Incidence and factors influencing neck shortening after screw fixation of femoral neck fractures with the femoral neck system

In this study, 113 patients with fresh femoral neck fractures met the inclusion criteria, of which 25 did not complete the follow-up and 1 had a re-fracture due to a postoperative fall, for a total of 87 patients who completed more than 12 months of follow-up. There were 49 male cases and 38 female cases; 46 left-sided cases and 41 right-sided cases. The fractures were classified by Garden and Pauwels: 23 cases of Garden I, 13 cases of Garden II, 12 cases of Garden III, and 39 cases of Garden IV; Pauwels I: 8 cases; Pauwels II: 30 cases; and Pauwels III: 49 cases.

For imaging data, 2 professionally trained orthopedic surgeons measured the imaging data separately, and any disputes were resolved by discussion and negotiation; the patient's radiographs were used to assess fracture staging (Garden Typing, Pauwels Typing).

This study complied with the Declaration of Helsinki as revised in 2013 and was approved by the Medical Ethics Committee of Fuzhou Second Hospital, Xiamen University. All patients gave informed consent and signed the informed consent form.

All patients were placed under general anesthesia with tracheal intubation in the supine “lithotomy” position with sterile towels. At the proximal end of the femur, 2 kerf pins are prepositioned laterally along the axis of the femoral neck, not exceeding the fracture line, according to the preoperative plan, a threaded bone traction nail can be inserted approximately 2–3 cm below the lateral apex of the greater trochanter for anterior and posterior positions that require correction of displacement, combined with longitudinal traction of the lower extremity, in the presence of femoral rotation, 2 kerf pins were placed in the femoral head to reset the femoral head rotation. After repositioning, anterior–posterior and table-piercing radiographs were taken using the intraoperative C-arm, and the position met the repositioning requirements and was fixed by inserting two preplaced kerf pins. A guide pin was inserted under fluoroscopy, parallel to the axis of the femoral neck, and adjusted so that it was located approximately 10% inferior to the mid-axis of the femoral neck in the anterior–posterior position and on the mid-axis of the femoral neck in the lateral position, and the corresponding length of FNS was selected for fixation after measurement, and the wound was irrigated and sutured.

Cefazolin sodium was given within 24 h after surgery to prevent infection, and low-molecular heparin calcium was injected on the first day after surgery to prevent deep vein thrombosis. Patients were instructed to perform foot flexion and extension exercises in a non-weight-bearing position, and all patients were kept in a non-weight-bearing position for at least 4 weeks after surgery. In patients with progressive growth of bone scabs on imaging observed at follow-up, patients were instructed to be partially weight-bearing, weighing approximately 20–30 kg, with the patient's weight-bearing status increasing at what rate determined by 2 physicians, with regular postoperative outpatient review at 1, 2, 3, 6, and 12 months.

The hip Harris score was used to assess the functional outcome of the affected limb during the follow-up of all patients included in the study. It includes four components: pain level, functional status, deformity, and range of motion. The total Harris score is 100, and the better the recovery of the affected limb after surgery, the higher the score.

These include non-healing fractures, femoral head necrosis, femoral neck shortening, and failure of internal fixation. Fracture healing is defined as a bone scab bridging the fracture end or a bone trabecula across the fracture end within 6 months, and vice versa is considered non-healing [10]. Femoral neck shortening was defined as shortening of ≥ 5 mm on final follow-up radiographs compared with the immediate postoperative period of the fracture [11, 12]. Femoral head necrosis: assessment based on Ficat and Arlet's radiological criteria [13]. Internal fixation device disorders include main nail cut out, fracture, screw loosening, etc.

Definition of femoral neck shortening: Femoral neck shortening 0–5 mm was defined as no shortening; 5–10 mm as mild shortening; and > 10 mm as severe shortening [14].

Measurement of femoral neck shortening: Both the femoral neck power cross-nailing system and the sliding hip screw are angled fixation devices that achieve fracture healing by a sliding compression mechanism. The difference between the sliding space compression distance on immediate postoperative radiographs and the sliding space compression distance on postoperative follow-up radiographs is based on the femoral neck shortening length of the pacs real-time system at the same magnification.

The quality of fracture repositioning in the patients of this study was classified according to the Garden alignment index [15]. The angle between the femoral head pressure trabeculae and the medial cortex of the femoral stem is 160° in normal hip orthopantomogram and 180° between the medial axis of the femoral head and the medial axis of the femoral neck in lateral radiograph. Based on the Garden Reset Index, the reset quality is classified into four levels.

