Archives of Physical Medicine and Rehabilitation
Volume 86, Issue 3 , Pages 498-504, March 2005

Fracture threshold in the femur and tibia of people with spinal cord injury as determined by peripheral quantitative computed tomography

  • Prisca Eser, PhD

      Affiliations

    • Institute for Clinical Research, Swiss Paraplegic Centre, Nottwil, Switzerland
    • Eser is currently affiliated with the Centre of Physical Activity and Nutrition Research, School of Health Sciences, Deakin University, Melbourne, Australia.
    • Corresponding Author InformationReprint requests to Prisca Eser, PhD, Centre of Physical Activity and Nutrition Research, School of Health Sciences, Deakin University, 221 Burwood Hwy, Burwood 3125, Victoria, Australia
    • ,
  • Angela Frotzler, MSc

      Affiliations

    • Institute for Clinical Research, Swiss Paraplegic Centre, Nottwil, Switzerland
    • ,
  • Yvonne Zehnder, MD

      Affiliations

    • Institute for Clinical Research, Swiss Paraplegic Centre, Nottwil, Switzerland
    • ,
  • Jachen Denoth, PhD

      Affiliations

    • Biomechanics Laboratory, ETH-Zürich, Schlieren, Switzerland

Article Outline

Abstract 

Eser P, Frotzler A, Zehnder Y, Denoth J. Fracture threshold in the femur and tibia of people with spinal cord injury as determined by peripheral quantitative computed tomography.

Objective

To determine bone traits of the femur and tibia with peripheral quantitative computed tomography (pQCT) that best distinguish between spinal cord injury (SCI) subjects with and without fractures.

Design

Cross-sectional study.

Setting

In- and outpatient paraplegic center in Switzerland.

Participants

Ninety-nine motor complete SCI subjects (duration of paralysis, 2mo–49y), 21 of whom had sustained fractures of the femur or tibia.

Interventions

Not applicable.

Main outcome measures

Subjects with SCI were questioned about the occurrence, location, and approximate date of fractures to their lower extremities. Trabecular and cortical bone mineral density (BMD), as well as bone geometric properties of distal epiphyses and midshafts of the femur and tibia, were measured by pQCT.

Results

Trabecular BMD of the femur and tibia distal epiphyses was found to distinguish best subjects with fractures from those without. Fractures occurred in subjects with trabecular BMD of less than 114mg/cm3 and less than 72mg/cm3 for the femoral and tibial distal epiphysis, respectively (corresponding to 46% and 29% of mean values of an able-bodied reference group). Approximately 50% of the subjects with chronic SCI (defined as time postinjury >5y for femur data and >7y for tibia data) had trabecular BMD values above the fracture threshold in the femur and about one third above the fracture threshold in the tibia.

Conclusions

By using pQCT, it may be possible to identify subjects with SCI who are at risk of sustaining fractures of the femur and tibia through minor trauma.

Key words:  Bone density , Fractures , Rehabilitation , Spinal cord injuries

 

SPINAL CORD INJURY (SCI) with a motor complete lesion leads to an extreme form of immobilization and disuse of the paralyzed extremities. The immediately initiated muscle loss is followed by a cascade of events: invasion of connective tissue,1 atrophy of the vascular system,2, 3 degenerative changes of cartilage,4 and bone loss. Bone loss has received much attention in medical literature because many people with SCI have fractures of the paralyzed legs as a consequence of minor trauma, such as falling out of a wheelchair or during transfer. The lifetime risk of a person with SCI suffering a bone fracture of the lower extremities is approximately twice the risk of an able-bodied person.5 Fractures may reduce the quality of life of people with SCI most severely and surgery for internal fixation is often necessary because of the prohibitive risk that casts will cause pressure ulcers.6, 7

Similar to postmenopausal women with fractures of the femoral neck or the lumbar spine, where fractures have been associated with low bone mineral density (BMD) at these locations,8, 9 the fracture rate in subjects with SCI has also been found to be related to BMD.10 In osteoporotic women, BMD is commonly measured at the sites of greatest fracture risk—the femoral neck and lumbar spine—with dual-energy x-ray absorptiometry (DXA). This technique results in 2-dimensional projections of the bones and BMD is calculated as mass per area (in g/cm2), also termed areal BMD. Results by this 2-dimensional method vary according to a person’s size and body composition (muscle/fat ratio).11, 12, 13 In contrast, bone assessment by quantitative computed tomography (QCT) makes it possible to calculate volumetric densities (in g/cm3). The QCT scans are placed orthogonally to the long bone axis and slice thickness is usually 1 or 2mm. The cross-sectional approach of QCT allows for the separation between cortical and trabecular bone compartments with the calculation of separate BMD, as well as the assessment of various bone geometric properties. Peripheral QCT (pQCT), by means of which arms and legs can be measured, also allows the measurement of true volumetric densities with a negligible exposure to radiographs.14, 15

