A novel anatomical titanium plate versus the reconstruction plate for quadrilateral plate fractures of the acetabulum: A propensity-matched cohort study

Background Because of the peculiar anatomy and this Many scholars consider that quadrilateral plate exist a absolute “dangerous zone” for screw placement and have designed a variety of buttress plates to control medial displacement of the quadrilateral surface. In order to achieve the safe and effective fixation in quadrilateral plate, we have developed a novel fixation technique using a special shaped reconstruction plate combined with several buttress screws of quadrilateral plate which was called Dynamic Anterior Plate-Screw System for Quadrilateral plate (DAPSQ) since 2005(RP group). And the long-term follow-up results have confirmed the effectiveness and safety of this technique. After 2016, a set of anatomical titanium plate(ATP group) of DAPSQ have been designed and applied for the treatment of quadrilateral plate fractures. The aim of this study was to compare the efficacy of anatomical titanium plate versus the reconstruction plate in the treatment of quadrilateral plate fractures. Methods We led a propensity-matched cohort study of quadrilateral plate fractures, and twenty-two patients were treated with anatomical titanium plate of DAPSQ in the inclusion period(2016-2018), and were matched to 22 cases in our database of RP group(2008-2016). The primary outcome measures were quality of reduction and functional outcome. Intraoperative conditions included surgical time, bleeding, and blood transfusion and the postoperative complications were also compared. Results Of these 22 consecutive patients in ATP group, the mean age was 46.7 years and the most common fracture pattern was a both-column fracture(12cases, 54.5%) according to Letournel-Judet classification. The mean follow up period was 23.1 months (range 12-37). There was no significant differences between the two groups with regards the quality of reduction using the Matta radiological criteria and functional outcomes evaluated by the modified Merle d’Aubigné score(P>0.05). Compared with the RP group, the ATP group has a short operation time(245.1 minutes vs. 286.8 minutes, P = 0.020), less intraoperative blood loss(1136.4

screws entering the hip or implant failure. Conclusions The fixation of anatomical titanium plate in quadrilateral plate fractures showed similar result to the reconstruction plate, in terms of quality of reduction and functional outcome. But the utility of anatomical titanium plate had the advantages of shorter operation time, less bleeding and transfusion which is worthy to recommend.

