Among lower limb amputees, 52.4% have reported at least one fall in the previous year.1 Therefore, it is surprising that lower limb fracture rates in this population have been reported at just 2.35% and 3%, although these recordings were taken more than 20 years ago when life expectancy after amputation was possibly shorter than it is now.2,3 It was suggested in one of these publications that fracture rate was likely to increase among amputees with improving life expectancy among dysvascular patients.3 There are reports of a decline in bone mineral density (BMD) at the hip of the amputated limb. This was initially suspected on plain x-rays4,5 and later confirmed by dual-energy x-ray absorptiometry (DXA).6 On the amputated limb, BMD has been reported to be an average of 10.4% to 12% less than the sound side, greater trochanter 14.9% less, and proximal tibia 45% less.7–9 When comparisons were drawn between BMD on the amputated side and controls, the mean differences at the neck of femur ranged between 10% and 28%, at the greater trochanter 8.8%, and at the proximal tibia 24%.7–9 It was also observed that the reduction in femoral neck BMD was significantly greater in transfemoral than in transtibial amputees.6
After fracture, a decline in mobility status has been reported in this patient population.2,3,10 In a population with such a high rate of falls and significantly lower BMD at the hip of the amputated leg, fracture prevention should form a part of long-term medical follow-up. Fracture prevention first entails finding patients who are at risk of fracture. BMD measurement using DXA at any specific site is the single best predictor of fracture at that site and provides the measures on which World Health Organization (WHO) diagnostic categories are based; the hip represents the site with greatest predictive power.11 In calculating fracture risk based on BMD, actual BMD measurements (gram per square centimeter) are expressed as relative values, T-score, and Z-score, T-score representing a comparison with the young adult reference mean and Z-score representing a comparison with the age-matched mean. According to WHO criteria, a T-score of less than or equal to 1 standard deviation (SD) below the young adult mean is normal, a T-score greater than 1 but less than 2.5 SD below the young adult mean is osteopenia, and a T-score of 2.5 or more SD below the young adult mean is classified as osteoporosis. Every SD in BMD below the age-matched mean correlates to an approximate doubling of fracture risk, a relationship best proven in postmenopausal women and men older than 50 years.11–14 To date, there is no published work examining fracture risk in lower limb amputees, based on the WHO diagnostic categories of low BMD. The objective of this study was to measure the prevalence of low BMD in lower limb amputees undergoing rehabilitation and to investigate for factors that affect and predict BMD in this population, with the aim of achieving some guidance on the patients we need to target for DXA assessment and fracture prevention.
SUBJECTS AND METHODS
This was a cross-sectional study of patients who were admitted to the National Rehabilitation Hospital, Dublin, Ireland, between August 2006 and March 2007. All patients had had a lower limb amputation performed at another hospital before referral for rehabilitation. Patients were eligible to participate if aged ≥16 years and at least 3 months postsurgery. The 3-month time period chosen was a practical one: by then, patients were expected to be medically stable and able to make an informed decision about participation in the study. In addition, ambulatory status is usually reestablished at 3 months. Exclusion criteria were a diagnosis of osteopenia or osteoporosis before onset of disability and any acute illness, which could have prevented DXA scanning. Ethical approval for the study was obtained from the ethics committee of the National Rehabilitation Hospital. Written patient information was provided, and written consent was obtained.
Subjects completed a questionnaire seeking information about demographics, level of and time since amputation, mobility status, history of fractures since onset, and risk factors for osteoporosis. Patient's body mass and Special Interest Group for Amputee Medicine (SIGAM) mobility grade were recorded. The SIGAM mobility grade is a 6-point measure of walking ability in lower limb amputees, grade A representing a cosmetic prosthesis user to grade F representing normal, unaided walking ability; within grades C and D, there are subgrades depending on type of walking aids used. Patient's mobility was also classified more simply as “Unable to walk at all,” “Able to walk indoors only, either with or without a walking aid,” or “Able to walk out of doors, with or without a walking aid.” The latter classification of mobility status resulted in larger numbers per category than the SIGAM mobility grade, for the statistical analysis. Females were questioned regarding menstrual history and menopausal status. Previous glucocorticoid steroid use was classified as long term, repetitive, short term, or never. Any medical conditions from which the patient suffered, known to be associated with low BMD, were recorded. Cigarette smoking was calculated as pack year history. Alcohol consumption was classified as none, low to moderate, or high; low to moderate included all females consuming less than or equal to 14 units per week and all males consuming less than or equal to 21 units per weeks; high was anything in excess of these figures.
