Intertrochanteric (IT) hip fractures are a common injury in the elderly population, and the incidence of IT fractures is increasing with the aging population^1,2,3). Operative fixation, most often using a cephalomedullary nail, is standard of care for this injury^4,5). Despite its relatively rarity, failure of a cephalomedullary nail can have significant effects on the patient. As the incidence of IT fractures continues to rise, a comprehensive understanding of implant failure as well as effective management are important.

Two distinct mechanisms, varus collapse of the head with an antero-superior cut-out (CO) of the implant’s lag screw or medial migration of the lag screw or cut-through (CT), are most often responsible for failure of a cephalomedullary nail^{6,7,8,9,10,11,12)}. While many studies examining the factors associated with the occurrence of implant failure have been reported, data regarding which factors may be predictive of one type of failure mechanism over the other are limited. Therefore, the purpose of this study is to compare clinical data from patients treated with a single lag screw design intramedullary nail in whom fixation failure occurred via the CT mechanism versus CO out in order to determine patient factors, injury details, reduction quality, or radiographic parameters that may be predictive of each mechanism.

This retrospective chart review study involving human participants was in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The Human Investigation Committee (IRB) of NYU Langone approved this study (approval No. i20-01766). Due to the approved IRB, informed consent was not required as study was retrospective in nature and information was anonymized.

In this retrospective cohort study, a database on hip fractures which was approved by an IRB was queried for a consecutive series of patients who presented to a single academic medical center (that includes four hospitals) with an IT hip fracture (AO/OTA classification 31A1, 31A2, 31A3) from 2014 to 2022. Inclusion criteria for this analysis included any patient who presented with an IT hip fracture during the study period and received treatment with a cephalomedullary nail, including both long and short options. Exclusion criteria included patients treated with a dynamic sliding hip screw and those who died within three months of surgery. The records for each patient were reviewed for demographic features and radiographic parameters. Demographic features included age, body mass index, comorbidity profile as compiled by the Charlson comorbidity index (CCI), baseline ambulatory status, osteoporosis diagnosis (as determined by previous dual energy X-ray absorptiometry [DEXA] scan), and smoking status at the time of injury (current vs. not current). Additional recorded clinical data included the respective time from surgery to fixation failure, the Score for Trauma Triage in the Geriatric and Middle Aged (STTGMA), a validated tool for assessment of orthopedic trauma risk, and the Fracture Risk Assessment Tool (FRAX) tool for each patient^{13,14,15,16,17)}. The STTGMA score utilizes various clinical variables including demographic data (age, comorbidities), injury details (AIS [Abbreviated Injury Scale] for the head/neck and chest), and baseline functional status for determination of a mortality risk score. Previous work by Konda et al.¹⁶⁾ was used to determine risk cohorts for each patient’s respective STTGMA score with the following breakdown: high risk, >9.0%; moderate risk, 3.4-8.2%; low risk, 0.8-3.4%; and minimal risk, <0.8%. The FRAX tool utilizes clinical variables including age, gender, height, weight, history of previous fracture, smoking, alcohol or glucocorticoid use, the presence of osteoporosis, and others to determine a risk score for the occurrence of future fracture. A FRAX score >20% was used as a cutoff for recommended medical therapy for intervention based on the US-based World Health Organization algorithm.

Review of radiographs included films of preoperative injury of the hip and pelvis for assessment of anteroposterior (AP) and traction/internal rotation¹⁸⁾. A review of intraoperative fluoroscopy was performed for both AP and lateral views. All fractures were classified according to the system of the OTA¹⁹⁾. Follow-up radiographs were reviewed to determine the mechanism of failure, and classified as either lag screw CO or medial migration/CT.