Background
With the development of industrialization, high-energy injuries caused by traffic or industrial construction accidents are gradually increasing in China, and the incidence of acetabular fractures is also increasing year by year [1]. Acetabulum is an important part of the hip, which is a complex anatomical structure with a lot of vital vessels and nerves surrounded. In order to allow early rehabilitation, improve the functional outcomes and decrease the risk of post-traumatic arthritis, operative reduction and internal fixation has become the "gold standard" for the treatment of displaced acetabular fractures [2,3].
Letournel-Judet classification is one of the widely used types of acetabular fractures. According to the acetabular double-column theory, acetabular fractures are divided into 10 categories, including 5 simple fractures and 5 complex fractures [2]. It is worth noting that except for the simple anterior or posterior wall fractures, the other 8 types of acetabular fractures can involve an important anatomical structure called quadrilateral plate. The quadrilateral plate is located in the medial surface of the acetabulum with the deep site and weak bones. When suffering from high-energy trauma, comminuted fracture is common in this area, which always leads to central dislocation of the femoral head. Chang et al. [4] suggested that the reduction of quadrilateral plate fracture played an important role for the surgical results, and more than 80% of implant failure occurred in this area. Although quadrilateral plate fractures are not formally considered as a separate parameter for the classification of acetabular fractures, it has gradually attracted the attention of orthopaedic surgeons and been considered as an important factor affecting the complexity of surgery [5,6].
Along with the renovation of internal fixation technologies, some scholars have proposed novel fixition strategies including 1/3 tubular plate [7], T-shaped [8] or L-shaped buttress plates [9], and infrapectineal plates [10].These devices are not entirely without any limitations, in most cases, the plate should be placed along the upper or medial edge of the pelvic boundary, and it is extremely difficult to fix the quadrilateral plate directly by screws, and slight improper manipulation can easily cause screws penetrating into the hip. Although many scholars have proposed several methods such as inserting screws under direct vision after hip capsulotomy, increasing fluoroscopic times to determine the location of screws [11,12]. The above-mentioned methods will inevitably prolong the operation time and increase the risk of bleeding and infection.
In previous studies, we have presented a novel technique, the Dynamic Anterior Plate-Screw System for Quadrilateral plate(DAPSQ) which was composed of a specially shaped reconstruction plate and several buttress screws of quadrilateral plate(quadrilateral screws) for the treatment of quadrilateral However, the reconstruction plate of DAPSQ requires special intraoperative shaping according to the patient's anatomical parameters, and repeated shaping may lead to a reduction in the elastic strength of the plate. For the reason, we have improved this internal fixation system, and designed a set of anatomical titanium plate of DAPSQ according to the acetabular anatomic parameters of Chinese.
The primary objective of this study was to compare the efficacy included quality of reduction and functional outcomes of anatomical titanium plate versus the reconstruction plate of DAPSQ in the treatment of quadrilateral plate fractures. Secondary objectives included surgical time, intraoperative bleeding and transfusion, in addition to comparison of postoperative complications between the two groups. Twenty-two consecutive acetabular fracture patients treated with anatomical titanium plate of DAPSQ(ATP group), with a minimum of 1-year of follow-up, were formed group 1. Then, the fractures in group 1 were selected and matched according to age, gender, fracture pattern and surgical approach, to similar cases from our dedicated acetabular plate fracture database which included more than 140 cases treated with the reconstruction plate of DAPSQ from January 2008 to January 2016. This pool of cases was then used to randomly choose individual cases to create a 1:1 ratio, and this matched cohort was group 2(RP group).

Surgical technique Preoperative Management
The initial management of acetabular trauma followed the principles of the Advanced Trauma Life Support (ATLS) and the most important was to keep the stable of vital signs [14]. Subsequently, all patients received routine physical and neurological examination. Radiological assessment included an anterior-posterior(AP) view and Judet views(iliac and obturator oblique views), along with twodimensional computed tomography(CT) and three-dimensional(3D) CT reconstruction. Fracture pattern were classified according to Judet and Letournel classification [2]. Skeletal traction was applied via the femoral condyles or tibial tubercle in all patients. Low molecular weight heparin was routinely used to prevent thrombosis, and the color Doppler ultrasonography of Lower Limb Vessels was performed to exclude deep venous thrombosis. One day before the operation, autologous blood transfusion machine and heterogeneous blood( 1000mL ) should be prepared.

Operation procedures
The corresponding author and another senior orthopedic surgeon performed all surgeries. All the surgeries was performed on a radiolucent table using a standard ilioinguinal approach described by Letournel[15] or combination with Kocher-Langenbeck approach. Through the "middle window" of the ilioinguinal approach, the acetabular anterior column, pelvic boundary and the upper part of quadrilateral plate could be directly exposed or touched. First, the medial dislocation of femoral head was repositioned under manual traction on the leg or the assistant of mechanical lateral traction via a Schanz pin in the proximal femur. Then restore the continuity of the pelvic ring followed a principle of proximal to distal and the fracture fragments of ilium wing and anterior column were fixed with screws or plates. Subsequently, push the fracture fragments of acetabular posterior column or quadrilateral plate into its bed with the help of L-shaped spike pusher until a smooth quadrilateral medial surface with no external stepoff was obtained and use a 2 or 3-claw reduction forceps to maintain the reduction.

