We completed a prospective cohort study with historical controls to compare the two-year postoperative outcomes of patients under the age of 65 undergoing FNF fixation (OTA/AO classification 31-B [17]) in a level I trauma center using three parallel FTHCS versus those who received three parallel PTS. Exclusion criteria for the prospective FTHCS cohort included patients with an immature skeletal system (age ≤ 18 years), pathological and old fractures, previous hip surgery, deformity or dysplasia of the ipsilateral hip or femur. Additionally, patients who received an open reduction, those with multiple injuries (injury severity scale, ISS > 16) or accompanying fractures of the ipsilateral lower extremity, femoral head fracture, pelvic or acetabular fracture, which might influence the process of rehabilitation and weight-bearing were also excluded.
We compared the prospective FTHCS cohort to a matched historic cohort that received the traditional common practice of three parallel PTS fixation for FNF. Matching criteria included sex, age (18–65 years), body mass index (BMI) within 3 kg/m2, comorbidities, injury laterality, and fracture classification. The primary exclusion criteria were identical to the FTHCS population. Additionally, patients without clinical assessment during follow-up, perioperative and follow-up radiographs, or lack of follow-up were excluded from the study. A complete medical record was available in the electronic medical record, including comorbidities, radiological images, operation details, and follow-up outcomes. These records provided valid information for the retrospective analysis.
The study was conducted in compliance with the principle of the Declaration of Helsinki, approved by the institutional review board of our center, and registered at www.Chictr.org.cn (ChiCTR-IPR-1900025851). Each patient in the prospective group who planned to follow-up at least 24 months, signed an informed consent form agreeing to participate in this study.
Fracture Management and Postoperative Management
In the Emergency Department, all patients underwent a standard radiological protocol of x-rays, including anteroposterior (AP) view of the pelvis and lateral view of the injured hip, as well as computed tomography (CT) scans and image reconstruction. Preliminary management included skin traction or bony skeletal traction to reduce and maintain limb alignment. On admission, demographics and mechanism of injury were recorded.
All surgeries in this study were performed by the team of authors (H.S. and W.Z.), orthopaedic traumatologists with at least 10 years of experience. The patients in both groups were given either general or regional anesthesia and positioned supine on a fracture table. Limb length was restored intraoperatively by gentle longitudinal traction under an image intensifier. Restoration of rotational malalignment was accomplished via internal or external rotation of the extremity. In each operation, the expectation of acceptable rotational reduction was slight valgus or anatomic reduction on the AP view (neck-shaft angle between 130-150°) and no posterior collapse or anterior angulation (less than 15° anteversion) on the real femoral lateral view [18, 19]. As a result of the lack in general consensus on grading the quality of FNF reduction, fracture reduction was graded on the amount of displacement and the degree of residual angulation, matching published criteria [7]. An excellent reduction is considered less than 2 mm of displacement and 5° of angulation in any plane; good reduction is 2-5 mm of displacement and/or 5°-10° of angulation; fair reduction is 5-10 mm of displacement and/or 10°-20° of angulation. Displacement exceeding 10 mm or an angulation of 20° is considered poor. After reduction, a guidewire was inserted up to the subchondral bone of the femoral head and was then measured and drilled. Three absolute FTHCS (Acutrak 6/7, Acumed, Hillsboro, OR, US) or two FTHCS with one PTS were implanted in parallel. If no obvious comminution on the neck cortex was seen, a PTS could be implanted first to compress the fracture site prior to FTHCS implantation. Additionally, prior to FTHCS implantation, the lateral cortex of the proximal femur was tapped to reduce twisting forces. PTS (Asnis III 6.5 mm, Stryker, Howmedic, Mahwah, NJ, US) were implanted via manufacturer instruction.
A standard postoperative rehabilitation protocol was followed regardless of the fixation technique performed. All patients were non-weight bearing for at least 8 weeks after surgery. When radiographic and clinical healing appeared to be progressing toward union, weight bearing was advanced slowly from toe touch to partial weight bearing as tolerated over the subsequent 6 weeks, at the discretion of the treating surgeon.
