. Murray KJ, Molyneux T, Le Grande MR, Castro Mendez A, Fuss FK, Azari MF. Association of Mild Leg Length Discrepancy and Degenerative Changes in the Hip Joint and Lumbar Spine Journal of Manipulative and Physiological Therapeutics. 2017; 40(5):320-329.
  10. ten Brinke A, van der Aa HE, van der Palen J, Oosterveld F. Is leg length discrepancy associated with the side of radiating pain in patients with a lumbar herniated disc? Spine. 1999; 24(7):684-6. 
  11. Harvey WF, Yang M, Cooke TD, et al. Association of leg-length inequality with knee osteoarthritis: a cohort study. Annals of internal medicine. 2010; 152(5):287-95. 
  12. Defrin R, Ben Benyamin S, Aldubi RD, Pick CG. Conservative correction of leg-length discrepancies of 10mm or less for the relief of chronic low back pain. Archives of physical medicine and rehabilitation. 2005; 86(11):2075-80. 
  13. Durnez W, Van Damme S. Let it be? Pain control attempts critically amplify attention to somatosensory input. Psychological research. 2017; 81(1):309-320. 
  14. Campbell TM, Ghaedi BB, Tanjong Ghogomu E, Welch V. Shoe Lifts for Leg Length Discrepancy in Adults With Common Painful Musculoskeletal Conditions: A Systematic Review of the Literature. Archives of physical medicine and rehabilitation. 2018; 99(5):981-993.e2. 
  15. Khamis S, Carmeli E. The effect of simulated leg length discrepancy on lower limb biomechanics during gait. Gait & posture. 2018; 61:73-80. 
  16. Song KM, Halliday SE, Little DG. The effect of limb-length discrepancy on gait. The Journal of bone and joint surgery. American volume. 1997; 79(11):1690-8. 
  17. Aiona M, Do KP, Emara K, Dorociak R, Pierce R. Gait patterns in children with limb length discrepancy. Journal of pediatric orthopedics. ; 35(3):280-4.

About Michael Ray

Dr. Ray is the founder of Shenandoah Valley Performance Clinic in Harrisonburg, VA. He obtained a M.S. in Exercise Science from the University of South Carolina and graduated Magna Cum Laude with his Doctorate of Chiropractic (D.C.) from Sherman College of Chiropractic.

Read More by Michael Ray