Introduction

Periprosthetic femur fractures are complex injuries that can be difficult to treat and recover from. With a growing number of total hip arthroplasties (THA) and revision arthroplasties being performed in an aging population, the incidence of these injuries is on the rise, topping 20% in some studies [1, 2]. Many risk factors for these injuries, which commonly happen after a fall, exist, including osteoporosis, osteolysis, female sex, and revision surgery [3, 4]. The combination of biological changes in bone density during aging combined with stress shielding of the implant at a high mechanical load point significantly increases the risk for periprosthetic fracture [5].

Periprosthetic femur fractures after hip arthroplasty have classically been classified using the Vancouver classification [6]. This classification, and ultimately treatment, is based on the site of fracture, the quality of the surrounding bone stock, and the stability of the pre-existing implant [7]. Treatment goals include early union, anatomical alignment, restoring length, a stable prosthesis, early mobilization, and a return to pre-injury functional status [8]. While each type of fracture has specific problems and treatment options [9,10,11,12,13], periprosthetic fractures in general can be challenging to manage due to their high complication rates and morbidity [14,15,16].

Multiple studies exist detailing outcomes that are associated with periprosthetic femur fractures after THA [10,11,12,13, 15, 16], but no study directly compares the post-operative course between fracture types as classified by the Vancouver classification system. This study compares the three Vancouver B fracture types to see if any type is associated with an increase in post-operative complications than others. The authors hope this information will help surgeons identify patients at higher risk for complications after surgically treated periprosthetic fractures and manage them accordingly.

Materials and methods

This retrospective chart review was conducted at a suburban orthopedic surgery department. Overall, 122 patients who presented to our hospital with periprosthetic proximal femur fractures after hip arthroplasty over the past 13 years were reviewed. Patients were included if they underwent surgical stabilization of their femur fracture. Patients were excluded if they sustained Vancouver A or C fracture types, underwent non-operative treatment, or had missing chart information.

For each patient, demographic information, fracture information, surgical information, and post-operative course were recorded. Each fracture was classified according to the Vancouver classification [6] by our senior author, a fellowship trained orthopedic trauma surgeon. We then compared union rate, infection rate, rate of re-fracture of femur, and rate of repeat surgery between Vancouver B classification fracture groups. Union was defined as radiographic healing of at least three cortices, seen using orthogonal femur films, by 9 months after surgery. Delayed union was defined as radiographic union after 9 months. Finally, for each patient, our state’s online drug monitoring program was reviewed. Post-operative opioid usage was recorded for each patient and compared between fracture groups. Data was cataloged using an electronic spreadsheet. Data between the two groups were compared using simple descriptive statistics and statistical significance tests (Chi-squared test and ANOVA test) as appropriate to compare post-operative outcomes between fracture groups. (IBM SPSS Version 23 Statistics for Windows, Armonk, NY: IBM Corp). For all analyses, p ≤ 0.05 denotes statistical significance.

Results

Overall, 88 patients who sustained Vancouver B periprosthetic proximal femur fractures and underwent surgical fixation were included in our study. Fifty-five (62.5%) were Vancouver type B1, 27 (30.7%) were Vancouver type B2, and 6 (6.8%) were Vancouver type B3. Most of our patients were female (n = 62, 70.5%) and older than 81 years of age (n = 53, 60.2%) with uncemented prosthesis (n = 83, 94.3%). Forty-eight (54.5%) of our patients had total hip arthroplasty implants in when they sustained their periprosthetic proximal femur fracture, and 40 (45.5%) had hemiarthroplasties (Table 1). Most patients sustained these fractures after low-energy falls (n = 80, 90.9%), while one patient sustained a fracture after a high energy injury (Vancouver B2). These fractures were sustained at an average of 4.68 years after an index hip replacement surgery. Most of our patients had their fractures stabilized using a plate and screw construct with cerclage wire (n = 57, 64.8%) or revision arthroplasty with cerclage wiring (n = 23, 26.1%) (Table 2).

Table 1 Demographic breakdown of our sample population
Full size table
Table 2 Types of surgical stabilization of periprosthetic fracture by Vancouver classification type
Full size table

When union rates were compared between the three subgroups, they were statistically similar to each other (p = 0.77), even though more than 90% of Vancouver B1 and B2 fractures achieved fracture union (Fig. 1). While Vancouver B2 fractures had the highest overall infection rate (n = 3, 11.1%), subsequent fractures after fixation (n = 3, 11.1%), and repeat surgery (n = 2, 7.4%), all three fracture subgroups were shown to be statistically similar in terms of infection rate (p = 0.32), subsequent fracture (p = 0.63), and repeat surgery (p = 0.64) (Fig. 2). Most of these repeat surgeries were due to infection or a subsequent periprosthetic femur fracture. Only one of these repeat surgeries happened in a patient with cemented prosthesis. Finally, when comparing post-operative opioid use (measured in milli-morphine equivalents) after surgical stabilization (Fig. 3), all three subgroups had statistically similar usages (p = 0.96).

