Contralateral Neck-shaft Angle Lower Than 130° Is Associated With Clinical Failure in Nongeriatric Individuals: Analysis of the National Femoral Neck Fracture Database of 1066 Patients

Abstract

Background

Treatment of femoral neck fractures in patients who are nongeriatric (≤ 60 years) is challenging because of high failure rates. Anatomic parameters influence the biomechanical environment for fracture healing, but their associations with clinical prognosis remains unclear.

Questions/purposes

(1) Which anatomic parameter that is identifiable on pelvic radiographs shows a statistical correlation with a higher risk of clinical failure defined as nonunion, avascular necrosis (AVN), reoperation, and functional failure (decrease in Harris hip score reaching the minimum clinically important difference) in the screw fixation of femoral neck fractures among nongeriatric patients? (2) How does the influence of anatomic parameters on clinical prognosis manifest: directly or mediated by additional mechanisms?

Methods

This retrospective, multicenter study used a nationwide database in China. Between January 2014 and December 2020, we evaluated 1066 patients with femoral neck fractures with a median age of 53 years (interquartile range 46 to 56) and median follow-up period of 62 months. Anatomic parameters including femoral neck-shaft angle (NSA), femoral head radius, femoral neck width, femoral offset, acetabular center-edge angle, and acetabular sharp angle were variables of interest. The primary outcome was clinical failure including nonunion, AVN, reoperation, and functional failure (decrease in Harris hip score reaching the minimum clinically important difference). Risk factors for failure were first filtered using the Bayesian information criterion and then assessed with multiple regression adjusting for confounders. The mediation effect was further explored using model-based causal mediation analysis with a quasi-Bayesian Monte Carlo method.

Results

Of all anatomic parameters we assessed, the contralateral NSA was associated with clinical failure, after adjusting for all potential covariates and confounding variables (adjusted odds ratio 0.92 [95% confidence interval 0.89 to 0.95]; p < 0.001). The optimal threshold for the NSA was 130°, with the highest Youden index of 0.27. Patients with an NSA < 130° (41% [441 of 1066]) demonstrated an increased occurrence of nonunion (15% [68 of 441] versus 5% [33 of 625]; p < 0.001), AVN (32% [141 of 441] versus 22% [136 of 625]; p < 0.001), functional failure (25% [110 of 441] versus 15% [93 of 625]), and reoperations (28% [122 of 441] versus 13% [79 of 625]). The impact of an NSA less than 130° on clinical failure was direct and substantially mediated by the type of displaced fracture (mediation proportion: 18.7%).

Conclusion

In our study of screw fixations for femoral neck fractures among nongeriatric patients, we identified that a contralateral NSA < 130° correlates with an increased risk of clinical failure including nonunion, AVN, functional failure, and reoperation. The effect is either direct or mediated through displaced fracture types. This is important for surgeons in order to recognize the elevated rate of clinical failure and nature of the challenging biomechanical environment, which should guide them in refining surgical details and selecting appropriate fixation and rehabilitation plans. Approaches to managing these fractures require further validation with large-scale clinical trials.

Level of Evidence

Level III, prognostic study.

Introduction

Femoral neck fractures in patients younger than 60 years, which constitute 3% to 10% of individuals [13, 29], are uncommon compared with those in older patients, but can be clinically challenging. Because these fractures often occur in the setting of high-energy trauma, they can result in an unfavorable biomechanical environment, tenuous vasculature [40], and severe fracture type [33]. Although a few patients treated with novel internal fixation devices claimed satisfactory outcomes [37], the prognosis for these fractures remains unsatisfactory, with reported rates of 21.8% for nonunion [38] and 14.3% for avascular necrosis (AVN) [36] and decreases in functional scores [35]. Disturbingly, 32.1% of patients in this category are treated with major reconstructive surgery, with 20.3% eventually undergoing hip arthroplasty [5]. In addition, patients in this group often have demanding functional requirements for work and recreational activities.

A better understanding of the factors influencing clinical prognosis can guide patient treatment after injury. Although well-known patient-related risk factors such as sex and initial fracture displacement have been extensively investigated based on large-scale database analyses [26], the role of hip anatomic parameters has received comparatively little attention. Current research primarily emphasizes the link between inherent hip anatomic features and clinical outcomes in patients with osteoporosis and fragility fractures [8, 15, 19]. Notably, factors such as hip axis length and femoral neck width are correlated with an elevated risk of femoral neck fragility fractures [8, 15, 19]. In contrast, there is a paucity of research exploring the impact of hip anatomic parameters on the clinical prognosis of nongeriatric patients suffering from femoral neck fractures and treated with internal fixations.

