1 Introduction

Interprosthetic femur fractures (IFFs) are specific femur fractures when there exist two prostheses on either side of the femur, as is the case with an ipsilateral total hip arthroplasty (THA) and total knee arthroplasty (TKA), additional stresses are placed on the femur which can result in the femur being more susceptible to fracture. This fracture in between a THA and TKA is termed an interprosthetic femur fracture¹ and studies have reported incidence rates of 1.25% and increasing.² IFFs are particularly difficult to treat and manage. Since they occur in between two prostheses, often, bone space is inadequate for fixation of the femur. In addition, blood supply is usually insufficient between implants causing bone quality to deteriorate. Patients that suffer from interprosthetic femur fractures are typically geriatric patients and have pre-existing co-morbidities such as osteopenia, osteoporosis, etc. that contribute to poor bone quality as well. The many methods of treatment of IFFs also makes them difficult to manage as there is no consensus on the best way to fixate IFFs.

Currently, there are several common methods for treating IFFs. These include open reduction internal fixation with the use of locking plate with or without cerclage wires or cables, prosthetic revision, nail/rod, Ilizarov external fixator (IEF), and other lesser used methods.^2,3 The method used in IFF treatment is dependent on a number of factors and tailored to the patient specifically.⁴ Fracture type and location, prosthetic type and stability, and remaining bone quality and quantity are all important considerations when deciding on treatment methods. Patient characteristics such as age, expectations, and activity level may also influence decision making.⁵ Utilization of a locking plate to stabilize a femur fracture is a common surgical treatment used when bone quality is adequate. Locking plates help to distribute stresses evenly across the femur and screws can be used for fastening of plates to bone. The types of screws most often used with locking plates are either unicortical or bicortical screws. Often, cerclage wires or cables can be supplemented for additional stabilization. Another treatment method is revision of prosthetics whether that be of the hip prosthetic, knee prosthetic, or both. Revision prostheses are often used when the patient's bone stock is inadequate for other treatment methods. Prosthesis revisions are often performed in the presence of implant loosening and usually with a longer stem to better stabilize the femur.⁶ When there is significant bone loss or when revision of hip and knee prostheses is not possible, the surgeon may opt to replace the entire femur. Total femur replacements are one option in this treatment category, in which the entire fractured femur is replaced with a prosthetic, though this is very rare. Interprosthetic femoral sleeves are another option in which a metal tube-like structure is placed in between the hip and knee prostheses, connecting the two and serving as a femur replacement.⁷ Intramedullary nails or rods are another treatment to fixate the fractured femur, but only when the primary knee replacement lacks a long stem. This treatment method utilizes a metal rod which is inserted into the fractured bone, across the fracture, to stabilize the femoral shaft.⁸ Herein we have described only surgical internal fixation methods which are the primary form of treatment for IFFs. However, a lesser used method, known as Ilizarov external fixator (IEF), has shown to have some promising results as well.^9–11 Instead of fixating and stabilizing the femur through contact with the bone, a metal ringed frame is positioned around the femur externally and applied to the femur through wires and/or pins inserted into the leg. This method is relatively minimally invasive and can achieve reliable fixation. Limitations with this approach include the bulkiness and inconvenience to the patient.^12–16

Although there exists a myriad of approaches to treating IFFs, there is currently no systematic reviews of the consensus on which method is the best. Current systematic reviews do not compare the many different treatment methods for IFFs specifically. The purpose of this review was to compare results of the most used methods to attempt to find out which approach is relatively advantageous in terms of healing time, less revisions, less complications, and better postoperative functional outcomes.

2 Methods

A comprehensive systematic literature review was conducted by two study team members following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) protocol. The research databases searched included PubMed (MEDLINE), Cochrane Central Register of Controlled Trials, and Embase. Reviewers used an assortment of combinations of keywords: “hip,” “knee,” “total,” “arthroplasty,” “replacement,” “ipsilateral,” “interprosthetic,” “femur,” “femoral,” “fracture,” and “periprosthetic” to identify studies published from inception of these databases through June 2021. Studies that involved the treatment of interprosthetic femoral fractures and included any outcome measurement or reported outcomes were included in this review. Case reports for treatment of IFFs specifically were also included. Studies involving periprosthetic fractures in which the outcome measurements were not able to be differentiated from IFF outcomes were excluded. In vitro and biomechanical studies as well as research articles not available in English were also excluded.

