The Effect of Cement Placement on Augmentation of the Osteoporotic Proximal Femur

Abstract

Femoroplasty, the augmentation of the proximal femur, has been shown in biomechanical studies to increase the energy required to produce a fracture and therefore may reduce the risk of such injuries. The purpose of our study was to test the hypotheses that: (1) 15 mL of cement was sufficient to mechanically augment the proximal femur, (2) there was no difference in augmentation effect between cement placement in the intertrochanteric region and in the femoral neck, and (3) cement placement in the femoral neck would predispose the proximal femur to an intertrochanteric fracture, whereas trochanteric placement would result in subtrochanteric fractures. In each of 18 pairs of osteoporotic human cadaveric femora, 15 mL of polymethylmethacrylate bone cement was injected into the trochanteric or femoral neck region of 1 femur, and the noninjected femur was used as the control. The augmentation effect of femoroplasty was evaluated under simulated fall conditions using a materials testing machine. Multiple linear regressions incorporating random effects were used to check for associations between covariates (bone mineral density, cement location, and treatment) and the parameters of interest (stiffness, yield energy, yield load, ultimate load, and ultimate energy). Significance was set at P < .05. It was found that femoroplasty with 15 mL of cement did not significantly increase stiffness, yield energy, yield load, ultimate load, or ultimate energy relative to paired controls. There were no significant differences in parameters of interest or fracture patterns in specimens augmented in the femoral neck versus the trochanter. It was concluded that 15 mL of cement was not sufficient to augment the proximal femur and that there was no biomechanical advantage to the placement of cement within the femoral neck versus the trochanter.

Femoroplasty (the use of bone cement in the proximal femur to augment femoral strength and prevent fracture) could offer a minimally invasive method to prevent fracture in osteoporotic patients with high fracture risk. Femoroplasty has experimentally been shown to increase fracture load and energy to fracture in quasistatic and dynamic loading conditions.^1,2 In those studies, large amounts of cement (approximately 40 mL) were used in an effort to fill the proximal femur. However, such large amounts of cement may create stress risers at the cement-bone interface,³ may risk thermal injury from the exothermic polymerization of the cement to the cancellous bone and adjacent blood vessels,⁴ and may make salvage in the proximal femur challenging.⁵ These undesirable outcomes may be mitigated by using less cement and controlling the specific location of cement placement in the proximal femur.

The purpose of our study was to test the hypotheses that: (1) 15 mL of cement was sufficient to mechanically augment the proximal femur, (2) there was no difference in augmentation effect between cement placement in the intertrochanteric region and in the femoral neck, and (3) cement placement in the femoral neck would predispose the proximal femur to an intertrochanteric fracture, whereas trochanteric placement would result in subtrochanteric fractures.

Materials and Methods

We obtained 18 pairs of fresh-frozen osteoporotic (T score < –2.5)⁶ human female femora (mean age, 87.0 ± 6.8 years at time of death) from the Maryland State Anatomy Board. Osteoporosis was confirmed with dual energy x-ray absorptiometry (DEXA; Discovery QDR Series; Hologic, Inc, Bedford, Massachusetts) to measure bone mineral density (BMD), by scanning the denuded femora in a water bath to compensate for the absent soft tissue. For the femora in the neck and trochanteric groups, the mean T scores were –3.54 ± 0.83 and –3.53 ± 0.82, respectively, and the BMD values were 0.51 ± 0.11 and 0.51 ± 0.10 g/cm², respectively. We used computed tomography to scan all specimens at 1-mm-thick slice intervals with axial orientation (Aquilion 64 CT System; Toshiba America Medical Systems, Tustin, California). All femora were stored in sealed plastic bags at –20°C until 1 day before testing, at which time they were removed from the freezer and allowed to thaw overnight to room temperature in the same sealed plastic bags.