Using a large-scale dataset of pelvic radiographs from a national database in China, we therefore asked: (1) Which anatomic parameter that is identifiable on pelvic radiographs shows a statistical correlation with a higher risk of clinical failure, defined as nonunion, AVN, reoperation, and functional failure (decrease in Harris hip score reaching the minimum clinically important difference) in the screw fixation of femoral neck fractures among nongeriatric patients? (2) How does the influence of anatomic parameters on clinical prognosis manifest: directly or mediated by additional mechanisms?

Patients and Methods

Study Design and Setting

This was a multicenter, retrospective study at the three largest local medical centers in China. The data were sourced from the National Femoral Neck Fracture Database of the Chinese population. This comprehensive database has been continuously gathering medical histories and imaging data of patients admitted to the three largest local medical centers in China since 2014. Each hospitalized patient is assigned a unique personal identifier, and their medical information, imaging data, and hip function status are recorded. The database undergoes annual updates to ensure its relevance and accuracy.

Participants

We included patients seen between January 2014 and December 2020 with data in the National Femoral Neck Fracture Database of the Chinese population at three medical centers. Eligibility criteria were age between 18 and 60 years, patients who were treated with partial-thread parallel screws, and patients who had standard bilateral postoperative femoral AP radiographs. To capture a standard AP radiograph, patients were placed in the supine position with 15° to 20° of internal hip rotation [30].

We contacted 1643 patients and excluded those treated with two or four cannulated screws (n = 281), those with pathologic or concomitant fractures (n = 51), those with prefracture Harris hip scores of 79 or less (n = 56), and those with missing data or follow-up duration of less than 2 years (n = 189) (Fig. 1).

Fig. 1.

The patient selection process is shown in this flowchart. NF-FNF = National Femoral Neck Fracture Database of the Chinese population.

Variables

Anatomic parameters were regarded as variables of interest, including femoral neck-shaft angle (NSA), femoral head radius, femoral neck width, femoral offset, acetabular center-edge angle, and acetabular sharp angle. NSA, femoral head radius, femoral neck thickness, and femoral offset were measured using the contralateral side as a template [32] (Fig. 2). To mitigate observer bias, all radiographic measurements were independently conducted by two senior surgeons (WJ and KW). In cases of discrepancies, the other senior surgeon (XL) made the final determination. The intraclass correlation coefficient (ICC) values for these measurements fell within the range of good-to-excellent agreement [22], with ICC values ranging from 0.78 to 0.94 (Table 1).

Fig. 2.

The measurement of hip anatomic parameters is shown. (A) This radiograph shows the determination of the femoral neck axis (solid red line) and femoral shaft axis (solid red line). The femoral neck axis is the line connecting the femoral head center and the femoral neck center. The femoral head center is determined by the center of the circle that best fits the femoral head (yellow circle), and the femoral neck center was the midpoint between the cutting points of the circle with the femoral neck. When fitting the femoral shaft axis, two best-fitting circles were drawn below the lesser trochanteric plane. The femoral shaft axis was defined as the line between the two circle centers (yellow circle, C1, C2). (B) This AP pelvis radiograph shows measurements of the neck-shaft angle, femoral offset, and femoral neck width. The neck-shaft angle is the angle between the femoral neck axis (yellow dotted line) and femoral shaft axis (yellow dotted line). Femoral offset is the distance (solid red line) between the femoral neck center and femoral shaft axis. The neck thickness is the shortest distance (red line) between the proximal and distal margin of the femoral neck perpendicular to the femoral neck axis. (C) This AP pelvis radiograph shows measurements of the acetabular center-edge angle and sharp angle. The center-edge angle is formed by a vertical line (solid yellow line) and a line (solid yellow line) connecting the femoral head center with the lateral edge of the acetabulum. The sharp angle is the angle between the teardrop line (black dotted line) and the line connecting the teardrop with the lateral edge of the acetabulum (solid yellow line). FNA = femoral neck axis; FSA = femoral shaft axis; FHC = femoral head center; NSA = neck-shaft angle; FO = femoral offset; NW = neck width; CE = center edge. A color image accompanies the online version of this article.