Fracture reduction and radiographic parameters of the implant were recorded. All thickness, length, or angular measurements were performed using PACS tools (ruler, goniometer PACS tool; Siemens). Measurement of the tip-apex distance (TAD) was performed according to measurements reported by Baumgaertner et al.²⁰⁾. Measurements of the lateral femoral wall were performed using the method described by Palm et al.²¹⁾ and Hsu et al.²²⁾. Assessment of reduction quality was performed using post-fixation angulation and cortical translation as measured on intraoperative fluoroscopy for comparison of any differences.

Of the 1,232 IT fractures treated during the study period, fixation failure occurred at an average of three months or 90 days from the initial surgery in 18 patients with 18 fractures (1.5%). Treatment with short cephalomedullary nails was administered in 17 patients, and one patient received treatment with a long cephalomedullary nail. Failed implants included: 17 Stryker Gamma nails (Stryker) and one Zimmer Biomet Nail (Zimmer). The average age of patients in the failure cohort was 77.67±10.35 years, younger than those in the non-failure cohort (P=0.168). The failure cohort included six males (33.3%) and 12 females (66.7%). Medial migration of the lag screw was detected in 10 patients (55.6%) (Fig. 1) while lag screw CO was detected in eight patients (44.4%) (Fig. 2). No difference in demographic parameters was observed between the CO, CT, and non-failure cohorts including injury mechanism (96% of patients had a low energy injury mechanism), fracture classification, percentage of patients in each respective STTGMA score risk cohort (P=0.540) or FRAX score >20% (P=0.647). The highest CCI was observed in patients in the CT cohort, demonstrating a larger comorbidity profile compared to the non-failure cohort (P=0.002). The highest rate of DEXA confirmed osteoporosis (25.0% vs. 60.0% vs. 11.4%) was observed in the CT group, which was not significant compared to the CO group (P=0.138) but significantly higher compared with the non-failure cohort with P<0.001 (Table 1).

In addition, no differences in radiographic parameters were observed between the CO, CT, and non-failure cohorts (Table 2), including but not limited to TADs; only one patient in each failure cohort had a TAD >25 mm.

The findings of this study demonstrated an association of both mechanisms of screw failure with similar clinical data including patient demographics, injury parameters, and radiographic data. It is more likely that the main predictive factor for patients in whom early failure occurs via the mechanism of medial migration is related to bone quality as described in previous biomechanical studies.

Although all patients who underwent treatment had some form of osteoporosis, no significant difference in the rate of osteoporosis diagnosis via DEXA scan and FRAX score >20% was observed between the CT and CO cohorts, and the CT cohort included a much higher percentage of patients with a diagnosis of osteoporosis at the time of injury. This was particularly notable in the comparison of the CT and non-failure cohorts. The mechanism of medial migration is thought to mimick the “Z effect” mechanism observed with two screw design nails due to relative differences in bone quality in the femoral head. According to a biomechanical study reported by Strauss et al.²³⁾, poor bone quality showed strong correlation with failure via a mechanism similar to the Z effect. Multiple causes for this incidence of the single screw Z effect were hypothesized in a case study on intrapelvic migration of the lag screw that included a subsequent review of possible mechanisms reported by Flint et al.⁷⁾. Many of these causes were related to bone quality, notably persistent instability of the fracture, avascular neecrosis of the femoral head, and detection of osteoporotic bone⁷⁾. Patients in both the CT and CO cohorts were significantly younger than those in the non-failure cohort, thus both failure mechanisms may be invovled in older patients with worse bone quality. Therefore, a patient’s bone quality is likely a factor in the mechanism of failure that may occur; however, conduct of a larger study will be required in order to provide conclusive evidence.