ATP group
After the reduction was accomplished in the ATP group, choose a appropriate mode of the anatomical titanium plate of DAPSQ. The anatomical titanium plate was placed on the superior arcuate line, and the ends extended toward the iliac wing and the superior pubic ramus. And according to the placement position, it was divided into three parts: the iliac region, the quadrilateral region and the pubic region (Fig.1). We have preliminary divided the anatomical titanium plate into three models according to the anatomical length of the DAPSQ trajectory and the different length proportion of the three regions measured in the Chinese (Fig.2). After placement, the both ends of the titanium plate were up-warped, and were not firmly against the bone surface. But through special nailing methods, the titanium plate could be fully adhered to the bone surface, and detailed nailing methods have shown in Figure 3. The key surgical procedures were as follows: Fistly, two or more fixition screws on the iliac and pubic region should be first fixed to stabilize the acetabular anterior column. Then with the help of a 4.5mm screwdriver, quadrilateral screws were inserted along the pelvic brim and parallel to the surface of quadrilateral plate under direct vision, and only the 1/3 to 1/2 transverse diameter of the quadrilateral screw into the bone of quadrilateral plate to avoid entering the hip. And during the process of nailing, the torsion and elastic recoil of the plate could provide a strong holding force for the quadrilateral screws to block the inward displacement of quadrilateral plate. In addition, make sure that the distal of the quadrilateral screws have exceeded the fracture line by at least 10 mm.

RP group
Patients in the RP group were used a reconstruction plate. The shaping steps were as follows: Firstly, select a 12 to 16 hole arc-shaped reconstruction plate according to the actual anatomical length.
Then the both ends of the plate were reverse twisted and upturned by using a Bender and screwdriver. The torsion angle of the iliac and pubic regions was higher than the radian of the bone surface, so that the both ends of plate were upturned after placed, and the plate in the quadrilateral region slightly incline to the intrapelvic about 15° (Fig.4). In order to achieve the best effect of screws placement, repeatedly adjustment of the torsion angle and proportion of the three regions was necessary. After well shaped, the nailing methods was similar to the ATP group.
Finally, if the reduction was unsatisfactory through a single ilioinguinal approach or patients complicated with the fracture of acetabular posterior wall, adding the Kocher-Langenbeck approach was a suitable choice. After the nailing process was completed in all patients, repeated fluoroscopy was necessary to confirm the good reduction and no screw entering the hip. After acquiring satisfied fluoroscopy, the operation area should be flushed completely and a drainage tube was placed before wound closed.

Postoperative Management
All patients regularly received an intravenous injection of prophylactic antibiotics for 5-7days after surgery. The drainage tube was removed within 3 days (24h drainage flow < 20 mL). Patients started early rehabilitation after awoke from anesthesia and were instructed to non-weight bearing exercises such as passive and active ipsilateral hip flexion or extension motion on the affected limb for 4-6 weeks. Then protected weight-bearing exercises was encouraged till 8-12 weeks, and gradually progress to full-weight bearing at 12 weeks.

Method of assessment and data collection
Gender, age, fracture pattern, mechanism of injury, fracture side, concomitant injuries, time between injury and surgery, surgical time, intraoperative bleeding, blood transfusion, hospital stay time were all collected. Complications analyzed included deep venous thrombosis, sciatic nerve problem, lateral femoral cutaneous nerve injury, surgical site infections, posttraumatic arthritis, heterotopic ossification, screws penetrating into the hip joint cavity and implant failure.
Quality of reduction was evaluated by 2 senior orthopedic surgeons using the Matta radiological criteria [16] according to immediate postoperative pelvic X-ray(anteroposterior, inlet, outlet, and Judet views) and 3D CT reconstruction. The scores were classified as anatomic(0-1mm), imperfect (2-3mm), or poor (>3 mm) based on the maximal displacement on all views.
Regular outpatient review and follow-up were performed 1, 2, 3, 6, 12 months after surgery and then yearly thereafter. Changes in clinical function, radiographic progress, fracture healing, and complications were recorded. Functional outcomes were assessed using the modified Merle d'Aubigné score [17] at the last follow-up and graded as excellent (18 points), good (15-17 points), fair (13 or 14 points), or poor (<13 points).

