Bilateral symmetry of the human metacarpal: implications for sample size calculations.

OBJECTIVE: The aim of this study was to assess the three-dimensional mechanical symmetry of the human second metacarpal and provide sample size estimates for future mechanical intervention studies of the metacarpal. DESIGN: Bone densitometry and digital image analysis were used to assess the morphometric, geometric and densitometric symmetry of the second human metacarpal. BACKGROUND: An assessment of the left-right mechanical symmetry of the human metacarpal is important in considering the suitability of using the contralateral metacarpal as a control and in providing sample size calculations for future studies involving a mechanical intervention to the metacarpal such as implantation of a metacarpophalangeal prosthesis. METHODS: Metaphyseal sectional areas, diaphyseal cortical sectional areas, second moments of area, average periosteal and medullary radii and bone densities were measured at nine transverse levels for each of seven pairs of index metacarpals using computed tomography and bone densitometry. Polar Fourier regression was used to assess the morphometry of sectional periosteal and endosteal boundaries. Differences between clinically important left-right parameters were assessed. RESULTS: Mean differences between clinically important left-right parameters were small (<3%) and similar to the degree of experimental precision. There were strong significant left-right correlations for the morphometric, geometric and densitometric parameters considered, indicating a high degree of bilateral mechanical symmetry. CONCLUSIONS: The contralateral bone is a suitable control for mechanical intervention studies of the human metacarpal, and the use of bilateral pairing results in an important reduction in sample size. RELEVANCE: Responses to mechanical interventions on the human metacarpal, such as implantation of a metacarpophalangeal prosthesis, are generally unknown. The degree of left-right mechanical symmetry in the human metacarpal provides a measure of the advantage of using paired design studies to address these questions.

Validation of a finite element model of the human metacarpal.

Implant loosening and mechanical failure of components are frequently reported following metacarpophalangeal (MCP) joint replacement. Studies of the mechanical environment of the MCP implant-bone construct are rare. The objective of this study was to evaluate the predictive ability of a finite element model of the intact second human metacarpal to provide a validated baseline for further mechanical studies. A right index human metacarpal was subjected to torsion and combined axial/bending loading using strain gauge (SG) and 3D finite element (FE) analysis. Four different representations of bone material properties were considered. Regression analyses were performed comparing maximum and minimum principal surface strains taken from the SG and FE models. Regression slopes close to unity and high correlation coefficients were found when the diaphyseal cortical shell was modelled as anisotropic and cancellous bone properties were derived from quantitative computed tomography. The inclusion of anisotropy for cortical bone was strongly influential in producing high model validity whereas variation in methods of assigning stiffness to cancellous bone had only a minor influence. The validated FE model provides a tool for future investigations of current and novel MCP joint prostheses.

Cement penetration and stiffness of the cement-bone composite in the proximal tibia in a porcine model.

PURPOSE: To assess the stiffness of the cement bone composite and the depth and uniformity of cement penetration into the surface of the tibial component during total knee reconstruction in a porcine model. METHODS: The effectiveness of 3 protocols were compared: 2 commonly used cementing techniques-finger-packing of cement on the cut surface followed by impaction, and coating of the undersurface of the prosthesis with cement followed by impaction-and a new method using a tibial cement-pressurising device. Cement penetration was measured by computed tomography; stiffness was determined by hydraulic penetration testing. RESULTS: Cement penetration at a depth of 1 mm was significantly greater following coating the undersurface of the prosthesis than following finger-packing (p=0.008). There was no significant difference at deeper levels or between the tibial-pressurising device group and either of the 2 other groups at any level (p>0.3 in all cases). Differences in surface stiffness by tibial plateau region were found in tibiae that had been cemented using finger-packing and in those that had had their undersurface coated, but not in tibiae that had been cemented using the tibial-pressurising device. CONCLUSION: The tibial cement-pressurising device eliminated regional differences in stiffness seen with other cementing methods. Elimination of these differences by using this device should reduce micromotion and the incidence of aseptic loosening of tibial base plates in total knee arthroplasty.

Changing incidence of primary total hip arthroplasty and total knee arthroplasty for primary osteoarthritis.