The reference lists of identified articles were reviewed to ensure there were no missed articles relating to IFF treatment and outcomes. The titles, abstracts, and full texts were screened and any discrepancies in included studies were discussed amongst team members with the senior author (JL) ultimately deciding on inclusion of studies.

Data collection was conducted and performed for all included studies. Extracted data were recorded and stored in a Microsoft Excel spreadsheet customized by the reviewers. Study data collected included author name, study title, year published, number of patients involved, and patient demographics (eg, age, sex). When reported, preoperative data of relevant diagnoses in hip and knee, mean time between THA/TKA and IFF, and fracture type were recorded. Intraoperative data were also recorded and included treatment method (eg, locking plate, plate with cerclage wires/cables, prosthetic revision, plate + revision, intramedullary nails/rods, plate + nails/rods, femur replacement, external fixator) and whether bone graft was utilized. If treatment method involved prosthetic revision, revision type (eg, proximal, distal, both) was recorded. If the femur was replaced for IFF treatment, the type of replacement (eg, TFR, interprosthetic femoral sleeves) was recorded. When reported, postoperative data (eg, number of unions, revisions, delayed unions, non-unions, mal-unions, hardware failures, infections, healing time, other complications, function scores) were recorded. Healing time for the purposes of this study referred to the duration of time needed to achieve union. Whereas union referred to the completed healing of the fracture. Subjective outcome scores included postoperative patient reported outcome measurements of Harris Hip Scores, Knee Injury and Osteoarthritis Outcome Scores, and Parker Scores. In addition, any additional information that reviewers felt relevant were recorded for future reference.

Mean values and corresponding standard deviation for age, time between THA/TKA and IFF, healing time, and functions scores were calculated and compared amongst treatment groups. Healing time was compared amongst treatment groups using independent ANOVA tests. Absolute values were calculated for non-unions, mal-unions, delayed unions, and other postoperative data. Infection rates and other complication rates for each treatment method were calculated for and compared using chi-squared tests. Statistical significance was considered with a standard p value of 0.05. Statistical analyses were performed in Microsoft Excel via Analysis ToolPak.

3 Results

Upon initial literature search of PubMed (MEDLINE), Cochrane, and Embase databases, 124 articles were identified for screening. After initial screening of titles and abstracts, 57 articles were excluded, and 67 articles were found to be relevant to our systematic review topic. Full-text articles were retrieved, and an additional 27 articles were excluded for reasons detailed in Fig. 1. The final assessment yielded 40 articles that were included in our systematic review for data collection. After manual searching of reference lists of identified articles, no additional articles were found to be included in the systematic review. Breakdown of the screening and inclusion/exclusion process is depicted in Fig. 1. The general information of the included studies is shown at Table 1.

Close

Fig. 1 Systematic Review Flow Chart: Shows the screening process, removing of papers based upon eligibility, and inclusion criteria to get final number of papers.

Export to PPT

Included in the 40 articles for our systematic review were 40 retrospective case reports/cohort studies, and no prospective case series/cohort studies. The 40 studies in this systematic review included 508 patients that underwent surgical treatment for interprosthetic femoral fractures specifically. Average reported age of patients was 78.7 years old and 403 (79.3%) were females, 90 (17.7%) were males, while 15 (3.0%) were unspecified. In the plate, prosthetic revision, femur replacement, nail/rod, and external fixator treatment groups there were 391 (77.0%), 57 (11.2%), 28 (5.5%), 11 (2.2%), 4 (0.8%) patients, respectively. Out of the plate group were 191 (48.8%) patients who also required cerclage wires/cables for fracture fixation.

Overall, mean time between total hip arthroplasty and interprosthetic femur fracture was 8.54 years and mean time between total knee arthroplasty and IFF was 8.19 years. In the plate, revision, replacement, nail/rod, and external fixator groups mean time between THA and IFF was 8.02, 7.36, 12.93, 6.73, and 11.13 years, respectively. Mean time between TKA and IFF was 7.39, 7.62, 12.28, 7.00, and 8.90 years, respectively.

