Total Hip Arthroplasty in Hip Osteoarthritis with Subtrochanteric Localized Periosteal Thickening: Preoperative Planning Using Finite Element Analysis to Determine the Optimal Stem Length

1. Introduction

The 2013 revision of the American Society for Bone and Mineral Research Task Force defines localized periosteal thickening of the femoral lateral cortex as a primary diagnostic criterion for atypical femoral fractures (AFFs) [1]. To avoid stress concentration, surgical interventions for an AFF generally involve using the longest possible intramedullary nails. While prophylactic intramedullary nailing is sometimes performed for localized cortical thickening, consensus regarding the appropriate surgical indications or timing is lacking. Moreover, careful monitoring is required as an AFF can develop with even minor trauma or a low-energy impact.

Total hip arthroplasty (THA) in the presence of ipsilateral osteoarthritis with localized cortical thickening is rare, with few reported cases. In a similar case report to ours, a stem of insufficient length was used, leading to the development of AFF postoperatively [2]. Although the stem tip might be positioned near the area of cortical thickening, potentially increasing the risk of an AFF due to concentrated stress, no studies have specifically addressed stem selection from the perspective of stress distribution.

Herein, we describe our experience in performing THA in a patient with ipsilateral osteoarthritis of the hip and localized cortical thickening in the subtrochanteric region of the femur. The procedure planning was guided by preoperative finite element method (FEM) simulation to assess femoral stress distribution.

In this study, we focused on stem length and conducted three-dimensional (3D) simulations using FEM to evaluate stress distributions in stems of varying lengths, ultimately guiding our stem selection. Moreover, we considered and emphasized treatment strategies for ipsilateral osteoarthritis with localized cortical thickening. We hypothesized that the longer the bridging length of the stem over the thickened cortex, the greater the stress reduction in that area. To the best of our knowledge, this study is the first to examine the relationship between stem length and stress distribution in local femoral cortical thickening using FEM.

2. Case Description

2.1. Case Presentation

A 78-year-old woman without a history of developmental hip dysplasia had been diagnosed with systemic lupus erythematosus at 46 years of age. The initial treatment involved steroid pulse therapy, followed by continuous oral prednisolone thereafter. Concurrently, treatment with 200 mg etidronate was initiated for glucocorticoid-induced osteoporosis. After 47 days, the medication was switched from etidronate to 35 mg alendronate, which was continued thereafter.

At 73 years of age, the patient experienced a fall and was diagnosed with a left subtrochanteric AFF based on plain radiography and computed tomography (CT) findings. Additionally, localized lateral cortical thickening was observed in the ipsilateral subtroschanteric region, along with osteoarthritis of the right hip (Figure 1). The patient underwent open reduction and internal fixation with an intramedullary nail for the left subtrochanteric AFF. Alendronate was discontinued, and treatment comprising 600 µg teriparatide (recombinant) was initiated, which was continued for 2 years. Due to nonunion at the fracture site, the patient underwent surgery for pseudarthrosis at 75 years of age (Figure 2). Given the gradual worsening of the right hip osteoarthritis, THA was planned when the patient turned 78 years of age.

2.2. Finite Element Analysis

Informed consent was obtained from the patient for using and publishing these data, and this retrospective study was approved by the ethics review committee of our institution (approval code: H27-041).

We performed preoperative 3D simulations using FEM to analyze stress distributions in the femur following stem insertion. We also evaluated the stress distributions at the site of cortical thickening for different stem lengths.

3. Methods

3.1. Three-Dimensional Modeling of the Femur and Implant

Preoperative CT was performed using a 256-slice multidetector CT scanner (Brilliance iCT; Philips Healthcare, Cleveland, Ohio, USA). The scan, which included a bone mineral density phantom (QRM-BDC/3 Phantom, QRM Quality Assurance in Radiology and Medicine GmbH, Möhrendorf, Germany), captured images from the pelvis to both knee joints (120 kV/166 mAs, 1-mm thin slice).

First, based on CT images, we simulated stem insertion using the ZedHip 3D preoperative planning software (Lexi Co., Ltd, Tokyo, Japan). The patient in this case had a narrow medullary canal and mild anterior bowing of the femoral shaft. As the selection of stems that could span the area of cortical hypertrophy and fit the femoral morphology of the patient was limited, we opted for the Arcos® One-piece Femoral Revision System (Zimmer, Warsaw, Indiana, USA) with a 9.5 mm diameter and 175 mm length, in the high offset configuration.

We then generated Standard Triangulated Language (STL) data for stems of varying lengths. The original stem STL data were imported into the 3D computer-aided design software Fusion 360 (Autodesk, San Francisco, CA, USA) and modified using the “cut” and “join” functions. The stem lengths included the original size of 175 mm, a 115-mm stem aligned with the level of cortical thickening at the stem tip, and additional lengths ranging from 90 mm to 180 mm in 10-mm increments (Figure 3). The 90-mm stem was the shortest length that could be generated without altering the proximal stem geometry, while 180-mm was the maximum length that could be inserted with the same alignment.