Blood samples were obtained from each patient, and the following assays were measured: total alkaline phosphatase, corrected serum calcium, intact parathyroid hormone (iPTH), 25-hydroxyvitamin D (25-OHD), follicle-stimulating hormone, luteinizing hormone, and testosterone (male patients only) or oestradiol and progesterone (females only). Although our laboratory recommended that Vitamin D deficiency be based on a value less than or equal to 25 nmol/L and insufficiency, a level between 26 and 52 nmol/L, we have chosen to base our classification on the recent, more stringent recommendation that a value of less than 50 nmol/L represents deficiency and up to 72 nmol/L represents insufficiency.15
BMD was measured using DXA, the Hologic Discovery A (Hologic Inc., MA), and all scans were performed by the same operator. Actual BMD measurements (gram per square centimeter) were expressed as relative values, T-score and Z-score. In all cases, WHO categories were applied to T-score results. Z-scores were classified as less than 1 SD below or greater than the mean (within normal range), greater than or equal to 1 but less than 2 SD below the mean, and greater than or equal to 2 SD below the mean.
All patients had densitometry measurements at the lumbar spine and both hips (unless precluded by hip prosthesis), and lateral vertebral assessment. Lumbar vertebrae numbers 1 to 4 inclusive were used to calculate the total lumbar BMD and to categorize according to WHO criteria. A vertebra was excluded from the analysis if its T-score was more than 1 SD greater or less than that of the adjacent vertebra. At the hip, neck of femur, greater trochanter, intertrochanteric area, and total proximal femur were measured, although only neck of femur or total proximal femur was used for diagnostic purposes. Two patients had bilateral total hip replacements, precluding BMD measurement at that site; one patient with amputation had had metal fixation to the lumbar spine as a result of a previous injury to the spinal column; in all three cases, forearms were assessed as the second diagnostic site. In six additional cases, only the sound hip was examined because of a metal prosthesis on the amputated side, or in one of those cases, hip disarticulation. Lateral vertebral morphometry was performed to eliminate false-negative results, because of vertebral fractures.
Daily quality control was performed by scanning a phantom spine, a human-like spine segment made of calcium hydroxyapatite, enclosed in a block of water-stimulant epoxy. Each daily quality control result is compared with 10 separate measurements taken at the time of the system's installation and provides the basis for continuous self-calibration by the system. As is the recommendation of the International Society for Clinical Densitometry, the operator carried out short-term precision error assessment by measuring proximal femur and lumbar spine BMD of 30 patients twice and entering results into the International Society for Clinical Densitometry calculator. The following precision results were obtained: total hip 1.4%, femoral neck 1.7%, and postero-anterior lumbar spine 1.1%, all within the accepted range.16,17
Statistical analysis of the data was performed using SPSS version 13 (SPSS Ireland, Kilmainham, Dublin, Ireland). Descriptive statistics are presented as mean and SD. Where the data were highly skewed, medians and interquartile ranges are provided. Independent sample t-test was used to compare BMD by gender. Paired t-test was used to compare differences between hips of sound and amputated limbs. One-way analysis of variance (ANOVA) was used to examine the relationships between BMD and each of mobility status, use of steroids, other medical conditions, and alcohol consumption. Univariate associations employing correlation coefficient were used to measure the association between BMD and each of age, duration of disability, vitamin D status, parathyroid hormone, oestradiol and testosterone levels, and cigarette smoking. Finally, a multiple linear regression analysis was performed to identify factors that may be predictors of BMD. Significance is expressed as p values <0.05.