Table 1.

Interobserver reliability of radiographic measurements

	ICC	95% CI	p value
Neck-shaft angle	0.84	0.77-0.88	< 0.001
Femoral head radius	0.94	0.93-0.95	< 0.001
Femoral offset	0.78	0.75-0.80	< 0.001
Femoral neck thickness	0.88	0.86-0.90	< 0.001
Center-edge angle	0.84	0.77-0.88	< 0.001
Sharp angle	0.88	0.87-0.90	< 0.001

Demographic information including age, sex, marital history, current smoker (defined as individuals who smoked 100 cigarettes or more in their lifetime and who smoked one cigarette per day for at least 6 months [12]), comorbidities (diabetes melletus or hypertension), and time to surgery (from the time of injury) were obtained during the hospital stay. Parameters describing fracture severity included the displacement degree and Pauwels classification (stable: Pauwels Type I and II; unstable: Pauwels Type III) [27]. Surgical variables included time to surgery, reduction method (open or closed), reduction quality (good or poor), screw configuration (invert or regular triangle), the application of washer, implant position (evaluated by implant positioning [IMPO] score), and surgical time. Per the Garden index [20], good reduction (anatomic or acceptable) is characterized by angles between the medial trabeculae and medial cortex (AP view) and between the central axis of the head and neck (lateral view) within 155º and 180º. Poor reduction (borderline acceptable or unacceptable) is defined if either angle is less than 155º or exceeds 180º. The IMPO score included six items: distance to the inferior calcar (1 to 6 mm), distance to superior cortex < 10 mm, tip-head distance (3 to 20 mm), distance to the superior or inferior caput ≥ 15 mm, distance to the anterior caput (25% diameter), and inter-implant angle < 5°. An IMPO score of 5 to 6 points indicates a good implant position, 3 to 4 points signifies fair, and 0 to 2 points suggests a poor implant position [21].

Clinical Outcomes

The primary endpoint was clinical failure, defined as a composite outcome including any of the following severe events: nonunion, AVN, patient-reported functional failure, and reoperation. All clinical complications were determined using radiographs. Nonunion was defined as fracture redisplacement or no radiographic evidence of union at 6 months postoperatively. Radiographic nonunion was assessed using the Radiographic Union Score for the Hip [3, 10], with scores less than 18 points reviewed by both surgeons (WJ and KW). AVN was evaluated based on the radiographic criteria of Ficat [2], and Type 2b or greater was defined as clinical failure. Patient-reported functional failure was defined as a decrease in the Harris hip score at the final follow-up compared with baseline. Although studies used the minimum clinically important difference of the Harris hip score to describe functional improvement after treatment such as hip arthroplasty [34] and hip-preserving surgeries for osteonecrosis [4, 9], there is a minimum clinically important difference value for internal fixation of femoral neck fractures. Therefore, we used the threshold of 18 points [14] used for nonunion treatment of femoral neck fractures in this study.

Descriptive Data

A total of 1066 patients with a median age of 53 years (interquartile range 46 to 56) and median follow-up period of 62 months were included in the analysis. None of these patients had missing data. Clinical failure occurred in 37% (391 of 1066) of patients (Table 2), including 10% (101 of 1066) with nonunion, 26% (277 of 1066) with AVN, 19% (201 of 1066) with reoperation, and 19% (203 of 1066) with patient-reported functional failure.

Table 2.