Next, we constructed a 3D model of the right femur. Preoperative CT data were imported into the FEM analysis software MECHANICAL FINDER (MF, version 13.0, Extended Edition, Research Center of Computational Mechanics, Tokyo, Japan), where bone was defined as an area with a density of ≥300 Hounsfield Units (HU). Only the right femur was extracted. We generated a model of the femur by performing an osteotomy at the proximal level 10 mm above the lesser trochanter and creating an isosurface mesh to represent the external shape of the femur.

By registering the acquired stem STL data with a 3D model of the right femur in MF, we were able to replicate femur post-stem insertion. The insertion depth and alignment of the stem were reproduced in MF based on the simulation conducted in ZedHip, ensuring consistency in 3D coordinates across the analysis of each stem (Figure 4).

3.2. Material Properties

We used four-node tetrahedron elements for the solid components. For stress analysis, shell elements with a thickness of 0.001 mm were applied to the bone surface to ensure that they did not affect bone strength.We conducted mesh convergence tests to determine the mesh size.

3.2.1. Bone Material (Femur)

We set the internal mesh size to minimum and maximum values of 1 mm and 2 mm, respectively, and employed a heterogeneous material model for the bone properties. The Young’s modulus was derived using bone mineral density (BMD, ρ (g/cm³)), calculated from the CT values (HU) based on an assumed linear relationship [3,4]. Then, the values were estimated using Keyak’s predictive transformation formula and incorporated into the model. Poisson’s ratio was set to 0.40 [5].

3.2.2. Other Materials (Stem and Artificial Head)

For these components, we set the internal mesh size to a minimum of 0.5 mm and a maximum of 1 mm. This study used a model with homogeneous material. The stem was modeled using the material properties of a titanium alloy (Ti-6Al-4V) with a Young’s modulus of 109 GPa and Poisson’s ratio of 0.28. For the artificial head, we used a ceramic (alumina) with a Young’s modulus of 350 GPa and Poisson’s ratio of 0.23.

3.3. Loading and Boundary Conditions

We designed the loading conditions to simulate the forces encountered during daily activities. We applied the maximum loads corresponding to “normal walking” and “stair climbing” based on previous reports [6,7]. The “stair-climbing” scenario allowed for the examination of higher load forces and stronger torques than “normal walking.” The magnitude of each vector was determined by the patient’s body weight (50.4 kg), with the load acting as shown in Table 1 and Figure 5.

For the constraint conditions, the distal part of the femur was fixed (Figure 4). The boundary conditions included a fixed connection between the stem and head, and a contact condition with a friction coefficient of 0.49 between the femur and stem [8].

3.4. Static Structural Analysis

We conducted simulations under normal walking and stair-climbing conditions using stems of varying lengths. Elastic analysis was performed with the load gradually increasing linearly, and linear static analysis was applied to the calculations. We used equivalent stress (MPa) as the measure of stress and calculated the overall stress distributions across the femur, as well as the average and maximum equivalent stress values in the area of cortical thickening. The hypertrophic region was standardized to 40 mm in width and 60 mm in length along the bone axis, centered around the apex of the hypertrophic area to encompass the entire hypertrophic region. The calculations considered only the stress values of the femoral surface shell elements (Figure 6).

4. Results

The number of elements comprising each model is shown in Table 2.

Figure 7 shows the stress distributions across the lateral surface of the femur for each model during normal walking and stair-climbing conditions. Figure 8 illustrates the mean and maximum equivalent stress values in the region of cortical thickening. The stress distribution across the lateral surface of the femur decreased following stem insertion. No new areas of stress concentration were observed during normal walking or stair-climbing conditions.

The mean equivalent stress at the thickened region decreased gradually as stem length increased, following an approximation curve. The stress nearly plateaued at a length of 120 mm during normal walking and 150 mm during stair climbing, with approximately the same level as the control. Similarly, the maximum stress decreased with increasing stem length, following a trend similar to the mean values. However, elevated stress levels were observed near a stem length of 115 mm, where the stem tip overlapped with the thickened region during normal walking and stair-climbing conditions. Beyond this length, where the stem tip extended past the thickened area, the stress followed a decreasing approximation curve. The stress values plateaued at a length of approximately 150 mm, which was almost identical to that of the control.

Based on the abovementioned findings, we performed THA using a posterior approach (Figure 9). Postoperatively, full weight bearing on the affected limb was allowed. The patient could walk with a cane and was transferred to another hospital for rehabilitation on postoperative day 14. At the final follow-up (3 months postoperatively), the patient could walk independently and experienced no difficulties in daily activities.

5. Conclusions

In the present case of femoral osteoarthritis with localized cortical hypertrophy in the subtrochanteric region, we performed FEM-based preoperative planning and used a long stem revision for THA. From the perspective of stress distribution, a stem length >150 mm (a bridging length of 1.4 femoral cortical diameters) is required to bridge the area of cortical hypertrophy. The 175 mm stem used in this case was sufficient for stress distribution.