Statistical analysis was performed by applying IBM.SPSS Statistics 26 statistical software. For the measurement data, the Shapiro–Wilk test was used to determine whether the data conformed to a normal distribution, where age and body mass index conformed to a normal distribution and the variance was chi-square, expressed as x ± s. Two independent samples t test was used for comparison between the two groups of patients. Harris score and femoral neck shortening length did not conform to a normal distribution and were expressed as M (P₂₅, P₇₅), and the Mann–Whitney U test was used to compare between the two groups of patients. Count data such as gender, mechanism of injury, smoking and alcohol history, laterality, fracture Garden's staging, Pauwels’ staging, quality of immediate postoperative reduction, reduction and complication rate were compared between the two groups using the χ² test. Pauwels typing and early complication rates were compared using Fisher's exact probability method. P < 0.05 was considered statistically significant. Whether the broken end cortex was comminuted, the type of fracture and the quality of immediate postoperative reduction were used as independent variables, and whether neck shortening occurred postoperatively was used as the dependent variable, which was first analyzed univariately using the χ² test, and then, the independent variables with statistically significant differences confirmed by univariate analysis were brought into a dichotomous logistic regression model for multivariate analysis, and differences were considered statistically significant at P < 0.05. The images were created using GraphPad Prism 9 software.

There was no statistically significant difference between the two groups in the comparison of general preoperative data such as age, sex, side, history of smoking and alcohol, and mechanism of injury (P > 0.05), while there was a statistically significant difference between the two groups in the comparison of fracture Garden typing, Pauwels typing, cortical comminution, and quality of reduction (P < 0.05).

Among the 87 patients with femoral neck fractures, 34 cases had neck shortening, with an incidence of 39.1%, and 15 cases had severe shortening, with an incidence of 17.2%; 84 cases had fracture healing, with a fracture healing rate of 96.5%. The hip Harris score at 12 months postoperatively was 83.99 (81.95, 89.20) in the neck shortening group and 90.87 (87.95, 94.80) in the group without neck shortening, which was statistically significant compared with the two groups (Table 1). There was no statistical difference between the two groups in terms of osseous nonunion, femoral head collapse necrosis, and failure of internal fixation (Table 2).

Multifactorial logistic regression analysis showed that cortical comminution of the severed end, fracture staging (Garden type III and IV), and postoperative quality of reduction (grade III and IV) were significantly associated with neck shortening (all P < 0.05). See Table 3.

The incidence of femoral neck shortening after femoral neck fixation with the femoral neck system (FNS) is poorly reported in the literature; in a case study of FNS for femoral neck fractures, the incidence of neck shortening was 28% [16], whereas in the current study, the incidence was 39.1%, and this may be related to the large proportion of Garden III and IV types in the current study. During the follow-up of this study, we found that the time period of postoperative shortening was concentrated at 3 months postoperatively (Fig. 1), and persistent shortening rarely occurred after 3 months. This on one hand may be related to the presence of cortical comminution, although the length of the femoral neck can be restored after intraoperative traction and repositioning, but because the FNS is an angular fixation device, it cannot be fixed to the comminuted cortex, and the fracture end will undergo continuous micromovement, combined with the mechanism of sliding compression healing, the FNS reserves 20 mm sliding compression space (Fig. 1A–C), and the resistance to restrict shortening comes from the cortical counteraction. Poor repositioning, bone defect, cortical comminution, etc., lead to cortical failure to form a buttress, which is prone to excessive postoperative shortening and no length-maintaining fixation of the femoral neck, and limiting the sliding space may reduce the distance of femoral neck shortening (Fig. 1D–F). On the other hand, it may be related to local bone resorption during the healing process of the fracture end after surgery.

The current literature focuses on the fields of osteonecrosis and femoral head necrosis, and less attention has been paid to femoral neck shortening and the factors affecting it. In this study, several variables were selected to explore their correlation with femoral neck shortening, and through binary logistic regression analysis, it was found that femoral neck shortening was most associated with cortical comminution, fracture type, and quality of postoperative repositioning.

Femoral neck fractures are often associated with cortical comminution, especially the posterior medial cortex, with an incidence of up to 90% [17], and more severe bone resorption after repositioning, with continued compression of the comminuted cortex under compression of the longitudinal load along the internal fixation after weight-bearing of the hip joint begins, leading to more severe femoral neck shortening [18]. Current studies have reported that in femoral neck fractures with posterior medial cortical comminution, placement of a single screw can compensate for the negative effects of cortical loss [19, 20], in the current study, there were unincluded patients with this type of fixation and postoperative femoral neck shortening < 10 mm (Fig. 1G–I).

