The influence of femoral component design on postoperative periprosthetic femoral fracture after uncemented direct anterior total hip arthroplasty in the elderly

Abstract

Introduction

Aging populations and expanding indications will greatly increase the volume of total hip arthroplasty (THA) in all age groups, including patients over 70 years old. Minimally invasive, uncemented direct anterior THA offers potential advantages for treating elderly patients. However, literature indicates higher risks of postoperative periprosthetic femur fractures (PPFFs) with both direct anterior THA and uncemented femoral stems. This retrospective study investigates the influence of femoral stem design on PPFF incidence in uncemented direct anterior THA among patients older than 70 years.

Methods

557 primary THAs in patients aged 70 or over were reviewed for PPFFs from a consecutive series of 2011 patients undergoing direct anterior THA from a fellowship-trained adult reconstruction surgeon from 2015 to 2020. Exclusion criteria included age (<70) and posterior approach. For the first cohort of 361 patients (79 of which passed exclusion criteria) the surgeon used a single-tapered, proximally porous coated, collarless titanium stem. For the next 1650, (478 of which passed exclusion), the surgeon used a dual-tapered, collared, hydroxyapatite-coated titanium stem. Included patients were carefully monitored until March 2021 for PPFFs. A Fisher's exact test was used to compare the incidence PPFFs between the 2 implant designs.

Results

2 of 79 (2.5%) patients had atraumatic PPFFs at an average of 19.5 days post-operatively in the first cohort. Both experienced a Vancouver type B2 periprosthetic fracture and required femoral revision. No patients (0/478, 0%) in the second group sustained a PPFF. (P = 0.0199).

Conclusion

In this comparison, the dual-taper, hydroxyapatite-coated implant had a significantly lower PPFF rate among elderly patients than a single-taper, proximally porous stem without a collar.

1. Introduction

The volume of total hip arthroplasty (THA) is expected to increase substantially due to aging populations and the expansion of THA into both older and younger patient demographics. The rate of THA is projected to grow from 498,000 surgeries annually in 2020 to 1,429,000 in 2040.¹ In the same period, the average patient age for the operation can be expected to increase past the current mean of 65 years.² Although THA is a safe and highly effective procedure, it remains a major surgical operation. Because complications and mortality rates have been shown to increase with age across the spectrum of elective surgery, special care must be taken to meet the needs of an older, often frailer, patient population in the face of this rise in surgical volume.³

Many surgeons have encouraged the use of the direct anterior approach for this elderly population citing a quicker recovery, lower dislocation rates, less postoperative pain, reduced postoperative opioid use, and reduced hospital stays in all patient populations with the anterior muscle sparing approach that preserves the hip abductors and the posterior hip capsule.⁴ Unfortunately, this approach has also been linked with an elevated risk of atraumatic postoperative periprosthetic femur fractures.⁴ Due to osteopenia, older patients are more susceptible to periprosthetic fracture, both intraoperatively and postoperatively.³ PPFFs are thought to occur following approximately 4% of total hip arthroplasties.⁵ They impose both serious human costs and significant financial burden to health systems. These fractures tend to occur in the first 3–6 months after surgery and have been shown to decrease patients’ hip functionality and increase pain in the long-term, as well as to significantly increase patient mortality, with one study finding a one-year mortality rate of 17.7% following PPFF in THA.⁵ Furthermore, treatment of PPFF can be very costly, both to payers and providers. Mean revision arthroplasty hospital charges are approximately $34,000, despite resulting in net financial losses to hospitals.⁵ Between increased demand, expanding indications, and two major trends in THA, namely, the increased prevalence of both the direct anterior approach and the use of uncemented stems, the incidence of postoperative periprosthetic femoral fractures is projected to increase by an average 4.6% every decade over the next thirty years.¹ This prospect necessitates greater research into how surgeons might best implement prevention strategies, including the use of clinical algorithms, surgical techniques, and prosthetic designs.