This study reports on the incidence of primary total hip arthroplasty (THA) and total knee arthroplasty (TKA) for primary osteoarthritis in Australia. Age-specific and gender-specific numbers for Australia, 1994 through 1998, and South Australia, 1988 through 1998, were obtained. Incidences were calculated per 100,000 population. In Australia, primary THA increased from 50.9/10(5) (1994) to 60.9/10(5) (1998). TKA increased from 56.4/10(5) to 76.8/10(5). Stratified by age and gender, changes in incidence for South Australia with respect to time were tested using regression analysis. South Australia showed a significant increase in the overall incidence of THA (P=.012) and TKA (P<.001), although this was not uniform across all age groups. No significant gender differences were found. The incidence of THA is increasing, and the incidence of TKA is increasing at a greater rate.

Mechanical and pathologic consequences of induced concentric anular tears in an ovine model.

STUDY DESIGN: Relations between induced concentric tears in the sheep disc and the mechanics of the intervertebral joint and vertebral body bone were analyzed. OBJECTIVE: To examine the effect of concentric disc tears on the mechanics of the spine. SUMMARY OF BACKGROUND DATA: Degeneration of the intervertebral disc results in changes to the mechanics and morphology of the spine, but the effect of concentric disc tears is unknown. METHODS: In this study, 48 merino wethers were subjected to surgery, and discs were randomly selected for either a needlestick injury or induction of a concentric tear in the anterior and left anterolateral anulus. Sheep were randomly assigned to groups for killing at 0, 1, 3, 6, 12, and 18 months. From each sheep, two spine segments were mechanically tested: one with a needlestick injury and one with a concentric tear. Macroscopic disc morphology was assessed by three axial slices of the disc. Sagittal bone slices were taken from cranial and caudal vertebral bodies for histologic analysis. RESULTS: Induced concentric tears decrease the stiffness of intact spine segments in left bending and the disc alone in flexion. In all other mechanical tests, the needlestick injury had the same effect as the induced concentric tear. In the isolated disc, the disc stiffness at 6 months was increased for right bending, as compared with the response at 1 month. This was associated with increased anterior lamellar thickening and increased vertebral body bone volume fraction. CONCLUSIONS: Concentric tears and needlestick injury in the anterior anulus lead to mechanical changes in the disc and both anular lamellar thickness and vertebral body bone volume fraction. A needlestick injury through the anulus parallel to the lamellae produces progressive damage.

Comparison of torsional strengths of bioabsorbable screws for anterior cruciate ligament reconstruction.

The goals of this study were to evaluate torsional strength and modes of failure in commercially available bioabsorbable interference screws and to test the effect of screw diameter on torsional strength when screws become jammed during insertion. We tested the Arthrex, BioScrew, Endo-Fix, Phantom, and Sysorb screws, all 20 mm in length. Four major modes of failure were encountered. Analysis of variance revealed that both screw type and diameter had a significant effect on failure torque. The Endo-Fix 7-mm screw had the lowest failure torque (1.07 +/- 0.18 N x m) and the Sysorb 8-mm screw had the highest (5.23 +/- 0.24 N x m). The Sysorb was significantly stronger than all the other screws. The failure torques were within the range that has been reported for manual screw insertion. We concluded that technical factors, which can affect insertion torque, assume particular importance with the use of bioabsorbable interference screws.

Hip fracture rates in South Australia: into the next century.

BACKGROUND: Fractures of the femoral neck already represent a major public health problem in Australia. This situation is set to worsen as the population ages. The present study estimates the number of patients over 50 years of age with femoral neck fractures that is expected to impact on the South Australian healthcare service into the next century. METHODS: Population projections from the Australian Bureau of Statistics 1996 census were combined with age- and gender-specific incidence rates for fractures of the femoral neck for persons over the age of 50 in South Australia. Projections for the expected number of hip fractures in this State were then calculated. RESULTS: Assuming there are no changes in the age- and gender-specific incidence of fracture rates, the number of fractures in South Australia is estimated to increase by approximately 66% by the year 2021 and 190% by 2051. CONCLUSION: Based on the population projections and the assumption that conditions contributing to hip fractures remain constant, the number of fractured neck of femurs will increase in far greater proportion than the overall population in the next century. The results of the present study indicate the serious implications for the South Australian healthcare system if there is no reduction in incidence rates.

Biomechanical analysis of the Medoff sliding plate.