Statistical analysis
Data was coded and analyzed with the statistical package SPSS version 19.0(IBM Corp, Armonk, NY).
Continuous variables were expressed as mean ± standard deviations and categorical variables with absolute frequencies and percentages. Independent-samples t test was used to compare quantitative variables. Chi-square test or Fisher exact test was used to compare categorical variables. While the ranked data were analysed with Mann-whitney U rank sum test. Cohen's Kappa Index was measured to estimate the inter-observer agreement. A value of P <0.05 was considered statistically significant.

Demographics
Twenty-two patients, including 18 males and 4 females with an average age of 46.7 years (range, 23 to 64 years) were treated with anatomical titanium plate of DAPSQ in the inclusion period. Fractures were classified as both columns in 12, anterior column in 2, ACPHT in 4, T-shaped in 2, transverse and transverse with posterior wall in 1, respectively. Randomization resulted in a matching ratio of 1:1 (Table 1).
Patients were evaluated in terms of their demographic variables. There were no significant differences between the groups with regards to gender (P = 0.176), age (P = 0.202), mechanism of injury (P = 0.646), fracture side (P = 0.353), concomitant injuries (P = 0.918) and mean time between injury and surgery (P = 0.799) ( Table 2).

Intraoperative conditions:
All patients underwent surgery through single ilioinguinal approach(77.3%) or combined with Kocher-Langenbeck approach(22.7%). There were 2 to 4 quadrilateral screws used to control the medial displacement of quadrilateral plate in all patients. For RP group, a 14 or 16 hole reconstruction plate was most commonly used, while type I anatomical plate was most used in ATP group.

Quality of reduction:
The reduction quality of the acetabulum was evaluated by two independent senior orthopedic surgeons according to the Matta radiological criteria. And the degree of agreement between them was 84.1%, with a kappa value of 0.68, which indicated substantial agreement. 16 cases (72.7%) were graded as anatomical reduction in the ATP group, 5 cases(22.7%) as imperfect, and 1 cases(4.5%) as poor. In the RP group, 14 patients (63.6%) showed anatomical reduction, while 6 (27.3%) and 2 (9.1%) patients showed imperfect and poor reduction, respectively. There was no statistically significant difference (P = 0.490) ( Table 2). Postoperative 3D CT reconstruction and X-ray had shown that there was no one quadrilateral screw entering the hip and no one case occurring early fracture displacement or implant failure.

Functional outcome:
Follow-up was greater than 12 months in all 44 patients. The mean time of follow-up was 23.1 months in ATP group vs 26.1 months in RP group (P = 0.260). According to the clinical and radiological evaluation at 12 weeks of follow-up, fracture healing was achieved in all patients. At the last followup, the functional outcome in the ATP group according to the modified Merle d'Aubigné score were excellent in 11 cases (50%), good in 9 cases (40.9%), fair in 1 case(4.5%) and poor in 1(4.5%). While the results in RP group were excellent in 9(40.9%), good in 9(40.9%) and fair in 4(18.2%). Although the excellent-to-good rate of ATP group was higher than the RP group(90.9% vs 81.8%), there was no significant difference between the two groups(P = 0.457). Detailed data was summarized in Table 3, and two typical cases were shown in Figs. 5 and 6.

Complications:
The mean complication rate was 36.4% in RP group which was higher compared with the ATP group(18.2%), but there was no significant difference(P = 0.176). In RP group, One patient presented with an iliac vein thrombosis and pulmonary embolism, and a vena cava filter was inserted on the thirth day after operation. Superficial wound infection was observed in 1 patients, and were treated with antibiotics(2 weeks) and superficial wound debridement. Two cases of lateral femoral cutaneous nerve injury had recovered within two months. Four cases had developed into mild post-traumatic arthritis according to the Kellgren-Lawrence osteoarthritis classification system [18]. In addition, lateral femoral cutaneous nerve injury(1 case) and post-traumatic arthritis(2 cases) were also observed in ATP group, and one case finally required THA at 48 months. Sciatica symptoms developed in 1 patient and recovered after nerve nutrition drugs were applied. None of the patients in two groups had heterotopic ossification or loss of reduction at the last follow-up.