In people with SCI, the sites where fractures most commonly occur are the distal epiphyses of the femur and tibia, the proximal epiphysis of the tibia, the femoral and tibia shaft, and, less commonly, the femoral neck and bones of the foot.6, 16 Other studies17, 18 have found fractures of the femoral neck to be as common as fractures of the femoral shaft. To predict whether a person with SCI is at a high risk to sustain a fracture of the paralyzed legs, it is important to measure BMD at the sites in which most of the fractures occur. If a patient wants to participate in a physical training program, such as functional electric stimulation walking or cycling, it is important to be able to assess the risk for fracture. DXA measurements are most commonly performed at the femoral neck and lumbar spine because the manufacturers provide software and reference values for these locations. Furthermore, whole-body DXA scanning provides the bone mineral content of the entire leg. The lumbar spine has not been reported as a fracture-prone area in people with SCI, and its DXA BMD has been found to increase rather than to decrease with time after lesion.19

We recently measured BMD and bone geometry of the femur and tibia distal epiphyses and shafts,20 and found an exponential decrease of bone mass and BMD in the epiphyses, which, apart from the thin cortical shell at their periphery, consist of trabecular bone. A new steady-state was reached at approximately 50% of the mean value of a reference group in the femur after the 3 years and at 40% in the tibia distal epiphysis after 5 years. The femoral and tibial shafts had exponential decay in cortical wall thickness but not in cortical BMD. The loss in wall thickness reached a new steady state after 5 years at 65% of reference values in the femur and after 7 years at 70% in the tibia. This bone loss uniformly affected all 99 measured subjects with SCI.

Our purpose in this study was to find the most common fracture sites (bone shaft or epiphyses) and to establish a fracture threshold of the bone parameters that best characterize those sites. The pQCT measuring technique allowed us to assess trabecular BMD of the epiphyses and cortical BMD of the shaft separately, as well as to measure bone geometric properties. The current practice of performing DXA measurements at the hip and predicting fracture risk for the lower extremities of chronic SCI patients is unsatisfactory because (1) T scores of −1 to −2 standard deviations (SDs) are the norm in this population10 and a finer scale is required, and (2) the measuring site is not congruent with the more distal sites in which fractures are most numerous.21 It would be invaluable for clinicians if, by means of pQCT measurements directly at the bone sites most prone to fractures, the risk to sustain a fracture from a low-energy trauma could be predicted by the most sensitive bone parameter. This would be a helpful tool in selecting people with SCI who could safely participate in a physical training program.

Back to Article Outline

Methods 

Participants 

Subjects were recruited from among ambulatory patients seen at the Swiss Paraplegic Centre, Nottwil, and in-house patients participating in the 4-month study. Inclusion criteria were motor complete para- and tetraplegia (American Spinal Injury Association grades A and B) as a result of trauma and a minimum age of 18 years. Exclusion criteria were menopausal or postmenopausal status for women and medical conditions other than paralysis known to influence bone mineral metabolism. Patients meeting the inclusion criteria were asked by telephone to volunteer for the study. Approximately 130 patients were contacted, and 103 agreed to participate. They reported to the radiology department at the Swiss Paraplegic Centre and were informed about the measuring procedure, after which they signed a written informed consent approved by the local ethical committee.

Bone measurements 

Subjects were asked about their handedness and the leg of the dominant arm side was measured accordingly. Exceptions were made if the selected leg had been fractured within the last 10 years; if so, the nonfractured side was measured. Subjects were then positioned supine on a couch. The length of the tibia was measured with a measuring tape from the distal end of the medial malleolus to the medial knee joint cleft. Because of poor accessibility, the femur length was estimated as being the same length as the tibia.

