Back Again? Tracking Possible Causes of Low Back Pain after Lower-Limb Amputation
January 2015 Issue
Periodic low back pain (LBP) has reported prevalence rates of 20 to 40 percent in the general population. For people with lower-limb amputations, it's an even bigger problem: Among those with transfemoral amputations, cited prevalence rates range from 50 to 87 percent.
For about half of these individuals, the occurrence is described as "occasional" or "a few times per month," but for a quarter to a third of the individuals with transfemoral amputations who suffer from LBP, the pain is described as "frequent" or "several times a week."1-4
The effect of the level of amputation as a risk factor for LBP is inconsistent, with some studies citing increased prevalence rates among individuals with amputations at the transfemoral level while others suggest comparable rates at the more distal knee disarticulation and transtibial levels.1-4
The studies examined in this article have largely focused their observations on subjects with transfemoral amputations, but likely have broader application for those with more distal amputation levels. This article presents different factors and considerations that may or may not influence LBP in individuals with lower-limb amputations.
Clinically, one of the first considerations associated with reports of LBP is the height of the prosthesis. If there is a leg-length discrepancy (LLD) between the sound and prosthetic sides, logic suggests this would create a nonlevel pelvis and associated strains on the lower back. However, while this is commonly suspected in clinical practice, there is limited published evidence to confirm this relationship.
To address the clinical utility of this theory, a 2009 study was performed through the VA Puget Sound Health Care System - Seattle Division.5 Drawing from a local database of people with transfemoral amputations, the authors excluded those individuals whose back pain might stem from outside influences. Thus, individuals undergoing tumor treatment, those with elevated BMI levels, those whose back pain predated their amputations, and those with significant depressive symptoms were excluded from the study.
The remaining 17 subjects who met the authors' inclusion criteria were then asked a simple question that would place them in one of two comparison groups: "Since your amputation, have you experienced persistent, bothersome back pain?" Those who said yes comprised the "amputee pain" group, while those who said no comprised the "amputee no-pain" group.8 The two groups were otherwise similar; average years since amputation, daily hours of prosthetic usage, age, and BMI demonstrated no significant differences.
Instead of relying on the rather subjective clinical technique of assessing the ideal height of a prosthesis and the resultant level pelvis by manually palpating bony landmarks on the pelvis, the authors relied on a more objective approach by assessing their subjects in the gait lab. As part of an additional study that will be described shortly, reflective markers were positioned over the hip joint centers of both the sound and affected limbs. A Vicon motion analysis system was used to measure the average vertical distance from the markers to the floor during a period of quiet standing.
As might be expected, LLD was identified in 14 of the 17 subjects. In the amputee pain group the average prosthetic side was 8.4mm shorter than the sound side. For the no-pain group the LLD was effectively no different from the pain group, with an average prosthetic side length that was 8.3mm shorter than the sound side. Thus, modest LLD during quiet standing failed to demonstrate any correlation to the presence of LBP.5
The Vicon system allowed further observation and analysis of dynamic LLD. How did the relative heights of the hip joint centers compare during ambulation? The maximum height of the sound side hip joint center at its peak during single leg stance was consistently higher than that measured on the prosthetic side during the corresponding moment in gait. Unlike the measurement of quiet standing, this differential was actually greater in the no-pain group than that measured in the pain group but again failed to reach any statistical significance.5 A similar analysis was performed during the double stance period of gait, with both the sound side and prosthetic side referenced as the leading or trailing limb. The hip joint center height of the intact limb remained longer than that of the prosthetic limb during these gait phases as well, with no significant differences in LLD observed between the two cohorts with transfemoral amputations.5 At least within this cohort of 17 patients, LBP could not be reasonably attributed to slight differences in leg length.
Lumbar Spine Mechanics
Following publication of the previous study, Morgenroth et al. published a second paper based on additional observations with the same cohort of individuals with transfemoral amputations. In this second effort, the researchers approached the problem from a different angle, seeking to record the amount of lumbar spinal movement experienced by people with transfemoral amputations and to determine if differences in this movement might be associated with the presence or absence of LBP.6
Expanding their narrative on the experimental protocols, the authors describe how the subjects, nine with LBP and eight without, were brought to the gait lab where a Vicon motion analysis system captured kinematic data using reflective markers placed on their hands, arms, heads, trunks, legs, and feet. In addition to the standard marker array, the authors placed three additional markers on the backs of the subjects to track spinal motion during gait. A smaller control group of able-bodied individuals participated in the study to provide comparison data.
