Distances between popliteal vessels to the femur and anthropometric impacts on distal femoral fracture surgery outcome: a retrospective study

The proximity of the popliteal vessels in the distal femur may increase the risk of iatrogenic vascular injury during cerclage wiring. In this study, the closest location and distance of the popliteal vessels to the femur was examined using magnetic resonance imaging (MRI). The associations between anthropometric factors and the distance that would guide the placement of wires safely during surgery were also identied.

complications associated with inadequate use of cerclage wires [1,4]. Indeed, occlusion of major vessels by cerclage wire could result in severe issues, such as below knee amputation [5].
Managing vascular injuries associated with distal femoral fracture surgery requires awareness of the anatomy of the popliteal vessel and the distal femur and adequate wire placement. The super cial femoral artery (SFA) crosses from anterior-superior to posterior-inferior in the distal third of the femur, then exits from the adductor hiatus (AH) to become the popliteal artery (PA), which comes closely to the cortex [6].
The proximity of the popliteal vessels to the distal femur may increase iatrogenic injury during cerclage wiring of the distal femoral fracture. Few studies focused on the vascular structure or AH in the distal third of the femur. In two cadaveric studies, it was reported the area up to 8 cm proximal to the adductor tubercle (AT) to be safe from vascular damages during surgery and localisation of the apex of the AH could be determined by a bony landmark [7]. However, the spatial resolution of the popliteal vessels in relation to the femur have not yet been examined and the effects of the anthropometric factors on such measures have not been elucidated.
This study aimed to determine (1) the closest location and distance of the popliteal vessels to the femur in adults on magnetic resonance imaging (MRI) and (2) the signi cance of associations between anthropometric factors (i.e., gender, age, body height, body weight, body mass index [BMI], thigh circumference, femoral length, and femoral width) and the distance that would guide the placement of wires to minimize the risk of vascular injuries during distal femur fracture surgery.

Patients
After obtaining institutional review board approval (CMUH109-REC3-106), we conducted a retrospective . All studies had a written report submitted by a musculoskeletal radiologist (HYC) at our institution.
All cases were on unilateral knees. The age of each patient at the time of the study was recorded. An electronic query and manual review of the medical records were completed to obtain patient anthropometric factors, including gender, body height, body weight, and thigh circumference, which was measured horizontally just distal to the gluteal fold [8]. BMI was calculated as weight in kilograms (kg) divided by height in meters (m) squared (kg/m 2 ). The femoral length and width, de ned as the distance from the tip of the greater trochanter (GT) to the AT, and the widest portion of the distal femur, respectively [9], were reviewed on lower limb scanograms.

Knee MRI
Patients were scanned with a 1.5 Tesla Signa MRI scanner (General Electric Medical Systems, Milwaukee, WI) in the supine position, with both lower extremities straight and knees extended. T1-weighted images in the axial, sagittal, and coronal planes with slice thickness of 2 mm were selected on each knee MRI for analysis. The distances (in millimetres, mm) were measured using a digital calliper tool within INFINITT's Picture Archiving and Communications System (PACS).
In the axial images, the shortest horizontal distance (d-H) from the femoral cortex to the popliteal vessels was measured after tracing nearby cuts of AH (Fig. 1a). In the coronal images, the vertical distance (d-V) from the axial cut of AT to the "d-H" axial level was measured (Fig. 1b). The posterior condylar axis (PCA), a line connecting the most posterior border of the medial and lateral condyle in the axial view, was used as a reference to set the sagittal plane of the femur (Fig. 1c). The posterior half of the femur was de ned by a line paralleling the PCA and crossing the centre of femoral canal, as described by Kim et al. [10]. At each "d-H" axial level, the posterior half of the femur was divided into eight sections that labelled "A to H" from posteromedial to posterolateral, and the position of the popliteal vessels was noted (Fig. 1d). One musculoskeletal radiologist (HYC) and two orthopaedic surgeons (HWC, TLL) recorded all measurements independently, and the mean between three physicians was used for data analysis.

