Bone Density among Lower-Limb Amputees: Prevalence, Risk Factors, and Implications
In the provision of prosthetic care for patients with lower-limb amputation, there are certain characteristics that are immediately apparent to treating professionals. Limb length, muscle strength, painful neuromas, and marked limitations in joint ranges are just some of the factors that we take into consideration as we develop our treatment plans and help our patients prepare for the life challenges following lower-limb amputation. Other factors are less apparent and may escape our immediate recognition. Among these is the phenomenon of decreased bone density. Several articles in the recent literature elucidate the prevalence of decreased bone density in this patient population. The impact that decreased bone density has on such things as fracture risk, pain, and activity levels are not as well understood but have also received some scrutiny. The purpose of this article is to bring to light some of these findings and help practitioners appreciate the significance of bone density in the well-being of their patients.
Defining the Problem
While there are multiple techniques for quantifying the relative bone-mineral density (BMD) in a given bony segment, the most commonly used is dual-energy x-ray absorptiometry (DXA). The finding of DXA can be reported in three ways:
- Measured density in grams per square centimeter (g/cm2).
- Z-score, or the number of standard deviations (SDs) above or below the bone density values of age-, gender-, and ethnically matched peers.
- T-score, or the number of SDs above or below the mean bone density values for a healthy 30-year-old of the same gender and ethnicity as the patient.
Of these, the T-score is the most readily referenced. A T-score that is less than or equal to a single SD below the young-adult mean is considered normal. T-scores greater than 1 SD but less than 2.5 SDs below those of matched 30-year-olds are considered osteopenic. T-scores greater than 2.5 SDs below the young adult mean are categorized as osteoporotic.1
In recent years, numerous studies have examined the relative bone density values of residual limbs following amputation, comparing these against the values of the sound-side limbs or those observed in matched controls. With only rare, individual exceptions, the density values of the amputated limbs are consistently below those of controls. Indeed, the question is no longer whether or not the BMDs of residual limbs are reduced, but rather, how much lower are they? The answer to the latter question is somewhat variable and depends upon several factors.
The work of Smith et al. provides a good starting point because their publication is one of the most recent and presents findings on one of the larger cohorts. Reporting on a sample of 52 subjects with an average age of 62 years and a rather heterogeneous sampling of subjects with both transfemoral, transtibial, unilateral, and bilateral lower-limb amputations, the authors report that based on T-scores, 50 percent of their subjects presented with osteopenia, and an additional 38.5 percent presented with osteoporosis.1 While the findings of other, earlier authors are not quite so dire, the prevalence warrants further consideration.
Does Amputation Level Matter?
The answer to this question depends on who you ask. The observations of Smith et al. suggest that amputation level has no effect on BMD values at either the femoral neck or the total proximal femur.1 However, this observation may be influenced by other factors, as their sample was somewhat older and less active. The observations from other samples suggest that amputation level does matter and that the residual limb of individuals with transfemoral amputations have markedly lower BMD than the residual limbs of individuals with transtibial amputations. Reporting on a cohort of relatively high-functioning, World War II-era subjects, Kulkarni et al. found that individuals with transfemoral amputations had significantly lower BMD values at the femoral neck than their transtibial peers.2 Similarly, Leclerq et al., in their analysis of 99 lower-limb amputees, found that hip BMD was lower in the affected side than it was in the intact limbs and the percentage difference in BMD between limbs was significantly greater for subjects with transfemoral amputations than it was for transtibial subjects.3
In the most carefully controlled trial, Sherk et al. compared the BMD values of 14 amputee subjects, carefully matched with respect to such variables as age, activity level, and time since amputation. Seven subjects had transfemoral amputations and the remaining seven had transtibial amputations. DXA bonedensity values at the femoral neck were reported at 1.015 and 0.704 g/cm2 for the transtibial and transfemoral cohorts respectively. (By way of comparison, comparable BMD values for matched able-bodied controls in the same study ranged from 1.078 to 1.158 g/cm2, suggesting fairly substantial BMD compromise within the transfemoral cohort.) In terms of categorization, among the transtibial cohort, two subjects (29 percent) presented with osteopenia in the femoral neck of their amputated limb, with a single additional subject (14 percent) presenting with osteoporosis. By contrast, among the transfemoral cohort, a single subject (14 percent) presented with osteopenia and five subjects (71 percent) presented with osteoporosis in the femoral neck of the amputated femur.4
Mars and Venus?