Overall union rate was 74.0% with 376 of 508 patients achieving fracture union after primary treatment of IFF. Only 271 patients had reported healing times of fractures with a mean of 5.15 months. The plate, nail/rod, external fixator, plate + nail/rod, and plate + revision groups had mean healing times of 4.69, 6.50, 5.10, 7.69, and 6.00 months, respectively. ANOVA single factor comparison of healing times between treatment groups yielded a p-value of 0.65. The replacement treatment group did not have an average fracture healing time due to the nature of the treatment involving removal of fractured bone and replacement with a prosthetic femur. Revision rates were highest in the femur replacement treatment group with 9 (32.1%) patients needing at least one reoperation surgery for any reason.

Overall, hardware failure and non-unions were the most reported complications in treatment of IFFs. There were 53 (10.4%) hardware failures and 32 (6.3%) non-unions reported. Twenty-six (49.1%) hardware failures belonged in the plate group, 19 (35.8%) in the revision group, 5 (9.4%) in the replacement group, 2 (3.8%) in the nail/rod group, and none in the external fixator group. Twenty-one (65.6%) non-unions reported involved the plate treatment group, 7 (21.9%) involved the revision group, 2 (6.3%) involved the nail/rod, and none was found in the external fixator group.

Other reported complications included infections, mal-unions, and delayed unions of which there were 18 (3.5%), 3 (0.6%), and 2 (0.4%), respectively. Six (33.3%) of the 18 infections were from the plate group and 3 (16.7%) were from the replacement treatment group. There was also 1 (5.6%) infection case found in each of the nail/rod and external fixator groups. All 3 of the mal-unions and both delayed and union cases occurred in patients who were treated with a plate. In this review, a total of 31 patients across all studies and treatment groups were reported to have died at some point postoperatively. Many of these cases were due to causes unrelated to treatment surgery. Twenty-four patients were from the plate treatment group, with 16 due to unspecified reasons, five reported as unrelated to surgery, and three due to cardiac complications related to surgery. Four of the patients that died postoperatively were in the femur replacement group, with two due to causes unrelated to surgery, one due to prosthetic joint infection (PJI) of affected extremity, and another due to persistent sepsis after a hip disarticulation. Two of the patients that underwent a prosthetic revision died postoperatively, one from unspecified reasons and another from severe COPD and concurrent pneumonia. Lastly, one patient in the external fixator group died due to complications related to femoral artery damage during surgery. This data is included in Table 2.

Rates of unions, revisions, hardware failures, non-unions, infections, mal-unions, and delayed unions were compared amongst the different treatment groups using the chi-square test. No statistical difference was found between treatment groups for rates of unions, non-unions, mal-unions, and delayed unions. However, differences between treatment groups for rates of revisions, hardware failures, and infections were found to be statistically significant. P-value was calculated at 0.000020, 0.0000018, and 0.00075 for revisions, hardware failures, and infections, respectively.

Postoperative functional outcome scores were available for 242 patients and included Harris Hip Scores, Knee Injury and Osteoarthritis Outcome Scores, Parker Scores, and Katz Scores. Overall, the average postoperative HHS, KSKS, KSFS, KOOS, Parker, and Katz scores across all treatment groups were 76.15, 81.36, 67.5, 71.8, 4.04, and 2.98. Harris Hip Scores and Knee Society Knee Scores were the function scoring systems that had the most patient data available. HHS for the plate, prosthetic revision, replacement, nail/rod, and plate + revision groups were 76.84, 77.14, 69.9, 77, and 78.4, respectively. The external fixator did not have any reported HHS scores. KSS for the plate, revision, replacement, and nail/rod groups were 87.14, 88, 42.5, and 88, respectively. The external fixator and plate + revision groups did not include any KSS data. Detailed scores for each treatment group are in Table 3 below.

4 Discussion

The management and treatment of interprosthetic femoral fractures is a challenging one. Differences in fracture types, bone quality, and treatment methods all contribute to numerous possible complications when treating IFFs. Furthermore, patients experience increased risk of additional surgery, mortality rates,⁴ and medical costs from IFFs.⁵⁷ While there have been many independent studies and case reports on the surgical treatments of IFFs, there exists no reported gold standard for IFF management and treatment. Current methods of treatment are largely up to the orthopedic surgeon's preference and available literature lacks data driven conclusions on which treatments are best. In this systematic review we have compared the surgical outcomes from 5 different IFF treatment methods across 40 published articles.