BACKGROUND: A new device for fixation of hip fractures the Medoff plate was biomechanically tested and compared with the Ambi plate and Gamma nail. METHODS: Six pairs of human cadaveric femurs were used. The characteristics evaluated were structural stiffness, strain distribution, and modes of failure. Results were compared with the Ambi plate and Gamma nail biomechanical studies from a previous study that used the same methods. RESULTS: The Medoff plate was stiffer than the other two implants for intertrochanteric fractures, and for segmental subtrochanteric fracture. The Medoff plate was more load sharing than the other implants in these fracture configurations. The mean load to failure was lower than for the Ambi plate or the Gamma nail. CONCLUSION: The Medoff plate is a better load-sharing device than the Ambi (DHS) or Gamma nail systems. The main concern is its structural weakness. The implant failed at loads 50% less than other devices. The greatest risk of implant failure is with unstable subtrochanteric fractures.

Effect of pin location on stability of pelvic external fixation.

Pelvic external fixators allow two locations of pin purchase: anterosuperior (into the iliac crest) and anteroinferior (into the supraacetabular dense bone, between the anterior superior and anterior inferior iliac spine). The purpose of this study was to compare the stability of these two methods of fixation on Tile Type B1 (open book) and C (unstable) pelvic injuries. Five unembalmed cadaveric pelves (mean age, 68 years; four males and one female) were loaded vertically in a servohydraulic testing machine in a standing posture. The AO tubular system and Orthofix were used. On each pelvis, a Type B1 injury was simulated. Each external fixator was applied in each location in random order. Cyclic loads were applied through the sacral body to a maximum of approximately 200 N while force and displacement of the pelvic ring were recorded digitally. Sacroiliac joint motion was quantified tridimensionally with displacement transducers, mounted on the sacrum and contacting a target fixed to the posterior superior iliac spine. A Type C injury was created and augmented with two iliosacral lag screws, and the tests were repeated. For the Type B1 injuries with anteroinferior pin purchase, the mean stiffness was 201.2 N/mm for the AO frame and 203.2 N/mm for the Orthofix. For the anterosuperior frames the mean stiffness was 143.9 N/mm for the AO frame and 163.3 N/mm for the Orthofix. For Type B1 and Type C injuries, the anteroinferior location of pin purchase resulted in significantly reduced sacroiliac joint separation. There were no significant differences between the frame types. Dissection of the preinserted anatomic specimen revealed no evidence of injury to the lateral femoral cutaneous nerve after blunt dissection and drilling with protective drill sleeves. It is concluded that the anteroinferior location of external fixation pins is a safe technique with the potential for increased stability of fixation.

The influence of screw omission on construction stiffness and bone surface strain in the application of bone plates to cadaveric bone.

Biological fracture repair is becoming an increasingly popular means of fracture fixation. This technique involves a reduction in the amount of surgical trauma thereby preserving vascular supply and soft-tissue integrity combined with the implantation of less hardware. The aim is the stimulation of callus formation as the means by which fracture union occurs. This paper describes the mechanical effect of the symmetrical omission of screws on construction stiffness and bone surface strain following the application of bone plates to cadaveric bone. The influence of the pattern of screw omission was evaluated in both intact and osteotomized bone specimens. The application of bone plates using certain patterns of screw omission did not significantly effect the stiffness of the final construction whilst inherent bone surface strain was increased. It was concluded that the omission of screws in certain defined patterns for a given plate-bone construction may meet the criteria for 'biological fixation'. There was no apparent deleterious effect on structural stiffness following the omission of 40% of the total screw complement from a plate-bone construction.

The influence of screw torque, object radius of curvature, mode of bone plate application and bone plate design on bone-plate interface mechanics.

Using Fuji Prescale Pressure Sensitive Film we evaluated the influence of screw torque, object radius of curvature, mode of bone plate application (compression or neutral loading) and bone plate design on bone-plate interface mechanics. Testing was performed using 4.5 mm and 3.5 mm dynamic compression (DC) (stainless steel) and limited contact-dynamic compression (LC-DC) plates (titanium). Plexiglass tubing, having different radii of curvature, was used to simulate bone but having more uniform geometric and structural properties. With the film interposed between the bone plate and the plexiglass, bone plates were applied at predetermined levels of screw torque. The resultant image was digitized and then underwent computer-assisted analysis to yield the interface contact area (%) and the average force (in Newtons N) between the bone plate and the object to which it was applied. In any given object, screw torque had the most significant influence on both interface contact area (P = 0.0001) and average force (P = 0.0001) at the interface. The bone plates responded differently to dynamization. The DC plate, when applied in compression loading mode, was lifted off the object between the two central screws when compared to the same plate applied in neutral loading mode (P = 0.0001). Conversely, in the LC-DC plate, an increase in the overall interface contact area was observed when applied in compression loading mode (P = 0.0002). At a given level of applied screw torque, the object radius of curvature appeared to be a major determining factor for interface contact area and average force (P = 0.0001, P = 0.0001). We conclude that variables other than bone plate design also influence the interface contact area and average force between a bone plate and object to which it is applied.