DEMOGRAPHIC DETAILS, CLINICAL FEATURES, AND LABORATORY RESULTS
Five lower limb amputees declined participation in the study, leaving 52 subjects. Demographic information, relevant clinical details, and laboratory results are summarized in Table 1. Vitamin D levels were available for 48 patients, of whom 33 (68.8%) had deficient levels less than 50 nmol/L, 5 (10.4%) had insufficient levels, between 50 and 72 nmol/L, and 10 (20.8%) had normal levels. Among those with vitamin D deficiency, two patients had levels of parathyroid hormone and alkaline phosphatase above the normal range. A further two had increased levels of parathyroid hormone only. In 6 (15.4%) of the 39 male participants, serum testosterone level was less than the normal range of 8 to 30 nmol/L. Laboratory results confirmed the postmenopausal status of the 13 females.
BONE MINERAL DENSITY RESULTS
Forearm BMD readings were normal in the three cases in whom this site was used for diagnostic purposes. Mean (SD) BMD at the lumbar spine, neck of femur, and total proximal femur on sound and amputated sides for both males and females are displayed in Table 2. When the BMD of all study participants was classified based on T-scores, 26 (50%) had osteopenia and 20 (38.5%) had osteoporosis. When BMD was classified based on Z-scores, 25 (48.1%) were greater than or equal to 1 SD but less than 2 SD below the mean and a further 10 (19.2%) were greater than or equal to 2 SD below the mean; in 1 case, Z-score was not supplied as the patient was too old. In Table 3, each classification is applied to the results from the lumbar spine, hip, and either site. Of the six males whose testosterone levels decreased below the normal range, four had osteopenia and two had osteoporosis. Of the 13 postmenopausal female participants, 4 had osteopenia and 9 had osteoporosis.
Paired sample t-test was used to examine for a difference between hip BMD of sound and amputated limbs. BMD was significantly lower on the amputated side than on the sound side at neck of femur (t = 6.17, df = 43, p < 0.001) and total proximal femur (t = 7.79, df = 43, p < 0.001).
RELATIONSHIP BETWEEN BMD AND OTHER VARIABLES
Independent sample t-test was used to compare BMD between genders; compared with males, females had significantly lower BMD at the lumbar spine (t = 2.87, df = 48, p = 0.006), neck of femur on the sound side (t = 2.76, df = 48, p = 0.008), total proximal femur on the sound side (t = 3.74, df = 48, p < 0.001), neck of femur on the amputated side (t = 2.05, df = 42, p = 0.047), and total proximal femur on the amputated side (t = 2.58, df = 42, p = 0.014).
One-way ANOVA showed that mobility status (unable to walk at all, indoor walking, or outdoor walking) had no effect on either lumbar spine or hip BMD in this group of patients. One-way ANOVA showed that the presence of any related medical condition had a significant effect on T-score and Z-score at the sound neck of femur and total proximal femur and on BMD, T-score, and Z-score at the neck of femur and total proximal femur of the amputated limb. However, post hoc analysis could not be performed because of inadequate numbers. One-way ANOVA followed by post hoc analysis did not show any effect of glucocorticoid steroid use or alcohol consumption on BMD at any site. Table 4 shows the results of correlation testing between BMD and each of duration of disability, age, body mass, 25-OHD, and iPTH. There was no correlation between BMD and cigarette pack year history.
To further explore the relationship between vitamin D, BMD, and mobility status, we performed an ANOVA to look for any association between mobility status and vitamin D level; none was found. To examine the relationship between amputation level and BMD, groups were collapsed into transfemoral (all transfemoral amputees, transtibial/transfemoral combination, and one hip disarticulation) and transtibial (all transtibial amputees and bilateral transtibial amputees). Independent samples t-test showed no effect of amputation level on BMD at the total lumbar spine, neck of femur, or total proximal femur.