Healing time of fractures was one method in which different treatment groups were compared amongst each other. The plate treatment group had the shortest average healing time of 4.69 months while the plate + nail/rod group had the longest average healing time of 7.69 months. ANOVA single factor comparison revealed no significant difference between the healing times of different treatment groups with a calculated p-value of 0.65. Chi-square tests ran with rates of unions, non-unions, mal-unions, and delayed unions between treatment groups also revealed no significant differences. P-values for these categories were 0.68, 0.13, 0.99, and 0.99, respectively. However, when it came to the rates of revisions, hardware failures, and infections, chi-square tests calculated p-values at 0.000020, 0.0000018, and 0.00075, respectively. These results show that the rates of these particular complications occurred less frequently than expected for the plate treatment group. It is important to consider, however, that the sample size for the other treatment groups were relatively smaller compared to the plate group and that may have skewed results.

Function scores were compared amongst treatment groups as well. The femur replacement group had the lowest HHS and KSS scores of 69.9 and 42.5, respectively. Revision and nail/rod groups each had the highest KSS score of 88. The plate group had the only reported Knee Society Function Score of 67.5. KOOS score was only found available from one study (insert study) and reported as 71.8 for the plate, revision, and plate + revision groups. Parker score was highest in the femur replacement group at 5 and lowest in the external fixator group at 3.37. Katz score was also only found in one study reported as 2.98 for the plate, revision, and external fixator groups. The different treatment groups were further classified under the bony union attempt group and revision arthroplasty group. Patients treated with prosthetic revision or total femur replacement were placed under the revision arthroplasty group. The rest of the patients with plate, nail/rod, or external fixator treatments were classified as the bony union attempt group because all of these methodologies involve some sort of attempt to achieve bony union postoperatively. Table 4 shows the results of unions, complications, and deaths from these different group classifications while Table 5 compares the function scores of the two groups.

Limitations to our study include the nature in which many of the studies used in this systematic review were conducted. Majority of the papers were retrospective and in addition, 13 of them were case reports. Many of these studies were relatively small in terms of numbers of patients reported and the sample size for some of our treatment groups were also limited. In addition, there were not many common controls between studies. When reporting on average function scores and healing times, some of the treatment groups have very few patients or no patients with specific function scores or healing times to report. The oldest study included in this review was the Dave et al. article published in February of 1995 while the most recent article was authored by Tibbo et al. and published in July of 2021. As the treatment of interprosthetic femoral fracture is a constantly evolving subject, there is a limitation of selection bias as newer and different treatment options become available over the years. This is another limitation to the study that was identified. The many discrepancies among scoring systems and methodologies given by the authors in the studies is also a limiting factor when attempting to compare outcomes across different treatment groups.

Almost all patients with IFFs were elderly patients. The female patients accounted for about 80%. The poor bone quality of elderly, and especially female, patients contributes to many of the intra- and postoperative complications that occur when treating interprosthetic femoral fractures. As previously mentioned, hardware failure and non-unions were the most reported complications in our review. Hardware failures can occur due to a combination of reasons not limited to the poor bone quality of elderly individuals, biomechanical stresses, and infections of hardware. Non-unions may result from hardware failures or other factors such as infections, mechanical, biological, metabolic, and patient-related factors. Complications relating to surgery can result in worse outcomes and may warrant further surgeries all of which increase mortality rate and decrease quality of life for patients. To decrease the rate of these complications and improve patient outcomes, it is imperative to understand the many different treatment options for IFFs and how the associated complications.⁵⁸

In conclusion, each treatment method should be carefully considered by the surgeon depending on the patient. Locking plate was the most common method for the treatment of the patients with IFFs. Half of them combined with cerclage wires/cables. Around two thirds’ patients could achieve union with the fastest mean healing time around 4.69 months. Other less common methods included prosthetic revision, femur replacement, nail/rod, external fixator, etc. A small number of patients treated with Ilizarov external fixator, and it has proven to be a viable option with few complications and high union rates. Almost all treatment methods could achieve around 70 of the Harris Hip Scores eventually.

Ethical approval

There is no need for IRB approval for literature review in our institution.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author contributions

Conception and design: JL & NE; Administrative support: NE; Provision of study materials or patients: All authors; Collection and assembly of data: BR & JL; Data analysis and interpretation: BR, PS, MT & JL; Manuscript writing: BR, PS, MT, & JL; Final approval of manuscript: All authors.

Data availability

The data will be shared if requested within one year after publication.