Baseline characteristics of patients

	Total (n = 1066)	Success (n = 675)	Failure (n = 391)	p value
Age	53 (46-56)	53 (45-55)	53 (49-57)	0.002
Sex
Female	57 (606)	60 (408)	51 (198)	0.002
Male	43 (460)	40 (267)	49 (193)
Marital history
Married	65 (688)	66 (445)	62 (243)	0.30
Unmarried	23 (243)	23 (152)	23 (91)
Divorced or widowed	13 (135)	12 (78)	15 (57)
Current smoker
No	75 (801)	78 (527)	70 (274)	0.004
Yes	25 (265)	22 (148)	30 (117)
Diabetes mellitus
No	93 (989)	94 (633)	91 (356)	0.10
Yes	7 (77)	6 (42)	9 (35)
Hypertension
No	81 (866)	83 (560)	78 (306)	0.06
Yes	19 (200)	17 (115)	22 (85)
Time to surgery
< 24 hours	9 (96)	9 (59)	10 (37)	0.88
24 to 48 hours	41 (435)	41 (274)	41 (161)
> 48 hours	50 (535)	51 (342)	49 (193)
In-hospital stay in days	4 (5-6)	4 (5-6)	4 (5-6)	0.66
Displacement degree
Nondisplaced	40 (431)	57 (384)	12 (47)	< 0.001
Displaced	60 (635)	43 (291)	88 (344)
Pauwels classification
Stable	53 (567)	61 (409)	40 (158)	< 0.001
Unstable	47 (499)	39 (266)	60 (233)
Reduction quality
Good	89 (948)	96 (649)	77 (299)	< 0.001
Poor	11 (118)	4 (26)	24 (92)
Reduction method
Closed	91 (966)	91 (614)	90 (352)	0.61
Open	9 (100)	9 (61)	10 (39)
Surgical time in minutes	58 (45-66)	56 (45-65)	60 (50-70)	0.001
Screw configuration
Invert triangle	58 (620)	60 (402)	56 (218)	0.23
Regular triangle	42 (446)	40 (273)	44 (173)
Washer
No	64 (683)	65 (442)	62 (241)	0.21
Yes	36 (383)	35 (233)	38 (150)
Implant position
Good (5-6)	45 (475)	46 (307)	43 (168)	0.03
Fair (3-4)	51 (539)	51 (344)	50 (195)
Poor (0-2)	5 (52)	4 (24)	7 (28)

Data presented as median (IQR) or % (n).

At baseline, patients in the clinical failure group demonstrated distinct characteristics from those in the success group (Table 2). There was an older age (53 years [IQR 49 to 57 years] versus 53 years [IQR 45 to 55 years]) in the clinical failure group than in the success group, as well as a higher male ratio (49% [193 of 391] versus 40% [267 of 675]) and proportion of current smokers (30% [117 of 391] versus 22% [148 of 675]). Fracture morphology indicated a predominance of displaced (88% [344 of 391] versus 43% [291 of 675]) and unstable types (60% [233 of 391] versus 39% [266 of 675]). There were longer surgical times in the clinical failure group (median 60 minutes [IQR 50 to 70 minutes] versus 56 minutes [IQR 45 to 65 minutes]) and higher proportions of poor reduction (24% [92 of 391]) and implant positioning (7% [28 of 391]) than in the success group (reduction: 4% [26 of 675]; implant positioning: 4% [24 of 675]).

Ethical Approval

This study was approved by the local institutional ethics review boards at Shanghai Sixth People’s Hospital Affiliated with Shanghai Jiao Tong University School of Medicine (No. 2022-KY-059[K]), the First Hospital of Harbin Medical University (202271), and the Second Affiliated Hospital of Harbin Medical University (KY2022-177), and it is registered in the Chinese Clinical Trial Registry (ChiCTR2200059592).

Statistical Analysis

We performed statistical analyses using SAS software (version 9.4 for Windows, SAS Institute) and the mediation package in R software (version 4.3.0). Categorical variables are presented as percentages and counts and were compared using the chi-square test or Fisher exact test. Continuous variables were checked for normality using the Kolmogorov-Smirnov test. Parametric continuous variables are expressed as means ± standard deviations and were compared using one-way analysis of variance. Nonparametric continuous variables are presented as median (interquartile range) and were compared using the Mann-Whitney test.

First, all recorded parameters were included in the binary logistic regression model and filtered using the backward selection method, applying Bayesian information criteria. The variables filtered from this selection model were considered as covariates for the primary outcome (clinical failure). All anatomic parameters were regarded as the variables of interest. Then, the effect of the variables of interest on outcomes was evaluated individually using multivariable regression models, adjusting for all possible covariates (variables filtered from the Bayesian information criteria test-based selection model) and confounders (other necessary variables according to an expert’s view). The effects are presented with odds ratios and 95% confidence intervals. The variable of interest that was statistically associated with clinical failure was then converted into categorical variables in the following analysis, according to the optimal threshold determined by the highest Youden index of the receiver operating characteristic curve.

Given the potential influence of anatomic parameters on clinical failure, potentially through factors such as fracture type or surgical quality, we established mediation models to clarify this issue. We first assessed whether other covariates differed across anatomy-based patient groups, identifying statistical differences as potential mediators. This allowed us to distinguish the direct effect of anatomic parameters from the indirect effect mediated through these variables [41]. Using model-based causal mediation analysis, we estimated the proportion of anatomic effect on clinical failure that occurred through the mediator pathway, along with 95% CI. Our analysis adhered to the assumption of confounding adjustment for the exposure-mediator-outcome relationships, and CIs were generated using a quasi-Bayesian Monte Carlo method.