In addition, comparison of the CT and CO cohorts found no difference in various radiographic parameters from the preoperative and intraoperative time periods. Findings from many studies have demonstrated that specific radiographic parameters, such as TAD, show high correlation with failure rate. A consistent approach to measurement of the TAD of sliding hip screw implants utilizing AP and lateral radiographs of the hip was developed by Baumgaertner et al.²⁰⁾. In their retrospective study, out of an overall cohort of 193 patients, there were 19 cases of fixation failure (9.8%). A strong statistical relationship was observed between increasing TAD and the rate of screw CO. According to another comparative study reported two years later by Baumgaertner and Solberg²⁴⁾, use of good surgical technique with an emphasis on minimizing TAD resulted in lower rates of screw CO and implant failure for both sliding hip screws and intramedullary devices. A biomechanical study conducted by Kuzyk et al.²⁵⁾ reported that an inferior placement of a lag screw showed the most axial and torsional stiffness, resulting in the highest load to failure. This was supported by the results of a computational analysis by Quental et al.²⁶⁾ which demonstrated that an inferior and deep placement of the screw would provide the greatest resistance to CO in treatment using a proximal femoral nail implant. A review article by Socci et al.⁵⁾ discussing recommendations for management of IT fractures noted the importance of proper reduction and minimizing TAD for stability and fixation in treatment with either a sliding hip screw or intramedullary device. Therefore, while the original hypothesis asserted that treatment with an intramedullary nail rather than an extramedullary device would have less impact on the failure rate, multiple studies have emphasized its importance. A retrospective study by John et al.²⁷⁾ demonstrated that a high TAD would predispose nails to higher rates of CO. A radiographic review by Stern et al.²⁸⁾ reported an association of a high TAD with higher rates of failure for both helical blade and lag screw intramedullary nails. Similarly, the quality of reduction, which can be approximated using intraoperative radiographic parameters such as post-fixation angulation and cortical translation has shown high correlation with failure rates. Turgut et al.²⁹⁾ reported that poor reduction, notably varus, was a strong predictor of failure when using the helical blade nail design. Of note, 16/18 of the cases of failure included in this study had an acceptable TAD <25 mm, suggesting that factors other than a TAD greater than 25 mm are predictive of implant failure.

This study has several limitations. First, the retrospective nature presents inherent biases commonly associated with this type of study. However, selection of patients was based on the occurrence of implant failure, thus it is likely that selection bias is minimal. Another challenge was the group of patients who were lost to follow-up, or those who may have experienced failure and presented for follow-up elsewhere, neither of whom were included in the analysis. Second, this analysis only included patients who were treated with intramedullary nail fixation using a single lag screw nail design; therefore, factors associated with each mechanism of failure cannot be predicted in cases where fixation is provided using alternative implants such as a helical blade nail or sliding hip screw. Third, due to the small failure cohort, this study may be underpowered, so that significant differences in demographic and radiographic factors associated with each mechanism of failure cannot be distinguished. Specifically, despite our suspicion that bone quality would be a significant factor in predicting the mechanism of failure, with such a small sample size, it may be that our study is simply underpowered so that the significance of this difference in bone quality cannot be determined. Future study may include multi-center collaboration in order to increase the size of the cohort.

As mechanisms of failure, neither migration of the medial lag screw through the femoral head nor screw CO showed an association with identifiable patient demographics, initial fracture reduction, or radiographic implant parameters. As reported in previous biomechanical studies, the main predictive factor for patients in whom early failure might occur via the CT mechanism may still be related to bone quality; however, conduct of larger studies will be required in order to determine whether there is truly a difference in bone quality between the mechanisms of failure.

(A) Anteroposterior (AP) injury film demonstrating a right intertrochanteric hip fracture (AO/OTA 31A1.2). (B) AP and (C) lateral intraoperative fluoroscopy demonstrating a right intertrochanteric hip fracture treated with an intramedullary nail. (D) AP radiograph at follow-up demonstrating implant failure via the cut-through mechanism with medial intrapelvic migration.

(A) Anteroposterior (AP) injury film demonstrating a right intertrochanteric hip fracture (AO/OTA 31A2.3). (B) AP and (C) lateral intraoperative fluoroscopy demonstrating a right intertrochanteric hip fracture treated with an intramedullary nail. (D) AP radiograph at follow-up demonstrating implant failure via the lag screw cut-out mechanism.