Discussion
The quadrilateral plate is located in the medial wall of acetabulum, which plays an important role in preventing the femoral head from moving into pelvic cavity. When high-energy injury occurs, the medial displacement of femoral head often results in the fracture of quadrilateral plate, and even central dislocation of femoral head[8-10]. If the hip joint surface can not be restored by good reduction and reliable fixation, it will inevitably result in high complication rates such as joint dysfunction or post-traumatic arthritis [2][3][4]. It is particularly important to reconstruct the anatomical structure of the acetabulum and femoral head. On the one hand, it is necessary to restore and coincide the rotation center of the femoral head and the acetabulum. On the other hand, the fixation of quadrilateral plate needs to overcome the inward and upward force of the femoral head. However, due to the deep site, weake bones and surrounded with many important blood vessels, nerves and tissues, quadrilateral plate fracture has always been a constant challenge to orthopaedic surgeons[5,6].
Currently, a number of fixation methods for the treatment of quadrilateral plate fractures are in use.
The most common methods include the reconstruction titanium plate of anterior column combined with lag screw in posterior column [19,20], reconstruction titanium plate of anterior and posterior column [21]. Letournel and Judet [2,15] described the fixation method of quadrilateral plate fractures with long screws inserting along the pelvic brim and through the thin quadrilateral plate. They also commented that this technique was limited if the medial wall of quadrilateral surface was comminuted or accompanied with severe osteoporosis. Although a biomechanical study carried out by Shazar et al. [22] have shown that periarticular long screws could increase the stability of plate fixation, but what can not be ignored is that this method always bears with a high risk of hip penetrating. In most cases, screw holes in the quadrilateral region of the plate are vacant. If insertion is necessary, the direction of the screw should deviate from the joint surface, or a short screw less than 12 mm, otherwise the screw has the risk of penetrating into the hip[10,13]. Abandoning the direct fixation or screws far from the quadrilateral plate means the less stability of the fixation [22].
Although the use of lag screw in posterior column can increase the stability, it requires the high integrity of anterior and posterior column fragments [19,20].
In order to avoid the screws mistakenly entering the hip, the technique of "buttress plating" for quadrilateral plate instead of screws was first described by Mears et al [7]. A one-third tubular plate used in this fashion was bent in an oblique manner and contoured over the pelvic brim to buttress the medial acetabular wall like a spring. However, subsequent studies have found that it was difficult to As the first propensity-matched cohort study that describes the use of anatomical titanium plate and the reconstruction plate of DAPSQ for quadrilateral plate fractures. Our data indicate that the utility of anatomical titanium plate had the advantages of shorter operation time, less bleeding and transfusion. These findings could be explained by the fact that repeated molding of the reconstruction plate was not required in the anatomical titanium plate group, and appropriate model could be individually selected during the operation. At the same time, the process of nailing was more standardized and sequencing, and the corresponding matching apparatus were more well-used and efficient. One could argue that the reconstruction plate can provide more reasonable design to all individuals, so that the plate could be tightly against the bone surface. However, the characteristics of this technique determine that it is not the key element of the internal fixation, but the repeated shaping process may lead to a reduction in the elastic strength of the plate.
In addition, no significant difference was found between the two groups with regards the quality of reduction using the Matta radiological criteria and the functional outcomes evaluated by the modified Merle d'Aubigné score. To our knowledge, there was only one report[28] that used a reconstruction plate placed on the pelvic brim with buttress screws rubbing on the medial surface of the quadrilateral plate to maintain the reduction of the quadrilateral fracture. In their study, 40 cases of quadrilateral plate fracture with mean age 35 years(range, 16-68 years) were enrolled. The anatomical reduction was achieved in 60% and the rate of excellent-to-good evaluated by the modified Merle d'Aubigné score was achieved in 90% of the patients. Compared with our study, the difference may come from the difference in fracture patterns and surgical approachs. But, more importantly, although the fixation concept was similar to ours [28], that was to control medial displacement of the quadrilateral surface on the basis of avoiding the risk of hip penetrating. The fixation method was obviously different from ours. Firstly, we have shaped the reconstruction plate with a special way, and the plate was not fully adhered to the bone surface before screws insertion.