All statistical analyses are two-sided and are considered statistically significant with a p value less than 0.05. We adjusted p values with the false discovery rate using the Benjamini-Hochberg method in the univariable analysis.

Results

Correlation Between Hip Anatomic Parameters and Clinical Failure

Of all anatomic parameters we assessed, NSA was the only parameter associated with clinical failure in the treatment of femoral neck fractures in nongeriatric patients, filtered according to Bayesian information criteria. After adjusting for covariates (age, current smoker, displaced type, reduction quality, and screw position) and confounding variables (sex, Pauwels classification, femoral offset, and surgical time) using the multivariable analyses (Table 3), NSA (adjusted OR 0.92 [95% CI 0.89 to 0.95]; p < 0.001) was statistically associated with clinical failure, while femoral head radius (adjusted OR 0.98 [95% CI 0.91 to 1.05]; p = 0.58), femoral offset (adjusted OR 1.02 [95% CI 0.99 to 1.04]; p = 0.25), femoral neck width (adjusted OR 1.01 [95% CI 0.98 to 1.05]; p = 0.45), center-edge angle (adjusted OR 1.00 [95% CI 0.98 to 1.02]; p = 0.93), and sharp angle (adjusted OR 0.98 [95% CI 0.95 to 1.01]; p = 0.18) were not (Supplemental Table 1; http://links.lww.com/CORR/B294).

Table 3.

Univariable and multivariable analyses of anatomic features associated with clinical failure

		Total (n = 1066)	Success (n = 675)	Failure (n = 391)	Crude OR (95% CI)	Adjusted OR (95% CI)	p value
Neck-shaft angle		131.3 (127.5-134.2)	131.9 (128.9-134.3)	128.6 (126.3-133.1)	0.90 (0.87-0.93)	0.92 (0.89-0.95)	< 0.001
Femoral head radius		28.6 (26.6-30.0)	28.5 (26.3-30.0)	28.7 (27.0-29.9)	1.03 (0.98-1.09)	0.98 (0.91-1.05)	0.58
Femoral offset		41.8 (37.7-46.1)	41.3 (37.6-45.3)	42.5 (38.0-47.5)	1.04 (1.02-1.06)	1.02 (0.99-1.04)	0.25
Femoral neck width		35.6 (33.1-38.4)	35.4 (32.8-38.1)	36.0 (33.7-38.9)	1.03 (1.00-1.06)	1.01 (0.98-1.05)	0.45
Center-edge angle		33.6 (29.9-38.2)	33.3 (29.6-38.1)	34.2 (30.4-38.4)	1.01 (0.99-1.03)	1.00 (0.98-1.02)	0.93
Sharp angle		38.3 (35.6-40.8)	38.3 (35.8-40.96)	38.3(35.3-40.6)	0.98 (0.95-1.00)	0.98 (0.95-1.01)	0.18

Data presented as median (IQR). Multivariable regression analyses adjusted for age, sex, current smoker, Pauwels classification, displaced type, surgical time, reduction quality, and screw position

Based on the receiver operating characteristic curve, the optimal threshold for NSA was 130°, with the highest Youden index of 0.27. Patients exhibiting an NSA < 130° demonstrated an increased occurrence of nonunion (15% [68 of 441] versus 5% [33 of 625]; p < 0.001) (Fig. 3) and AVN (32% [141 of 441] versus 22% [136 of 625]; p < 0.001) (Fig. 4). The prevalence of functional failure was statistically (p < 0.001) greater in the low NSA group (25% [110 of 441]) than in the high NSA group (15% [93 of 625]) (Fig. 5). Additionally, we observed a substantially (p < 0.001) elevated proportion of reoperations in the low NSA group (28% [122 of 441]) compared with the high NSA group (13% [79 of 625]) (Table 4).

Fig. 3.

(A-C) A 55-year-old man with a displaced femoral neck fracture and a neck-shaft angle of 125.6° was treated with three cannulated screws. (D-E) Radiographs taken 6 months after surgery showed femoral neck shortening and nonunion. A color image accompanies the online version of this article.

Fig. 4.