Fig. 1
figure 1

Union rate after surgical fixation of periprosthetic fractures by Vancouver classification type

Full size image
Fig. 2
figure 2

Infections, subsequent fracture, and repeat surgery after surgical fixation of periprosthetic fracture by Vancouver classification type

Full size image
Fig. 3
figure 3

Post-operative opioid use after surgical stabilization of periprosthetic fractures by Vancouver classification type

Full size image

Discussion

Periprosthetic femur fractures present a challenge for surgeons to treat and patients to recover from. With the expected increase in the incidence of these injuries, patients and surgeons must be wary of the many complications associated with these injuries [14,15,16]. While various treatment options are utilized based on the Vancouver classification fracture type, each fracture and treatment option present its difficulties [6]. This study compares the post-operative course after surgical stabilization of three Vancouver B periprosthetic fracture type to see if there is a difference in the post-operative course.

Most of our fractures occurred in elderly females, both of which are known risk factors for periprosthetic fractures [3, 4]. Furthermore, most of our Vancouver B1 fractures were treated with a plate and screw construct, while our B2 and B3 fractures were treated with revision arthroplasty with or without plates and screws, both of which are the most common treatment options used for these fracture types [9,10,11,12,13, 17, 18]. Our results showed a statistically similar union rate between fracture types. While each fracture type presents its own challenges to stabilization, this is a testament to the use of proper surgical treatment for each specific fracture, keeping in mind surgical goals [8]. This also holds true for the statistically similar opioid consumption between groups. Still, these similarities, along with the statistically similar infection rate between groups, are an important distinction to make. Unlike many other classification systems, including the Garden, Winquist and Hansen, and AO/OTA classification systems in the native femur, where injury severity usually worsens with each subsequent type, each Vancouver fracture type seems to have an independent identity in comparison to the other types with similar post-operative courses.

A previous study by Lindahl et al. in 2005 noted a higher failure rate after open reduction and internal fixation (ORIF) compared to revision arthroplasty [3]. Cohen et al. echoed this finding 13 years later, finding less mortality in patients with periprosthetic femur fractures treated with revision arthroplasty than those treated with open reduction and internal fixation (ORIF) [19]. However, these treatment types are used to stabilize more than one Vancouver fracture type and cannot be extrapolated to an outcome comparison of the different fracture types. Our results suggesting that there are no differences in the post-operative course between fracture types are therefore not contradictory, but supplementary to these previous results. While complications are high after periprosthetic fractures in general, no Vancouver B fracture type has a lower union rate, infection rate, subsequent fracture rate, repeat surgery rate, or higher opioid consumption than other types. Finally, while we did not compare the distribution of Vancouver fractures between cemented and uncemented implants or the complication rate between the two types of implants after surgical stabilization of periprosthetic fractures due to low numbers, it is important to note the documented rate in the literature of periprosthetic fractures was lower in cemented hip arthroplasties [20, 21], a finding that has impacted surgeon decision making in hopes of decreasing rates of all Vancouver fractures.

The weaknesses of our study are inherent to most retrospective chart reviews. Our sample cohort is considered a convenience sample, which makes our results subject to bias. While our sample number is low, it compares to previous studies on periprosthetic fractures. Because of this, no power analysis was done; rather, all periprosthetic femur fractures after hip arthroplasty were reviewed. Operative Vancouver A and C fractures are especially low because of their relative rarity compared to Vancouver B fractures, and because of this they were not included in this comparison. Results are also subject to the accuracy of physician charting. Finally, this study involved patients from only one geographic area. Future studies should aim to include multiple centers from multiple locations to include a large sample size with a variety of patients and surgeons.

Conclusion

While Vancouver B periprosthetic femur fractures after hip arthroplasty are associated with high complication rates and poor outcomes, there is no difference in union rate, infection rate, subsequent fractures, repeat surgery rate, and opioid usage between the different types of fractures. Surgeons should continue to treat each injury appropriately, as dictated by the fracture type, bone stock, and implant stability while being vigilant for complications post-operatively.