The effects of foot position and load on tibial nerve tension.

Patients with tarsal tunnel syndrome of unknown etiology do poorly after surgical decompression. Although surgical decompression addresses the soft tissue constraints, it ignores the role of osseous support. Some authors have suggested that a pes planus deformity (i.e., valgus hindfoot and abducted forefoot) is an unrecognized cause of tarsal tunnel syndrome due to increased tibial nerve tension. An in vitro study was performed on nine cadaveric feet to determine the effects of foot position and load on tibial nerve tension. Tensile forces placed through the tibial nerve were measured when the foot was placed in dorsiflexion, eversion, combined dorsiflexion-eversion, and then under cyclical load and increasing internal rotation at 5 degrees increments from 0 degrees to 20 degrees. The nerve tension was reassessed after the creation of a pes planus deformity under the previous conditions. Tibial nerve tension in the stable and unstable foot was significantly increased by eversion, dorsiflexion, and combined dorsiflexion-eversion. Tibial nerve tension was significantly greater in an unstable foot when compared with a stable foot during eversion, dorsiflexion, and combined dorsiflexion-eversion. In the stable foot, tibial nerve tension was significantly increased during axial loading with increasing internal rotation when compared with 0 degrees rotation. The increased tibial nerve tension in the stable foot was significant with increasing internal rotation when 0 degrees was compared with 10 degrees, 15 degrees, and 20 degrees. In the unstable foot, the tibial nerve tension was significantly increased with increasing internal rotation compared with the nerve tension at 0 degrees of rotation. The increased tibial nerve tension in the unstable foot was significant with increasing internal rotation when 0 degrees was compared with 5 degrees, 10 degrees, 15 degrees, and 20 degrees. When stability of the foot and internal rotation were compared independently, each factor increased tibial nerve tension. However, these factors acting together did not significantly compound the increase in nerve tension. This study demonstrates that tibial nerve tension is increased in an unstable foot compared with a stable foot during eversion, dorsiflexion, combined dorsiflexion-eversion, and cyclical load with increasing internal rotation.

Stiffness of trabecular bone of the tibial plateau in patients with rheumatoid arthritis of the knee.

Stiffness of subchondral proximal tibial trabecular bone is a factor in the stability of prostheses implanted into that bone. The stiffness of trabecular bone in osteoarthritis (OA) has been documented. Trabecular bone in rheumatoid arthritis (RA) is osteopenic in numerous sites and morphologically abnormal in the proximal tibia. Reliable data on proximal tibial bone in RA are lacking, although 1 study failed to identify abnormalities. The purposes of this study were (1) to document the stiffness of the proximal tibial cancellous bone in patients with RA, (2) to determine the effect of angular deformity on bone stiffness in rheumatoid patients, and (3) to compare RA stiffness values with those in published reports for OA. Fifteen tibial plateau were obtained from patients with RA during surgery. Each plateau was horizontally seated in a mold and covered with cement. The plateau was divided into 6 regions, which were used to facilitate comparison between specimens and the existing literature. Indentation tests were conducted with a 4-mm-diameter cylindrical indentor controlled by an MTS machine. The indentor descended at a rate of 2 mm/min to a maximum depth of 1.0 mm; load and displacement data were digitally recorded. Stiffness was calculated from the slope of the linear region of the curve using best-fit linear regression. Where varus deformity was present, stiffness in the medial plateau was higher overall than for the other compartment; whereas in the case of valgus deformity, stiffness of the lateral side was significantly higher (P < .05 for each observation). In comparison to older normal specimens, both the medial compartment of the varus RA specimens (P < .01) and the posterolateral compartment of the valgus RA specimens (P < .01) had significantly lower stiffness. Comparison with OA specimens showed that in varus RA, the posteromedial region had significantly lower stiffness than in varus OA at the same site (P < .01). In valgus RA, the lateral region had significantly lower stiffness than in valgus OA at the same site (P < .01). The mean stiffness ratio of the valgus RA was significantly (P < .01) altered from normal, and for the varus RA, it was significantly (P < .01) different from normal posteriorly. The stiffness ratios for the varus RA were significantly (P < .01) different from those for varus OA; there was no difference between valgus RA and valgus OA. It is concluded that RA affected bone has significantly lower stiffness than normal and osteoarthritic bone. The loaded plateau is stiffer than the unloaded plateau in angular deformity, but is still less stiff than normal bone and osteoarthritic plateaus with corresponding deformities.