(A) These preoperative radiographs are from a 46-year-old woman with a contralateral femoral neck shaft angle of 126.2° (B) who sustained a displaced femoral neck fracture and (C) was treated with three cannulated screws. (D) A radiograph taken 2 years postoperatively revealed femoral neck shortening and avascular necrosis. (E) The patient was subsequently treated with arthroplasty. A color image accompanies the online version of this article.

Fig. 5.

(A) A 53-year-old man with a low contralateral femoral neck shaft angle (128.9°) had (B) a displaced femoral neck fracture. (C) The patient was treated with three cannulated screws. (D) Six-month and (E) 2-year postoperative radiographs revealed femoral neck shortening. Despite the absence of nonunion or avascular necrosis, the Harris hip score decreased from 100 before the fracture to 65 at the final follow-up interval, reaching the minimum clinically important difference. A color image accompanies the online version of this article.

Table 4.

Comparison of covariates and outcomes between the low NSA and high NSA groups

	NSA > 130° (n = 625)	NSA < 130° (n = 441)	p value
Clinical failure	26 (164)	51(227)	< 0.001
Age	53 (47-56)	53 (45-55)	0.46
Displacement degree
Nondisplaced	49 (307)	28 (124)	< 0.001
Displaced	51 (318)	72 (317)
Reduction quality
Good to fair	91 (569)	86 (379)	0.02
Unacceptable to poor	9 (56)	14 (62)
Implant position
Good (5-6)	43 (271)	46 (204)	1
Fair (3-4)	52 (323)	49 (216)	0.44
Poor (0-2)	5 (31)	5 (21)	0.72
Nonunion
No	95 (592)	85 (373)	< 0.001
Yes	5 (33)	15 (68)
Avascular necrosis
No	78 (489)	68 (300)	< 0.001
Yes	22 (136)	32 (141)
Functional failure
No	85 (532)	75 (331)	< 0.001
Yes	15 (93)	25 (110)
Reoperation
No	87 (546)	72 (319)	< 0.001
Yes	13 (79)	28 (122)

Mediation Analysis of the Relationship Between Low NSA and Clinical Failure

Statistically different displacement degrees (p < 0.001) and reduction quality (p = 0.02) were observed between the low NSA (< 130°) and high NSA (> 130°) groups (Table 3). The association between low NSA and clinical failure was substantially mediated by the displacement degree (p < 0.001), but not by reduction quality (p = 0.42) (Fig. 6). The direct effect of low NSA on clinical failure was 0.17, and the indirect effect through displaced degree was 0.04, accounting for a mediation proportion of 18.7% (Table 5).

Fig. 6.

This is the directed acyclic graph for mediation relationships.

Table 5.

Mediation effects between NSA < 130° with incidence of clinical failure

Exposure: NSA < 130°	Mediator: displaced fracture type Incidence of clinical failure	p value	Mediator: reduction quality Incidence of clinical failure	p value
Controlled direct effect	0.17 (0.12 to 0.23)	< 0.001	0.17 (0.12 to 0.23)	< 0.001
Indirect effect	0.04 (0.02 to 0.06)	< 0.001	0.01 (-0.01 to 0.02)	0.42
Total effect	0.21 (0.15 to 0.27)	< 0.001	0.18 (0.12 to 0.23)	< 0.001
Proportion mediated	18.7% (8.6% to 30%)	< 0.001	2.9% (-4.7% to 10%)	0.42

Sensitivity Analyses

In two studies [5, 6], the definition of “young patients” is typically those younger than 50 years. Consequently, to assess the generalizability of our findings, we conducted a subgroup analysis stratified by age, dividing the cohort into those younger or older than 50 years. Remarkably, our analysis revealed that an NSA less than 130° consistently demonstrated a statistical association with clinical failure across both age groups (Supplemental Table 2; http://links.lww.com/CORR/B294).

Discussion

Consistent with previous research [23, 36], the treatment of femoral neck fractures among patients younger than 60 years was a challenge in this study, with 37% of patients experiencing clinical failure and 19% undergoing reoperation. Identifying risk factors for clinical failure is crucial to reduce its incidence. Although factors such as reduction quality and displacement degree have been extensively studied, the association between anatomic parameters and clinical prognosis remains unknown. For the first time, based on our multicenter database of 1066 patients with a mean follow-up period of 5 years, it was revealed that patients with femoral neck fractures with an NSA less than 130° faced an elevated risk of clinical failure. The effect of a low NSA on clinical prognosis is either direct or mediated by creating more-displaced fracture types.