It is commonly recognized that women are at greater risk for osteoporosis than men. Given this reality, it is surprising that only one published report striated its data by gender. The findings are quite pronounced. Reporting on 52 subjects, of which 13 (25 percent) were female, Smith et al. identified the following: DXA examination of the lumbar spine found that while the male cohort presented with normal BMD (T-score = -0.33), the mean density of the female cohort equated to a T-score of -1.65, placing them soundly in the osteopenic category irrespective of the BMD of their affected femurs. Similarly, the observed BMD values of the proximal sound-side femur in the two groups were markedly different, with the mean T-score of the male subgroup reported at -0.57, consistent with normal BMD values and that of the female subgroup reported at -1.68, identifying osteopenia in the soundside femur. These differences persisted in the BMD analysis of the proximal femur of the amputated side. Here, the mean BMD of the male subgroup categorized them as osteopenic (T-score = -1.38), while that of the female subgroup identified them as osteoporotic (T-score = -2.65).1 Thus, the elevated risk of osteoporosis among able-bodied women appears to carry over into those with amputations, with women demonstrating significantly lower BMD values than their male counterparts at all tested bony locations, with particularly concerning values at the affected femur.
Does Activity Level Impact BMD?
Again, the answer to this question depends on who you ask. Smith et al. failed to demonstrate a relationship between functional activity level and BMD values.1 However, only two of their 52 subjects were capable of "normal or near-normal walking," with the majority being incapable of ambulation beyond 50 meters without walking aides. In contrast, the data of Leclercq et al. suggest that those subjects who wore a prosthesis at least six hours a day had significantly greater BMD than those who wore their prosthesis less than that.3
The data of Yazicioglu et al. seem to support this later premise. 5 In their evaluation of very young patients with traumarelated, transtibial amputations, the only attribute found to significantly correlate with decreased BMD values at the proximal tibia was the presence of residual-limb pain. However, another correlate that fell just shy of statistical significance was that of exercise level (p = 0.07). This data would seem to suggest that those subjects whose residual-limb pain precluded elevated activity levels were more likely to have reduced BMD in their affected proximal tibia.5
A Product of Time?
Time appears to have an impact on BMD values, both with respect to overall age as well as the lapsed time between amputation and evaluation. In regards to the former, Smith et al. reported a negative correlation between advancing age and hip BMD.1 Bearing in mind that BMD decreases 3 to 7 percent with each decade following age 40, it is not surprising that the BMD values of the older cohorts of Smith1 and Kulkarni2 were substantially lower than those of the comparatively younger cohorts of Yazicioglu5 and Sherk.4
Smith et al. also identified a negative correlation with affected hip BMD and the duration of disability.1 This is consistent with trends observed in both the stroke and spinal cord injury (SCI) populations, reinforcing a relationship between decreased activity levels over time and progressive compromise to BMD.6—7
Therefore…What?The greatest concern with decreased BMD values among this population is elevated fracture risk. The relationship between T-scores and fracture risk within the able-bodied is well established. With each declining SD at a given site, the likelihood of a fracture at that location roughly doubles. So for the amputee cohort examined by Smith et al., even with the relatively young mean age of about 62 years, 48 percent carry a two-fold risk of hip fracture, and 33 percent face at least a four-fold risk of hip fracture.1 For the older cohort of Kulkarni et al., with a mean age of 73, the mean T-score at the affected femoral neck was reported at 2.26, suggesting that on average, these patients carried a fourfold risk of femoral neck fracture.2
Fortunately, the literature has thus far failed to support the assumption that the generally observed compromises to BMD among amputees has resulted in elevated fracture rates. In fact, what little has been published in this area would suggest that such fractures are quite rare, that they are more common in those with transtibial rather than transfemoral amputations, and that they tend to be the result of higher activity levels rather than reduced BMD. In their review of 341 lower-limb amputees, Gonzalez and Matthews reported only eight femoral fractures, with seven occurring to transtibial amputees and only a single involving a transfemoral amputee.8 Similarly, in their chart review, authors from the Mayo Clinic documented only 14 femoral fractures in 22 years, 13 of which occurred in transtibial patients with the remaining fracture in a patient with a through-knee amputation.9 In a third retrospective review of some 1,063 lower-limb amputees, only 23 fractures were identified, with the majority of these occurring in transtibial amputees.10