Bone plate fixation: an evaluation of interface contact area and force of the dynamic compression plate (DCP) and the limited contact-dynamic compression plate (LC-DCP) applied to cadaveric bone.

OBJECTIVES: To determine the effects of bone morphology on the interfacial mechanics of bone plate fixation. DESIGN: A comparison of two bone plate designs using measures of contact area and force in three conditions of bone type. INTERVENTION: The influence of bone plate design, bone type, and implant location on interfacial characteristics of contact area and average force were evaluated using pressure-sensitive film in cadaveric bone. Data were quantitated using computer-assisted image analysis. MAIN OUTCOME MEASUREMENTS: Means of the two dependent variables in each of the three conditions were compared using analysis of variance. RESULTS: In human femora and equine third metacarpi, there were no apparent differences in interface contact area attributed to bone plate design, although there were significant differences for interface contact force. In humeri, significant differences in both interface contact area and force were found between bone plate designs. CONCLUSIONS: Results show that bone surface topography and implant location are major determinants of the interfacial elements of contact area and average force between a plate and the bone to which it is applied.

Insertional torque and pull-out strengths of conical and cylindrical pedicle screws in cadaveric bone.

STUDY DESIGN: Insertion torque and pull-out strengths of conical and cylindrical pedicle screws were compared in human cadaveric vertebral bodies. OBJECTIVES: To compare the performance of the conical design with the cylindrical design, and to determine whether insertional torque correlates with pull-out strength. SUMMARY OF BACKGROUND DATA: A tapered pedicle screw design may lessen the likelihood of implant failure. Its effect on thread purchase is not known. Previous studies of cylindrical designs on the relation between insertion torque and pull-out strength have been conducted in bovine and synthetic bone. METHODS: Seventy-eight pedicles were assigned randomly to one of the following pedicle screw: Texas Scottish Rite Hospital (Sofamor-Danek, Memphis, TN), Steffee VSP (Acromed, Cleveland, OH), Diapason (Dimso, Paris, France), AO Schanz (Synthes, Paoli, PA), or Synthes USS (Synthes, Paoli, PA). Pedicle screws were inserted with a torque screwdriver. Each screw was extracted axially from the pedicle at a rate of 1.0 mm/sec until failure using an MTS machine (Bionix 858, Minneapolis, MN). Force data were recorded. RESULTS: The conical design had the highest insertion torque. There were no significant differences in pull-out between any of the screw types. Correlation between insertional torque and pull-out strength was statistically significant only with the Texas Scottish Rite Hospital and Steffee VSP in L4 and AO Schanz in L5. CONCLUSIONS: A conical screw profile increases insertion torque, although insertional torque is not a reliable predictor of pull-out strength in cadaveric bone. Screw profile (with similar dimensions) has little effect on straight axial pull-out strengths in cadaveric bone.

Fixation of segmental subtrochanteric fractures. A biomechanical study.

Segmental subtrochanteric fractures have been associated with high rates of malunion, nonunion, and implant failure. Although the use of second generation intramedullary nails with proximal interlocking in the femoral head has been advocated for these fractures, shaft fractures at the tip of short second generation intramedullary rod nails in clinical studies raise concerns about the mechanical suitability of these implants. No biomechanical data are available on the strength of fixation and mode of failure of these newer implants. This study compares the strength and failure mode of segmental subtrochanteric fractures stabilized with 3 current implants. Eighteen anatomic specimen femurs were obtained. The 3 implants tested were the Russell-Taylor reconstruction nail, a short intramedullary hip screw, and a long intramedullary hip screw. Each femur was instrumented and a segmental subtrochanteric fracture was created. The femurs were loaded in a Materials Testing System. The strength of the reconstruction nail group (2869 +/- 210 N) was significantly greater than for the short intramedullary rod hip screw (2330 +/- 490 N), and the long intramedullary rod hip screw (2181 +/- 244 N). The failure modes for the implants were: screw cut out of the femoral head for the reconstruction nail; fracture of the femoral shaft for the short intramedullary hip screw, and implant bending for the long intramedullary hip screw. This study suggests that the reconstruction nail is the superior implant for segmental subtrochanteric fractures of those tested. Implant bending and shaft fractures at lower loads make the 2 intramedullary rod hip screw implants less suitable.