As far as we are aware, there have been no previous studies classifying BMD of lower limb amputees using WHO diagnostic categories. In this group of amputees, we found that 50% of patients had osteopenia and 38.5% had osteoporosis at either lumbar spine or hip. Although lumbar spine fractures cause a great deal of pain and eventually a decline in respiratory function, it is hip fractures that arouse most interest in the study of osteoporosis, because they are associated with much greater morbidity and are a huge economic burden. In addition, the accuracy of hip BMD (as measured by DXA) in the prediction of hip fracture is greater than the corresponding relationship at any other site.
At the hip, we found 48.1% of our amputees to be osteopenic and 32.7%, osteoporotic. A total of 67.3% of our study participants had a Z-score of −1 or less at any site, 28.8% at the lumbar spine, and 51.9% at the hip, which probably reflects the influence of disability on their bone status. It has also been postulated that there is an association between the development of atherosclerosis and low BMD, a factor which may be relevant in this patient population.18,19 The mean age of our study population was 61.7 years. In this age group, the relationship between T-score and fracture risk is well established: every SD decline in BMD at a particular site is associated with an approximate doubling of fracture risk at that site, a relationship best proven at the hip. Therefore, 88.5% of our patients carry, at least, a two-fold risk of hip fracture. The reported high rates of falls both in the community (52.4%)1 and during rehabilitation (28% to 54.4%)20,21 heighten our concerns for these patients. Listed complications of lower limb fractures among amputees include the need for prosthetic modification, increased need for walking aids, less time spent in walking, and loss of walking ability or ability to wear a prosthesis, resulting in wheelchair use.3,10
Inadequate levels of vitamin D occurred in 79.2% of our patients. Although we are unaware of any similar studies examining vitamin D deficiency in lower limb amputees, we can draw comparison with a study of 53 patients participating in in-patient rehabilitation because of a range of medical conditions, of whom 83% had low levels of vitamin D (33.9% insufficient and 49.1% deficient) based on the same ranges as we have used.22 It is worth noting that in our patients, vitamin D measurements were obtained from 42 (87.5%) during the winter months with the remaining 6 (12.5%) having been taken during the summer months. The seasonal variation in 25-OHD levels is well established in the Northern hemisphere.23 In this study, there were four patients with low BMD whose vitamin D deficiency was associated with increased levels of iPTH; in two of these cases, alkaline phosphatase levels were also above the normal range; it is valid to suggest that osteomalacia may have contributed to low bone mass in these cases; although as a histological diagnosis, it was not possible for us to confirm this.23
Surprisingly, there was no correlation between BMD and 25-OHD levels, suggesting that other factors may be more influential on BMD in this population. The absence of any relationship between vitamin D levels and mobility status may suggest that seasonal variation in Vitamin D levels affected all patients equally, not just those who could not walk outdoors, as patients who use wheelchairs for mobility will go outside when the weather is fine and there are more hours of sunlight.
We did not find any effect of mobility status or amputation level on lumbar spine or hip BMD. The lack of effect of ambulatory ability could be due to the small number of patients. In a much larger group of disabled patients, examined by the same authors, ambulatory status was found to be strongly associated with and an independent predictor of hip BMD, in that those patients who were able to walk out of doors had significantly greater BMD than those who could only walk indoors or could not walk at all.24 Similarly, the lack of effect of amputation level may be due to small numbers although a higher level of amputation was found to be associated with lower hip BMD among just 44 subjects in a previous study.6 In this study, we did not perform linear regression analysis to examine for variables that independently predict BMD, as the number of study participants was too small to achieve reliable results.