All measurements were performed with a Stratec XCT 3000a scanner. This pQCT apparatus measures attenuation of radiographs that are linearly transformed into hydroxyapatite (HA) densities. Unlike other pQCT scanners, the Stratec XCT 3000 is calibrated with respect to fat, which is set at 0mg HA, water hence results in 60mg HA.22 HA equivalent densities are automatically calculated from the attenuation coefficients by using the manufacturer’s phantom, which itself is calibrated with respect to the European Forearm Phantom.b,22 The radiation dose as whole-body equivalence is stated as .001mSv for pQCT, whereas it is .01mSv for DXA and 0.4mSv for axial computed tomography.23

A scout view of the distal ends of the femur and tibia was performed to locate the joint cleft and place the reference line on the distal end of the more proximal condyle of the considered bone. Scans were performed in the distal epiphyses of the femur and tibia at 4% of the bone’s length (length of the tibia was also used for the femur). Diaphyseal scans were performed at 25% of total bone length from the distal end for the femur and at 38% for the tibia. A scan location of 38% would have also been desirable for the femur; however, because of insufficient abduction of the hip in most patients, the scan location of 25% was more accessible. Slice thickness was set at 2mm, and voxel size was set at 0.3mm for edge length in the femur and 0.5mm in the tibia. For the tibia, the measuring protocol as recommended by the manufacturer was followed; there were no such recommendations for the femur at the time of the study.

Measuring parameters 

Four percentage scans 

The periosteal surface of each bone’s epiphysis was found by a contour algorithm based on thresholding at 180mg/cm3 for the tibia and at 150mg/cm3 for the femur. Total BMD was determined from the mass and bone cross-sectional area (total CSA). Concentric pixel layers were then peeled from the bone’s perimeter until a central area covering 45% of the total bone CSA was left. Trabecular BMD was determined from this central area.

Twenty-five percent (femur) and 38% (tibia) scans 

The periosteal surface of the bone’s diaphysis was found by a contour algorithm. Total CSA was calculated. Cortical bone was selected by using a threshold at 710mg/cm3. Of the selected area, cortical CSA, cortical BMD, and the polar bone strength strain index24 (polar SSI) were calculated.

Fracture assessment 

Subjects were asked about the time and location of fractures that had occurred after their SCI. Details of reported fractures of the lower extremities were verified through their medical records.

Data analysis 

To avoid erroneously low cortical BMD because of partial volume effect,25 mean cortical thickness of the diaphyseal scans was calculated so that a limit could be set for cortical thickness below which cortical BMD was discarded. Cortical thickness of the diaphysis was calculated based on the assumption that bone shaft be cylindrical by calculating the radius of total CSA (CSA=πr2), the radius of the bone marrow area (total CSA − cortical CSA), and subtracting the marrow radius from the total bone shaft radius.

Automatic forward stepwise discriminant analyses were performed with Systat softwarec to find the bone parameters that best distinguished groups with and without fractures. Two separate analyses of parameters were performed for the femur and tibia: bone parameters of the femur for groups with nil or at least 1 fracture of each and bone. The enter and removal probabilities were set at .15, tolerance at .001, and force at 0. In a related study,20 we found exponential decreases for all femur and tibia bone parameters, except for total CSA and cortical BMD of the shaft, and for total CSA of the epiphyses. After this initial extensive bone mass loss, all parameters reached a new steady state at values ranging from 25% to 75% lower than the values of an able-bodied reference group. The bone parameters with the most extensive decrease were trabecular and total BMD of the distal femur and tibia epiphyses. Bone parameters of the femur reached a steady state at less than 5 years postinjury, whereas those of the tibia took less than 7 years.20 Therefore, for the discriminant analysis of the femur, only subjects with a lesion duration of at least 5 years were included, and for the tibia only subjects with a lesion duration of at least 7 years were included. All subjects entered the analysis with values of their bone parameters measured at the time of this study and the date of their first fracture after injury.

In the femur and tibia shaft, only cortical thickness decreased at a concomitantly stable cortical BMD.20 When cortical thickness falls below 1.6mm,20 cortical BMD shows erroneously low values caused by what is called the partial volume effect.25 Because subjects with femur fractures had very thin femoral cortices and consequently artificially low cortical BMD values, cortical BMD was excluded from the discriminant analysis. Alternatively, we could have excluded subjects from analysis who had cortical thickness of less than 1.6mm. However, this would also have excluded subjects with fractures and hence would have lowered the statistical power. Because cortical BMD did not decrease after SCI, excluding this parameter introduced no bias. Cortical BMD did not have to be excluded from the discriminant analysis for the tibia because cortical thickness of the tibia was greater than 1.6mm in all subjects.

Back to Article Outline

Results 

Participants 

There were 103 subjects in the study. Data from 99 subjects were included in the analysis; 4 subjects had to be excluded because of movement artifacts in the pQCT measurements caused by muscle spasms. Subject characteristics are shown in table 1.