Currently, Garden and Pauwels typing is most commonly used in femoral neck fractures; Garden typing is classified according to the degree of fracture displacement; Garden I and II are stable fractures; Garden III and IV are displaced fractures. Garden III and IV fractures with large initial fracture displacement and severe bone destruction are considered to be factors that affect the stability and healing of femoral neck fractures. Stockton DJ et al. performed a retrospective study analyzing the imaging data of 65 patients and showed that displaced fractures were on average 5.9 mm more shortened than undisplaced fractures [8]. Displaced fractures were also noted as an independent predictor of femoral neck shortening in a prospective multicenter trial published by Gerard P. Slobogean et al. in 2017 [21]. In addition, it has also been noted in a large body of the literature that cortical comminution and displaced fractures correlate with femoral neck shortening [7, 14, 22]. Pauwels typing is based on the angle between the distal fracture line and the horizontal line (Pauwels angle), type I: < 30°; type II: 30–50°; type III: > 50°. Pauwels typing and femoral neck shortening did not constitute independent risk factors in the present study, probably related to the predominance of Pauwels type III in the present study, with large Pauwels angles and a fracture end dominated by shear forces and pressure loads not acting well to pressurize the fracture end.

Femoral neck fractures are treated by striving for anatomical repositioning and restoring the concentric matching of the femoral head to the acetabulum. In 1971 Garden et al. performed a minimum of 3 years of clinical follow-up after internal fixation treatment of more than 400 femoral neck fractures and found that there was a graded relationship between the quality of reduction of femoral neck fractures and postoperative complications; as the quality of reduction decreased, the incidence of postoperative complications increased, and proposed the Garden reduction alignment index, which was divided into four grades and is used today [15]. Yoram A. Weil et al. found in a study of 41 cases of femoral neck fractures treated with three hollow nails that the incidence of postoperative neck shortening could be as high as 51% with up to 75% non-anatomic reduction [23]. This suggests that when the quality of the repositioning is unsatisfactory, partial insertion of the fracture end aggravates shortening when subjected to human loading, and even necrosis of the femoral head can occur under long-term stress stimulation [12, 24]. Since collapsed femoral head necrosis usually does not show clinical symptoms until 1 year, or even longer, after surgery, and most patients have a low detection rate in the early stage of the lesion, there are also some patients who may be missed or misdiagnosed, etc. This also means that the signs of femoral head necrosis are not yet fully detected based only on the 12-month postoperative follow-up of patients, which is the main reason why the incidence of femoral head necrosis in this study is lower than the results reported in the literature. Therefore, we still need to follow the patients for a long time and continue to observe the occurrence of postoperative femoral head necrosis.

The effect of postoperative neck shortening on fracture healing and postoperative functional recovery after FNS fixation of femoral neck fractures is one of the current hot issues of concern. The current study showed a healing rate of (32/34) 94.1% in the neck shortening group and (52/53) 98.1% in the group without neck shortening, with no statistically significant difference between the two groups (P > 0.05). Although neck shortening has no significant effect on fracture healing, it can cause a decrease in hip function. In an analysis of a prospective multicenter study on the prognosis of hip fractures in China published by Slobogean et al. in 2017, it was noted that femoral neck shortening ≥ 1 cm was associated with a poorer functional prognosis after surgery in young patients with femoral neck fractures [21]. A multicenter study of 660 patients by Zlowodzki showed a gradient decrease in SF-36 scores in the no-shortening group (< 5 mm), moderate shortening group (5–10 mm), and moderate shortening group (≥ 10 mm) and that neck shortening was the only predictor of physical function scores [12]. In this study, we used the hip Harris score as an index to evaluate the recovery of hip function, and at the 12-month postoperative follow-up, the score was 83.99 (81.95, 89.20) in the neck shortening group and 90.87 (87.95, 94.80) in the neck non-shortening group, and the difference between the two groups was statistically significant, and the gap on the Harris score was mainly in the use of aids, abnormal gait, etc. The reason for this analysis may be the shortening of the force arm and increased load on the hip abductor muscle group due to neck shortening, thus failing to balance the vertical stress from the pelvis and causing varying degrees of claudication. In addition, the lateral plate placed on the lateral side of the femur may cause irritation to the surrounding soft tissues, thus causing discomfort to the patient and affecting the functional activity of the hip joint.

In conclusion, the rate of femoral neck shortening after FNS internal fixation is high, and femoral neck shortening affects the hip function of patients, and the cortical comminution of the broken end, the type of fracture, and the quality of fracture reduction are the influencing factors of femoral neck shortening. However, the limitations of this study are the radiographic set affected by position and the individual incomplete follow-ups, which may be altered by radiographic lower limb position despite efforts to standardize the measurement of femoral neck shortening values; in addition, femoral neck shortening was measured after repositioning and fixation, and for patients with shortening after repositioning, subsequent measurements may be more distant compared to the true shortening. The measurement of femoral neck shortening in this study was measured by two researchers to group the measurements, and any inconsistencies were coordinated and resolved. In addition, to reduce bias errors, data measurement was not performed by the same personnel as the hip Harris score (Additional file 1).