One femur from each pair was randomly chosen for femoroplasty, and the contralateral femur was used as the control specimen. In femora chosen for augmentation, an 11G cannula with trocar (Stryker Instruments, Kalamazoo, MI) was inserted under fluoroscopic guidance (Phillips BV 300 C-Arm, Eindhoven, the Netherlands). All procedures and measurements were carried out jointly by 2 of the authors (EGS, SJW). We chose 2 positions within the proximal femur for cement placement: (1) the femoral neck and proximal portion of the trochanter (neck group) and (2) the entire trochanter (trochanter group). There was no significant difference in BMD between the neck and trochanter groups. In both augmentation groups, a single entry point in the lateral aspect of the greater trochanter was made. Spineplex bone cement (Stryker Instruments) was prepared, immediately loaded into 5-mL syringes, and manually injected under fluoroscopic guidance into the assigned regions of femora slated for augmentation until leakage occurred or the cement began to deviate from the desired region. In the neck group, cement was injected to fill the region distal to the femoral head and just distal of the intertrochanteric line, whereas in the trochanter group, the fill region was distal to the femoral neck and superior to the lesser trochanter (Figure 1 ). The mean cement-filling volumes for the neck and trochanter groups were 15.2 ± 1.5 mL and 15.1 ± 1.6 mL, respectively.

Figure 1.

Schematic of femoral neck augmentation (left) versus trochanteric augmentation (right). Dotted lines indicate injection needle trajectories.

The temperature of the exterior-posterior aspect of the femoral neck cortex was recorded to determine the potential for damage to the posterior femoral neck vessels, and thus femoral head vascular supply, caused by augmentation. After the temperature began to increase, temperature readings were taken every 60 seconds with a logging thermocouple (K-Type; Omega Engineering, Inc, Stamford, Connecticut). Temperature measurements were terminated once the temperature began to decrease. In the neck and trochanter groups, the mean differences between starting and peak temperatures were 6.4°C ± 4.0°C and 3.5°C ± 1.9°C, respectively.

All femora were transected, potted, and tested as reported previously.² The femurs were positioned so that the axis of the femoral shaft was 10° relative to the horizontal plane and the shaft was rotated internally 15° (Figure 2 ). This configuration simulates a fall on the greater trochanter.⁷ The greater trochanter was supported by a dollop of bone cement to distribute reaction loads. A preload of 40 N was applied through a nylon surrogate acetabulum attached to the actuator of a materials testing machine (MTS Bionix 858 Test System; MTS, Eden Prairie, Minnesota). The actuator was displaced downward at a rate of 100 mm/s until failure occurred.⁷ Load and displacement data were recorded at 1024 Hz.

Figure 2.

A schematic of the test setup.

Yield load was defined as the first inflection point on the load-versus-displacement curve (Figure 3 ), and ultimate load was defined as the load immediately preceding a sharp drop in load that was not recovered.² Yield energy and ultimate energy were calculated as the area under the load-versus-displacement curve to yield load and ultimate load, respectively. Stiffness was calculated as the slope of the linear portion of the elastic region of the load-versus-displacement curve.

Figure 3.

Sample plot of a femur augmented in the trochanteric region and its contralateral control: yield load (a, c) and ultimate load (b, d).

After fracture, all specimens were again scanned with computed tomography to document fracture location and, in the femoroplasty specimens, whether the fracture occurred through cement. Fracture locations were classified according to the Orthopaedic Trauma Association system⁸ and were summarized into 4 categories (intertrochanteric, subtrochanteric, basicervical, and subcapital).

Multiple linear regressions incorporating random effects were used to check for an effect of cement placement (neck versus trochanteric) and treatment (femoroplasty versus control) on the parameters of interest (yield and ultimate load, yield and ultimate energy, and stiffness). We also included BMD as a covariate. All data were analyzed using Stata10 software (StataCorp LP, College Station, Texas). Significance was set at P < .05.

Results

We found no significant differences in stiffness, yield energy, yield load, ultimate load, or ultimate energy between the control and augmented specimens (Table 1 ) or between the neck and trochanter groups.

Table 1.

Summary of Parameters of Interest^a

Group	Stiffness, kN/m	Yield Load, N	Ultimate Load, N	Yield Energy, J	Ultimate Energy, J
Neck
Control	635 (485 – 784)	1918 (1677 – 2158)	1762 (1530 – 1994)	4.9 (3.8 – 6.0)	10.2 (6.0 – 14.3)
Femoroplasty	591 (461 – 721)	2044 (1642 – 2445)	1977 (1588 – 2367)	5.4 (3.3 – 7.6)	11.6 (7.5 – 15.6)
Trochanteric
Control	599 (510 – 689)	1821 (1395 – 2247)	1912 (1470 – 2353)	3.5 (2.2 – 4.7)	5.2 (3.6 – 6.8)
Femoroplasty	645 (434 – 856)	2037 (1389 – 2685)	1975 (1488 – 2462)	3.6 (2.8 – 4.5)	11.9 (7.7 – 16.2)

^a Values are means (95% confidence interval).