Table 1. Subject Characteristics
ParameterAll SCI SubjectsSubjects With FracturesSubjects Without Fractures
No. of subjects992574
Women/men10/894/216/68
Tetraplegic/paraplegic27/723/2224/50
Flaccid/spastic10/894/216/68
Age (y)41.4±13.7 (19–83)47.5±10.5 (26–68)39.3±14.1 (19–83)
Height (cm)175.8±8.0 (150–196)174.1±8.9 (150–196)176.3±7.6 (153–190)
Weight (kg)72.4±12.2 (39–103)74.4±13.6 (39–102)71.7±11.7 (41–103)
Years postinjury12.3±11.6 (0.2–48.5)20.0±11.9 (1.1–43)9.7±10.3 (0.2–48.5)
Age at lesion (y)29.1±12.6 (8–65)27.4±13.6 (8–60)29.6±12.3 (13–65)

NOTE. Values are n or mean ± SD (range).

In the 99 subjects, 99 tibia and 97 femur measurements were completed. Two femur measurements could not be completed because of difficulties in placing the limb in the gantry. Furthermore, of a total of 194 femur scans, 21 had to be excluded because of movement artifacts detected by visual inspection.

Occurrence of fractures and bone parameters 

We counted the following fractures and locations: 5 proximal femur; 2 femur shaft; 7 distal femur (in 6 people); 1 patella; 5 proximal tibia (in 3 people); 1 tibial shaft; 8 distal tibia (in 6 people); 3 fibula; 4 foot; and 9 toes (in 6 people). Twenty-seven subjects had sustained a fracture of the lower extremities, of which 18 had sustained a fracture to either the femur (n=9) or tibia (n=7) or both (n=2). Because we measured bone parameters only in the femur and tibia, fractures in these bones are described in more detail.

Because there were only 2 femur shaft and 1 tibia shaft fractures in our subjects, we did not perform separate analysis for shaft fractures with shaft bone parameters or epiphyseal (or metaphyseal) fractures with epiphyseal bone parameters. Instead, we only distinguished between fractures of the femur and tibia with corresponding bone parameters.

Ten subjects had sustained fractures of the femur in the time after SCI of which 1 subject had sustained 4 femur fractures. For the discriminant analysis, only subjects with a fracture duration of at least 5 years were included. Of these 47 subjects, 6 had 1 fracture and 1 had 4 fractures of the femur. The bone parameter that best discriminated between the subjects with or without femur fractures was trabecular BMD of the epiphysis (1-way analysis of variance [ANOVA], F-to-enter 11.11, P=.002). This parameter correctly allocated subjects to the fracture or the nonfracture group in 79% of all cases. Table 2 shows the F-to-enter values of all included parameters. When trabecular BMD was removed and entered as a covariate into ANOVA of the remaining factors, the F values of almost all other parameters were reduced considerably (table 2). This shows a high correlation of the various bone parameters among each other. There was little change in the F values of total CSA of the epiphyses and cortical BMD of the tibial shaft (table 2). These 2 parameters did not change after SCI20 and therefore did not correlate with the other parameters. Figure 1A shows a scatterplot of trabecular BMD of the femur distal epiphysis versus time after injury, with the number of fractures in each subject.

Table 2. Discriminant Analysis Between the Fracture and Nonfracture Groups
Scan SiteBone ParameterFemurTibia
F-to-Enter*F-to-Enter After Removing of Trabecular BMD†F-to-Enter*F-to-Enter After Removing of Trabecular BMD†
Epiphysis
Mass5.590.096.450.45
Total CSA0.420.020.310.27
Total BMD8.690.043.380.03
Trabecular BMD11.11 8.95
Diaphysis
Mass1.890.044.510.63
Total CSA0.60.142.520.5
Cortical CSA5.710.464.530.59
Cortical THI10.611.533.70.42
Cortical BMDNANA0.930.98
Polar SSI1.140.033.230.33

Abbreviations: NA, not applicable; THI, thickness.

* F value by 1-way ANOVA.

† F value by 1-way analysis of covariance with trabecular BMD as covariate.

  • View full-size image.
  • Fig 1. 

    (A) Trabecular BMD of the distal epiphysis of the femur versus time after injury of the 99 subjects. Legend: ○, subjects who had never had a fracture of the femur; ▴, subjects who had had 1 fracture of the femur; ■, 1 subject who had had 4 fractures of the femur. The shaded area shows the mean ±2 SDs of a reference group of able-bodied people. The lower line indicates the highest trabecular BMD value where a fracture was found. (B) Trabecular BMD of the distal epiphysis of the tibia versus time after injury of the 99 subjects. Legend: ○, subjects who had never had a fracture of the tibia; ▴, subjects who had had 1 fracture to the tibia or fibula (1 case); ■, subjects who had had 2 fractures of the tibia. The shaded area shows the mean ±2 SDs of a reference group of able-bodied people. The lower line indicates the highest trabecular BMD value where a fracture was found.