During the screws placement, the change of torsion angle on the plate could provide quadrilateral screws with the sufficient force of lateral dynamic compression to block the medial displacement of quadrilateral plate. The quadrilateral screws were not only used as buttress screws but also dynamic pressurized buttress screws. Secondly, We have drilled semi-"U" type holes parallel to the surface of quadrilateral plate under direct vision, which not only enable the quadrilateral screws partially fixed the fracture fragments, but also make the screws adhere to the quadrilateral surface tightly.
Meanwhile, the quadrilateral screws were placed through a special process, which could make use of the elastic retraction of the plate and prevent the screws migration or floating on the bone surface.
Finally, we have designed a anatomical titanium plate with the immobilization length and proportion of three regions, which is more convenient to use. Furthermore, no significant difference was found between the two groups compared with the mean complication rate. As expected, no one had occurred the quadrilateral screws entering the hip. And there was also no one case occurring early fracture displacement or implant failure. In addition, it is worth noting that the fixation method was effective in 6 elderly patients whose age above 60 years.
This may suggested that the technique was effective for osteopenic quadrilateral plate fractures, but larger sample sizes were required. In clinical practice, we found that the anatomical titanium plate of DAPSQ has the following advantages compared with the reconstruction plate: Firstly, additional prebending procedure of the plate is not needed during the operation, which greatly simplifies the operation process. Secondly, the pre-moulding design of the anatomical titanium plate can provide more stable elastic strength for quadrilateral screws and reduce the loss of elastic strength due to repeated shaping of the reconstruction plate during operation. Finally, the design of the additional screw hole in the quadrilateral region can not only assist the quadrilateral screws nailing, but also can be used to insert the lag screw in the posterior column to improve the fixation effect if necessary. And the process of nailing in the anatomical titanium plate is more standardized and the matching apparatus are more complete, which greatly improves the efficiency of nailing.
There are several limitations in this study. Firstly, this is a propensity-matched cohort study, where the study group originate from a prospective cohort and are matched with the cases from a retrospective database, which may have selection bias. The selection bias can be minimised by matching all the basic data such as age, fracture types and surgical approaches as much as possible but can not be eliminated. One possible bias worth mentioning is that the learning curve for the new technique may affect the results. Finally, the most common fracture type included in this study is double column fracture, which is different from the distribution types of acetabular fractures in clinical practice. This selection bias is mainly caused by the fact that our cases come from a level I trauma center where the most complex acetabular fractures are referred. Therefore, the results of this study need to be interpreted with caution and confirmed with a large sample.

Availability of data and materials
The datasets generated and/or analyzed during the current study are available from the corresponding author by reasonable request. Tables Table 1 Fracture types and surgical approach in the anatomical titanium plate(ATP) group and reconstruction plate(RP) group    Figure 1 The structure and components of DAPSQ DAPSQ was placed on the superior arcuate line, and the ends extended toward the iliac wing and the superior pubic ramus, respectively.
According to the placement position, it was divided into three parts: the iliac region, the quadrilateral region and the pubic region. And the screws placed in the quadrilateral region were called "quadrilateral screws".  Then move the plate of the iliac region away from the pelvic rim, so that the first quadrilateral screw can fit snugly against the medial surface of quadrilateral plate and insert a screw in the iliac region(e). Subsequently, two fixation screws were inserted into the pubic and the iliac region, respectively, to stabilize the acetabular anterior column(f). Then, a special method of quadrilateral screw placement was adopted. Firstly, a 4.5mm screwdriver was inserted into the adjacent screw hole to lift the internal edge of the plate, so that the 1/3 to 1/2 of the screws hole were exposed to the

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download. stability of quadrilateral screws.MTS The placement sequence of screws .mp4