Duration of disability was found to be negatively correlated with hip BMD on the amputated side. This finding has similarities with other groups of disabled patients, in whom lower limb activity is compromised. Negative correlations have been confirmed in stroke patients, between duration since stroke onset and BMD at the paretic hip,25–29 and in spinal cord injured patients, between duration since onset of cord lesion and lower limb BMD.30–39 Given the relationship between duration of disability and BMD at the hip of the amputated limb, early identification and management of low hip BMD in amputees may help to reduce fracture risk. In this respect, falls prevention should form part of these patients' rehabilitation program, in addition to addressing their vitamin D deficiency and calcium intake. Although there are pharmacological agents available for the treatment of osteoporosis, there have been no randomized controlled trials looking at their effect in lower limb amputees. A few studies have examined the use of bisphosphonates on BMD in the spinal cord injured population; the most valuable gains were found in those patients who had regained walking ability, whereas none of the studies referred to a lowering of fracture rates.40–45 Extrapolating from these results, it could be suggested that the desired effect of this treatment might not be seen at the hip of an amputated limb, because of reduced weightbearing forces. Clearly, a multicenter randomized controlled trial is needed to examine the effect of bisphosphonates and strontium ranelate on BMD at both hips and on hip fracture rates in the amputee population.
In conclusion, a very high proportion of lower limb amputees had low BMD, leaving them at increased risk of fracture, particularly of the hip. With increasing duration since amputation, hip BMD, particularly on the amputated side, is likely to continue to decline. Therefore, bone health monitoring should form part of the long-term medical follow-up of this group of patients.
1. Miller WC, Speechley M, Deathe B. The prevalence and risk factors of falling and fear of falling among lower extremity amputees. Arch Phys Med Rehabil
2. Gonzalez EG, Matthews MM. Femoral fractures in patients with lower extremity amputations. Arch Phys Med Rehabil
3. Denton JR, McClelland SJ. Stump fractures in lower extremity amputees. J Trauma
4. Burke MJ, Roman V, Wright V. Bone and joint changes in lower limb amputees. Ann Rheum Dis
5. Benichou C, Wirotius JM. Articular cartilage atrophy in lower limb amputees. Arthritis Rheum
6. Kulkarni J, Adam J, Thomas E, Silman A. Association between amputation, arthritis and osteopenia in British male war veterans with major lower limb amputations. Clin Rehabil
7. Leclercq MM, Bonidan O, Haaby E, et al. [Study of bone mass with dual energy x-ray
absorptiometry in a population of 99 lower limb amputees.] Ann Readapt Med Phys
8. Rush PJ, Wong JSW, Kirsh J, Devlin M. Osteopenia in patients with above knee amputation. Arch Phys Med Rehabil
9. Royer T, Koenig M. Joint loading and bone mineral density in persons with unilateral, transtibial amputation. Clin Biomech
10. Bowker JH, Rills BM, Ledbetter CA, et al. Fractures in lower limbs with prior amputation. A study of ninety cases. J Bone Joint Surg Am
11. Kanis JA, Gluer C-C. An update on the diagnosis and assessment of osteoporosis with densitometry. Committee of scientific advisors, International Osteoporosis Foundation. Osteoporos Int
12. Kanis JA. Diagnosis of osteoporosis and assessment of fracture risk. Lancet
13. Kanis JA, Johnell O, Oden A, et al. Ten year probabilities of osteoporotic fractures according to BMD and diagnostic thresholds. Osteoporos Int
14. Kanis JA, Johnell O, Oden A, et al. Diagnosis of osteoporosis and fracture threshold in men. Calcif Tissue Int
15. Holick MF. Vitamin D deficiency. N Engl J Med
16. Binkley N, Kendler D, Leib E, et al. Official positions of the International Society for Clinical Densitometry. Position Development Conference, Vancouver, British Colombia, Canada, 2005.
17. Baim S, Wilson CR, Lewiecki EM, et al. Precision assessment and radiation safety for dual-energy x-ray absorptiometry (DXA). J Clin Densitom
18. Jorgensen L, Joakimsen O, Rosvold Berntsen GK, et al. Low bone mineral density is related to echogenic carotid artery plaques: a population based study. Am J Epidemiol
19. Schulz E, Arfai K, Liu X, et al. Aortic calcification and the risk of osteoporosis and fractures. J Clin Endocrinol Metab
20. Vlahov D, Myers AH, Al-Ibrahim MS. Epidemiology of falls among patients in a rehabilitation hospital. Arch Phys Med Rehabil
21. Gooday HMK, Hunter J. Preventing falls and stump injuries in lower limb amputees during inpatient rehabilitation: completion of the audit cycle. Clin Rehabil
22. Shinchuk LM, Morse L, Huancahuari N, et al. Vitamin D deficiency and osteoporosis in rehabilitation in-patients. Arch Phys Med Rehabil