Five subjects had sustained 1 fracture of the tibia, and 5 had had 2 such fractures. Of the 10 subjects, 9 had a lesion duration of more than 7 years and were hence included in the discriminant analysis that included 49 subjects (40 without fractures of the tibia). Trabecular BMD of the epiphysis was also the parameter that best discriminated between the subjects with or without fractures of the tibia (F-to-remove 8.95, P=.004). This parameter correctly allocated subjects to the fracture or nonfracture group in 67% of all cases. Table 2 shows the F-to-enter values of all parameters. Figure 1B shows a scatterplot of trabecular BMD of the tibial distal epiphysis versus time after injury, with the number of occurred fractures in each subject.

All subjects with femur fractures had a femoral trabecular BMD of less than 114mg/cm3 and the subject with 4 fractures had a trabecular BMD of 49mg/cm3. The 10 subjects with femur fractures had a mean femur trabecular BMD of 84.8±23.8mg/cm3. In contrast, the 40 subjects who never had a femur fracture and who were at least 5 years postinjury (which was the time required to reach the new steady state20) had a mean femur trabecular BMD of 116.8±26.0mg/cm3.

Subjects who had sustained tibia fractures had a trabecular BMD of the distal tibia of less than 71.4mg/cm3, and the people with 2 fractures had less than 53.2mg/cm3. Mean trabecular BMD of the distal tibial epiphysis was 46.9±21.8mg/cm3 in subjects with tibial fractures. The 43 subjects who never had a fracture of the tibia and who were at least 7 years postinjury (again, the time required to reach the new steady state20) had a mean tibia trabecular BMD of 68.4±22.4mg/cm3.

Concerning the lesion duration after which fractures occurred, in 3 subjects a femur fracture occurred at 2 and 4 years postinjury, whereas in the other 7 subjects the fracture(s) occurred at least 10 years postinjury. Among the subjects with a lesion duration of at least 5 years, just over half of the subjects (58%) had fallen below the fracture threshold of 114mg/cm3 for the femur. The subject with the earliest tibia fracture sustained it at 4 years postinjury, 2 subjects at 9 and 14 years postinjury, and the other 7 subjects had their earliest tibia fracture at more than 20 years postinjury. All of these subjects had a trabecular BMD of the distal tibia of less than 75mg/cm3. Two thirds of the subjects with a lesion duration of at least 7 years had fallen below the fracture threshold of 75mg/cm3 found in this study.

Back to Article Outline

Discussion 

We found trabecular BMD of the epiphyses is the most sensitive parameter to differentiate between SCI patients at high or low risk of a fracture of the paralyzed lower extremities. Trabecular BMD is the parameter experiencing the most extensive loss: in the distal femur 54% was lost within the first 5 years after the SCI, and in the distal tibia 73% was lost within the first 7 years after injury.20 Thereafter, a new steady state was reached that showed the same intersubject variation as the variation found in a group of able-bodied subjects with normal values of trabecular BMD.20 Despite the magnitude of this loss, our study shows that the individual level of the new steady state is important in terms of fracture risk. Our data imply that there is a fracture threshold at approximately 110mg/cm3 in the distal femur and 70mg/cm3 in the distal tibia, above which no fractures have occurred and below which fractures due to minor trauma are common. A shortcoming of our study is that the bone status was not measured at the time of the fracture, but often the fracture(s) had occurred years before the measurements made in this study. However, among the subjects measured at a lesion duration of greater than 5 years, there was only 1 fracture within the first 5 years postinjury (at 4y postinjury); all other fractures happened thereafter, when trabecular BMD can be assumed to be stable. In the other 4 subjects who suffered a fracture within the first 5 years postinjury, the bone measurements were also taken within that period. Hence, the trabecular BMD shown in figures 1A and 1B should reflect bone status at the time of fracture.

Fractures of the toes were not associated with low trabecular BMD of the femur or tibia and traumas causing them were either unknown or crushing, except in 1 case in which the person fell backward out of his wheelchair. We believe that bone strength plays a minor role with respect to the risk of fracturing a toe, because the toes are the most peripheral skeletal site and are easily crushed in the absence of sensibility and pain. There is also a high probability that some toe fractures were not recorded in this study because they were never diagnosed.