Another important consideration revolves around the optimal fixation method for total hip arthroplasty. Joint registry data indicates that use of cementless femoral implants is increasing across all age groups, including the elderly.² Surgeon preference for uncemented implants stems from reduced operative times, reduced risk of cardiopulmonary collapse, less complex revisions following failure or infection, and difficulties experienced with cementing technique when using the increasingly popular direct anterior approach.⁶ Use of cementless stems also generally increases the likelihood of postoperative periprosthetic fracture, both intraoperatively and post-operatively.⁶ Trends toward the direct anterior approach and the use of cementless implants suggest increased prevalence of future PPFFs.

As such, more research is needed to determine both the possibility of reducing PPFF risk among uncemented, direct anterior THAs and the relative merit of using biologic fixation with this approach with older patients, if significant risk reduction proves impossible. One area in which practitioners may effectively mitigate the incidence of PPFFs is through implant selection. Registry studies have suggested that certain aspects of uncemented femoral component designs may be associated with differential risk for PPFFs, including stem geometry, presence of a calcar collar, and surface coating.⁷ The purpose of this study is to compare the prevalence of postoperative periprosthetic femur fractures in an older patient population between two uncemented femoral component designs used in the direct anterior approach. This study analyzes incidence of PPFFs between a collarless flat, single-taper, wedged femoral implant and a dual-taper, hydroxyapatite-coated implant which was used with a calcar whenever possible.

2. Methods

2.1. Study participants and exclusion criteria

The current study is an IRB-approved retrospective review of patients undergoing primary THA performed via a direct anterior approach between February 2015 and November 2020. All procedures were performed consecutively by a single, fellowship-trained adult reconstruction surgeon. Prior to the study period, the operating surgeon had performed over 800 direct anterior THAs. Patients excluded from the current analysis were those younger than 70 years old, as well as those who received THA with a posterior approach, or had femoral neck fractures, pre-existing proximal femoral hardware, or leg length discrepancy greater than 10 mm. Severe osteoporosis was assessed by the senior surgeon viewing preoperative AP and lateral hip radiographs, and patients diagnosed with severe osteoporosis were also excluded from the first cohort.

2.2. Group characteristics

Beginning in February 2015, the surgeon performed 361 consecutive primary direct anterior total hip replacements using a flat, single-taper, wedged femoral implant (TaperLoc; Zimmer-Biomet, Warsaw, IN; Fig. 1A and B). A standard length, reduced distal design was used in all cases. Both high offset and standard offset proximal geometries were utilized. Of these patients, 121 met the age criteria. 42 of these patients were then excluded due to posterior approach or the aforementioned medical reasons. In June 2016, that same surgeon, using the exact same surgical technique and postoperative weight-bearing protocol, began using a dual-taper, hydroxyapatite-coated implant (Origin, DJO Orthopedics, Austin, TX; Fig. 2A and B) for 1650 consecutive direct anterior THAs. Of these patients, 498 met exclusion criteria for age and the presence of a collar. Twenty of these patients were excluded due to posterior approach or the reasons mentioned above. It should be noted that indications for direct anterior THA were expanded under the use of the second stem since no patients were excluded due to osteoporosis, hence the higher number of exclusions in the first cohort. Standard offset, high offset, and coxa vara implants were utilized in order to match the patient's native proximal femoral geometry. A collared design was used whenever possible.

Fig. 1.

The figure shows (a) a Taperloc Stem and b) an X-ray of a Taperloc Stem.

Fig. 2.

The figure shows (a) Origin Stem and (b) an X-ray of an Origin Stem.

2.3. Description of procedure

A Hana table was used for all direct anterior hip operations, and femoral exposure involved the release of the anterior capsule initially. This was followed by femoral hyperextension and full external rotation using the femoral lift. Then, using electrocautery, the soft tissue attachments on the medial aspect of the greater trochanter (including the piriformis) were released, progressing from the anterior to the posterior corner of the greater trochanter while the hyperextended and fully externally rotated proximal femur was gently elevated using the lift. If exposure still remained inadequate, the obturator internus and the soft tissue attachments on the posterior border of the femoral neck were released. Fluoroscopy was utilized in all cases, and, other than the change in the femoral implant, no modifications were made to the surgical technique used for both cohorts. Flexible reaming was performed in selected cases due to a tight femoral diaphysis in both cohorts. Templating was completed by a single implant representative for the entire study, and no changes were made to the templating process.