Limitations

First, reported NSA values showed high variance across different studies, indicating low interobserver reproducibility. To address this, we exclusively used standard AP radiographs. The measurement protocol we chose was easily applicable and highly repeatable with minimal system error. The assessment was conducted by two experienced surgeons who underwent uniform training. This approach enhanced interobserver consistency in our research. Second, our dataset does not encompass specific details of the rehabilitation process. Although our centers adhered to the same clinical guidelines, the exact duration of weightbearing differed for each individual and was not factored into our analysis. This represents a potential confounding factor that should be further addressed in future studies. Third, for enhanced statistical efficacy, we used a composite primary endpoint. Although the four components had similar clinical significance, they may involve distinct biological processes, potentially complicating result interpretation. Fourth, our cohort exclusively underwent screw fixation, the most widely applied method in such patients. This approach inherently limits the generalizability of our findings. Future research should explore and compare the efficacy of different fixation techniques in this specific patient population, thereby expanding the applicability of our study's findings.

Correlation Between Hip Anatomic Parameters and Clinical Failure

Among a series of anatomic measurements obtained from pelvic radiographs, we identified that a contralateral NSA < 130º is a crucial indicator of clinical failure. Interestingly, prior research [11, 15, 31] suggested that an increased NSA is linked to a higher risk of hip fractures. This relationship is because a higher NSA results in an increased bending moment in the proximal femur during low-energy lateral falls, thus increasing stress and raising the fracture risk. Conversely, in the scenario of postoperative weight-bearing, external force is mainly vertical to support the body in the gravitational field [1]. Consequently, a lower NSA results in a greater bending moment, which contrasts with the effect observed during lateral falls [11, 15, 31]. Additionally, according to the classic Pauwels theory [28] and cantilever beam formula [18], in patients with low NSAs, the proximal femur will experience more prominent varus deformation force and shearing stress, which is unfavorable for bone healing [24]. Another potential cause of clinical failure in patients with low NSAs is increase in interface stress because of a larger shearing angle between the resultant force and screw axis. This leads to excessive bone yielding around screws and causes screw migration and loosening [7]. Additionally, anatomically, the region of the Ward triangle, where the bone tissue is scarce, is larger in a femur with a low NSA [28]. Fixation in these areas might reduce structural stability and exacerbate screw loosening. Moreover, according to previous biomechanical studies [25, 28], a decrease in NSA elevates femoral head stress. This accelerates the progression of AVN and results in early femoral head collapse.

Mediation Analysis of the Relationship Between Low NSA and Clinical Failure

Beyond its direct influence, the effect of a low NSA on clinical failure was further mediated by the presence of displaced fracture types, accounting for 18.7% of the total effect. Femoral neck fractures in younger individuals are often the result of high-energy traumas that are typically vertically oriented [1]. In individuals with low NSAs, fractures because of vertical impacts tend to be more susceptible to displacement and varus deformation. The presence of a displaced fracture poses a high risk of vascular damage [40] and subsequent clinical failures such as AVN, a finding corroborated by the current study and previous research [36].

Conclusion

In screw fixations of femoral neck fractures among nongeriatric patients, a contralateral NSA < 130° is associated with higher risk of clinical failure including nonunion, AVN, functional failure, and reoperation. The impact of a contralateral NSA < 130° on clinical prognosis can be direct and mediated (accounting for 18.7% of the total effect) by creating a more-displaced fracture type. These findings highlight the importance for surgeons to evaluate contralateral anatomic structures preoperatively to anticipate clinical prognosis. For patients with contralateral NSA < 130º, given the elevated proportion of clinical failure and challenging biomechanical environment, it is necessary for surgeons to ensure there is an acceptable reduction quality and good or fair screw position (IMPO score > 2) intraoperatively, explore novel biomechanically stable internal fixation implants [16, 17], and prevent premature weightbearing [39]. Our study raises potentially important issues, such as what internal fixation skills or strategy are needed to alter the clinical outcome for these patients. Further validation with solid clinical evidence, particularly in a large-scale randomized clinical trial, is warranted.

Supplementary Material

abjs-482-1801-s001.docx ^{(118.3KB, docx)}

abjs-482-1801-s002.docx ^{(28.5KB, docx)}