Most subjects reported that fractures resulted from falling out of their wheelchair or during transfers or bumping into an object. Only 2 subjects had suffered fractures in car collisions; however, both had also had fractures after falling from their wheelchair. The force required to cause the different fractures may have varied greatly; however, the events causing them are events that are likely to happen to most people with SCI at some stage in their lives. Assuming that our subjects above the fracture threshold are unlikely to break a bone in a minor accident such as a fall out of the wheelchair, about one third of people with long-standing SCI are at a minimal risk of bone fracture, whereas two thirds are at an increased risk and should try to avoid falls altogether.

Because bone mass decreases sharply within the first few years after SCI, fracture rate has increased from 1% within the first 3 years after SCI to 2% thereafter.5 In our study, we found 3 lower extremity fractures had occurred within the first 2 years after injury, which equates to a fracture rate of 15% (15 fractures per 100 person years, with person years calculated as sum of each person’s years postinjury until the first fracture occurred or until measurement for those without fractures). Within the first 5 years postinjury, which is the mean time to complete bone loss after SCI,20 6 of 45 subjects had sustained a fracture of the lower extremities—a fracture rate of 6%. Among subjects with lesion durations of greater than 5 years, 20 of 54 subjects had sustained a fracture—a fracture rate of 2%. If we only consider subjects with a lesion duration of more than 5 years with a trabecular BMD below the fracture threshold of 114mg/cm3 for the femur and 72mg/cm3 for the tibia, then 11 of 20 subjects had sustained a fracture, which equates to a fracture rate of 4%, Of the 13 subjects with trabecular BMD of the femur and tibia above the fracture thresholds, only 2 subjects had fractures of the toes, which equates to a fracture rate of 0.7%. We believe that the high fracture rate within the first 5 years after injury may be because some subjects (especially those with fractures in this study) lose bone more rapidly than others and may reach values around the fracture threshold already after 1 to 2 years (see figs 1A, 1B). Calculated on 100 person years after injury, the percentage is large. When subjects with a lesion duration of greater than 5 years postinjury are considered, older subjects who have never had a fracture (possibly because they are above the fracture threshold) reduce the fracture rate by contributing many person years.

Among subjects who sustained fractures of the lower-extremity bones and who had a lesion duration of at least 5 years, 12 had suffered only 1 fracture, whereas 7 had suffered more than 1, all from separate events. This shows a high incidence of suffering multiple fracture events among people with SCI with low bone density.

Previous studies have found that fracture rates were associated with areal BMD10 or shaft areal moment of inertia.26 Lazo et al10 found a significant relation between hip areal BMD and time since injury and found significantly more fractures in the osteoporotic group that had longer lesion durations. Szollar et al27 found that the fracture threshold defined for DXA measurements at the hip was reached at 1 to 5 years postinjury. However, our study is the first to propose numeric fracture thresholds for the population with chronic SCI.

Of 41 subjects with SCI in the Lazo10 study, 34% had sustained at least 1 fracture of the lower extremities, compared with 25% in our study. We counted 14 fractures above and 29 fractures below the knee, similar to the distribution found by Lazo10 and Garland and Adkins.16 We found the same number of femur fractures compared with tibial fractures. This disagrees with other studies5, 6, 17, 18 that have generally found more fractures in the femur compared with the tibia. Garland and Adkins16 found about 25% of all fractures in the femur and tibia affecting the shafts. We recorded only 3 shaft fractures versus 24 fractures in the tibial and femoral epiphyses. Although the loss in bone mass resulted from a reduction in BMD in the epiphyses, bone mass was lost in the shaft by a reduction in cortical thickness at a rate of approximately .25mm per year within the first 5 to 7 years after injury.20 In terms of bone mass, the diaphyses lost 14% (femur) and 33% less than the distal epiphyses.20 This may make the femur or tibia shaft a lesser fracture risk compared with the epiphyses. However, fractures are a result of a combination of factors, such as material strength, bone geometry, and impacting forces.

Ragnarsson and Sell6 found that fractures of the lower extremities were more common in subjects with paraplegia than in those with tetraplegia, probably because of their higher activity level. We can confirm this notion, with only 2 tibia fractures occurring in subjects with tetraplegia and all other fractures occurring in people with paraplegia.