In the neck group, the specimens sustained 8 intertrochanteric fractures and 1 intracapsular fracture (none through the cement); their respective controls sustained 7 intertrochanteric fractures and 2 intracapsular fractures. In the trochanter group, the specimens sustained 7 intertrochanteric fractures and 2 intracapsular fractures (4 at the cement boundary and 5 through the cement); their respective controls sustained 6 intertrochanteric fractures, 2 intracapsular fractures, and 1 subtrochanteric fracture. We found no significant association between fracture pattern and loads, energies, or stiffness. None of the fracture lines passed through the cannulae holes.

Discussion

This study did not support our hypothesis that femoroplasty with a limited amount (15 mL) of cement would increase stiffness, yield load, ultimate load, yield energy, and ultimate energy. Our results indicated that 15 mL of polymethylmethacrylate cement is not sufficient to augment the proximal femur. In 2 previous studies,^1,2 larger amounts of cement resulted in an increased resistance to fracture in a cadaveric model. In our current study, we chose to inject only 15 mL of cement to limit the deleterious effects of potential thermal injury to femoral neck vessels during cement polymerization: the temperature did not rise above the threshold for thermal necrosis, which is considered to be exposure to temperatures greater than 50°C for more than 1 minute.^9–11 Temperature measurements were made on the periosteal surface. It is unknown what the temperatures were in the intramedullary canal, but they are anticipated to be higher than those measured on the exterior cortex. Lesser amounts of cement would also be advantageous; should revision surgery be necessary, they also theoretically pose a reduced risk for embolization.

Based on cement placement in the neck or trochanter of the proximal femur, we found no significant differences in augmentation or site-specific fracture patterns; the primary pattern in all specimens (control and augmented) was intertrochanteric. This finding may indicate that fracture pattern is governed more by the test protocol than by augmentation. A study using a similar test setup resulted in most neck fractures occurring in unaugmented specimens.¹² Yet in another study, the test resulted in a majority of intertrochanteric fractures.¹ In the former study, only 2 specimens were osteoporotic; the balance were osteopenic or normal. In the later study, the specimens were all osteopenic or osteoporotic. It is unknown what confounding role bone density may have on fracture pattern. Furthermore, BMD as measured by DEXA is an area-based “apparent” density and not a true volumetric-based density. As such, specimens with the same BMD may exhibit different mechanical behavior because of the unaccounted for geometric differences. For example, in an earlier study,² the specimens had BMDs similar to those in the current study, yet the current control specimen failure strengths were notably lower. This result may be caused in part by the fact that all the specimens in the current study were female (typically smaller) rather than the even-gender mix of the previous study. There may be other gender-related differences yet unidentified.

Our study did have limitations. First, we used paired cadaveric specimens to minimize the variability between femoroplasty and control groups; however, differences in mechanical properties and BMD between paired femora can be as high as 15%¹³ and 20%,¹⁴ respectively. The greatest difference in BMD values between paired femurs in our study was 17%, and the average percentage difference across all pairs was 7%. Second, we did not measure the intramedullary canal pressure during injection of cement and therefore cannot comment on the actual effect that reducing the volume of cement has on the potential for embolization of fat, bone marrow, or bone cement.¹⁵ Third, we did not investigate the augmentation effect of larger volumes of cement. We used 15 mL of cement because we sought to minimize the cement volume to allow more precise control of placement. Greater quantities of cement have been shown to result in augmentation.^1,2 Furthermore, potential deleterious effects of PMMA cement in the proximal femur, such as osteolysis and osteonecrosis, are unknown.

In conclusion, proximal femoral augmentation with 15 mL of polymethylmethacrylate cement was ineffective in improving strength, stiffness, or energy to fracture of the proximal femur. Injecting the cement in the trochanteric or femoral neck areas did not affect the variables measured nor result in different fracture patterns. Future studies with finite element analysis are needed to explore the use of different filling patterns, agents, and potential rebar additives, and additional computer modeling studies using such data could lead to improved materials or injection locations that would augment proximal femurs in the clinical setting. If not, as the current study suggests, femoral augmentation may not be indicated for clinical use.