Because the two test groups were similar in demographics other than the presence or absence of LBP, this experiment revealed that the only significant difference from one group to the other was the average amount of lumbar spine transverse rotation that occurred during gait.6 Compared to the able-bodied controls, transverse plane rotation at the lumbar spine was out of sync for all participants with amputations, with peak rotation occurring earlier in the gait cycle. However, for the amputee no-pain cohort, the amount of motion observed at the lumbar spine was nearly identical to that seen among able-bodied walkers. By contrast, the amount of motion experienced by the amputee pain cohort was nearly 40 percent more than that observed in the other two groups. While the study's observations can't be used to assign causation (i.e., that elevated lumbar spine movement causes back pain), the study does suggest a likely relationship between the mobility of the lumbar spine, particularly in the transverse plane, and the presence of back pain.6
Given that the effect of amputation levels on the presence and intensity of LBP has proven somewhat inconsistent, it stands to reason that at least some of the causative agents are common across lower-limb amputation levels. The absence of the physiologic foot and ankle complex and the resulting compensatory gait strategies appear to place those with lower-limb amputations at risk for the development of LBP. Additionally, these underlying pathomechanics appear to create unusual forces across the lumbar spine, especially in the transverse plane.
Speculatively, the problem may begin with the inadequate push-off provided by almost all modern lowerlimb prosthetic components. In the absence of the physiologic pushoff event to change the body's trajectory from forward and down to forward and up, the sound side limb experiences a much more abrupt loading moment. (Author's note: For more information, read "Collision Course: Understanding the Role of Prosthetic Feet in Sparing the Sound Side Limb," The O&P EDGE, October 2014.) Such moments are likely translated from the heel of the sound side limb up to the lower back with attendant, repetitive damage potentially experienced with every step.
But there are additional considerations to this push-off problem. When propulsion does not come from the foot and ankle, it has to be generated somewhere else. Across lower-limb amputation levels, this appears to come from elevated sound side hip extension coupled with augmented hip flexor activity on the prosthetic side.
In an older study about conduction on individuals using transfemoral prostheses, Seroussi et al. provide some quantitative figures on the extent of the compensations.7 During able-bodied locomotion, the hip extensors fire eccentrically during loading response, aiding the body's forward propulsion. It's a supplemental propulsive activity, generally timed as it is to coincide with the push-off that is generated by the contralateral limb. Among able-bodied walkers, the magnitude of this eccentric hip extensor power burst averages 3.6 joules, supplementing the 34 joules produced by contralateral push-off.7 But what happens when the propulsive activity at the foot and ankle are replaced with a much less effective prosthetic substitute? The power burst generated by the hip extensors of the sound side leg nearly triples, going from an average of 3.6 joules to 9.9 joules.7
During this same moment in gait, the hip flexors of the trailing leg eccentrically "pull off" while the ipsilateral plantarflexors push off. When a reduced push-off occurs through prosthetic substitution, the magnitude of the prosthetic side hip flexors' "pull-off" increases, despite the fact that the prosthetic limb weighs substantially less than the physiologic leg.7
Viewed collectively, these three phenomena may begin to explain some of the pathomechanics that generate LBP. In the presence of a compromised prosthetic push-off, there is a compensatory power burst at the sound side hip extensors coupled with an augmented pull-off by the affected side hip flexors, both of which occur as the body's momentum brings the leading sound side limb into abrupt contact with the ground. It is possible that some individuals are able to better withstand these altered pathomechanics, such that increased lumbar spine movement and the associated LBP affect some more than others.
Back Strength and Endurance
Irrespective of the specific pathomechanics that may contribute to instability and pain in the low back, data from Friel et al. suggest that these events may be mitigated when there is adequate low back strength and endurance.8 Their work is unique among the papers considered in this article in that their study cohort comprised individuals with transfemoral (n=8) or transtibial (n=11) amputations. The subjects were asked to rate their back pain within the last week on a visual analog scale ranging from no pain (0) to severe pain (10), with individual responses ranging from 0 to 7. In addition to establishing perceived pain levels, each subject was evaluated with respect to such variables as the length of the iliopsoas and hamstring muscles, the strength of the abdominal and back extensor muscles, and the endurance of the back extensor muscles.