Only 1 previous study18 found more fractures in flaccid compared with spastic subjects. In subjects with lesion duration of greater than 5 years, we found a fracture rate of 5.6% and 1.4% in flaccid and spastic subjects, respectively. In a related study,28 we found that bone loss in the femur after SCI was reduced in subjects with stronger spasticity, compared with subjects with weaker or absent spasticity. It seems probable that spasticity is effective at preserving bone mass and reducing fracture risk. In the same study,28 we found age does not independently influence trabecular BMD or any of the other measured bone parameters (only through its covariance with time postinjury). The first fracture in subjects was sustained at an average of 8.7±7.1 years postinjury, at which time the decrease in trabecular BMD and cortical thickness had stabilized on a new steady state.20

Back to Article Outline

Conclusions 

In this study, we have shown that volumetric trabecular BMD of the epiphyses of the lower extremities is the most sensitive bone parameter in determining a fracture threshold. Trabecular BMD can only be determined by QCT or pQCT. With DXA, the trabecular bone compartment cannot be separated from the thin cortical shell on the bone’s perimeter. Furthermore, areal BMD as measured by DXA is dependent on the exact positioning of the limb on bone size, and on the distribution and composition of soft tissue.11, 12, 29 In the able-bodied population, people with areal BMD values below 2 SDs of a healthy reference group are classified as osteoporotic, with increased risk for fracture.30 It can be seen in figures 1A and 1B that all chronic SCI subjects measured in this study were below 2 SDs of our reference group. We believe that simply classifying most people with chronic SCI as osteoporotic is inadequate with regard to predicting fracture risk. A classification system appropriate for the population of chronic SCI needs to consider that all people with chronic SCI have greatly reduced trabecular BMD and cortical thickness, and fractures are the result of different types of trauma than they are in the able-bodied population. Furthermore, because differences in trabecular BMD among people with chronic SCI are small (the variance of the chronic SCI subjects was smaller than the variance of the reference group20), equipment used to determine a fracture threshold must be as accurate as possible. We therefore postulate that bone status measurements in people with SCI should be performed with pQCT in the tibia and/or femur in which most of the fractures from minor trauma occur. We will collect additional data to confirm or adapt the fracture thresholds found in this study, so that in the future, pQCT measurements can be used as a diagnostic tool in the risk assessment for SCI intervention studies involving functional mobility.

Suppliers

Back to Article Outline

Acknowledgments 

We gratefully acknowledge the help of Peter Wüseke from Stratec Medical, Pforzheim, Germany, for the programming of special software for some of our analyses. Hans Schiessl and Johannes Willnecker provided many helpful comments to the manuscript. The supportive collaboration of the Radiology Department of the Swiss Paraplegic Centre was invaluable.