23. Francis RM, Selby PL. Osteomalacia. Ballieres Clin Endocrinol Metab 1997;11:145–163.
24. Smith EM, Comiskey CM, Carroll AM. A study of bone mineral density in adults with disability. Arch Phys Med Rehabil
25. Sato Y, Kaji M, Honda Y, et al. Abnormal calcium homeostasis in disabled stroke patients with low 25-hydroxyvitamin D. Bone
26. Iwamoto J, Tsukimura T, Takeda T. Bone mineral density of metatarsus in hemiplegic subjects. Am J Phys Med Rehabil
27. del Puente A, Pappone N, Mandes MG, et al. Determinants of bone mineral density in immobilisation: a study on hemiplegic patients. Osteoporos Int
28. Demirbag D, Ozdemir F. Kokino S, Berkarda S. The relationship between bone mineral density and immobilisation duration in hemiplegic limbs. Ann Nuc Med
29. Hamdy RC, Krishnaswamy G, Cancellaro V, et al. Changes in bone mineral content and density after stroke. Am J Phys Med Rehabil
30. Lazo MG, Shirazi P, Sam M, et al. Osteoporosis and risk of fracture in men with spinal cord injury. Spinal Cord
31. Bauman WA, Spungen AM, Wang J, et al. Continuous loss of bone during chronic immobilisation: a monozygotic twin study. Osteoporos Int
32. Dauty M, Perrouin Verbe B, Maugars Y, et al. Supralesional and sublesional bone mineral density in spinal cord-injured patients. Bone
33. Clasey JL, Janowiak AL, Gater DR. Relationship between regional bone density measurements and the time since injury in adults with spinal cord injuries. Arch Phys Med Rehabil
34. Sabo D, Blaich S, Wenz W, et al. Osteoporosis in patients with paralysis after spinal cord injury. Arch Orthop Trauma Surg
35. Garland DE, Adkins RH, Stewart CA, et al. Regional osteoporosis in women who have a complete spinal cord injury. J Bone Joint Surg
36. Finsen V, Indredavik B, Fougner KJ. Bone mineral and hormone status in paraplegics. Paraplegia
37. Bauman WA, Spungen AM, Morrison N, et al. Effect of a vitamin D analog on leg bone mineral density in patients with chronic spinal cord injury. J Rehabil Res Dev
38. Demirel G, Yilmaz H, Paker N, Onel S. Osteoporosis after spinal cord injury. Spinal Cord
39. Kaya K, Aybay C, Ozel S, et al. Evaluation of bone mineral density in patients with spinal cord injury. J Spinal Cord Med
40. Gilchrist NL, Frampton CM, Acland RH, et al. Alendronate prevents bone loss in patients with acute spinal cord injury: a randomised, double-blind placebo-controlled study. J Clin Endocrinol Metab
41. Nance PW, Schryvers O, Leslie W, et al. Intravenous pamidronate attenuates bone density loss after acute spinal cord injury. Arch Phys Med Rehabil
42. Shapiro J, Smith B, Beck T, et al. Treatment with zoledronic acid ameliorates negative geometric changes in the proximal femur following acute spinal cord injury. Calcif Tissue Int
43. Zehnder Y, Risi S, Michel D, et al. Prevention of bone loss in paraplegics over 2 years with alendronate. J Bone Miner Res
44. Pearson EG, Nance PW, Leslie WD, Ludwig S. Cyclical etidronate: its effects on bone density in patients with acute spinal cord injury. Arch Phys Med Rehabil
45. Baumann WA, Wecht JM, Kirshblum S, et al. Effect of pamidronate administration on bone in patients with acute spinal cord injury. J Rehabil Res Dev