Before correlations could be drawn, the researchers had to define LBP for their cohort. They did so using two conventions, describing them as the pain-2 and pain-5 variables. In the pain-2 convention, the ten subjects who reported their pain levels below 2 were defined as the "minimally painful group," with the remaining nine subjects with pain scores above 2 defined as the "painful group." However, recognizing the value of further scrutinizing those individuals with more severe back pain, the pain-5 convention raised the cutoff value such that only those four individuals reporting pain values of 5 or more were considered "painful" while the remaining 15 subjects were considered "minimally painful." For the purposes of this article, the data observed with the pain-5 convention will be presented.
Back extensor strength was assessed using a trunk-raising test in which subjects were placed in a prone position with straps and manual stabilization across their buttocks, thighs, and tibial regions. To score a grade 3 on the 5-point scale, subjects were asked to place their hands behind their low back and raise their xiphoid process off of the table. Those who were able to do so were invited to attempt a grade 4 performance, with hands behind their back, bringing their mid- to lower-abdominal region off of the table. If this was achieved, subjects attempted a grade-5 performance, placing their hands behind their head and bringing their mid- to lower-abdominal region off of the table. For those subjects with LBP rated below 5, the average strength score was 3.4.8
Subjects who were unable to perform a grade 3 on this test were instructed to attempt a grade-2 performance, in which they placed their arms by their sides and lifted their head and upper sternum off of the table. The average strength score for those four individuals with the most severe back pain was 2.25.8
Back extensor endurance was evaluated using the Sorensen test. In this planking exercise, subjects assume a prone position with their pelvis and lower limbs strapped to a table. Their trunk, arms, and head extend past the table, where they lean on a chair until the test begins. Subjects are asked to bring their arms across their chest and hold the unsupported upper body in a horizontal position for as long as possible. Healthy participants are able to sustain this position, on average, for about three minutes. By contrast, among those subjects with LBP less than 5, the average time was 30 seconds. For those with LBP reported at 5 or more, the average time was only 13.5 seconds.8
Taken collectively, these studies provide some initial insights into possible contributors to the elevated occurrences of LBP experienced by people with lowerlimb amputations. LLD, though long suspected as a source of the problem, appears to be less concerning provided it is fairly small. Rather, it is excessive movement in the lumber spine that contributes to elevated LBP rates and severities, which are likely due to a combination of weakness and fatigue in the back extensor muscles that are unable to withstand the forces created with compensatory gait mechanics.
Phil Stevens, MEd, CPO, FAAOP, is in clinical practice with Hanger Clinic, Salt Lake City. He can be reached at .
- Hammerlund, C. S., M. Carlstrom, R. Melchior, and B. M. Persson. 2011. Prevalence of back pain, its effect on functional ability and health-related quality of life in lower limb amputees secondary to trauma or tumour: A comparison across three levels of amputation. Prosthetics and Orthotics International 35 (1):97-105.
- Behr, J., J. Friedly, I. Molton, D. Morgenroth, M. P. Jensen, and D. G. Smith. 2009. Pain and pain related interference in adults with lower-limb amputation: Comparison of knee disarticulation, transtibial, and transfemoral surgical sites. Journal of Rehabilitation Research and Development 46 (7):963-72.
- Stam, H. J., A. V. Dommisse, and H. J. Bussmann. 2004. Prevalence of low back pain after transfemoral amputation related to physical activity and other prosthesis related parameters. Disability and Rehabilitation 26 (13):794-7.
- Kulkarni, J., W. J. Gaine, J. G. Buckley, J. J. Rankine, and J. Adams. 2005. Chronic low back pain in traumatic lower limb amputees. Clinical Rehabilitation 19 (1):81-6.
- Morgenroth, D. C., A. Shakir, M. S. Orendurff, and J. M. Czerniecki. 2009. Low-back pain in transfemoral amputees: Is there a correlation with static or dynamic leg-length discrepancy? American Journal of Physical Medicine and Rehabilitation 88 (2):108-13.
- Morgenroth, D. C, M. S. Orendurff, A. Shakir, A. Segal, J. Shofer, and J. M. Czerniecki. 2010. The relationship between lumbar spine kinematics during gait and low-back pain in transfemoral amputees. American Journal of Physical Medicine and Rehabilitation 89 (8):635-43.
- Seroussi, R. E., A. Gitter, J. M. Czerniecki, and K. Weaver. 1996. Mechanical work adaptations of above-knee amputee ambulation. Archives of Physical Medicine and Rehabilitation 77 (11):1209-14.
- Friel, K., E. Domholdt, and D. G. Smith. 2005. Physical and functional measures related to low back pain in individuals with lower-limb amputation: An exploratory pilot study. Journal of Rehabilitation Research and Development 42 (2):155-66.