Back to Article Outline

References 

  1. Järvinen T , Jozca L , Kannus P , Järvinen T , Järvinen M . Organization and distribution of intramuscular connective tissue in normal and immobilized skeletal muscles . J Muscle Res Cell Motil . 2002;23:245–254
  2. Olive JL , Dudley GA , McCully KK . Vascular remodeling after spinal cord injury . Med Sci Sports Exerc . 2003;35:901–907
  3. Hopman MT , Nommensen E , van Asten WN , Oeseburg B , Binkhorst RA . Properties of the venous vascular system in the lower extremities of individuals with paraplegia . Paraplegia . 1994;32:810–816
  4. Vanwanseele B , Eckstein F , Knecht H , Stüssi E , Spaepen A . Knee cartilage of spinal cord-injured patients displays progressive thinning in the absence of normal joint loading and movement . Arthritis Rheum . 2002;46:2073–2078
  5. Vestergaard P , Krogh K , Rejnmark L , Mosekilde L . Fracture rates and risk factors for fractures in patients with spinal cord injury . Spinal Cord . 1998;36:790–796
  6. Ragnarsson KT , Sell GH . Lower extremity fractures after spinal cord injury (a retrospective study) . Arch Phys Med Rehabil . 1981;62:418–423
  7. Nottage W . A review of long-bone fractures in patients with spinal cord injuries . Clin Orthop . 1981;155:65–70 ;Mar-Apr
  8. Jergas M , Glüer C . Assessment of fracture risk by bone density measurements . Semin Nucl Med . 1997;27:261–275
  9. Ryan P . Overview of role of BMD measurements in managing osteoporosis . Semin Nucl Med . 1997;27:197–209
  10. Lazo M , Shirazi P , Sam M , Giobbie-Hurder A , Blacconiere M , Muppidi M . Osteoporosis and risk of fracture in men with spinal cord injury . Spinal Cord . 2001;39:208–214
  11. Tothill P , Laskey MA , Orphanidou CI , van Wijk M . Anomalies in dual energy X-ray absorptiometry measurements of total-body bone mineral during weight change using Lunar, Hologic and Norland instruments . Br J Radiol . 1999;72:661–669
  12. Farrell T , Webber C . The error due to fat inhomogeneity in lumbar spine bone mineral measurements . Clin Phys Physiol Meas . 1989;11:77–84
  13. Lochmuller EM , Miller P , Burklein D , Wehr U , Rambeck W , Eckstein F . In situ femoral dual-energy X-ray absorptiometry related to ash weight, bone size and density, and its relationship with mechanical failure loads of the proximal femur . Osteoporos Int . 2000;11:361–367
  14. Braun MJ , Meta MD , Schneider P , Reiners C . Clinical evaluation of a high-resolution new peripheral quantitative computerized tomography (pQCT) scanner for the bone densitometry at the lower limbs . Phys Med Biol . 1998;43:2279–2294
  15. Müller A , Rüegsegger E , Rüegsegger P . Periphral QCT (a low-risk procedure to identify women predisposed to osteoporosis) . Phys Med Biol . 1989;34:741–749
  16. Garland DE , Adkins RH . Bone loss at the knee in spinal cord injury . Top Spinal Cord Inj Rehabil . 2001;6:37–46
  17. Ingram R , Suman R , Freeman P . Lower limb fractures in the chronic spinal cord injured patient . Paraplegia . 1989;27:133–139
  18. Comarr A , Hutchinson R , Bors E . Extremity fractures of patients with spinal cord injury . Am J Surg . 1962;103:732–739
  19. Szollar SM , Martin E , Parthemore JG , Sartoris DJ , Deftos LJ . Demineralization in tetraplegic and paraplegic man over time . Spinal Cord . 1997;35:223–228
  20. Eser P , Frotzler A , Zehnder Y , et al.   Relationship between the duration of paralysis and bone structure (a pQCT study of spinal cord injured individuals) . Bone . 2004;34:869–880
  21. Garland DE , Maric Z , Adkins RH , Stewart CA . Bone mineral density about the knee in spinal cord injured patients with pathologic fractures . Contemp Orthop . 1993;26:375–379
  22. Augat P , Gordon C , Lang T , Lida H , Genant HK . Accuracy of cortical and trabecular bone measurements with peripheral quantitative computed tomography . Phys Med Biol . 1998;43:2873–2883
  23. Rüegsegger E . Bone density measurement . Rheumatology . 1996;18:103–116
  24. Schiessl H , Ferretti J , Tysarczyk-Niemeyer G , Willnecker J . Noninvasive bone strength index as analyzed by peripheral quantitative computed tomography (pQCT) . In:  Schönau E editors. Paediatric osteology (new development in diagnostics and therapy) . Amsterdam: Elsevier Sci; 1996;p. 141–146
  25. Hangartner TN , Gilsanz V . Evaluation of cortical bone by computed tomography . J Bone Miner Res . 1996;11:1518–1525
  26. de Bruin ED , Herzog R , Rozendal RH , Michel D , Stussi E . Estimation of geometric properties of cortical bone in spinal cord injury . Arch Phys Med Rehabil . 2000;81:150–156
  27. Szollar SM , Martin EM , Sartoris DJ , Parthemore JG , Deftos LJ . Bone mineral density and indexes of bone metabolism in spinal cord injury . Am J Phys Med Rehabil . 1998;77:28–35
  28. Eser P , Frotzler A , Zehnder Y , Schiessl H , Denoth J . Assessment of anthropometric, systemic, and lifestyle factors influencing bone status in the legs of spinal cord injured individuals . Osteoporos Int . 2005;16:26–34
  29. Pors Nielsen S . The fallacy of BMD (a critical review of the diagnostic use of dual x-ray absorptiometry) . Clin Rheumatol . 2000;19:174–183
  30. World Health Organization . Assessment of fracture risk and its application to screening for postmenopausal osteoporosis . Geneva: WHO; 1994; WHO Technical Report Series.
  • a Stratec Medizintechnik GmbH, Durlacher Str 35, D-75172 Pforzheim, Germany.
  • b QRM GmbH, Dorfstr 4, D-91096 Möhrendorf, Germany.
  • c Systat, version 8; Systat Software Inc, 501 Canal Blvd, Ste C, Point Richmond, CA 94804-2028.

 Supported by the Swiss Paraplegic Foundation.No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the author(s) or on any organization with which the author(s) is/are associated.

PII: S0003-9993(04)01241-9

doi:10.1016/j.apmr.2004.09.006

Archives of Physical Medicine and Rehabilitation
Volume 86, Issue 3 , Pages 498-504, March 2005

Article Tools

* Image Usage & Resolution