Biomechanical analysis of hemipelvic deformation after corticospongious bone graft harvest from the posterior iliac crest.

STUDY DESIGN: The stiffness of the hemipelvis during simulated physiologic loading and the bone deformation in the remaining posterior ilium after the harvest of increasing sizes of corticospongious bone graft were compared with values in intact bone. OBJECTIVES: To quantify the biomechanical effects of the removal of bone grafts from the posterior ilium and to relate the size of graft removed to the stiffness of the hemipelvis and deformation of the remaining bone. SUMMARY OF BACKGROUND DATA: Bone fractures and pelvic instability have been reported to complicate graft harvest from the posterior iliac crest. There is no quantitative data relating graft size to the mechanical properties of the remaining ilium. METHODS: Seven cadaveric hemipelves were loaded with a materials testing machine through the superior sacrum while supported at the acetabulum and stabilized with a cable fixed to the ilium. Force and displacement histories and deformation in the greater sciatic notch were recorded for the intact bone and after removal of corticospongious bone graft in 1.5-cm increments from the posterior iliac crest. RESULTS: If the length of the removed corticospongious bone graft exceeded 3.0 cm, the stiffness of the posterior pelvic ring decreased, and deformation in the remaining bone increased substantially. CONCLUSIONS: Removal or bone graft in excess of 3 cm from the posterior ilium increases the risk of iatrogenic fatigue fracture.

Sample size estimates for the use of human bone in the experimental study of cancellous screw extraction mechanics.

Cancellous bone screw extraction strengths are determined in four matched anatomical locations in five pairs of unembalmed human distal femora. The results are used to calculate sample size estimates for the experimental detection of differences in screw extraction mechanics. While unmatched designs require prohibitively large numbers of bone specimens, matched designs require between 71.6 and 88.5% fewer, depending on the correlation between matched observations and the homogeneity of sample variances.

The strength of plate fixation in relation to the number and spacing of bone screws.

Our purpose was to study the relationship between the number of plate holes filled and the spacing between the screws and the resultant strength of plated constructs. Broad regular DC plates were anchored with 4.5-mm cortical screws to blocks of polyurethane foam. Six constructs were tested: (a) screws in holes 1, 2, and 3; (b) screws in holes 1 and 3; (c) screws in holes 1 and 4; (d) screws in holes 1 and 5; (e) screws in holes 1 and 6; (f) screws in holes 1, 3, and 5. The strength was quantified using a material-testing system. In cantilever and four-point bending, the constructs were loaded in both gap-closing and gap-opening modes. Screws in holes 1, 2, and 3 were tested against other constructs. For cantilever bending (gap opening and gap closing), construct (a) was stronger than construct (b), as strong as construct (c), but weaker than the constructs with more widely spaced screws (p < 0.0001). In terms of four-point bending, for gap opening, the standard fixation (construct (a) was stronger than construct (b) but weaker than the more widely spaced constructs. For gap closing, construct (a) was stronger than constructs (b) and (c) but weaker than the rest. Regardless of the spacing of screws and the plate length, strength in torsion was dependent on the number of screws securing the plate. In a laboratory fracture model of plate-bone constructs tested to failure by screw pullout, wider spacing of bone screws increases the bending strength of screw-plate fixation and can be more effective than increasing the number of screws. Torsional strength is independent of screw placement in plates of a given width and depends on the number of screws used.

Characterization of vertebral strength using digital radiographic analysis of bone structure.

Bone mineral densitometry (BMD) is useful in predicting fracture risk, but, unfortunately, there is a significant degree of overlap in BMD measurements of patients who have a high risk of fracture and patients with a low risk of fracture. In this study, a method of characterizing trabecular bone structure in digitized radiographs of vertebrae is proposed and assessed. A significant correlation between bone "structure" and the compressive strength of vertebral bodies was found. The utility of the parameter for distinguishing between "weak" and "strong" bone samples was assessed using receiver operating characteristic (ROC) analysis. Using this analysis, the structural parameter produced an area under the ROC of 0.88 +/- 0.05, while a bone density measure produced an area of 0.79 +/- 0.07. The results suggest that the addition of a measure of bone structure to the conventional measures of bone density may prove useful in predicting the quality of bone when considering surgical or medical intervention for osteoporotic conditions.