2.4. Postoperative plan of care

The patients were carefully monitored for 3 months after surgery to identify the frequency of acute periprosthetic femur fractures. Both cohorts participated in a standard postoperative rehabilitation program, where they were out of bed on the day of surgery and were maintained with limited weight bearing on a walker for 2 weeks after surgery. After 2 weeks, patients were encouraged to progress to walking without assistive devices, according to their comfort. Most patients were maintained on aspirin for deep vein thrombosis prophylaxis, though all patients were risk stratified, and some received more aggressive anticoagulation. Patients were encouraged to wean pain medicines according to comfort.

2.5. Analysis

Patient charts were retrospectively reviewed for the current study in March 2021. Descriptive statistics were used to summarize the demographic data of the 2 cohorts. Risk factors examined in this study included age, gender, and body mass index (BMI). Welch's two-sample t-test was utilized to test for statistically significant differences in BMI and age, while a Fisher's exact test was used to compare gender distributions. A Fisher's exact test was then used to determine if the prevalence of periprosthetic femur fractures differed between the consecutive series of patients. An alpha-level of P < 0.05 was considered statistically significant for all statistical tests. All analyses were performed using R software.

3. Results

A total of 557 cases were included in the analysis from a series of 1650 total cases. 64.6% of THAs (n = 360) were performed in female patients, whereas 35.4% (n = 197) were in males. The mean age was 76.0 years (range: 70–96) and the average BMI was 29.1 kg/m2 (range 17–50.3). There were no statistically significant differences between the 2 cohorts for all demographic characteristics, and the characteristics stratified by implant type are shown in Table 1. Age did not vary significantly across implant groups with mean age being 75.9 and 76.0 years for the first and second cohort, respectively (P = 0.8621). Mean BMI also did not vary significantly between cohorts with the first cohort average BMI being 28.7, and the second cohort BMI average being 29.2 (P = 0.516). Finally, there were no statistically significant differences in the distribution of gender across the 2 cohorts (P = 0.2059). In the first group treated with a flat, single-tapered femoral implant, 2 (2.5%) patients sustained atraumatic periprosthetic femur fractures at an average of 19.5 days postoperatively (fractures occurred at 23 days and 16 days). Both patients sustained a Vancouver type B2 periprosthetic fracture, and both patients required femoral revision. In the second cohort treated with a dual-taper, hydroxyapatite-coated implant, no patients (0/478, 0%) sustained an atraumatic postoperative, periprosthetic fracture (P = 0.0199). The rate of fractures stratified by cohort is shown in Table 2. No patients in either cohort sustained greater trochanteric fractures that required intraoperative fixation or a return to the operating room. Patients in both cohorts who had nondisplaced calcar cracks intraoperatively were treated routinely with a single cable if deemed necessary. One patient in the first cohort had an intraoperative fracture at the index arthroplasty that extended below the lesser trochanter that was judged to be stable and was treated with limited weight bearing for 6 weeks. One patient in the second cohort had an intraoperative fracture during primary THA that extended below the lesser trochanter and were treated with open reduction and a single cable. Neither of these patients went on to have postoperative femur fractures or required additional surgery. All cases that met the exclusion criteria, totaling 61 cases, were done using the posterior approach. Of the 61 cases excluded from the study, 41 cases were excluded from the first cohort and 20 cases were excluded from the second cohort. An additional case was excluded from the second cohort involved an 82-year-old female who fell out of her bed on the same day of her operation while unsupervised in the hospital, sustaining a trauma-induced fracture that required revision.

Table 1.

Demographic data organized by femoral type.

Result	Taperloc, Zimmer-Biomet (Flat, single-taper, proximally porous titanium)	Origin, DJO (Dual-tapered, collared, extensively hydroxyapatite-coated)	P-Value
Total Number of Consecutive Cases	361	1650
Number of Patients Aged 70+	120	498
Number of Elderly Direct Anterior THA Patients for Analysis	79	478
Mean Age (years)	75.9	76	0.8621
Gender (%Female)	0.582278481	0.6569037657	0.2059
Mean Body Mass Index (kg/m2)	28.7	29.2	0.8621

Table 2.