In the largest and most informative of the existing reviews, Bowker et al. collected caseload information from 20 orthopedic surgeons in the United States and Canada.11 This effort ultimately identified 85 fractures in individuals with lowerlimb amputations, of which only 35 occurred in patients with a transfemoral amputation. Several interesting points were further identified in the review. First, younger, more active amputees seemed to have the largest fracture risk, suggesting one of the possible reasons for the high number of fractures among transtibial amputees. Second, the possible influence of the prosthetic socket as either a causative or protective agent can be reasonably questioned. About two-thirds of the reported fractures occurred while amputees were wearing their prostheses. However, nine transfemoral subjects sustained femoral neck or intertrochanteric fractures despite the use of a hip joint and pelvic band as part of their prosthesis. Similarly, distal femur fractures occurred in eight transtibial subjects while wearing metal knee joint/thigh corset prosthetic systems.11 Thus, the role of the prosthesis in promoting or guarding against femur fracture is uncertain. Significantly, all of these reports are at least 20 years old. The amputee population is now much older than it was at the time of these publications. Given the statistical elevations in fracture risk that accompany the aging process that would seem to compound the ultimate fracture risk facing older amputees, the recent silence in the literature regarding amputee fracture rates is somewhat frustrating. Simply put, as a profession we don't know whether the decreases in BMD due to amputation and aging have ultimately lead to elevated fracture rates.
Following amputation, patients experience a substantial localized reduction in the BMD of their affected limb. This compromise to bone density appears to be considerably greater among females and those with transfemoral amputation levels and is aggravated with increasing time since amputation, reduced activity levels, and general aging. While the published observations regarding BMD levels would suggest sizeable elevations in ultimate fracture risks, the older literature fails to support large fracture rates in this population, and there is no recent literature to confirm or refute actual fracture rates in the past few decades. In the absence of such literature, it becomes the responsibility of treating clinicians to be mindful of the relative BMD compromises their patients face and plan their treatment strategies accordingly.
Phil Stevens, MEd, CPO, is in clinical practice with Hanger Prosthetics & Orthotics, Salt Lake City, Utah. He can be reached at
- Smith, E., Comiskey, C., and Carroll, A. A study of bone mineral density in lower limb amputees and a national prosthetics center. J Prosthet Orthot. 2011:14-20.
- Kulkarni, J., Adams, J., Th omas, E., and Silman, A. Association between amputation, arthritis and ostoepenia in British male war veterans with major lower limb amputations. Clin Rehabil. 1998;12:348-53.
- Leclercq, M. M., Bonidan, O., Haaby, E., Pierregean, C., and Sengler, J. Study of bone mass with dual energy x-ray absorptiometry in a population of 99 lower limb amputees. Ann Readapt Med Phys. 2003;46:24-30.
- Sherk, V. D., Bemben, M. G., and Bemben, D. A. BMD and bone geometry in transtibial and transfemoral amputees. J Bone Miner Res. 2008;23(9):1449-57.
- Yazicioglu, K., Tugcu, I., Yilmaz, B., Goktepe, A. S., and Mohur, H. Osteoprosis: A factor on residual limb pain in traumatic transtibial amputations. Prosthet Orthot Int. 2008;32(2):172-8.
- Demirbag, D., Ozdemi,r F., Kokino, S., and Berkarda, S. Th e relationship between bone mineral density and immobilization duration in hemiplagic limbs. Ann Nuc Med. 2005;19:695-700.
- Clasey, J. L., Janowiak, A. L., and Gater, D. R. Relationship between regional bond density measurements and the time since injury in adults with spinal cord injuries. Arch Phys Med Rehabil. 2004;85:59-64.
- Gonzalez, E. G., and Mathews, M. M. Femoral fracture in patients with lower extremity amputations. 1980;61:276-80.
- Lewallen, R. P., and Johnson, E. W. Fractures in amputation stumps: Review of treatment of 16 fractures. Mayo Clin Proc. 1981;56:22-26.
- Denton, J. R., and McClelland, S. J. Stump fractures in lower extremity amputees. J Trauma. 1985;25(11):1074-78.
- Bowker, J. H., Rills, B. M., Ledbetter, C. A., Hunter, G. A., and Holliday, P. Fractures in lower limbs with prior amputation. A study of ninety cases. J Bone Joint Surg Am. 1981:63-A(6):915-20.