Statistical Analysis
Statistical analyses were performed using SPSS for Windows, version 21.0 (SPSS Inc., Chicago, IL, USA). The reliability of each measurement was examined by the intra-class correlation coe cient (ICC). Continuous data are presented in the form of mean ± standard deviation. Groups were compared using a t-test for independent samples. The effects of gender, age, body height, body weight, BMI, thigh circumference, femoral length, and femoral width on each measurement were evaluated using multivariate linear regression analysis. The coe cient of determination, R 2 , was used to check the goodness of t of the statistical models and as a measure of how much of the original uncertainty in the data was explained by the multivariate analysis. R 2 varied between 0 and 1, with 0 indicating no bene t gained by applying multivariate analysis, and 1 indicating bene t. The correlation between the most in uential anatomical factor and distance measurements was analysed using Pearson's correlation coe cient and signi cant differences with Games-Howell post-hoc analysis. Statistical signi cance was set at P<0.05.

Results
The study group included 110 men and 96 women with a mean age of 47.  19.4%, and 0.5%, respectively). The ICC of d-H and d-V were 0.915 (range 0.883-0.947) and 0.923 (range 0.897-0.961), respectively. The d-H was 7.38 ± 3.22 mm. The d-V was 57.01 ± 11.14 mm. There was no signi cant difference in the distances between the groups of different pathologic diagnosis of knees (P = 0.721).
There was a signi cant difference between genders in their d-H and d-V as well as body height, body weight, BMI, thigh circumference, femoral length, and femoral width (P < 0.001, < 0.001, 0.008, 0.002, < 0.001, and < 0.001, respectively; Table 1), Because the gender difference might account for changes in d-H and d-V, multivariate analysis was performed.   Fig. 2a) d-V (mm) = 0.269 * femoral length (mm) -54.184 (Fig. 2b) These formulas predicted that patients with smaller thigh circumference (especially smaller than 399 mm) have smaller d-H (Table 3), and those with shorter femoral length (especially smaller than 369 mm) have shorter d-V (Table 4).  The proximity of the vascular structures traversing the AH in the distal femur may increase the risk of iatrogenic popliteal vascular injury during cerclage wiring. In the current study, reference values for safe distances from injury and the closest location of popliteal vessels to the femur were established using MRI in adult knees. The closest locations of popliteal vessels were at posteromedial aspect of the femur. The d-H and d-V were 7.38 ± 3.22 mm, and 57.01 ± 11.14 mm, respectively. We also assessed the effect of anthropometric factors on these distances and found thigh circumference and femoral length to be the most important indicators for the d-H and d-V, respectively.
Apivatthakakul et al. evaluated computed tomography angiography (CTA) of 80 patients, which divided the whole femur into eight equal segments (7 levels) from the tip of the GT to the lateral tibiofemoral joint line in the coronal plane and eight equal directions from anterior to posterior of the medial femur in the axial plane. They found that when the SFA was at levels 6 and 7, it was located between sectors F and H (posteromedial and posterior to the femur) and at a distance of about 13.63 ± 3.59 mm and 10.08 ± 3.09 mm, respectively [18]. Their result was similar to the current study, which revealed closest point of popliteal vessels to situate posteromedial and posterior to the femur. During cerclage wiring, either from anterolateral or posterolateral direction, surgeons should be cautious of posteromedial and posterior aspects of the femur. The present study demonstrated the precariousness of popliteal vessels and that any distance shorter than the closest one shown here between the vessels and femur cortex could prove more detrimental than previously thought; we would also strongly suggest subperiosteal over blunt dissection during wiring, where the wire passer tips are as close to the bony cortex as possible or at lowrisk position from the popliteal vessels, to avoid vascular injury.
To our best knowledge, this was the rst study of d-H in the literature, which demonstrated smaller d-H in small thigh circumference patients. The explanation for this association is still uncertain. While the in uence of obesity on the anatomical relationship between the PA and tibial nerve in the popliteal fossa had been reported, no direct evidence between the popliteal vessels and the femur cortex were described [19]. Even though we had hypothesised the d-H was relevant to the thickness of the fatty tissue around the popliteal fossa, no correlation between the measurement and body weight or BMI was noted. Therefore, body fat percentage and regional distribution should be included for evaluation in future studies.
There is a transition zone in the hiatal area, from the adductor canal to the popliteal fossa. In comparison with the more exible fatty tissue of the popliteal fossa, the region of AH is more rigid and xes the junction of SFA and PA close to the femur cortex [7]. Kwon et al. reported the level of AH to be over 59.8 mm proximal to the superior border of the patella [20]. Cadaveric studies with 24 and 28 thighs described the level of AH to be above 10 cm (range 8.0-13.5 cm) and 7.4 cm (range 5.6-9.2 cm) from the AT, respectively [21,22]. Kanawati et al. assessed 41 limbs using CTA to describe the relationship between the SFA and the whole femoral shaft and warned of the "danger zone" from 239.6 mm to 172.5 mm proximal to the AT [9]. In the current study, the d-V (57.01 ± 11.14 mm) was shorter than the distance between AT and AH described in the literature [21,22]. This suggests that the closest level of vascular bundle occurs slightly distal to the AH, at the point where the PA crosses posteriorly to the distal femur.
Kanawati et al. proposed doubled width of the femoral condyles as an estimated safe distance proximal to the AT for intervention [9]. They also mentioned the danger zone where SFA crossed inferiorly by halving the distance between GT and AT. Both predictors could be measured on a true anteroposterior (AP) radiograph preoperatively or intraoperatively. In the current study, surgeon could estimate d-H and d-V preoperatively based on the thigh circumference and the femoral length, respectively, without CTA or MRI. The thigh circumference could easily be computed directly below the gluteal fold [8]. The femoral length could be calculated from length between GT and AT on the AP view of a whole femur radiograph. Although there is a little difference between bilateral thigh, the thigh circumference and the femoral length could be used clinically by measuring the normal contralateral instead of the fractured limb.