Postoperative periprosthetic fracture incidence data organized by stem type.

Result	Taperloc, Zimmer-Biomet (Flat, single-taper, proximally porous titanium)	Origin, DJO (Dual-tapered, collared, extensively hydroxyapatite-coated)
Fracture	2	0
No Fracture	77	478
Postoperative PFF Rate (% Fractures)	0.0253164557	0

4. Discussion

The systemic challenge of providing total hip arthroplasty to an increasingly older patient population mandates clinical research into achieving optimal results for older patients. As noted above, both the direct anterior approach and the use of uncemented stems are increasing as preferred methods among surgeons. Many aspects of the direct anterior approach are particularly useful for treating an older demographic, especially minimal soft tissue damage, faster recovery time, and diminished dislocation risk. However, the increased risk of early postoperative periprosthetic fractures remains a significant issue with direct anterior THA especially when using cementless implants. This is particularly true for the elderly, where it has potentially devastating consequences.⁵ A number of studies indicate differing rates of complications using the direct anterior approach with different implant designs, such as variations in stem size, presence of a collar, or coating materials.^7,8

Our study focused specifically on atraumatic postoperative periprosthetic fractures following uncemented direct anterior THA in patients older than 70 years and found significant improvements in fracture rate following a switch to a dual-tapered, collared, extensively HA-coated stem (Origin, DJO Orthopedics, Austin, TX). Several studies have found similar results when comparing the PPFF rates between flat, single-tapered, proximally porous coated titanium and dual-tapered, collared, HA-coated stems designs in the general population, as well as when examining the PPFF rates of the various design variables individually.^7,8

There are numerous independent, yet interacting, variables involved in the design of an uncemented femoral stem. All designs rely on the principle of biologic fixation, wherein bone will grow onto or into the surface of a metal implant if it has a certain topography, either rough for osseointegration or porous for ingrowth, respectively. To achieve long-term stability in uncemented total hip arthroplasty through biologic fixation, an implant must have enough primary stability to maximize prosthesis-bone contact and resist micromotion.⁹ It is here that surgical technique and stem design greatly influence the likelihood of implant loosening and periprosthetic fracture. It is likely that a combination of design aspects contributed to the superior fracture-related outcomes observed when using the dual-tapered, collared, and HA-coated stem. First, a dual-tapered geometry, in which the stem narrows distally in both the medial-lateral and anterior-posterior planes, may confer greater initial stability than a flat single-tapered stem, which narrows only in the medial-lateral plane.⁸ Dual-tapered stems are designed to be wider in the anterior-posterior dimension of the femoral metaphysis, thereby increasing the circumferential contact between prosthesis and cortical bone, and have been associated with lower rates of fracture, loosening, and revision.⁸ In addition to maximizing cortical contact for osseointegration, Ma et al. further hypothesize that the abrupt mediolateral tapering found in single-tapered designs may contribute to fractures by resulting in increased shear forces along the medial aspect of the femoral metaphysis.⁸

Next, the presence of a collar appears to aid in achieving primary stability.¹⁰ By reducing early micromotion, collared stems are thought to facilitate rapid and thorough osseointegration. In a cadaver study, Demey et al. measured the average forces necessary to induce subsidence and fracture in collared and collarless stems following primary fixation, finding significantly higher force thresholds for both events in collared stems.¹¹ In a thirty-four study meta-analysis comprising 6825 patients in direct anterior THA, Panichkul et al. find lower postoperative PPFF incidence among patients who received collared stems, and attribute the improved outcomes to reduced subsidence, better rotational stability, and reduced propagation of intraoperative calcar fractures.¹²

Finally, it is likely that osteoconductive hydroxyapatite coatings accelerate the osseointegration process, resulting in faster, stronger secondary fixation.¹³ Extensive HA coating is thought to improve contact between prosthesis and bone.¹⁰ As hydroxyapatite is itself a critical component of bone matrix, it offers a natural, bioactive porous surface for thorough bone ingrowth. This effect may be especially pronounced in elderly patients, who tend to have decreased osteoblast activity and consequently lower natural rates of osseointegration. Ulivi et al. find excellent ten-year survivorships among elderly patients using extensively HA-coated stems.¹³