All patients had routine follow-up. Physical examination was performed and standard radiographs were obtained at each follow-up visit. Postoperative CT or magnetic resonance imaging (MRI) was utilized at the discretion of the treating surgeon to evaluate for nonunion or ANFH. The time to radiographic union, Harris Hip Score (HHS), and any complications observed at any time during follow-up were recorded.
Fracture Classification
The Garden [18], modified Pauwels [20], and VN angle [21] classifications of FNF in each patient were assessed by two independent investigators (L-Y. S. and J-W. L.). Disagreements were settled by a third, trauma-trained orthopaedic surgeon (Y.Z.). All three investigators were blinded to treatment. The radiological images were obtained using picture archiving and communication system workstations. Measurements were performed using Kingstar Winning TV view software (Shanghai Kingstar Winning Medical Information Technology Co. Ltd., Shanghai, China). Due to poor intra-observer and inter-observer reliability by using the various classifications [21], both modified Pauwels classification and VN angle were applied. Low-energy fractures were defined as Garden I-II, Pauwels I-II, or VN < 15° patterns and high-energy fractures were defined as Garden III-IV, Pauwels III, or VN ≥ 15° patterns [20, 21].
Outcome Measures
Our primary outcome was fixation failure, defined as varus collapse (>10°), fracture displacement (>5 mm) or femoral neck shortening (>10 mm vertically). To evaluate fixation failure, the immediate postoperative radiographs were compared with follow-up radiographs. For varus collapse, the change in neck-shaft angle between the postoperative and follow-up radiographs were measured on pelvic AP radiographs (Figure 1 A, C, F). Femoral neck shortening was evaluated on pelvic AP view of postoperative and follow-up radiographs. On these images, two horizontal lines were drawn perpendicular to the femoral anatomical axis from the top of the femoral head and the tip of the greater trochanter on bilateral hips. Vertical femoral neck shortening was defined as the bilateral difference between the two lines (Figure 1 C-F) [14]. Although other methods of measuring shortening using the contralateral hip have been validated in prospective case series [11, 15], they are not suitable for historical control cases where leg rotation could not be standardized.
Secondary outcomes included fixation loosening, nonunion, and ANFH. Fixation loosening was identified if there was any screw penetration or withdrawal on radiographs. Screw withdrawal can manifest clinically as soft tissue irritation in the greater trochanter region. If the fracture line was grossly visible at 6 months postoperatively, a fracture nonunion was considered to be present [22]. The radiographic criteria of Ficat and Arlet were used to diagnose ANFH [23].
Statistical Methods
Based on previous results from the treatment of vertical FNF [10], the rate of fixation failure in the group treated with PTS alone was 41.9%, while fixation failure rate in the group treated with FTHCS was 14.3%. A two-tailed test was performed with an α level of 5% (α=0.05) and power of 80% (β=0.20) to determine that a sample size of 43 would be required in each group to gain significant results. Considering a certain proportion of non-vertical fractures (40%) in adult patients under the age of 65, a rate of lost to follow-up of 10%, and the rate of open reduction of 10%, we decided to enroll 100 patients in the prospective FTHCS fixation group.
The sample size was calculated using G*Power 3.1.9.3 software (Institute for experimental psychology in Dusseldorf, Germany). Statistical analyses were performed by an independent statistician blinded to clinical outcomes using SPSS 19.0 (SPSS, Inc., Chicago, IL, US). Continuous variable were presented as mean ± SD, and tested by Student’s t-test. Categorical variables were shown as number and percentages (%) and tested by the chi-squared test. Fisher’s exact test was implemented when 20% of the cells had expected values less than 5.
After controlling for important confounders in the cohorts, a multivariable logistic regression model was used to determine the independent risks of complication associated with fixation failure, nonunion, and ANFH. Logistic regression analysis results were presented as odds ratios (OR) with 95% confidence intervals (95% CI). Stratified analysis using the same regression models was then performed to characterize differences in the strength of the fixation methods across the fracture classification (injury severity). A value of p < 0.05 indicated statistical significance.