Limitations
This study has several limitations. First, our distances were measured on MRIs and not intraoperatively.
Given that the MRIs were taken with the patient in a supine position, while some distal femoral fracture surgeries were performed with the patient in a lateral decubitus position, certain anatomical relationships may differ. Second, the direction of the femoral artery in the current study was dependent on the neutral position of the intact femur. Although the position was similar to most of the clinical conditions that used lateral approach in the distal femur, there were several factors including medial approach, extremity location, traction, soft tissue retractor, and the displacement of bony fragment that could in uence the distance between the popliteal vessels and the femoral cortex. Third, instead of measuring the distances on 2D MRI scans, it would be more reliable to make 3D reconstructions using software such as Mimics® (Materialise, Leuven, Belgium) of the soft tissue of interest, i.e. the vessels, and relate those distances to the 3D bony landmarks. Furthermore, given that all subjects presented here were of unrelated Han Chinese ethnicity, it would be interesting to conduct this work in other populations.
We believed MRI provided more valuable information about relationships between soft tissues, including neurovascular bundle, muscles, and fatty tissues. In addition, the innovation of non-contrast-enhanced MR technique improved the resolution of soft tissue anatomy without the risk of contrast-induced complication as seen in angiography or CTA [26,29]. Moreover, the sample size in our study was larger and the analysed MRIs were used to assess structurally pathologic knees, which was re ective of the anatomical reality in patients who were actually undergoing surgery. All distances were independently measured by three physicians. Therefore, the results provided could be viewed as reliable reference data in future work.

Conclusions:
By measuring a large number of adult knee MRIs, this study found the closest location and distance of the popliteal vessels to the femur to provide useful information for wire placement during distal femoral fracture surgery while minimising risk of vascular injury. Surgeons should strive to perform subperiosteal dissection and pass the wire passer tips as close to the bony cortex as possible during wiring to avoid vascular injury based on the patient's thigh circumference. The vertical positions of the cerclage wire should be checked intraoperatively based on the AT and the femur length to estimate the low-risk position from the popliteal vessels. Given that patients with smaller thigh circumference and shorter femoral length are more likely to have smaller d-H and shorter d-V, respectively, cautious measure should be taken in such cases.

Declarations
Ethics approval and consent to participate CMUH109-REC3-106 Consent for publication Written informed consent was obtained from the patient for the publication of this report and any accompanying images.