This study has certain limitations that should be considered. First, the time-dependent assignment criteria of the two patient cohorts may suggest the influence of surgical factors such as changes in technique or surgical skill. This is not the case, as the intraoperative technique and postoperative protocols remained consistent across the two cohorts, making the implant design the only variable across the two cohorts. Moreover, a comparison of the surgeon's prior experience with direct anterior THA with literature on learning curves suggests that he had thoroughly surpassed the curve. For example, De Steiger et al. found that the 100-case revision rates for direct anterior THA stabilize for surgeons after 50 procedures on average.¹⁴ At the most conservative end of estimates, Stone et al. found a learning curve of 400 cases.¹⁵ Here, the surgeon had already completed over 800 cases of direct anterior THA prior to this study, indicating that a learning curve can be dismissed as a potential source of bias.

However, switching from a flat, single-tapered, proximally porous coated femoral implant to a dual-tapered, collared, HA-coated design did require subtle changes in femoral preparation technique to comply with manufacturer's instructions. The initial implant utilized a bone-removing broach system while the second used bone impaction broaching. In general, broaching was less strenuous in the second cohort. While we doubt that this difference impacted results, we note that the broaching technique utilized is itself determined by the implant design employed by the surgeon. Nevertheless, slight differences in femoral preparation should be considered a limitation of this study.

Next, there is a large discrepancy between the sizes of the patient cohorts ultimately available for analysis, with 79 patients remaining in the first cohort after exclusion criteria were applied to an initial population of 361 patients. In the second cohort, 478 patients remained from a series of 1650. While an ideal study would have two large, relatively equally-sized cohorts, we believe our data still provide valuable insights into the relative performance of the two stems. Most notably, the series of 478 at-risk patients without atraumatic postoperative fractures (0.0%) in the dual-tapered, hydroxyapatite-coated, collared stem compares very favorably not only with this study's results using the other stem, but also with the notably high rates of PPFFs associated with direct anterior THA throughout the literature. The low fracture rate suggests that stem design may play an important role in optimizing uncemented direct anterior THA for use with older patients.

Additional limitations that should be noted include the retrospective, non-randomized nature of the study. The transition to a different implant design was motivated by clinical considerations by the surgeon, and future randomized trials will be necessary to verify these findings. There remains some potential for selection bias, as certain patients in both cohorts were deemed unsuitable candidates for direct anterior THA and were instead treated with a posterior approach and excluded from the study. These include rare cases where the surgeon operated on patients with femoral neck fractures, proximal femoral hardware, abnormal anatomy requiring greater than 10 mm of lengthening, or cases with severe osteoporosis. The posterior approach was used more often during the initial cohort, as indications for direct anterior THA were expanded during the second series and significant osteoporosis was no longer used as a reason for exclusion.

4.1. Conclusions

Despite its limitations, this study provides useful insight into the role of implant design in mitigating postoperative periprosthetic femoral fracture in the elderly population. As demographic transitions and expanding indications for total hip arthroplasty drive increases in THA volume among the elderly, solutions must be found that optimize clinical results specifically for older patients. Certain stem design factors may confer additional stability and better protect against postoperative periprosthetic femoral fracture, allowing for surgeons to provide the clinical advantages of direct anterior THA and cementless implants-including improved short-term outcomes, reduced theater time, and easier revisions-while diminishing the risk of increased PPFF. While the results of the current study should not be generalized to other implant designs or techniques, it appears that using a dual-tapered, collared, and extensively hydroxyapatite-coated femoral stem reduces the incidence of postoperative fracture in elderly patients compared to flat, collarless, proximally porous stems.

Ethical review statement

Each author certifies that the study presented in the submitted article was approved by an Institutional Review Board in accordance with the ethical standards in the 1964 Declaration of Helsinki, as well as the relevant regulations of the US Health Insurance Portability and Accountability Act (HIPAA).

Declaration of competing interest

None.